QUEENSLAND UNIVERSITY OF TECHNOLOGY SCHOOL OF NURSINGeprints.qut.edu.au/43715/1/Qunyan_Xu_Thesis.pdf · 2011. 7. 28. · Xu, Q., Anderson, D., & Lurie-Beck, J. The effect of central

QUEENSLAND UNIVERSITY OF TECHNOLOGY

SCHOOL OF NURSING

An examination of the relationships between lifestyle factors and mental health

among Australian midlife and older women

by

Qunyan Xu

RN, BNurs, MNurs

A thesis submitted in fulfilment of the requirement for the degree of

Doctorate of Philosophy

School of Nursing and Midwifery

Queensland University of Technology

2010

I

KEYWORDS

Lifestyle factors, BMI, smoking, alcohol use, physical activity, mental health,

anxiety, depression, midlife and older women, diabetes, longitudinal

II

STATEMENT OF ORIGINAL AUTHORSHIP

The work contained in this thesis has not been previously submitted to meet

requirement for an award at this or any other education institution. To the best of my

knowledge and belief, the thesis contains no materials previously published or

written by another person except where due reference is made.

Signature Date

III

RELATED PUBLICATIONS AND PRESENTATIONS

Portions of this thesis have been published in the following journal article:

Xu, Q., Anderson, D., & Courtney, M. (2010). A longitudinal study of the

relationship between lifestyle and mental health among midlife and older women in

Australia: findings from the Healthy Aging of Women Study. Health Care for

Women International,31(12), 1082-1096.

Xu, Q., Anderson, D., & Lurie-Beck, J. The effect of central obesity on

depression in general population: a systematic review and meta-analysis. Obesity

Research and Clinical Practice (under review).

Portions of this thesis have been presented at the following conferences:

Xu, Q., Anderson, D., & Barr, J. (2009). The relationships between mental

health and chronic disease risk factors among midlife and older women in Australia.

Paper presented at the 4th International Congress on Innovations in Nursing: Perth,

Australia.

Xu, Q., & Anderson, D. (2009). Psychological symptoms in cardiovascular

disease: a women’s study. Paper presented at Heart Foundation Conference:

Brisbane, Australia.

Xu, Q., & Anderson, D. (2009). Mental well-being and psychological factors

in relation to diabetes in midlife and older women: results from the Queensland

Healthy Ageing Women Study. Paper presented at the Australian Diabetes Educators

Associations, Queensland Branch Conference: Brisbane, Australia.

Xu, Q., Anderson, D., & Lang, C.P. (2010). The relationship between

diabetes and mental health in Australian midlife and older women. Paper presented

at Women’s Health 2010: the 18th Annual Congress: Washington D.C., America.

IV

Seminar presentation:

Xu, Q. (2008, Feb.). Mental health, quality of life and chronic disease in

midlife and older women: a literature review. Postgraduate Research Development

Forum, School of Nursing and Midwifery, Queensland University of Technology,

Brisbane.

Xu, Q. (2008, Jul.). Mental health, chronic disease and quality of life in

midlife and older women: methodology of the study. Postgraduate Research

Development Forum, School of Nursing and Midwifery, Queensland University of

Technology, Brisbane.

V

ACKNOWLEDGEMENT

My deepest gratitude goes first to my supervision team: principle supervisor,

Professor Debra Anderson, and associate supervisor, Professor Mary Courtney, for

their constant support and guidance at every stage of the PhD journey.

Second, I feel grateful to Queensland University of Technology and Chinese

Scholarship Council, who provided me with financial assistance to my doctoral

study. Without their support, studying a PhD would have been impossible.

Third, many thanks to my dear friends: Amy Mitchell, Amanda McGuire,

Cathryne Lang, Reimei Hong, Ralph Tramm and Yan Lou. I feel extremely fortunate

to have their company and moral support during the PhD journey.

Finally, my thanks would go to my beloved husband Yang for his loving

consideration through the three years.

VI

ABSTRACT

Background

It is well known that lifestyle factors including overweight/obesity, physical

inactivity, smoking and alcohol use are largely related with morbidity and mortality

of chronic diseases including diabetes and cardiovascular diseases. The effect of

lifestyle factors on people’s mental health who have a chronic disease is less defined

in the research. The World Health Organisation has defined health as “a state of

complete physical, mental and social well-being”. It is important, therefore to

develop an understanding of the relationships between lifestyle and mental health as

this may have implications for maximising the efficacy of health promotion in

people with chronic diseases.

Objectives

The overall aim of the research was to examine the relationships between

lifestyle factors and mental health among Australian midlife and older women.

Methodology

The current research measured four lifestyle factors including weight status,

physical activity, smoking and alcohol use. Three interconnecting studies were

undertaken to develop a comprehensive understanding of the relationships between

lifestyle factors and mental health. Study 1 investigated the longitudinal effect of

lifestyle factors on mental health by using midlife and older women randomly

selected from the community. Study 2 adopted a cross-sectional design, and

compared the effect of lifestyle factors on mental health between midlife and older

women with and without diabetes. Study 3 examined the mediating effect of

VII

self-efficacy in the relationships between lifestyle factors and mental health among

midlife and older women with diabetes. A questionnaire survey was chosen as the

means to gather information, and multiple linear regression analysis was conducted

as the primary statistical approach.

Results

The research showed that the four lifestyle factors including weight status,

physical activity, smoking and alcohol use did impact on mental health among

Australian midlife and older women. First, women with a higher BMI had lower

levels of mental health than women with normal weight, but as women age, the

mental health of women who were overweight and obese becomes better than that of

women with normal weight. Second, women who were physically active had higher

levels of mental health than those who were not. Third, smoking adversely impacted

on women’s mental health. Finally, those who were past-drinkers had less anxiety

symptoms than women who were non-drinkers as they age.

Women with diabetes appeared to have lower levels of mental health

compared to women without. However, the disparities of mental health between two

groups were confounded by low levels of physical activity and co-morbidities. This

finding underlines the effect of physical activity on women’s mental health, and

highlights the potential of reducing the gap of mental health by promoting physical

activity. In addition, self-efficacy was shown to be the mediator of the relationships

between BMI, physical activity and depression, suggesting that enhancing people’s

self-efficacy may be useful for mental health improvement.

VIII

Conclusions

In conclusion, Australian midlife and older women who live with a healthier

lifestyle have higher levels of mental health. It is suggested that strategies aiming to

improve people’s mental health may be more effective if they focus on enhancing

people’s self-efficacy levels. This study has implications to both health education

and policy development. It indicates that health professionals may need to consider

clients’ mental health as an integrated part of lifestyle changing process. Furthermore,

given that lifestyle factors impact on both physical and mental health, lifestyle

modification should continue to be the focus of policy development.

IX

TABLE OF CONTENTS

KEYWORDSKEYWORDSKEYWORDSKEYWORDS............................................................................................................................................................................................................................................................................................................................................................................................................................................................ IIII

STATEMENTSTATEMENTSTATEMENTSTATEMENT OFOFOFOF ORIGINALORIGINALORIGINALORIGINAL AUTHORSHIPAUTHORSHIPAUTHORSHIPAUTHORSHIP.................................................................................................................................................................................................................................................................... IIIIIIII

RELATEDRELATEDRELATEDRELATED PUBLICATIONSPUBLICATIONSPUBLICATIONSPUBLICATIONS ANDANDANDAND PRESENTATIONSPRESENTATIONSPRESENTATIONSPRESENTATIONS........................................................................................................................................................................................................................IIIIIIIIIIII

ACKNOWLEDGEMENTACKNOWLEDGEMENTACKNOWLEDGEMENTACKNOWLEDGEMENT........................................................................................................................................................................................................................................................................................................................................................................................VVVV

ABSTRACTABSTRACTABSTRACTABSTRACT............................................................................................................................................................................................................................................................................................................................................................................................................................................................VIVIVIVI

TABLETABLETABLETABLE OFOFOFOF CONTENTSCONTENTSCONTENTSCONTENTS............................................................................................................................................................................................................................................................................................................................................................................................IXIXIXIX

LISTLISTLISTLIST OFOFOFOF TABLESTABLESTABLESTABLES.................................................................................................................................................................................................................................................................................................................................................................................................................... XIIIXIIIXIIIXIII

LISTLISTLISTLIST OFOFOFOF FIGURESFIGURESFIGURESFIGURES........................................................................................................................................................................................................................................................................................................................................................................................................ XVIIXVIIXVIIXVII

LISTLISTLISTLIST OFOFOFOF APPENDICESAPPENDICESAPPENDICESAPPENDICES....................................................................................................................................................................................................................................................................................................................................................................................XIXXIXXIXXIX

ChapterChapterChapterChapter 1:1:1:1: IIIIntroductionntroductionntroductionntroduction................................................................................................................................................................................................................................................................................................................................................................................1111

OverviewOverviewOverviewOverview ofofofof thethethethe ResearchResearchResearchResearch............................................................................................................................................................................................................................................................................................................................................................................................1111

BackgroundBackgroundBackgroundBackground....................................................................................................................................................................................................................................................................................................................................................................................................................................................................................1111

DefinitionsDefinitionsDefinitionsDefinitions andandandand TermsTermsTermsTerms....................................................................................................................................................................................................................................................................................................................................................................................................................8888

AimsAimsAimsAims ofofofof thethethethe ResearchResearchResearchResearch................................................................................................................................................................................................................................................................................................................................................................................................................11111111

ResearchResearchResearchResearch PlanPlanPlanPlan............................................................................................................................................................................................................................................................................................................................................................................................................................................................ 11111111

SignificanceSignificanceSignificanceSignificance ofofofof thethethethe ResearchResearchResearchResearch.................................................................................................................................................................................................................................................................................................................................................................... 12121212

StructureStructureStructureStructure ofofofof thethethethe ThesisThesisThesisThesis........................................................................................................................................................................................................................................................................................................................................................................................................12121212

ChapterChapterChapterChapter SummarySummarySummarySummary....................................................................................................................................................................................................................................................................................................................................................................................................................................13131313

ChapterChapterChapterChapter 2:2:2:2: LiteratureLiteratureLiteratureLiterature ReviewReviewReviewReview.................................................................................................................................................................................................................................................................................................................................... 15151515

BriefBriefBriefBrief IntroductionIntroductionIntroductionIntroduction.................................................................................................................................................................................................................................................................................................................................................................................................................................... 15151515

WomenWomenWomenWomen andandandandMentalMentalMentalMental HealthHealthHealthHealth........................................................................................................................................................................................................................................................................................................................................................................ 15151515

TheoreticalTheoreticalTheoreticalTheoretical BackgroundBackgroundBackgroundBackground................................................................................................................................................................................................................................................................................................................................................................................................18181818

ConceptualConceptualConceptualConceptual andandandand TheoreticalTheoreticalTheoreticalTheoretical FrameworkFrameworkFrameworkFramework............................................................................................................................................................................................................................................................................................ 21212121Health Promotion.................................................................................................................. 21The Health Belief Model........................................................................................................ 23The Theory of Planned Behaviour.......................................................................................... 25The Transtheoretical Model................................................................................................... 27The Social Cognitive Theory................................................................................................... 29Appraisal of Social Cognitive Theory and Other Psychological Models.................................... 34

MultipleMultipleMultipleMultiple LifestyleLifestyleLifestyleLifestyle Factors,Factors,Factors,Factors, DiabetesDiabetesDiabetesDiabetes andandandandMentalMentalMentalMental HealthHealthHealthHealth............................................................................................................................................................................................ 36363636Obesity and Mental Health.................................................................................................... 38Physical Activity and Mental Health....................................................................................... 46Smoking and Mental Health...................................................................................................52Alcohol Use and Mental Health..............................................................................................57Diabetes and Mental Health...................................................................................................63

ExaminingExaminingExaminingExamining thethethethe LimitationsLimitationsLimitationsLimitations ofofofof thethethethe PreviousPreviousPreviousPrevious ResearchResearchResearchResearch................................................................................................................................................................................................................ 67676767

ConceptualConceptualConceptualConceptual FrameworkFrameworkFrameworkFramework....................................................................................................................................................................................................................................................................................................................................................................................................69696969

X

AimsAimsAimsAims ofofofof thethethethe ResearchResearchResearchResearch................................................................................................................................................................................................................................................................................................................................................................................................................70707070Aim 1: The Relationships Between Lifestyle Factors and Mental Health.................................. 70Aim 2: Examining the Mediating Role of Self-efficacy............................................................. 71

ResearchResearchResearchResearch QuestionsQuestionsQuestionsQuestions........................................................................................................................................................................................................................................................................................................................................................................................................................ 72727272

ChapterChapterChapterChapter SummarySummarySummarySummary....................................................................................................................................................................................................................................................................................................................................................................................................................................72727272

ChapterChapterChapterChapter 3:3:3:3: MethodologyMethodologyMethodologyMethodology................................................................................................................................................................................................................................................................................................................................................................75757575

IntroductionIntroductionIntroductionIntroduction........................................................................................................................................................................................................................................................................................................................................................................................................................................................................75757575

AnAnAnAn OverviewOverviewOverviewOverview ofofofof thethethetheWomenWomenWomenWomen’’’’ssss HealthHealthHealthHealth StudiesStudiesStudiesStudies............................................................................................................................................................................................................................................................75757575

LinkageLinkageLinkageLinkage ofofofof thethethethe CurrentCurrentCurrentCurrent ResearchResearchResearchResearchWithWithWithWith HOW,HOW,HOW,HOW,WWPWWPWWPWWPandandandand CDWWPCDWWPCDWWPCDWWP.................................................................................................................................... 77777777

MethodologyMethodologyMethodologyMethodology ofofofof StudyStudyStudyStudy 1111............................................................................................................................................................................................................................................................................................................................................................................................ 78787878Design................................................................................................................................... 78Sample.................................................................................................................................. 78Ethical Clearance................................................................................................................... 80Measures...............................................................................................................................80Statistical Analysis..................................................................................................................89

MethodologyMethodologyMethodologyMethodology ofofofof StudyStudyStudyStudy 2222............................................................................................................................................................................................................................................................................................................................................................................................ 91919191Design................................................................................................................................... 91Sample.................................................................................................................................. 91Ethical Clearance................................................................................................................... 92Measures...............................................................................................................................93Statistical Analysis..................................................................................................................97

MethodologyMethodologyMethodologyMethodology ofofofof StudyStudyStudyStudy 3333............................................................................................................................................................................................................................................................................................................................................................................................ 99999999Design................................................................................................................................... 99Sample.................................................................................................................................. 99Ethical Clearance................................................................................................................... 99Measures.............................................................................................................................100Statistical Analysis................................................................................................................106

ChapterChapterChapterChapter SummarySummarySummarySummary............................................................................................................................................................................................................................................................................................................................................................................................................................109109109109

ChapterChapterChapterChapter 4444 RRRResultsesultsesultsesults ofofofof StudyStudyStudyStudy 1:1:1:1: thethethethe relationshipsrelationshipsrelationshipsrelationships betweenbetweenbetweenbetween lifestylelifestylelifestylelifestyle factorsfactorsfactorsfactors andandandandmentalmentalmentalmental healthhealthhealthhealth amongamongamongamong AustralianAustralianAustralianAustralian midlifemidlifemidlifemidlife andandandand olderolderolderolder womenwomenwomenwomen................................................................................................................................ 111111111111

IntroductionIntroductionIntroductionIntroduction................................................................................................................................................................................................................................................................................................................................................................................................................................................................111111111111

ResultsResultsResultsResults ofofofof StudyStudyStudyStudy 1111............................................................................................................................................................................................................................................................................................................................................................................................................................ 111111111111Characteristics of Sample.....................................................................................................112The Correlations Among Lifestyle Factors.............................................................................115Multiple Linear Regressions: Lifestyle Factors Predicting Mental Health at Baseline..............117Multiple Linear Regressions: the Prospective Relationships Between Lifestyle Factors andMental Health......................................................................................................................128Reflection on Research Questions........................................................................................ 142

ChapterChapterChapterChapter 5555 RRRResultsesultsesultsesults ofofofof StudyStudyStudyStudy 2:2:2:2: thethethethe relationshipsrelationshipsrelationshipsrelationships betweenbetweenbetweenbetween lifestylelifestylelifestylelifestyle factorsfactorsfactorsfactors andandandandmentalmentalmentalmental healthhealthhealthhealth amongamongamongamong AustralianAustralianAustralianAustralian midlifemidlifemidlifemidlife andandandand olderolderolderolder womenwomenwomenwomen withwithwithwith andandandand withoutwithoutwithoutwithoutdiabetesdiabetesdiabetesdiabetes............................................................................................................................................................................................................................................................................................................................................................................................................................................................ 145145145145


ResultsResultsResultsResults ofofofof StudyStudyStudyStudy 2222............................................................................................................................................................................................................................................................................................................................................................................................................................ 146146146146Description of Sociodemographic Characteristics of the Sample........................................... 146Number of Co-morbidities................................................................................................... 150

XI

The Description and Comparison of Mental Health and Lifestyle Factors of WomenWith andWithout Diabetes.................................................................................................................150The Difference of Eight Scales and Composite Score in SF-36 Between WomenWith andWithout Diabetes.................................................................................................................157The Binary Analysis Between Independent Variables and Dependent Variables.................... 160Predicting Mental Health: Hierarchical Regression Analysis.................................................. 165Reflection on Research Questions........................................................................................ 175

ChapterChapterChapterChapter 6666 RRRResultsesultsesultsesults ofofofof StudyStudyStudyStudy 3:3:3:3: analysisanalysisanalysisanalysis ofofofof tttthehehehe mediationmediationmediationmediation ofofofof self-efficacyself-efficacyself-efficacyself-efficacy inininin thethethetherelationshipsrelationshipsrelationshipsrelationships betweenbetweenbetweenbetween liefstyleliefstyleliefstyleliefstyle factorsfactorsfactorsfactors andandandand mentalmentalmentalmental healthhealthhealthhealth inininin midlifemidlifemidlifemidlife andandandand olderolderolderolderwomenwomenwomenwomen withwithwithwith diabetesdiabetesdiabetesdiabetes........................................................................................................................................................................................................................................................................................................................................................................ 177177177177


ResultsResultsResultsResults ofofofof StudyStudyStudyStudy 3333............................................................................................................................................................................................................................................................................................................................................................................................................................ 179179179179The Correlations Between Mental Health and Lifestyle Factors.............................................179The Correlations Between Mental Health and Self-Efficacy in Managing Chronic Disease......180Associations Between Self-Efficacy in Managing Chronic Disease and Lifestyle Factors......... 190The Associations of Self-EfficacyWith Sociodemographic Factors and Other Confounders....192Mediation Analysis...............................................................................................................194Reflection on the Research Questions.................................................................................. 201

ChapterChapterChapterChapter 7:7:7:7: DiscussionDiscussionDiscussionDiscussion............................................................................................................................................................................................................................................................................................................................................................................203203203203


TheTheTheThe LifestylesLifestylesLifestylesLifestyles ofofofof AustralianAustralianAustralianAustralianMidlifeMidlifeMidlifeMidlife andandandand OlderOlderOlderOlderWomenWomenWomenWomen............................................................................................................................................................................................203203203203

TheTheTheThe EffectEffectEffectEffect ofofofof LifestyleLifestyleLifestyleLifestyle FactorsFactorsFactorsFactors ononononMentalMentalMentalMental HealthHealthHealthHealth ininininMidlifeMidlifeMidlifeMidlife andandandand OlderOlderOlderOlderWomenWomenWomenWomen............................................206206206206Overweight and Obesity and Mental Health.........................................................................206Physical Activity and Mental health......................................................................................209Smoking and Mental Health.................................................................................................211Alcohol Use and Mental Health............................................................................................213Age and Mental Health........................................................................................................ 216

TheTheTheThe EffectsEffectsEffectsEffects ofofofof LifestyleLifestyleLifestyleLifestyle FactorsFactorsFactorsFactors onononon thethethetheMentalMentalMentalMental HealthHealthHealthHealth ofofofofWomenWomenWomenWomenWithWithWithWith andandandandWithoutWithoutWithoutWithoutDiabetesDiabetesDiabetesDiabetes........................................................................................................................................................................................................................................................................................................................................................................................................................................................................................216216216216

TheTheTheTheMediatingMediatingMediatingMediating EffectEffectEffectEffect ofofofof Self-Efficacy,Self-Efficacy,Self-Efficacy,Self-Efficacy, MentalMentalMentalMental HealthHealthHealthHealth andandandand LifestyleLifestyleLifestyleLifestyle FactorsFactorsFactorsFactors....................................................................219219219219

Self-Efficacy,Self-Efficacy,Self-Efficacy,Self-Efficacy, DurationDurationDurationDuration ofofofof DiabetesDiabetesDiabetesDiabetes andandandandUseUseUseUse ofofofof AntidepressantsAntidepressantsAntidepressantsAntidepressants.................................................................................................................................... 221221221221

TheoreticalTheoreticalTheoreticalTheoretical ReflectionReflectionReflectionReflection....................................................................................................................................................................................................................................................................................................................................................................................................222222222222

StrengthsStrengthsStrengthsStrengths andandandand LimitationsLimitationsLimitationsLimitations............................................................................................................................................................................................................................................................................................................................................................................227227227227

ChapterChapterChapterChapter 8:8:8:8: ConclusionsConclusionsConclusionsConclusions....................................................................................................................................................................................................................................................................................................................................................................231231231231


SummarySummarySummarySummary ofofofof MajorMajorMajorMajor FindingsFindingsFindingsFindings................................................................................................................................................................................................................................................................................................................................................................231231231231Effect of Lifestyle Factors on Mental Health in Midlife and Older Women.............................231Effect of Lifestyle Factors on Mental Health in Midlife and Older WomenWith and WithoutDiabetes.............................................................................................................................. 232The Mediating Role of Self-Efficacy in the Associations Between Lifestyle Factors and MentalHealth AmongMidlife and Older WomenWith Diabetes...................................................... 233

ImplicationsImplicationsImplicationsImplications................................................................................................................................................................................................................................................................................................................................................................................................................................................................ 233233233233Implications of the Study..................................................................................................... 233Implications for Health Education........................................................................................ 235Implications for Policy Making..............................................................................................236Implications for Future Research..........................................................................................237

XII

ConclusionConclusionConclusionConclusionssss....................................................................................................................................................................................................................................................................................................................................................................................................................................................................238238238238

ReferencesReferencesReferencesReferences............................................................................................................................................................................................................................................................................................................................................................................................................................................ 240240240240

XIII

LIST OF TABLES

Table 2.1 The Transtheoretical Model Constructs 28

Table 2.2 Classification of Overweight and Obesity by BMI, Waist Circumference

and Associated Disease Risk 39

Table 4.1 Descriptive Analysis of Study Sample 113

Table 4.2 The Correlation Matrix Table of Lifestyle Factors, the Significance Level

Tests (p values) 117

Table 4.3 Multiple Linear Regression, Anxiety at Baseline as Dependent Variable (N

= 433) 118

Table 4.4 Multiple Linear Regression, Depression at Baseline as Dependent Variable

(N = 439) 120

Table 4.5 Multiple Linear Regression, Psychological Symptoms at Baseline as

Dependent Variable (N = 420) 122

Table 4.6 Multiple Linear Regression, the Mental Health Inventory at Baseline as


Table 4.7 Multiple Linear Regression, the Mental Composite Scores at Baseline as


Table 4.8 Multiple Linear Regression, Anxiety at Follow up as Dependent Variable

(N = 405) 129

Table 4.9 Multiple Linear Regression, Depression at follow up as Dependent

Variable (N = 417) 131

Table 4.10 Multiple Linear Regression, Pyschological Symptoms at Follow up as


Table 4.11 Multiple Linear Regression, the Mental Health Inventory Scores at

Follow up as Dependent Variable (N = 433) 135

XIV

Table 4.12 Multiple Linear Regression, the Mental Composite Scores at Follow up

as Dependent Variable (N = 346) 137

Table 4.13 A Comparison of the Changing Scores of Mental Health Among Women

With Different BMI Categories (N = 385 ~ 492) 139

Table 4.14 Paired t-Test of Depression and Psychological Symptoms Between

Baseline and 5 years Follow up (N = 454 ~ 493) 141

Table 5.1 The Characteristics of Sociodemographic Factors of Women With and

Without Diabetes (N = 176) 148

Table 5.2 The Number of Co-morbidities (excluding diabetes) in Women With and

Without Diabetes (N = 177) 150

Table 5.3 The Statistics of Normality Examination of Mental Health Variables (N =

177) 151

Table 5.4 The Differences in Mental Health Between Women With and Without

Diabetes (N = 177) 152

Table 5.5 The Differences in Lifestyle Factors Between Women With and Without


Table 5.6 Scoring Check: the Correlations Between Eight Scales and Composite

Scores of SF-36 (N = 172 ~ 175) 157

Table 5.7 The Differences in Quality of Life Between Women With and Without


Table 5.8 Level of Significance (p values) of the Relationships Between

Sociodemographic Factors and Mental Health (N = 176) 161

Table 5.9 The Relationships Between Lifestyle Factors and Mental Health Variables

(N = 170 ~ 177) 163

XV

Table 5.10 The Correlations Between Number of Co-morbidities, Menopausal Status

and Mental Health Variables (N = 168 ~ 176) 164

Table 5.11 Hierarchical Regression Analysis to Predict Anxiety (N = 163) 167

Table 5.12 Hierarchical Regression Analysis to Predict Depression (N = 164) 169

Table 5.13 Hierarchical Regression Analysis to Predict Psychological Symptoms (N

= 163) 171

Table 5.14 Hierarchical Regression Analysis to Predict the Mental Health Inventory

Scores ( N = 163) 172

Table 5.15 Hierarchical Regression Analysis to Predict the Mental Composite Scores

(N = 159) 174

Table 6.1 The Correlations Between Lifestyle Factors and Mental Health (N = 83)

180

Table 6.2 Mean Score of Self-efficacy in Managing Chronic Diseases (N = 80) 181

Table 6.3 The Description of Anxiety and Depression and Measured by the Hospital

Anxiety and Depression Scale (N = 80) 182

Table 6.4 The Frequency Distribution of Mental Health Domain in SF-36 (N = 80)

183

Table 6.5 The Frequency Disctribution of Vitality Domain of SF-36 (N = 80) 183

Table 6.6 The Frequency Distribution of Social Function of SF-36 (N = 83) 184

Table 6.7 The Frequency Distribution of Role Emotional of SF-36 (N = 83) 185

Table 6.8 The Frequency Distribution of Physical Function of SF-36 (N = 83) 185

Table 6.9 The Frequency Distribution of Role Physical of SF-36 (N = 83) 186

Table 6.10 The Frequency Distribution of Bodily Pain of SF-36 (N = 83) 187

Table 6.11 The Frequency Dsistribution of General Health of SF-36 (N = 83) 187

XVI

Table 6.12 The Correlations Between Eight Scales of SF-36 and Two Composite

Scores (N = 83) 189

Table 6.13 Associations Between Self-efficacy in Managing Chronic Diseases and

Mental Health (N = 83) 190

Table 6.14 Associations Between Self-efficacy in Managing Chronic Diseases and

Lifestyle Factors (N = 80) 191

Table 6.15 The Differences of Self-efficacy in Managing Chronic Diseases in

Relation to Sociodemographic Factors and Other Confounders (N = 77 ~ 83)

193

Table 6.16 Multiple Linear Regression, Using Depression to Predict Physical

Activity (N = 70) 197

Table 6.17 Multiple Linear Regression, Using Depression to Predict BMI (N = 70)

198

Table 6.18 Multiple Linear Regression, Using BMI to Predict Depression (N = 72)

200

Table 6.19 Multiple Linear Regression, Using Physical Activity to Predict

Depression (N = 70) 201

XVII

LIST OF FIGURES

Figure 2.1. The individual as a structural coupling of three systems: organism, mind

and social status. 18

Figure 2.2. Schematic model of the Global Strategy on Diet, Physical Activity and

Health. 22

Figure 2.3. The Health Belief Model. 25

Figure 2.4. Conceptual framework of theory of planned behaviour. 27

Figure 2.5. Social cognitive theory. 31

Figure 2.6. Structural paths of self-efficacy theory. 32

Figure 2.7. Self-efficacy theory - structural paths of influence. 34

Figure 2.8. The conceptual framework of the study. 70

Figure 3.1.The research design and its relation to women’s health studies. 78

Figure 3.2. The flow chart of sample recruitment for Study 1. 79

Figure 3.3. Basic causal chains of mediation model. 107

Figure 3.4. Mediation model using lifestyle factors as outcome variables. 107

Figure 3.5. Mediation model using mental health as outcome variables. 108

Figure 4.1. The impact of alcohol use on the change of anxiety scores over 5 years.

140

Figure 4.2. The impact of smoking on the change of MHI scores over 5 years. 140

Figure4.3. The impact of smoking on the change of MCS scores over 5 years. 141

Figure 6.1. Sample recruitment procedure of Study 3. 178

Figure 6.2. Mediating model l: using depression to predict physical activity. 195

Figure 6.3. Mediation model 2: using depression to predict BMI. 196

Figure 6.4. Mediation model 3: using BMI to predict depression. 196

Figure 6.5. Mediation model 4: using physical activity to predict depression. 196

XVIII

Figure 7.1. A conceptual framework for the relationships between lifestyle factors

and mental health among Australian midlife and older women. 226

XIX

LIST OF APPENDICES

Appendix A: A Comparison of Physical Activity Among Different BMI Categories

280

Appendix B: Medical Outcomes Study Short Form (SF-36) 281

Appendix C: Greene’s Climacteric Scale 284

Appendix D: The Hospital Anxiety and Depression Scale 285

Appendix E: Self-efficacy in Managing Chronic Diseases 287

Appendix F: The Seattle Physical Activity Questionnaire 288

Appendix G: Ethical Approval 289

1

CHAPTER 1: INTRODUCTION

Overview of the Research

The goal of this research is to develop a comprehensive understanding about

relationships between lifestyle risk factors and mental health among Australian

women in midlife and older. Popular understanding variably places ‘midlife’ at 45 to

64 years of age, ‘old’ at 65 to 84 years and ‘oldest old’ at 85 years and beyond. In

the current research, the sample population are mostly consisted of women at their

middle life, and it also incorporates a small percentage of old women. It is expected

that the knowledge generated from this research will contribute to the development

of strategies that are effective in facilitating change of unhealthy lifestyle for such

women.

Background

The last century witnessed a considerable increase in the proportion of

mature age Australians (defined as 50 to 59 years of age) amongst the overall

national population. As revealed by the sixth edition of Australian Social Trend

(People in their 50s: then and now, p. 7): the number of middle aged Australians has

climbed up from 1.5 million to 2.2 million, with an equivalent increase rate for men

and women (Australian Bureau of Statistics, 2006d). In Australia, due to the low

fertility rate and increased life expectancy, the number of Australians within this age

group is projected to increase continuously. The life expectancy of women at age 50

has increased by 4 years in 2002-2004, up from 81 years in 1980-1982 (Australian

Bureau of Statistics, 2006d); however, the life expectancy without disability has not

increased proportionally (Australian Institute of Health and Welfare, 2004). The

2

above described circumstances make middle aged population significant in

government policy development in terms of promoting healthy ageing of the society.

Prior to the description of relevant health issues among midlife and older

women, a brief overview of the socioeconomic context where these women live is

provided to allow a rich understanding of their health profile. Along with the overall

improvement of Australians’ qualifications, midlife women have also become better

educated. The proportion of midlife women holding a bachelor’s degree and beyond

has increased from 3% in 1984 to 17% in 2005 (Australian Bureau of Statistics,

2006d). The advancement of women’s education levels certainly has a profound

impact on their participation in work force, and subsequently on their income and

wealth. In 1984, it was reported that 37% midlife women were in employment, and

by 2005, the corresponding figure goes up to 47%. With respect to financial

circumstance, the Australian Survey of Income and Housing 2003-2004 indicated

that middle aged group (45-54 years) had the highest income of $1,400 per week in

all age groups and gradually levelled off as people get older. In spite of this,

household net wealth continued to increase until reaching the summit of $740,000 in

people aged from 60 to 64 years of age (Australian Bureau of Statistics, 2006d, see p.

146). In short, middle aged women live in a financially stable environment which

allows them to have better access to services and products required in daily life.

The prevalence of chronic diseases has been increasing with an alarming rate

among the aged population. As revealed by the National Health Survey (2007-2008),

the average number of the National Health Priority Area (NHPA) conditions, which

include arthritis and osteoporosis, asthma, cancer, diabetes, cardiovascular disease,

injury, mental health and obesity, increased steadily from 0.2 within the cohort aged

from 0 to 14 years to 2.7 in the age group of 75 years and over (Australian Bureau of

3

Statistics, 2006f). Cardiovascular disease, cancer and neurological diseases account

for half of the disease of burden amongst adults aged from 45 to 64 years (Begg, Vos,

Barker, Stevenson, Stanley, & Lopez, 2007). Thus, how to tackle the health

problems or delay the occurrence of diseases among middle aged population

becomes critically important to the health practice of older population (Healy, 2004).

As reported by the World Health Organisation (WHO), chronic diseases

including cardiovascular disease, diabetes, obesity, certain types of cancer and other

chronic respiratory diseases accounted for 60% of the 58 million deaths in 2005

worldwide (World Health Organisation, 2005b). Non-communicable diseases are

strongly related with unhealthy lifestyle factors, constitute a significant financial

burden to the health system and largely impair people’s health related quality of life.

Type 2 diabetes is one of the typical lifestyle diseases that affect a large number of

individuals. The statistics revealed that the prevalence of Type 2 diabetes in

Australia is 7% based on measured blood sugar level, and 3% on self-report

diagnosis (Australian Bureau of Statistics, 2006e; Australian Institute of Health and

Welfare, 2008). In addition, the prevalence of Type 2 diabetes keeps increasing with

an annual increase rate of 0.8% (Barr et al., 2005). Studies consistently show that the

incidence of Type 2 diabetes is strongly related with unhealthy lifestyle factors.

Physical inactivity is found to increase the risk of developing Type 2 diabetes by

86% among black women (Krishnan, Rosenberg, & Palmer, 2009), while people

who perform regular physical activity have 30% less risk of having this chronic

condition (Jeon, Lokken, Hu, & Van Dam, 2007). In terms of weight status, the risk

of having Type 2 diabetes is over four times for people with obesity as that of those

with normal weight (Cameron et al., 2009; Guh et al., 2009). Smoking is another

strong risk factor, which presents a positive dose-response relationship with the risk

4

of having Type 2 diabetes (Patja et al., 2005). It is revealed that the increased risk

ranges from 23% for past smokers to 61% for heavy current smokers (> 20 cigarettes

per day) (Willi, Bodenmann, Ghali, Faris, & Cornuz, 2007). Healthy lifestyle is not

only related with incidence of chronic disease, but also associated with better quality

of life, less disability (Motl & McAuley, 2010), and lower mortality (Capewell et al.,

2009). Lifestyle modification is; therefore, ultimately important for both general and

clinical populations.

In the past two decades, many studies have been carried out to modify

people’s lifestyles and improve their health related quality of life, yet, the results of

the efficacy of interventions (normally including dietary advice and instruction about

physical activity) on reducing multiple lifestyle risks are mixed (Ebrahim, Beswick,

Burke, & Davey Smith, 2009; Orozco et al., 2008). At the population level, the

increasing prevalence of overweight or obese people continues showing no sign of

reduction. As reported by the Australian Bureau of Statistics, the percentage of

people who are overweight or obese has actually risen by 7% from 57.9% in 1995 to

62% between 2007-2008, and studies also show that the trend of increasing

proportion of overweight and obese people trend is significantly driven by a rising

proportion of obesity, which poses a greater risk of chronic diseases (Australian

Bureau of Statistics, 2009). In terms of physical activity, the proportion of people

with a sedentary or low exercise level climbed up from 69.4% in 2001 to 72.8%

(Australian Bureau of Statistics, 2009). In addition, number of deaths with diabetes

as an underlying cause has doubled from 1984 to 2004 (Australian Bureau of

Statistics, 2006e; Barr et al., 2005). These data clearly show that unhealthy lifestyle

factors not only increase the incidence of Type 2 diabetes, but also the mortality of

people with this disease (Al-Delaimy, Willett, Manson, Speizer, & Hu, 2001; Hu et

5

al., 2005; Tanasescu, Leitzmann, Rimm, & Hu, 2003). Therefore, concerted effort is

required to continue exploring the most effective strategies in improving unhealthy

lifestyle for both people who have and have not developed Type 2 diabetes.

The existing research has demonstrated that the process of lifestyle changing

involves complicated psychological adjustments, and a balance between logical

decision making and emotional fluctuation (Prochaska & Velicer, 1997; Prochaska et

al., 1994). In addition, a proportion of people living with an unhealthy lifestyle may

have already suffered from psychological problems before behavioural change under

the pressure of societal perceptions (Wott & Carels, 2010). Therefore, understanding

the psychological features of people living with an unhealthy lifestyle may be critical

in terms of providing a psychological perspective to lifestyle intervention, especially

for those having the greatest challenge of changing (e.g. the morbidly obese).

The literature has included discussions about the psychological features of

people who live with an unhealthy lifestyle. Studies examining the correlations

between lifestyle risk factors and common mental health problems, including

depression and anxiety, provided valuable information in this regard. For example,

some studies found that women who are overweight or obese tend to have a higher

risk of depression or depressive symptoms (Eunkyung, 2009; Heo, Pietrobelli,

Fontaine, Sirey, & Faith, 2006; Scott et al., 2008). Physically active women have

fewer mental health problems than those who do not perform physical activity, and

physical activity is also used as a conjunctive therapy for depression (De Moor,

Beem, Stubbe, Boomsma, & De Geus, 2006; Galper, Trivedi, Barlow, Dunn, &

Kampert, 2006). Furthermore, smoking has long been connected to having more

depressive and anxiety symptoms within women (Khaled, Bulloch, Exner, & Patten,

2009), even after controlling for alcohol consumption (Massak & Graham, 2008).

6

Apart from these, patterns of alcohol consumption and dietary habits have also been

researched, although inconsistent results were yielded (Sanchez-Villegas, Henriquez,

Bes-Rastrollo, & Doreste, 2006).

The majority of the studies have the advantages of using a large and

representative study sample with a broad age scope from 18 and beyond, which

provides more confidence in the ability to generalise research findings. But on the

other hand, the epidemiology of the relationships between lifestyle factors and

mental health for different age groups cannot be extracted due to a lack of reporting

on age-stratified results. Understanding age specific characteristics is essential in

developing tailored lifestyle intervention programs for different age groups, as the

prevalence rates of both lifestyle risk factors and mental health varies with age

(Australian Bureau of Statistics, 2007, 2009). What is more, the socioeconomic

context where people live also differs largely among age groups, and has the

potential for affecting their choice of lifestyles. For example, young adults are often

confronted with stress from the pressures of career development, child raising, and

financial strain, while adults from middle age and beyond tend to have more spare

time for themselves, a stable income, and are free from the obligation of looking

after children.

Among women with Type 2 diabetes, an elevated risk of depression has been

demonstrated, suggesting that females with Type 2 diabetes have more depression

symptoms as compared to those without. Yet, many studies failed to control for

factors that are associated with depression as well, for example, one’s physical

condition and unhealthy lifestyle factors (Ali, Stone, Peters, Davies, & Khunti, 2006).

Therefore, whether the elevated risk of depression is contributed to by diabetes or

other potential factors remains unknown. If lifestyle factors are the factors that

7

contribute to the disparities, there is a large potential to improve mental health for

people with diabetes.

Furthermore, the literature in this area also reveals a shortage of studies on

anxiety and lifestyle risk factors, with a primary focus on the dimension of

depression. Although the adverse consequences of depression for multiple aspects of

health are well demonstrated, anxiety as the most common mental health disorder

(Kessler et al., 2005) was often overlooked by many of the previous researchers.

More importantly, results of some research have indicated a high likelihood of

depression and anxiety co-morbidity (Australian Bureau of Statistics, 1997;

Wolitzky-Taylor, Castriotta, Lenze, Stanley, & Craske, 2010), which suggests an

integrated research approach aimed at developing a better understanding of the

correlations between mental health and lifestyle risk factors. A French study (Bonnet

et al., 2005) included various risk factors and both anxiety and depression. In

exploring the relationships between lifestyle risk factors and mental health, the

authors created an unhealthy lifestyle index which was used to indicate to what

extent individuals’ habits deviated from the national recommendation. However, the

prerequisite of this approach is that the effects of lifestyle risk factors were equally

weighted against mental health, whereas this fact has not been proved.

Middle life is a significant time when women experience significant changes

in both biological and social perspectives (Ballard, Kuh, & Wadsworth, 2001), and it

is perceived as a great opportunity for the clinicians and researchers to promote a

healthy lifestyle, prevent occurrence of chronic diseases and improve the quality of

life. This research is dedicated to identifying the psychological aspects correlated

with lifestyle risk factors among different groups of midlife and older women in

Australia.

8

Definitions and Terms

Dictionaries define lifestyle as ‘the habit, attitudes, tastes, moral standards,

economic level, etc., that together constitute the mode of living of an individual or

group’. As seen from the definition, lifestyle is a very broad term in the sense of

covering multiple aspects of life, with health related lifestyle being one of the many.

While health related lifestyle seems to have a narrower meaning, health related

lifestyle can mean anything ranging from preventative behaviour (e.g. screening tests)

to daily activities affecting health such as diet and physical activity, depending on

the research context. In the area of chronic disease prevention and management,

several specific aspects of unhealthy lifestyles including weight status, physical

activity, smoking, alcohol drinking, and diet have received considerable attention

from researchers. The emergence of research is primarily driven by the strong

correlations between the above stated unhealthy lifestyle factors and chronic disease

development. Furthermore, these lifestyles are adopted, not preordained, which

implies the potential for them to be modified, thus improving the health outcomes

subsequently.

Different terms were used in the literature to refer to these five factors in

general for the purpose of conciseness. For example, lifestyle risk factor has been

used in the Lifescript Initiative by the Department of Health and Aging in Australia

(Department of Health and Ageing, 2008). In the Lifescript resources kit, lifestyle

risk factors included smoking, poor nutrition, alcohol misuse, physical inactivity,

and unhealthy weight. Research using lifestyle risk factors including the Lifescript

Initiative have a focus on the negative aspects of unhealthy lifestyles. For example,

overweight or obesity and physical inactivity were the areas of interest of the

research. Modifiable risk factor has also been a common term to emerge in the

9

literature: smoking, obesity and physical inactivity were combined with other risk

factors, and generally termed as modifiable risk factors (Yusuf et al., 2004). The

word modifiable is included to underline their potential for change, which is absent

for other chronic disease risk factors such as age and family history. Furthermore,

the use of lifestyle behaviour is not uncommon as well. Different from lifestyle risk

factors and modifiable risk factors, the term lifestyle behaviour means action and the

way of doing something. Smoking and alcohol drinking are behaviours, but being

overweight is a status, while losing weight is behaviour.

In the current research, the term lifestyle factor is chosen as the overall term

for the investigated variables, which included weight status, physical activity,

smoking, and alcohol drinking. The word ‘risk’ is not included because the purpose

of the current research is not only to examine the effects of an unhealthy lifestyle on

mental health, but also the protective effect of a healthy lifestyle against poor mental

health. In addition, factor is favoured to behaviour, as stated in the previous

paragraph, being overweight or obese is not an action, so it is not appropriate to refer

to it as behaviour.

In spite of the common use of ‘mental health’ in research, practice and

political areas, articulating the meaning of mental health in this research study is

somewhat challenging. It is helpful to review and compare different meanings of

mental health among previous studies before giving the definition for this research.

One definition of mental health (Gaylord, Gruener, Rodgers, & Zalice, 2008, p. 4) is

“an ability to see oneself as others do and to fit into the culture and society where

one lives”. The WHO (2005a) illustrated mental health as “a state of well-being in

which every individual realises his or her own potential, can cope with the normal

stresses of life, can work productively and fruitfully, and is able to make a

10

contribution to her or his community”. Although described somehow differently,

both definitions emphasise mental health as a concept of positive mental status,

which is characterised not only by the absence of mental illness, but also a presence

of psychological well-being. In contrast to mental health, mental illness is described

as “an inability to see oneself as others do and not having the ability to conform to

the norms of the culture and society”. Mental health and mental illness are viewed as

end points on a continuum, with movement back and forth through life (Gaylord, et

al., 2008, p.4).

However, within the literature, it is not uncommon to see that mental health

is used as a general term for mental status, without an indication of being negative or

positive. For example, there is a recent publication titled “Collective resources or

local social inequalities? The social determinants of mental health in rural areas”

(Riva, Bambra, Curtis, & Gauvin, 2010). Although the authors used the term, mental

health, in the title, what was investigated was common mental disorders. So, under

this situation mental health does not possess the meaning of an optimal status of

mental health or, in other words, well-being. Rather, it is a neutral term which is

used to describe mental status in contrast to physical health. Similarly, the Australian

Survey of Mental Health and Well-being, which was undertaken in 2007, had a

strong focus on the prevalence of mental disorders within adult Australians and

related factors (Australian Bureau of Statistics, 2007). While investigating these

mental disorders, some positive aspects of mental status were also examined in this

national survey. This is another example of using mental health as an unbiased term

to represent general mental status, which could convey either negative or positive

meanings, or sometimes both.

11

For the current research, the concept of mental health as a neutral term was

applied. This is because the main purpose of the research was to examine the

relationships between lifestyles and two common mental health problems, anxiety

and depression. Also, the overall mental health and its correlation with lifestyle

factors were investigated. In brief, mental health in this research does not mean

healthy mental status, but it is a general term describing mental status.

Aims of the Research

The research aimed to develop a comprehensive understanding about the

correlations between lifestyle risk factors and mental health among Australian

women in midlife and older. This broad aim was achieved by three individual studies

designed to describe the relationships, with different research approaches in both

general and clinical populations. Subsequently, the prospective relationships between

lifestyle risk factors and mental health in midlife women from the general population

was investigated first; then the differences of mental health between general and

diabetic populations was examined, together with the contribution of lifestyle factors

to mental health; and lastly, the mediating role of self-efficacy in the correlations

between lifestyle factors and mental health among midlife and older women with

diabetes was evaluated.

Research Plan

The relationships between lifestyle risk factors and mental health among

midlife and older women were investigated in both apparently healthy and clinical

populations using a number of research designs. Study 1 utilised a large sample of

midlife women from the community, and analysed the ability of baseline lifestyle

risk factors in predicting mental health status after five years of follow up. Study 2

12

was of cross-sectional design. It compared the correlation in women attending

diabetic clinics with those from the community, and particularly examined the effect

of diabetes on women’s mental health status. Study 3 expanded the research on the

basis of the previous two. This study examined the mediating effect of self-efficacy

in the relationships between lifestyle factors and mental health among midlife and

older women with diabetes by using a questionnaire survey. The varied research

designs with different study samples allow the understanding of the multiple aspects

of the association between lifestyle risk factors and mental health among midlife and

older women.

Significance of the Research

The research study is one of the few studies with a focus on health issues for

midlife and older women in Australia. Understanding the relationships between

lifestyle factors and mental health among midlife and older women and their

underlying mechanism is the central goal of the research. The research is undertaken

based on the social cognitive theory (Bandura, 1997, 2004), which not only allows

the examination of behavioural factors on personal factors, but also conveys the

implications of how to improve mental health. Therefore, this research has the

potential to enhance researchers’ and health professionals’ knowledge of individuals’

mental health in relation to lifestyle factors, and more importantly, has a strong

implication for the development of theoretically driven design.

Structure of the Thesis

This thesis has eight chapters. Chapter 1 is introduction of the research,

which lays out the background of the study, clarifies the definitions and terms, and

states the significance of the study. Chapter 2 is the literature review. This chapter

13

has a dual focus, one is the examination of theories on behaviour change in health

promotion, and the other is an overview of the studies on the correlations between

lifestyle factors and mental health.

Chapter 3 describes the methodology used in the research study. The data of

the current study is based on previous women’s health studies, so, the linkage of the

current study with these women’s health studies is explained. The current research

involves three studies: Study 1, Study 2 and Study 3. The design, sample and

measures of each study are outlined.

Chapters 4 to 6 present the results of Study 1, Study 2 and Study 3,

respectively. At the end of each study, a brief reflection of the study results as related

to the research questions is provided.

The discussion about the study results and further thoughts on study strengths

and limitations are presented Chapter 7. Finally, the conclusions of the research as

well as the implications generated from the study are described in Chapter 8.

Chapter Summary

This chapter introduced the broad background of this research, and

highlighted the importance of research in the area of chronic disease prevention and

lifestyle improvement. The ageing population combined with the increasing rate of

chronic diseases demand more research on this area. The changing of lifestyles (e.g.

physical inactivity) serves as a strong driver of the rising prevalence of chronic

disease, thus, it needs to be investigated. A better understanding of the relationships

between mental health and lifestyle factors may have the potential to contribute to

the development of effective strategies in solving these problems.

The aims of the research were briefly outlined (see Chapter 2 for detailed

research questions), followed with research plan. The key definitions and terms were

14

explained and clarified, which included listing of the four lifestyle factors examined

in the current study as well as the meaning of mental health. Finally, the structure of

the thesis was described. Chapter 2 begins with an overview of women and mental

health, continues with the examination of behavioural theories, and finishes with a

thorough literature review on lifestyle factors and mental health.

15

CHAPTER 2: LITERATURE REVIEW

Brief Introduction

This chapter has two aims: to examine the common theories in health

promotion, and to conduct a comprehensive review of the literature on lifestyle

factors and mental health. A brief description on women and mental health precedes

the review of theories. Common behavioural theories in the field of health promotion

were compared including the rationale for choosing social cognitive theory as a basis

for this research. Following the examination of theories, a constructive review about

the studies on the relationships between lifestyle factors and mental health was

presented to provide the context of the current research and to identify research gaps

in this particular area. For each lifestyle factor, the relevant literature on its

relationship with mental health was presented, and the modifying effect of gender on

the relationships was highlighted along the way.

Women and Mental Health

Gender appears to play a significant role when it comes to the issue of mental

health. National surveys in the U.S. and Australia show a higher prevalence of

affective and anxiety disorders in women, but lower substance use disorders as

compared to men (Australian Bureau of Statistics, 2007; Hasin, Goodwin, Stinson, &

Grant, 2005; Somers, Goldner, Waraich, & Hsu, 2006). Research from North

America revealed that among adults aged 18 years or over, the past-year and lifetime

prevalence rates of major depression were 5.28% and 13.23%, respectively between

2001 and 2002, and women were twice as likely as men to have major depression

(Hasin, et al., 2005). As revealed by the most recent Australian National Survey of

Mental Health and Well-being (Australian Bureau of Statistics, 2007), women

16

experienced higher rates of any 12-month mental disorder (anxiety disorder,

affective disorder & substance use disorder) than men (22% vs. 18%). Specifically,

women also experienced a higher rate of anxiety disorders (18% vs. 11%) and

affective disorder (7.1% vs. 5.3%) than men, but not substance use disorders.

The gender disparities in risk for mental illnesses was thought to stem from

both biological (Deecher, Andree, Sloan, & Schechter, 2008) and psychosocial

differences between men and women (Klose & Jacobi, 2004; Zender & Olshansky,

2009). Women have a different hormone profile from men and respond differently to

stressful activity. Moreover, the varying sexual hormone levels and hormone

secretion patterns across women’s lifespan also contribute to women's vulnerability

to mental disorders. Endocrine change occurs during the major reproductive periods,

which include premenstrual, postpartum, and perimenopausal periods (Zender &

Olshansky, 2009).

Anxiety has been commonly reported among women experiencing

menopausal transition. 24% of women in early menopausal transition reported

having anxiety symptoms, as compared to 19% among women in premenopause

(Freeman, Sammel, Lin, Gracia, & Kapoor, 2008). In the Study of Women's Health

Across the Nation (U.S.A.), 52% of women presented anxiety symptoms described

as “feeling tense” (Avis et al., 2001). The risk factors of having anxiety during the

menopausal stage were identified as being associated with premenstrual syndrome,

history of depression, higher perceived stress (Maki, 2008) and sleep disturbance

(Parry, 2007).

The issue of depression in relation to menopause seems rather controversial.

The Harvard Study of Moods and Cycles followed 420 non-depressed women aged

from 36 to 45 years for 6 years (Cohen, Soares, Vitonis, Otto, & Harlow, 2006).

17

Using telephone interviews and questionnaires, the authors found that women

entering perimenopause were nearly twice as likely to develop clinical depression as

compared to those remaining in premenopause, regardless of age and negative life

events. Likewise, in the study by Freeman and his colleagues, an increased risk of

depressive symptoms during the transition to menopause was detected after

controlling for important confounders including age, history of depression, sleep

quality and sociodemographic factors (Freeman et al., 2004). This study additionally

pointed out that after menopause, the risk of depression diminished.

However, studies suggesting a non significant relationship between

depression and menopausal status are not uncommon (Gallicchio, Schilling, Miller,

Zacur, & Flaws, 2007; Kaufert, Gilbert, & Tate, 2008; Lu, Tseng, Lin, Luh, & Shu,

2009; Smith-DiJulio, Woods, & Mitchell, 2008). These studies commonly indicated

that depression in menopause is more likely to be contributed to by health habits (e.g.

smoking and physical inactivity) and the number of menopausal symptoms

(Gallicchio et al., 2007; Lu et al., 2009), negative life events (Lu et al., 2009;

Smith-DiJulio et al., 2008) and health status including the number of chronic

diseases (Kaufert et al., 2008). Nevertheless, the literature on menopause and mental

health problems suggested that many women do not develop mental health problems

at this particular stage, but there may be a subset of women who are at an elevated

risk of depression and anxiety. The association between menopausal status and

mental health remains unclear.

18

Theoretical Background

The WHO constitution (1948) proposed that “health is a state of complete

physical, mental and social well-being and not merely the absence of disease or

infirmity”. As described, health has multiple dimensions and these are inseparable

and interlinked components of the totality when describing health. Based on the

WHO’s definition of health and some other social theories, Pelican demonstrated

that an individual person must be understood as the structural and operational

coupling of three different kinds of systems (Pelican, 2007), which are the body or

organism, the mind or mental system, and the social status of a person (Figure 2.1).

The interaction between these three systems is illustrated, suggesting that the

organism and social status of the person can have an impact on the mind, but equally

the mind can also influence the organism and social status of the person. Therefore,

exploring the effect of lifestyle factors on individuals’ mental health can also be

essential in terms of developing a comprehensive understanding of lifestyle factors

on health as a whole.

Figure 2.1. The individual as a structural coupling of three systems: organism, mind

and social status. Information source: Pelican, J. M. (2007). Understanding

differentiation of health in late modernity by use of sociological systems theory. In D.

V. McQueen, I. Kickbusch & L. Potvin, Health and modernity: the role of theory in

health promotion. New York: Springer.

This figure is not available online.Please consult the hardcopy thesisavailable from the QUT Library.

19

At the time when chronic diseases have increasingly become the challenge of

health promotion, the role of lifestyle factors in the development of various chronic

diseases and some kinds of cancers has attracted considerable attention from

researchers and clinicians. While the understanding of the relationships between

lifestyle factors and chronic diseases is well established and most people know that

maintaining a healthy lifestyle helps them to live longer and in better health, the

proportion of people who are committed to habit or behaviour change might be

lower than anticipated. Among a group of well-functioning, community dwelling

elderly people (N = 2,708), 67% had an indication of the need to lose weight, yet

only 27% participants reported an intention to do so (Lee et al., 2004). In 2004-2005,

70% of Australians aged 15 years and over were still classified as sedentary or

having low exercise levels (Australian Bureau of Statistics, 2006g), and this number

has not changed in the past ten years. These data probably convey the information

that behaviour change is not a simple and linear process. It involves social and

psychological factors which can be critical in designing effective lifestyle

intervention. However, when compared to the volume of studies examining the

effect of lifestyle factors on people’s physical condition, the research in regards to

the correlations between lifestyle factors and mental health is relatively limited.

Investigating the associations between lifestyle factors and mental health provides

health professionals and researchers with information that may contribute to

effective behaviour modification programs in the future.

Based on the model, it is thought that one’s mental health can be influenced

by lifestyle factors in two potential pathways. Firstly, lifestyle factors may alter

mental health via the latter’s influence on physical health. For example, while doing

exercise, the human brain is stimulated to release endorphin, a natural pain killer,

20

and thus boosts a sense of well-being and pleasance (Bender et al., 2007). Secondly,

an individual’s lifestyle in relation to health-promoting factors is socially embedded.

Making a lifestyle choice is the process that frequently engages self-comparison with

other people. As demonstrated by the theory of planned behaviour (Ajzen, 1991), the

motivation of meeting the “social norm” of a particular behaviour is one of the

important factors in predicting behaviour. By doing so, a sense of belonging is

fulfilled and thus a better status of mental health is created, as the WHO’s definition

of health suggests. In the current society where slimness is highly regarded, being

overweight or obese deviates from the social norm. Research has shown that people

with obesity are stigmatised and suffer from enormous social and psychological

stress (Carr & Friedman, 2005; Rogge, Greenwald, & Golden, 2004).

In all, health is a comprehensive concept. In health promotion, it is essential

that people not only research the effect of lifestyle factors on physical health but also

investigate the mental health sphere. Considering the well-established effect of

lifestyle factors on chronic conditions and the fact that the high prevalence of

unhealthy lifestyles is continuing, it may be worthwhile to examine mental health

status in relation to lifestyle factors. A number of common theories in the field of

health promotion were introduced and compared to select the most suitable theory as

the conceptual framework for this research. It is expected that the knowledge

generated from this study would enhance people’s understanding of the effect of

lifestyle factors on individuals’ mental health. In addition, it is also anticipated that a

model be produced upon which more effective lifestyle intervention could be

developed.

21

Conceptual and Theoretical Framework

Health Promotion

Health promotion is initially defined by the WHO as “the process of enabling

people to increase control over the determinants of health and thereby improve their

health” (World Health Organisation Regional Office for Europe, 1986). Confronted

with the increasing prevalence of chronic and non-communicable diseases in most

developed countries and more recently in some developing countries, modification

of individuals’ unhealthy lifestyle factors that are related to premature mortality, and

total burden of disease has drawn enormous attention since the 1970s. It is reported

by the WHO (2005b) that 80% of heart disease, stroke, and Type 2 diabetes can be

prevented by appropriate intervention on aetiological factors like unbalanced diet

and physical inactivity. Hence, modifying lifestyle factors is considered as the most

efficient way of preventing disease morbidity and premature mortality. A number of

health promotion programs have been conducted within a broad spectrum, such as

smoking cessation, cancer screening, and lifestyle change. From a global perspective,

the WHO (2005b) developed a model of Global Strategy on Diet, Physical Activity

and Health (DPAS) aiming to guide efforts across the world in the field of chronic

disease prevention (Figure 2.2).

22

Figure 2.2. Schematic model of the Global Strategy on Diet, Physical Activity and

Health (WHO, 2006b).

This model highlights the role of social and environmental factors in the

effectiveness of health promotion, and specifies that supportive environment,

policies and programs should be developed to facilitate behaviour change at the level

of the individual. In spite of the emphasis on external factors that influence the

effectiveness of health promotion, a person’s behaviour change remains the critical

component in this model. This change is affected by external factors and further

affects a variety of outcomes in social, economic and heath domains. Therefore, the

understanding of the factors that influence an individual’s behaviour cannot be

overemphasised.

A substantial amount of work has been undertaken to understand the

phenomenon of behaviour change by the development and application of theoretical

models in order to achieve maximal health gain (Lorig, Doak, Doak, & Giloth, 2001).

These theories can be classified into individual, interpersonal and community

23

categories in terms of their operating levels. Theories operating at the individual

level include, but are not limited to the health belief model, the theory of planned

behaviour, and the transtheoretical model. Those incorporating social and

environmental factors as well as the personal ones include the social cognitive theory

(Bandura, 2004). In this section, a close examination of commonly applied

behavioural theories was conducted in combination with the research questions of

this research. This process allowed the emergence of an appropriate theoretical

framework for the current research study.

The Health Belief Model

The health belief model originated in the 1950s and is the oldest model in

health promotion. It was developed by a group of U.S. social psychologists who

attempted to work out the factors that motivated people to attend a free tuberculosis

screening program. In the early formulation of this theory, the core concept was that

individuals’ health behaviour was motivated by four factors, which are: 1) perceived

susceptibility, the degree to which a person feels at risk for a health problem; 2)

perceived severity, the degree to which a person believes the consequence of the

health problem will be severe; 3) perceived benefits of an action, the positive

outcomes a person believes will result from the action; and 4) perceived barriers to

taking that action, the negative outcomes that a person believes will result from the

action (Rosenstock, 1974). As more research was carried out and the utility of the

health belief model was increasingly examined, two new components were added

into the original framework to enhance its ability to predict behaviours (Cummings,

Jette, & Rosenstock, 1978). The latterly introduced two components are: 1) cues to

action, an external or internal event that motivates a person to act (e.g. a

consultation), and 2) self-efficacy, a person’s belief in his or her ability to take action.

24

It was considered at the time that a person needs something external or internal to

make them start thinking about behaviour change. If such an event is absent, the

reasoning and balancing process may not be executed. Furthermore, a person’s

confidence in actually being able to carry out the task is also valued, because even

when a person favours a healthy behaviour and is fully aware of the risk of an

unhealthy behaviour, she/he may not perform this action due to low confidence in his

or her ability to do so. As shown in Figure 2.3, the health belief model assumes that

individuals’ behaviour can be changed if their health beliefs are modified

accordingly by the provision of necessary information. Despite its long history, the

application of this theory has not always been successful (Medina-Shepherd, 2008).

This model well identifies the factors affecting an individual’s internal process of

decision making; however, it does not take much consideration of external factors

(e.g. other beliefs), which can equally influence people’s behaviours. In addition,

research being conducted recently has shown that belief formation does not always

precede behaviour modification. In some situations, they are actually formed after a

behavioural change (DiClemente, Crosby, & Kegler, 2009).

25

Figure 2.3. The Health Belief Model in Glanz K., Lewis F.M., & Rimer B.K., (Eds.).

Health Behavior and Health Education: Theory, Research and Practice. San

Francisco: Jossey-Bass.

The Theory of Planned Behaviour

The theory of planned behaviour is an extension of reasoned action theory,

which was developed by Fishbein and Ajzen in the early 1980s (Ajzen & Fishbein,

1980; Fishbein & Ajzen, 1975). This theory has been applied not only in health

promotion, but has also been used widely in other social activities such as

advertising, parenting and so forth. The theory of planned behaviour assumes that a

person’s intention to perform certain behaviour is the immediate factor that can

predict that action mostly accurately. Therefore, the theory of planned behaviour has

a strong focus on the predictors that form the intention. According to theory of

planned behaviour, an individual’s intention is a function of two factors: 1) a

person’s attitude towards a specific behaviour; and 2) his/her perception of the

subjective norms associated with that behaviour. A person’s attitude to a specific

behaviour is the degree to which performance of the behaviour is positively or

negatively valued. Subjective norm is the perceived social pressure to engage or not

to engage in a behaviour. Based on this model, an individual’s intention of carrying


26

out a specific behaviour could be well explained by one’s attitudes and subjective

norms related with that behaviour, if there are no other barriers that are likely to stop

intention from transferring to behaviour.

But in the real world, it is common that one’s intention to undertake certain

action is interrupted by unexpected factors that are not under one’s control.

Therefore, the model with two elements of attitude and subjective norm may not be

predictive to behaviours in some situations. To address this problem, a new element,

perceived behavioural control, was added to the theory of planned behaviour under

the influence of self-efficacy theory (Ajzen, 1991). Perceived behavioural control

refers to people's perceptions of their ability to perform a given behaviour. In brief,

the theory of planned behaviour explains human behaviour by investigating the

factors that are likely to determine the intention to conduct that behaviour. It is

hypothesised that behavioural intention is formed under the interactive influence of a

person’s attitude, subjective norms, and perceived behavioural control (Figure 2.4).

One criticism that theory of planned behaviour often receives is that behavioural

intention does not necessarily lead to behaviour change, as something unexpected

could occur ahead of subsequent behaviour and alter the intention. Furthermore, the

theory of planned behaviour assumes the process of decision making is cognitive and

rational, which overlooks the effect of emotion. The systematic review by Hardeman

et al. found that among interventions developed by using the theory of planned

behaviour, in half of them change in intention was reported and in two thirds change

in behaviour was reported (Hardeman et al., 2002).The effect sizes were general

small, which may be attributable to the limitations.

27


Figure 2.4. Conceptual framework of theory of planned behaviour. Source: Ajzen, I.

(1991). The Theory of Planned Behaviour Organizational Behaviour and Human

Decision Processes, 50(2), 179-211.

The Transtheoretical Model

The transtheoretical model was initially developed in the early 1980s,

primarily applied in the area of smoking cessation. The most distinctive feature in

which the transtheoretical model differs from other theories (e.g. health belief model

and the theory of planned behaviour) is its emphasis on behaviour change as a

dynamic process of change involving multiple stages, rather than an event at one

time point (Prochaska & Velicer, 1997; Prochaska et al., 1994). The key concept of

this model is that people complete behaviour change by several stages, where

different psychological processes are presented. This argument suggests that not

everyone is ready to make a behaviour change, thus, interventions must take the

stages into account to be able to address people’s needs to the largest extent possible.

The stage of change does not constitute the transtheoretical model alone, along with

it decisional balance, process of change and self-efficacy also serve as crucial

components in this model, co-occurring dynamically at different stages of change

(Table 2.1).

28

Table 2.1

The Transtheoretical Model Constructs

The proposed stages of change are: pre-contemplation, contemplation,

preparation, action, maintenance and termination. Moving from one stage to the next

involves different processes of change. In brief, people in the first stage,

pre-contemplation, are not yet at the point that they think of a specific behaviour

change, which suggests they have not considered change or they think change is not

necessary (Resnicow, McCarty, & Baranowski, 2003). When people move from

pre-contemplation to contemplation, they begin to think about assuming a behaviour

change. This could be triggered by any number of cues, such as conversation with

friends or the TV news. Once people make the decision to change, the stage of

preparation follows. People normally do lots of preparatory work during this short

period to get them ready for real action. After preparation is completed and people

begin to perform certain actions (e.g. exercise), they are considered to be at the stage

of action. At this time, people work actively to modify the problem they think needs

to be addressed. In the maintenance stage, people have made considerable change to

This table is not available online.Please consult the hardcopy thesisavailable from the QUT Library.

29

a specific behaviour and struggle to keep that behaviour going and to prevent

regression. This is a process which often requires substantial hard work, as it is very

likely that they will encounter negative factors that will stop them from carrying out

that action. In addition, this is also the stage where self-efficacy becomes largely

important in determining the outcomes of maintenance. If a person has high

self-efficacy, he/she is more likely to persist in behaviour change; otherwise he/she

is at high risk of regressing to early stages. But meanwhile, self-efficacy is also

influenced by many other factors. It could be a preceding factor of behaviour change,

but also a result of concrete behaviour change. Nevertheless, successful completion

of the maintenance stage leads people to the final stage of contemplation. At this

point, people are regarded to have finished the whole process of behaviour change

and command a high level of self-efficacy.

As mentioned in the beginning of this section, the transtheoretical model

recognises people could be at different stages of change, thus research intervention

should target those people’s needs in a manner that matches the stages they are at.

Research has reported the efficacy of transtheoretical model based programs in

enhancing participants’ adherence to treatment regimens (Johnson et al., 2006),

however, accurately differentiating each stage has often been a difficult issue for

many researchers (Plotnikoff et al., 2009).

The Social Cognitive Theory

The social cognitive theory was developed by Albert Bandura in the 1970s,

and has produced a significant effect in the discipline of psychology. As the name

suggests, the social cognitive theory proposes that the cognitive process is the

mechanism of learning. Before that, the view of behaviourism was pervasive at that

time. Many behaviourists believed that a new behaviour was shaped by its

30

performance effects (stimuli), which could either be reinforcement or punishment. In

contrast to the mainstream of behavioural psychology, Bandura presented a view in

his early publication that changes were achieved by different methods derived from a

common cognitive mechanism (Bandura, 1977). Meanwhile, he commented that a

person’s cognitive ability is not only a product of the environment, but also the

producer of the environment. Later in 1986, he further emphasised the fact that

human functioning is explained in terms of a model of triadic reciprocality in which

behaviour, cognition, and other personal and environmental factors operate as

interacting determinants of each other (Bandura, 1986).

This triadic reciprocal determinism is an essential feature in which the social

cognitive theory differs from many other theories described above. Those theories

attempted to explain human behaviour as an outcome of a number of personal or

environmental factors, which depicts a unidirectional relationship from personal

and/or environmental factors to behaviour. This approach denies a person’s proactive

capacity in influencing environment and other personal factors such as education. As

shown in Figure 2.5, the reciprocal determinism acknowledges a constant interplay

between these three dimensions, yet, reciprocality does not mean symmetry in the

strength of bidirectional influences, nor is the patterning and strengths of mutual

influences fixed in reciprocal causations.

31

Figure 2.5. Social cognitive theory, Bandura, A. (1977). Self-efficacy: toward a

unifying theory of behavioural change. Psychology Reviews, 84(2), 191-215.

The social cognitive theory also specifies a core set of determinants, which

includes knowledge of health risks and benefits of different health practices,

perceived self-efficacy that one can exercise control over one’s health habits,

outcome expectations about the expected costs and benefits for different health

habits, the health goals people set for themselves and the concrete plans and

strategies for realising them, and the perceived facilitators and social and structural

impediments to the changes they seek (Bandura, 2004). As seen from Figure 2.6, a

person’s outcome expectation has a direct impact on behaviour. If people think

changing behaviour will not produce much benefit for them, then the possibility of

adopting a new pattern of behaviour is low. Goals setting procedure also play an

important role. By self-evaluating one’s behaviour against some standards, people

will be motivated to make the next step, thus moving towards completion of

behaviour change. In this model of causal structure, self-efficacy is at the central

position, because it not only influences behaviour directly, but also shapes other

factors that affect behaviour.

This figure is not avaible online.Please consult the hardcopy thesisavailable from the QUT Library.

32

Figure 2.6. Structural paths of self-efficacy theory (Bandura, 2004).

With regard to the concept of self-efficacy itself, it is recognised that

self-efficacy is a multifaceted conception which contains three dimensions including

magnitude, generality and strength (Bandura, 1997). Magnitude means the level of

task difficulty that a person believes that they are capable of completing. Generality

refers to the extent that self-efficacy expectancy can be generalised into another

similar area. Strength reflects the degree of confidence with which one can perform

the tasks. Despite its property of generality, self-efficacy under many situations is a

situation specific variable. A high sense of self-efficacy in one domain is not

necessarily accompanied by high self-efficacy in another realm (Perkins & Jenkins,

1998). For example, a person who has a high level of self-efficacy in refraining from

alcohol misuse will not necessarily be equally confident in smoking cessation.

Therefore, although self-efficacy is regarded as an important factor in influencing

behaviour change, a global estimate of self-efficacy is likely to sacrifice its power of

predicting behaviour.

In his later publication (Bandura, 1986), it was proposed that self-knowledge

about one’s self-efficacy relied on four primary sources of information, whether or


33

not one’s perceived self-efficacy was accurate or faulty. These terms regarding the

four sources of information were modified slightly in a later publication (Bandura,

1997). They are enactive mastery experiences, described as the most influential

source of self-efficacy information; vicarious experiences that alter self-efficacy

belief through transmission of competencies and comparison with the attainments of

others; verbal persuasion and allied types of social influences reflecting the view

that one possesses certain capabilities (e.g. consultation with health professionals);

and self-appraisal from which people judge their capability, strength, and

vulnerability to dysfunction.

In 2005, Bandura’s models published in 1997 and 2004 were adapted and

combined into one, which is presented in Figure 2.7. As described in the model,

personal characteristics are related with self-efficacy, which partially influences

outcome expectation and socio-structural factors. Meanwhile, self-efficacy’s direct

effect on the control of behaviour change is also recognised. The course of forming a

new patterning of behaviour is then accomplished via a procedure of gradual goal

setting, which is co-influenced by self-efficacy, outcome expectations and

socio-structural factors. In turn, the result of behaviour change, regardless of success

or failure, contributes significantly to the continuing self-efficacy as an important

information source. Moreover, the effect of behaviour change also spreads to

personal factors and outcome expectations, which produces an effect on behaviour

change in a reciprocal manner.

34


Figure 2.7. Self-efficacy theory - structural paths of influence (adapted from

Bandura, 1977, 2004).

Appraisal of Social Cognitive Theory and Other Psychological Models

The above theories pay considerable attention to the role of psychological

variables in the process of behaviour change. In particular, the concept of

self-efficacy has been constantly thought to be critical to fostering health related

behaviour change (Shortridge-Baggett, 2001). Although named differently, the

concept of perceived control integrated in the theory of planned behaviour overlaps

with self-efficacy, as commented by Bandura himself (2004). Moreover, the

expectation of behavioural change outcome has also been split into seemingly

different determinants across various models. In the health belief model, perceived

benefits toward behaviour change are positive outcome expectations, while perceived

barriers are negative outcome expectations. In the theory of planned behaviour or

reasoned action, one of the key elements that influence behaviour is attitudes, which

are actually formed by a person’s assessment and his/her expectation of the outcome

resulting from behaviour change. Similarly, subjective norms result from society’s

expectations of the individual to perform a specific behaviour and how keen that

person is to comply with that expectation. In this case, both attitude and subjective

norms are outcome expectations. So, attitude refers more to the expectation of a

35

physical outcome, while subjective norms are social outcomes of behaviour

according to Bandura’s social cognitive theory (Bandura, 2004). In all, these

psychological models largely overlap in their essential constructs. Research has also

shown that after considering self-efficacy and self-evaluation, other psychological

factors made trivial contributions in explaining behaviour change (Dzewaltowski,

Noble, & Shaw, 1990).

Based on the reciprocal mechanism between personal, environmental and

behavioural factors, the social cognitive theory not only provides a framework in

explaining behaviour change, but more importantly, it presents guidance for

modifying behaviours in an effective way. The Chronic Disease Self-Management

Program developed by the Stanford Patient Education Research Centre is a

successful example of a peer-led program which has also been applied in many

countries outside the United States. By teaching participants the skills of problem

solving and using modelling strategies, people’s self-efficacy is greatly enhanced,

which permitted them in turn to achieve a satisfactory level of self-management

(Lorig, Ritter, Laurent, & Plant, 2006; Mancuso, Rincon, McCulloch, & Charlson,

2001; Roach et al., 2003). In particular, it was revealed that participants in the

program had a significant reduction of HA1C level (-0.4%), health distress, hypo-

and hyper-glyceamia symptoms and an increase in self-efficacy enhancement (p

<0.05) (Lorig, Ritter, Villa, & Piette, 2008).

Taken all together, social cognitive theory was chosen as the theoretical

model for the current research for two reasons. First, social cognitive theory

recognises the reciprocal relationships between personal, behavioural and

environmental factors. This provides the research study with theoretical support for

the examination of the associations between lifestyle factors (behavioural factors)

36

and mental health (personal factors). This assumption aligns with the DPAS

proposed by the WHO in terms of the awareness of the role that environmental

factors play in facilitating or impeding behaviour change. Second, the key constructs

in other psychological models overlap with the concept of self-efficacy and outcome

expectation in the social cognitive theory. It is hypothesised in the current research

that self-efficacy acts as a mediating factor in the relationships between lifestyle

factors and mental health. If approved, corresponding strategies may be implemented

to promote self-efficacy, with the additional consideration of individuals’ mental

health level.

Multiple Lifestyle Factors, Diabetes and Mental Health

It is well known that unhealthy lifestyle factors including the condition of

being overweight or obese, physical inactivity, smoking and alcohol over

consumption contribute substantially to the burden of disease in Australia (Bauman

& Owen, 1999; Social Research Centre, 2006; Thorburn, 2005). A large scale (N =

16, 043) and cross-sectional study undertaken in Spain pointed out that women with

four lifestyle risk factors were about 3 times more likely to experience

non-compliance of blood pressure and cholesterol assessment, 6 times more likely

not to perform cytology, and 10 times more likely not to have a mammography

screen (Galán et al., 2006). It is known that lifestyle risk factors are not only related

with physical health, but also mental health.

Studies in this field generally illustrate a positive relationship between a

healthy lifestyle and mental health (Bonnet et al., 2005; Rohrer, Rush Pierce, &

Blackburn, 2005). For example in a study conducted in France, individuals’ cigarette

smoking, diet and physical activity habits were contrasted against corresponding

guidelines to generate a score, which indicated the extent of deviation of the

37

individual’s lifestyle from established guidelines. The results showed that people

living with unhealthy lifestyles presented significantly more symptoms of depression

and anxiety (Bonnet et al., 2005). For people living with chronic diseases (e.g. Type

2 diabetes), unhealthy lifestyle factors are even more prevalent. Given the adverse

effect of these factors on the clinical outcomes of people with Type 2 diabetes,

investigating the relationships between lifestyle factors and mental health among this

population is important. One advantage of this study was that it examined lifestyle

factors as a whole group, recognising the correlations among lifestyle factors

themselves. This was regarded to be important in examining the relationship between

lifestyle factors and mental health, as the interactive effect among lifestyle factors

may confound their individual relationships with mental health (Pisinger, Toft,

Aadahl, Glümer, & Jørgensen, 2009).

A growing body of evidence suggests that unhealthy lifestyle factors often

are combined together in one individual, with 60% of the populations having two or

more unhealthy lifestyle factors (Fine, Philogene, Gramling, Coups, & Sinha, 2004;

Poortinga, 2007). In women aged from 18 to 55 years (N = 394), 80% of overweight

women had multiple lifestyle risk behaviours including physical inactivity,

percentage of calories from fat, insufficient daily servings of fruit and vegetables,

and daily sedentary time (Sanchez et al., 2008). Previous studies also showed that

people leading a sedentary lifestyle are twice as likely to be in the obesity category,

however defined (BMI or waist circumference). In addition, a dose-response

relationship between relative weight and physical activity was identified (Stamatakis,

Hirani, & Rennie, 2009). A positive correlation between extra weight gain and

alcohol consumption has again been illustrated in the literature (Wannamethee &

Shaper, 2003), as well as an inverse relationship between physical activity and

38

smoking (Kaczynski, Manske, Mannell, & Grewal, 2008). The co-occurrence of

smoking and alcohol (Degenhardt & Hall, 2003), and less intake of vegetables and

fruit in smokers (McClure et al., 2009) were demonstrated. Concerning the

modifiable effect of one lifestyle factor on the other, Widome et al. (2009) examined

the role of BMI categories in the association between depression and smoking. This

study only found a significant relationship between smoking and depression among

the obesity category, but not other categories.

Obesity and Mental Health

Epidemiology of Obesity

Obesity is a disease in which excess body fat has accumulated to such an

extent that health may be negatively affected (World Health Organisation, 2000b).

There are two common ways of assessing an individual’s relative body weight,

which are the body mass index (BMI) and waist circumference (WC). The former

reflects one’s overall body fat distribution, while the latter is a better indicator of

abdominal fat (Janssen, Katzmarzyk, & Ross, 2002).

BMI is calculated by dividing weight in kilogram (kg) by height in meters

squared. This measure is strongly correlated with total body fat content, yet is not the

direct way of measuring body fat. Therefore, it has the limitation of differentiating

muscle content from fat tissue in special populations such as when one is an athlete

(National Heart Lung and Blood Institutes, 1998). WC or waist-to-hip ratio (WHR)

is a good indicator for abdominal fat intensity. The National Cholesterol Education

Program’s Adult Treatment Panel III Standard (NCEP-ATPIII) defined abdominal

obesity as a WC greater than 88 cm in women and 102 cm in men. It is measured as

the circumference at the middle way between the iliac crest (hip bone) and the costal

margin (lower rib). WC has been identified as an independent risk factor of chronic

http://en.wikipedia.org/wiki/Disease

39

diseases and morbidity, since it measures the visceral adipose tissue, which has been

termed an endocrine organ partly due to its function of secreting adipocytokines and

other vasoactive substances that can influence the risk of developing metabolic traits

(Fox et al., 2007). The sex based classifications of both BMI and WC are listed in

Table 2.2, and this classification is supported by the Australian Heart Foundation as

well.

Table 2.2

Classification of Overweight and Obesity by Body Mass Index, Waist Circumference

and Associated Disease Risk

Disease risk* relative to normal weight and waist

circumference

BMI

(kg/m2)

Obesity

class

Men≤102 cm (≤40 in)

Women≤88 cm (≤35 in)

MenMenMenMen >102>102>102>102 cmcmcmcm (>40(>40(>40(>40 in)in)in)in)

WomenWomenWomenWomen >88>88>88>88 cmcmcmcm (>35(>35(>35(>35 in)in)in)in)

UnderweightUnderweightUnderweightUnderweight <18.5

Normal+Normal+Normal+Normal+ 18.5-24.0

OverweightOverweightOverweightOverweight 25.0-29.9 Increased HighHighHighHigh

ObesityObesityObesityObesity 30.0-34.9 I High VeryVeryVeryVery HighHighHighHigh

35.0-39.9 II Very High VeryVeryVeryVery HighHighHighHigh

ExtremeExtremeExtremeExtreme

ObesityObesityObesityObesity

≥≥≥≥40404040 IIIIIIIIIIII ExtremelyExtremelyExtremelyExtremely HighHighHighHigh ExtremelyExtremelyExtremelyExtremely HighHighHighHigh

* Disease risk for Type 2 diabetes, hypertension and cardiovascular disease

+ Increased waist circumference can also be a marker for increased risk even in persons of normal weight.

The prevalence of obesity and overweight has been climbing substantially in

an increasing number of countries in the past two decades, replacing traditional

health problems such as malnutrition as the new threat to the health of populations.

Like many western countries, Australia encounters the epidemic of obesity as well.

Research showed that in Australia, 58% of women aged from 55 to 64 years are

classified as overweight or obese, this proportion decreased slightly to about 55% in

40

women aged from 65 to 74 years, and dropped further to 44% among women who

are 75 years older (Australian Bureau of Statistics, 2008a). The statistics clearly

shows that midlife women were the heaviest among all of the female age groups

(Australian Bureau of Statistics, 2008a). A substantially increased cost relating to

obesity has been revealed by recent studies (Cai, Lubitz, Flegal, & Pamuk, 2010;

Colagiuri et al., 2010). It was claimed (Colagiuri et al., 2010) that the direct annual

cost (health & non-health) for people who are obese was $2,788 (95%CI,

$2542-$3035), which was nearly twice as much as that for people with normal

weight ($1,472, 95%CI, $1204-$1740). Obese midlife women (defined as 45 years

in the study) are found to incur significantly higher average lifetime Medicare

costs than normal weight midlife women if they survive to 65 years old (Cai et al.,

2010). Moreover, literature has been constantly showing that overweight/obesity is a

risk factor for cardiovascular disease, certain types of cancer, diabetes mellitus and

kidney diseases (Chu et al., 2007; Dhaliwal & Welborn, 2009; Emmanuel & Jatkin,

2007; Jee et al., 2006) and related with high mortality (Cai et al., 2010). The

literature about the effect of obesity on mental health is reviewed in the next section.

Obesity and Depression

Despite a higher prevalence of anxiety disorders than affective disorders

(14.4% vs. 6.2%) (Australian Bureau of Statistics, 2007), depression received much

more attention from researchers than anxiety. The research studies investigating the

effect of obesity on depression, both cross-sectional and longitudinal, have been

growing rapidly in the past decade. These studies commonly have a large sample

size (such as national surveys) and most of them cover a broad age range (e.g. 18+

41

years), thus providing valuable information on the effect of obesity and overweight

on depression.

Findings from cross-sectional studies have not been consistent, and the role

of gender in this correlation remains undetermined. Several studies found that people

with obesity had an increased risk of having depression as compared to people with

normal weight (Barry, Pietrzak, & Petry, 2008; Bruffaerts et al., 2008; Carpenter,

Hasin, Allison, & Faith, 2000; Chen, Jiang, & Mao, 2009; Dong, Sanchez, & Price,

2004; Eunkyung, 2009; Heo et al., 2006; Mather, Cox, Enns, & Sareen, 2009; Scott

et al., 2008; Zhao et al., 2009), while other studies found a non-significant

correlation between obesity and depression (Goldney, Dunn, Air, Dal Grande, &

Taylor, 2009; Hach et al., 2007; Hach, Ruhl, Klotsche, Klose, & Jacobi, 2006; Jorm

et al., 2003; Turley, Tobias, & Paul, 2006). As mentioned earlier, these studies

generally have large and nationally representative sample sizes. But important

confounders such as physical illnesses and lifestyle factors were not necessarily

controlled for.

It is noticed that controlling for important confounders was not as prevalent

in studies demonstrating a positive relationship than in studies discovering a

non-significant relationship (50% vs. 80%). For example, a study carried out in six

European countries found that people with obesity are 30% (95%CI, 1.0-1.8) more

likely to have mood disorder and 40% (95%CI, 1.0-2.2) more likely to have more

than one mental disorder (Bruffaerts et al., 2008). Yet, this study did not control for

physical health conditions, only sociodemographic factors. This may have masked

the true relationship between the two variables. In the fourteen reviewed

cross-sectional studies, eight studies found an increased risk of having depression

within women (Barry et al., 2008; Carpenter et al., 2000; Chen et al., 2009; Dong et

42

al., 2004; Eunkyung, 2009; Mather et al., 2009; Scott et al., 2008; Zhao et al., 2009),

while only three studies revealed similar findings in men (Barry et al., 2008; Dong et

al., 2004; Zhao et al., 2009). The moderating effect of gender is further complicated

when it was claimed that obesity had no association with depression in women, but

was a protective factor for depression in men, after adjusting for physical illnesses

and sociodemographic factors (Goldney et al., 2009). Nevertheless, a more robust

correlation between obesity and depression was seen in women, which lends support

to the explanation of a higher prevalence of depression in women than in men to

some extent.

The longitudinal prospective correlation between obesity and depression is

also debated. The most recent meta-analysis of longitudinal studies indicated that

both obesity and overweight at baseline increased the risk of depression at follow up

(OR, 1.55; 95% CI, 1.22-1.98; OR, 1.27; 95%CI, 1.07-1.51), with no gender

difference being found (Luppino et al., 2010). However, this meta-analysis (Luppino

et al., 2010) was only able to adjust for age and sex. As acknowledged by the authors

themselves, the inability of controlling for potential covariates such as medication

use may have inhibited uncovering of the true effects and magnitude of obesity on

depression. Several more longitudinal studies were published after this meta-analysis.

One used a cohort at the age range from 51 to 61 years, and followed participants for

6 years (Carroll, Blanck, Serdula, & Brown, 2010). The results of the study showed

that women who remained obese during follow up were 40% (95% CI, 1.07-1.76)

more likely to develop depression as compared to those who remained non-obese.

For men, a non-significant relationship was found (OR, 1.10; 95%CI, 0.79-1.51). A

range of significant confounders were controlled in the study (Carroll et al., 2010).

The other study was undertaken in elderly people (70-79 years), with a 5-year follow

43

up (N = 2,547). It was discovered that after adjusting for sociodemographic factors,

chronic diseases and medications, obesity increased the risk of depression by 20% in

men, but no effect was seen among women (Vogelzangs et al., 2010). The

correlation was debated even more when the results of the analysis of the National

Population Health Survey (NPHS) in Canada were published (Gariepy, Wang,

Lesage, & Schmitz, 2010). In this study, past-year major depression was measured

by the Composite International Diagnostic Interview-Short Form for Major

Depression (CIDI-SFMD), and the measurement of BMI was self-reported.

Controlling for sociodemographic, health and lifestyle variables, the authors found

that obesity at baseline did not predict subsequent depression in women (HR, 1.03,

95%CI, 0.84-1.26), but was a protective factor for depression in men (HR, 0.71,

95%CI, 0.51-0.98). What needs to be considered is that Gariepy et al. (2010),

combined normal and overweight groups as a reference group, with which obesity

was compared. Therefore, the results may not necessarily be comparable to other

longitudinal studies.

Taken all together, it is clear that previous studies examining the effect of

obesity on depression have the advantages of using a large and representative sample.

However, the evidence supporting the hypothesis of the association between obesity

and depression is still weak (Atlantis & Baker, 2008). For women in particular, it

was argued that the correlation between obesity and depression may vary with age.

This is because when people who are older than 65 years are analysed individually,

the significant correlation between obesity and depression among women

disappeared (Heo et al., 2006). In addition, when people are stratified into different

age groups, obesity is correlated with depression in women aged from 18 to 39 years

(OR, 1.67; 95% CI, 1.29-2.15), but not with depression in women at the 40 to 59

44

years age range (Chen et al., 2009). Last, when a cohort of older people is

investigated, no correlation is observed in women, but a significant one is found in

men (Vogelzangs et al., 2010).

Obesity and Anxiety

The prevalence of anxiety is higher in many developed countries; studies

examining the effect of obesity on anxiety were also identified. Researchers from

Canada conducted a systematic review and meta-analysis with regard to the

associations between obesity and anxiety disorders in adults from community

samples (Gariepy, Nitka, & Schmitz, 2009). This collection of review articles

included fourteen cross-sectional and two prospective studies. The pooled estimates

of cross-sectional studies was 1.4 (95% CI, 1.2-1.6), with no moderating effect of

gender being observed. However, it was highlighted by the authors that half of the

studies included in the review had a poor control of confounders (Gariepy et al.,

2009), which may artificially inflate the magnitude of the correlation. As stated in

the paper, the pooled estimate of good quality studies was 1.2 (95% CI, 1.1-1.5),

which was lower than that of poor quality studies (OR, 1.5; 95% CI, 1.3-1.8). One

study carried out in Australia was not included in the review (Jorm et al., 2003). This

study reported anxiety scores among obese and non-obese groups, adjusting for

physical ill health, lack of physical activity, social support, education and financial

difficulty. The results suggested that obese women had significantly lower anxiety

scores than women with normal weight. Meanwhile, no difference was found in men

(Jorm et al., 2003).

The findings from prospective studies have not been consistent, partly due to

a shortage of studies. The two prospective studies reviewed by Gariepy et al. (2009)

45

generated different results, particularly for women. Another study followed 544

child-bearing age women for three decades and controlled for chronic disease, social

support and socioeconomic factors in the analysis. Results showed that women who

are obese were 6 times (95% CI, 1.39-28.16) more likely to develop anxiety (Kasen,

Cohen, Chen, & Must, 2008). In contrast, a Norwegian study only found an

increased risk of having anxiety in men (OR, 1.50; 95% CI, 1.23-1.83), not in

women (OR, 0.99; 95% CI, 0.85-1.15). This study was on a large scale (N = 33,777),

and well controlled for other unhealthy lifestyle factors, medication use and

sociodemographic factors (Bjerkeset, Romundstad, Evans, & Gunnell, 2008).

In brief, cross-sectional studies generally supported a positive relationship

between obesity and anxiety and the strength of the relationship is generally mild.

Gender does not seem to have a modifiable effect on the relationship as suggested by

cross-sectional studies. However, most of the evidence was derived from

cross-sectional studies; no conclusion can be drawn in terms of the longitudinal

effect of obesity and anxiety.

Obesity and General Mental Health

The measure of general mental health is commonly included in health related

quality of life assessment, of which the Medical Outcomes Study short form (SF-36)

is a widely applied measurement. In relation to relative weight categories, the

literature has consistently revealed that as compared to people with normal weight,

the physical health of people who are obese was often compromised, however, their

general mental health remained undisturbed (Mond & Baune, 2009; Renzaho,

Wooden, & Houng, 2010; Vasiljevic et al., 2008; Wee, Wu, Thumboo, Lee, & Tai,

2010). All the studies adopted cross-sectional designs, used an adequate sample size

46

ranging from 2,732 to 9,771, and controlled for essential confounders such as

chronic condition, social support and family functioning. For example, the study by

Renzaho et al. (2010) from Australia found that in women, physical functioning

declined continuously as the BMI increased, but deterioration of general mental

health (both MHI & MCS) was only observed for morbidly obese women (BMI ≥40),

not other obesity categories (BMI, 30.00-39.99). The difference in general mental

health was not significant across all weight categories (p = .0356). In short, the

studies provided strong evidence that women with obesity have impaired physical

functioning, but their general mental health remained as good as those with normal

weight.

Physical Activity and Mental Health

Epidemiology of Physical Activity

Physical activity is defined as any bodily movement produced by skeletal

muscles that result in energy expenditure significantly beyond resting level,

particularly involving continuous actions of large muscles (Briffa et al., 2006). It has

been long regarded as a simple and inexpensive way to maintain individuals’ health,

functioning and quality of life. Evidence from epidemiological studies and clinical

trials demonstrates that physical activity reduces the risk of developing coronary

heart disease, diabetes mellitus, stroke, hypertension, cancer, metabolic syndromes,

falls and osteoporosis (Breslow, Ballard-Barbash, Munoz, & Graubard, 2001;

Knowler et al., 2002; Li et al., 2006; Ma et al., 2008; Parker, Jacobs Jr, Schreiner,

Schmitz, & Dengel, 2007; Rothenbacher, Koenig, & Brenner, 2006). While being

physically active has proved to be beneficial to individuals’ health, individuals

leading a sedentary lifestyle had a higher risk of all-cause mortality (Lee & Skerrett,

2001).

47

In spite of the well demonstrated benefits of physical activity, the prevalence

of physical inactivity (defined as sedentary lifestyle and low level of physical

activity) in Australian adults declined by only a little (Australian Bureau of Statistics,

2006g; Bauman & Owen, 1999). Globally, over 60% of adults are not active enough

to benefit their cardiac health (World Health Organisation, 2003). The proportion of

Australian adults who are sedentary or taking infrequent physical activities (these

levels were defined by an estimation of daily energy expenditure based on the

frequency and duration of reported physical activity) were 30% and 36%

respectively (Bauman & Owen, 1999), which was associated with 7% of disease

burden in the country (Mathers, Vos, Stevenson, & Begg, 2000). By gender, it

accounts for 6.0% of total disease and injury burden in males and 7.5% in females,

respectively. Some subgroups like women, the low-income, and the elderly are even

less likely to take part in physical activities which provide basic health benefits

(Mathers, Vos, & Stevenson, 1999).

Physical Activity and Depression

The correlation between physical activity and depression has been long

discussed. It was believed that physical activity has a protective effect against

depression based on both biological and psychosocial mechanisms (Donaghy, 2007).

Physical activity increases the blood flow to the brain, stimulating the release of

endorphins and other chemicals depleted during depression. Psychosocially, physical

activity improves mental well-being by enhancing self-esteem and improving

body-image through the process of making plans, setting goals and achieving them.

For women approaching, at or after menopausal transition, physical activity may be

of greater importance. As they were reported to be less active (Wen et al., 2002),

48

they are at the stage of life when the risk of having many chronic diseases may

increase (American Heart Association, 2010).

The finding of cross-sectional design studies mostly suggested a favourable

effect of physical activity on depression (De Moor et al., 2006; Galper et al., 2006;

Goodwin, 2003; Hassmén, Koivula, & Uutela, 2000; Taylor-Piliae et al., 2010;

Vallance, Murray, Johnson, & Elavsky, 2010), except for two studies (Jackson, 2006;

Rakovac, Baric, & Heimer, 2007). While controlling for potential covariates such as

chronic disease, in studies demonstrating positive findings social support and

sometimes lifestyle factors were frequently observed. Although most of the studies

agreed on the protective effect of physical activity, it was noticed that the variation

in the measurement of physical activity was fairly considerable. In addition, the

majority of studies reviewed actually measured physical activity as exercise, not

physical activity which did not involve exercise. For example, a postmenopausal

women study used both subjective (meeting a physical activity guideline,

self-reporting) and objective (pedometer) measurements of physical activity

(Vallance et al., 2010). It was found that women who met the physical activity

recommendation had a significantly lower score of depression (effect size d = .27, p

= .022) than those who did not, while women achieving 7,500 steps per day did not

have less depression symptoms (effect size d = .20, p = .078).

Galper et al. (2006) also measured physical activity in both subjective

(self-report of a physical activity inventory) and objective (treadmill) methods, but

the results yielded from the two methods showed both were beneficial (Galper, et al.,

2006). Other studies used questions to identify the frequency of exercise (Goodwin,

2003; Hassmén et al., 2000). One of the studies claiming a non-significant

relationship between physical activity and depression was flawed in failing to control

49

for any potential covariates (Rakovac et al., 2007). In addition, it defined physical

activity as going to the gym regularly for at least 5 years. This method does not

necessarily allow comparison between physically active and inactive women, as

women can do exercise outside of the gym (Rakovac et al., 2007).

Longitudinal prospective studies carried out in the U.S., Australia, the U.K.

and Japan provided additional support for the protective effect of physical activity

against depression (Bhui & Fletcher, 2000; Brown, Ford, Burton, Marshall, &

Dobson, 2005; Fukukawa et al., 2004; Strawbridge, Deleger, Roberts, & Kaplan,

2002; Van Gool et al., 2007). For instance, the last mentioned study found that

people who reported themselves as having more than 30 minutes light activity per

day had about 50% reduced risk of having depression within the six year period of

the study. A number of confounders were controlled during analysis. These included

age, marital status, education, function and number of chronic diseases (Van Gool et

al., 2007). In addition, in a study on Australian women in midlife it was reported that

compared to women who reported less than 60 minutes of moderate physical activity

per week, those who reported physical activity beyond this level had 30% to 40%

less risk of having depression within a five year period (Brown et al., 2005). A

Japanese study, however, suggested that the protective effect of physical activity

might be limited to age, as the study showed that walking had a protective effect

against depression for people aged from 65 to 79 years, but not for the mid-aged

(40-64 years) (Fukukawa et al., 2004). Furthermore, one study claimed that physical

activity only benefited men, but not women (Bhui & Fletcher, 2000).

In all, previous studies achieved agreement of the protective effect of

physical activity against depression. Men and women are equally likely to gain

psychological benefits from physical activity, while the moderating effect of age still

50

requires further research. However, it should be noted that most of the studies

examined physical activity as exercise; hence the relationship of depression with

physical activity unrelated to exercise remained unanswered.

Physical Activity and Anxiety

Evidence about the influence of physical activity on anxiety was primarily

drawn from cross-sectional studies, the results of which constantly showed a

beneficial effect of physical activity on anxiety (De Moor et al., 2006; Goodwin

2003; Rakovac et al., 2007; Vallance et al., 2010). For example, the study of 217

postmenopausal women showed that women whose physical activity level meeting

the guidelines reported significantly fewer symptoms of anxiety than those who did

not meet the guidelines by conducting a univariate ANOVA analysis (Vallance et al.,

2010). In addition, this result remained unchanged after adjusting for age, BMI and

co-morbidities.

Longitudinal studies in this area were rather limited as compared to

depression studies. One U.K. study undertaken in 2000 investigated anxiety by

means of a General Health Questionnaire (GHQ). The results showed that, after

adjusting for age, smoking, disability, income, employment and baseline GHQ score,

there was no relationship between exercise and anxiety in women, only a favourable

effect observed in men (Bhui & Fletcher, 2000). The relationship between physical

activity and anxiety has received much less attention when compared to the

relationship between physical activity and depression. Similar variance in the

measurement of physical activity existed in the literature on physical activity and

anxiety. Therefore, it was recognised that the beneficial effect of physical activity

mainly referred to exercise. No long-term effect of physical activity on anxiety could

51

be determined at this stage due to the lack of longitudinal prospective studies

(Martinsen, 2008).

Physical Activity and General Mental Health

As measured by the mental health scale or mental composite score of the

SF-36, all of the studies indicated a better general mental health in people who are

more physically active than inactive individuals (Aoyagi, Park, Park, & Shephard,

2010; Vallance et al., 2010). In postmenopausal women, it was also found that

women meeting the recommended guidelines of physical activity had a mental

composite scores (MCS) of 52.1 (SD = 6.8), which was significantly higher than that

of women who did not meet the guideline (49.8 ± 8.3, p = .011). The result of this

study was believed to be reliable, as the study took consideration of confounders

including socioeconomic factors, smoking, BMI and co-morbidities (Vallance et al.,

2010). A further study has made an additional contribution to the area, examining the

interaction between physical activity not involving exercise and physical activity as

exercise on general mental health (Aoyagi et al., 2010). In the end, the authors found

that after adjusting for the former and other confounders, older people engaging in

activity > 3 Metabolic Equivalent Tasks (MET) had a significant higher score on a

mental health scale as well as increased physical functioning, vitality and body pain,

in spite of the absence of gender specific results in this study report.

Longitudinal prospective effect of physical activity was less explored.

Australian researchers evaluated the correlation between physical activity and mental

health among midlife women over five years. Controlling for sociodemographic

factors, smoking, BMI, menopause and chronic conditions, the study found a

52

30%-40% lower risk of having a MCS score under 52 for women reporting more

than 60 minutes of moderate physical activity per week (Brown et al., 2005).

In general, it was believed that physical activity has a positive relationship

with general mental health, regardless of gender. What needs to be noted is that the

evidence was mainly based on cross-sectional studies; hence, further research is

required to permit exploration of the long-term effect of physical activity on general

mental health.

Smoking and Mental Health

Epidemiology of Smoking

Smoking is strongly correlated with morbidities and mortality (Majid Ezzati,

Henley, Lopez, & Thun, 2005; Ezzati & Lopez, 2004; Warren, Jones, Eriksen, &

Asma, 2006) and is also the leading preventable cause of death around the world,

including in Australia. In spite of the relatively stable downward trend of smoking

prevalence in Australia from 22.9% in 1997 to 18.4% in 2005 (Social Research

Centre, 2006), tobacco smoking was still responsible for 7.8% of the total burden of

disease and injury in the country in 2005 (Australian Bureau of Statistics, 2006b),

ranking as the first among all health risk factors. As revealed by the latest release of

smoking estimates from the WHO, in many developing countries, the prevalence of

current female smokers was much lower than that among males (Storr et al., 2010).

For example, in China, the percentage of current smokers was 55.5% for men and

5.5% for women. But, in Australia, women were almost equally likely to be a current

smoker. The prevalence of current smokers was 26% in men and 20% in women,

respectively (Keizer & Eytan, 2005). Therefore, reducing smoking should always be

a routine in chronic disease prevention among women. Research has shown that

smoking is a pertinent issue for individuals with mental disorders, with 33% of the

53

cigarettes being consumed by people with 12-month duration of mental disorders

(Tobias, Templeton, & Collings, 2008).

Smoking and Depression

It has been long regarded that there is a strong linkage between smoking and

depression in general populations, as indicated by a number of cross-sectional

studies (Hämäläinen et al., 2001; Husky, Mazure, Paliwal, & McKee, 2008; Khaled

et al., 2009; Lawrence, Mitrou, & Zubrick, 2009; Massak & Graham, 2008; Nakata

et al., 2008; Pasco et al., 2008; Pomerleau, Zucker, & Stewart, 2003; Trosclair &

Dube, 2010) , except for two studies that were undertaken in Norway (Mykletun,

Overland, Aarø, Liabø, & Stewart, 2008) and Chile (Araya, Gaete, Rojas, Fritsch, &

Lewis, 2007). Smoking status which was classified never, past (ex- or quitter), and

current, was commonly seen across the literature as a relevant indicator. Studies

usually targeted a broad age range of participants, and an even distribution of males

and females in sample population was observed. The sample sizes were fairly large,

ranging from 931 in an exclusively female study (Pomerleau et al., 2003) to 73, 024

in an American national survey (Trosclair & Dube, 2010).

Different effect sizes were seen among studies supporting a positive

correlation between smoking and depression. For instance, in one study, the authors

used both clinical interview and the Centre for Epidemiologic Study Depression

Scale (CESD) (Lewinsohn, Seeley, Roberts, & Allen, 1997) to measure depression

(Massak & Graham, 2008). After adjustment for age, education, alcohol and gender,

the analysis showed that the odds of having depression for current smokers were

1.10-3.24 as compared to non-smokers. In a Canadian study, among women older

than 12 years, the proportion of depression among current smokers was 14.8%,

54

which was much higher than 4.0% for the category of former smokers, and 4.6% for

those who have never smoked (Khaled et al., 2009). Scores on a depression scale

(CESD) were also reported, with a significantly higher depression score being found

in current smokers than non-smokers (13.63 ± 9.14 vs. 10.73 ± 9.75, p < .001). The

difference accounted for several confounders including age, education and marital

status (Pomerleau et al., 2003). Regardless of the effect sizes expressed in these

studies, smoking is believed to be a risk factor for depression in most of the studies.

In addition, being a female was indicated to be another strong risk factor for

depression, beyond the effect of smoking (Hämäläinen et al., 2001).

However, in some cultures, smoking may not be a risk factor for depression.

The HUNT study (N = 60, 814) conducted in Norway did not find an elevated risk of

depression among current smokers as compared to non-smokers (OR, 1.10; 95% CI,

0.99-1.21), after adjusting for sociodemographic factors, somatic symptoms, alcohol,

physical activity and other covariates (Mykletun et al., 2008). This study utilised the

Hospital Anxiety and Depression Scale (HADS) (Zigmond & Snaith, 1983) as the

instrument of depression measurement, and applied 8 points (in a range of 0 to 21) as

the cutoff score. Also in Chile, when adults from a community sample were studied

(N = 3,870), the odds of having depression was also non-significant (OR, 1.49; 95%

CI, 0.9-2.5), controlling for sociodemographic factors.

Despite the strong evidence from cross-sectional studies, whether smoking

initiates depression remains unclear. Some studies agree that smoking increased the

risk (2 to 4 times) of developing depression (Klungsoyr, Nygard, Sorensen, &

Sandanger, 2006; Pasco et al., 2008), some do not agree (Cuijpers, Smit, Ten Have,

& De Graaf, 2007; Takeuchi, Nakao, & Yano, 2004), and some authors argued the

risk only existed in males (Korhonen et al., 2007). It was noticed that all the

55

longitudinal studies performed a good control of potential confounders including

sociodemographic factors, somatic diseases and sometimes lifestyle factors;

therefore, it was less likely that the true association was biased. The length of follow

up did vary to a large extent, ranging from one year to eleven years. Given the effect

of smoking on the human body usually accommodated by the body’s homeostatic

response, long-term sequelae such as depression may only be reliably demonstrated

over an extended time frame (Pasco et al., 2008). However, the role of gender in

cohort studies remained unclear.

Smoking and Anxiety

The U.S. National Survey on Drug Use and Health surveyed the prevalence

of smoking among individuals with various mental disorders between 2005 and 2006

(Trosclair & Dube, 2010). The prevalence of smoking was highest among

individuals with both anxiety and depression (41.3%), followed by those having

anxiety 35.2%, and was lowest among those having depression (29.9%). But the

correlation between smoking and anxiety was less examined than that between

smoking and depression. Nevertheless, all of the existing literature suggested a

positive relationship between anxiety and smoking, regardless of the study design

and measurement of anxiety (Araya et al., 2007; Cuijpers et al., 2007; Mykletun et

al., 2008; Tselebis, Panaghiotou, Theotoka, & Ilias, 2001).

The previously listed study also examined anxiety using the HADS.

Although they failed to find a non-significant association of smoking in relation to

depression, results of the study did suggest an increased risk of having anxiety

among current smokers (OR, 1.22; 95% CI, 1.13-1.30), as compared to non-smokers.

A significant, though lower risk of anxiety was also found among quitters (OR, 1.08;

95%CI, 1.00-1.62) (Mykletun et al., 2008). The other study of 114 female nurses

56

(mean age: 33 ± 6 years) used the State Trait Anxiety Inventory (STAI) to compare

anxiety symptoms according to smoking status. Authors found a significantly higher

anxiety score in current smokers than in non-smokers (43.04 ± 8.48; 38.94 ± 6.45, p

< .02), although confounding control was missing (Tselebis et al., 2001).

A prospective study from The Netherlands examined whether baseline

smoking predicted depression in two years (Cuijpers et al., 2007). In this study,

mental disorders were assessed by the Composite International Diagnostic Interview

(CIDI) (Patten, Brandon-Christie, Devji, & Sedmak, 2000). In relation to anxiety, it

was found that the risk of developing any anxiety disorder was increased by 70%

(95%CI, 1.10-2.86). More importantly, the risk of having general anxiety disorder

was the highest, being nearly four times (IRR, 3.80, 95% CI, 1.09-13.21).

Due to being limited to the number of studies on the association between

anxiety and smoking, the effect of gender could not be determined based on the

current literature. In all, it was considered that people who are current smokers have

a higher risk of anxiety than non-smokers, yet more longitudinal research is

warranted to validate the longitudinal relationship and possibly, to investigate the

role of gender.

Smoking and General Mental Health

The correlation between smoking and general mental health has achieved less

consistency, particularly in the longitudinal correlation. An Iranian study examined

water pipe smoking and health related quality of life in a general population (N =

1,675) via a cross-sectional design (Tavafian, Aghamolaei, & Zare, 2009). The

authors compared the odds of having poor mental health (MCS < 50) for water pipe

smokers and non-smokers. The results showed that current smokers were twice as

likely to have poor mental health (OR, 1.88; 95% CI, 1.36-2.60). Sociodemographic

57

factors including age, education and employment were adjusted for in the analysis,

and being female increased the risk of having poor mental health as well (OR, 2.26;

95% CI, 1.55-3.30). These results corresponded to an earlier Japanese study (Mino et

al., 2001).

However, the long-term effect of smoking on general mental health was less

clear. The Japanese study incorporated a prospective design with a follow up period

of two years. As measured by the General Health Questionnaire, mental health was

constantly found to be worse among current smokers over 24 months, particularly in

women (Mino et al., 2001). The Nurses’ Health Study in America was comprised of

158,736 nurses aged from 29 to 71 years. The relationship between smoking and

mental health (SF-36) was investigated in 1992/1993 and 2000/2001 (Sarna, Bialous,

Cooley, Jun, & Feskanich, 2008). The baseline analysis showed that current smokers

had a significantly worse general mental health than non-smokers (MCS -2.0; MHI

-3.7) after adjustment for age, BMI, physical activity, living alone and

co-morbidities. Moreover, it was found in this nursing cohort that nurses’ mental

health improved steadily as they age across all smoking statuses, but the disparities

of mental health as related to smoking status remained (Sarna et al., 2008). These

studies may suggest that smoking does have a strong negative impact on people’s

general mental health, yet further research is required to verify the correlation.

Alcohol Use and Mental Health

Epidemiology of Alcohol Use

A rise in alcohol consumption is seen around the world, especially in many

developing countries (World Health Organisation, 2008). In Australia, alcohol

dependence and harmful use was ranked 17th in the 20 leading causes of the burden

58

of disease and injury in 2003. The proportion of people drinking at a high risk level

has kept increasing over the last three National Health Surveys, from 8.2% in 1995

to 10.8% in 2001 and 13.4% in 2004-2005, with greater increases in women than in

men, after adjustment for age differences (Australian Bureau of Statistics, 2006c).

Globally, alcohol causes 3.2% of total deaths and contributes to 4% of the disease

burden (World Health Organisation, 2008). Varied detrimental health effects of high

risk alcohol use have been identified, ranging from long-term risk like the

development of chronic diseases to short-term consequences including an increasing

number of hospital admissions (Hart & Smith, 2009). Alcohol consumption not only

affects physical health, but also mental health status too. Some research studies on

alcohol consumption and mental health are detailed in the next few sections. This

review focused on alcohol consumption rather than alcohol dependence or

alcoholism.

Alcohol Use and Depression

Summarising the correlation between alcohol consumption and depression is

somewhat challenging. Unlike alcohol dependence, which could be diagnosed by

CIDI or another standardised instrument, there is the lack of a standard means for

measuring alcohol consumption. In the literature, the measurement of alcohol varied

considerably, which contributed to the difficulty of effective comparison of the

studies. In general, the cross-sectional studies suggested a correlation between

alcohol consumption and depression. For example, Chan and his colleagues

conducted a study in 1,594 middle-class people aged 50-97 years in the U.S., and

they measured alcohol consumption by quantifying the number of drinks in the past

two weeks and the frequency of drinking in the last week (Chan, Von Muhlen,

59

Kritz-Silverstein, & Barrett-Connor, 2009). Controlling for age, chronic diseases and

lifestyle factors, the results indicated that in females, the prevalence of depression

decreased gradually as the alcohol consumption increased (non-drinkers: 15.4%;

occasional: 8.3%; light: 7.3%; moderate: 2.4%, p<0.001), yet the results were

non-significant in males (Chan et al., 2009). This negative and linear relationship

actually suggested that alcohol drinking was protective against depression, however,

it should be noticed that among this midlife and older population, the prevalence of

alcohol abuse was very low, the adverse effect of which may therefore be difficult to

detect.

However in an Australian study, a U-shaped correlation was found in 2,725

adults, both male and female (Rodgers et al., 2000). In this study, participants were

categorised into five levels including non-drinking, occasional, low-level, high-level

and hazardous drinking based on the Australian National Health and Medical

Research Council. The analysis showed that women drinking at a low level (< 7

standard drinks per week) had the least depression symptoms (Rodgers et al., 2000)

after adjustment for sociodemographic factors, childhood adversity, social support

and personality factors.

In addition, another Australian study, which classified young adults (20-24

years) into light, moderate, hazardous and harmful drinking levels, found a linear

one in women controlling for similar confounders as in the study by Rodgers was

found (Caldwell et al., 2002). These two studies paralleled in study design,

measurement of alcohol and confounding effect control, only the age of the samples

differed. Therefore, this may suggest that the pattern of the correlation varies in the

course of a life time (Alati et al., 2005), as the prevalence of alcohol consumption is

lower in older people (Australian Bureau of Statistics, 2006c). Lastly, analysis based

60

on the Norwegian HUNT study revealed a U-shaped correlation too, without

identifying a gender effect (Skogen, Harvey, Henderson, Stordal, & Mykletun,

2009).

The temporal association between alcohol consumption and depression also

remained uncertain. Canadian researchers conducted a follow up study over the

duration of two years in a national general population sample (Wang & Patten, 2001).

This study used four different ways to differentiate levels of alcohol consumption.

Basically, no difference was detected with regard to the incidence rate of depression

regardless of the measurements of alcohol. Only women who reported having more

than five drinks on a drinking occasion had a higher incidence rate of depression

than those who drank less than this level (6.4% vs. 3.7%). Furthermore, it has been

demonstrated that excessive and hazardous drinking was not related with the onset of

depression, and abstinence was associated with a reduced risk of developing

depression (Haynes et al., 2005). The follow-up of the study was one and half years

and important confounders included sociodemographic factors, life events, social

support, smoking and mental health at baseline. It was noticed that both studies

followed participants for no more than two years, which may not be long enough for

depression to occur.

In brief, evidence from cross-sectional studies confirmed the correlation

between alcohol consumption and depression, despite a substantial variance in the

measurement of alcohol consumption. The pattern and significance level of the

correlation may be affected by age and gender. No conclusion can be drawn

regarding the longitudinal effect of alcohol on depression.

61

Alcohol Use and Anxiety

The issue of alcohol measurement remained among studies examining its

effect on anxiety. The study by Rodgers et al. described above also examined

depression in 2,725 Australian aged from 18-59 years, finding a U-shaped

relationship between alcohol use and anxiety assessed by the Goldberg Depression

and Anxiety scale (GDAS) in both men and women after controlling for

sociodemographic factors, childhood adversity, social support and personality

(Rodgers et al., 2000). In addition, in women at child-bearing age, a variation of the

correlation was observed: a J-shaped relationship when women were in their 30s,

and a linear relationship when women were at 25 and 40 years (Alati et al., 2005).

Furthermore, a study from Norway also defined a U-shaped relationship among

abstainers and low-level alcohol consumers. The results suggested that abstention

was related to an increased chance of anxiety (OR, 1.34; 95% CI, 1.19-1.52) as

compared to people drinking moderately, after controlling for socioeconomic status,

social network, somatic illness and gender (Skogen et al., 2009).

For the prospective association, only a study from the U.K. was found

(Haynes et al., 2005). The results of the study did not support a significant

correlation between baseline excessive alcohol use and anxiety, while for frequency

of binge drinking, the men who binge drank weekly were found to be three times

more likely to have anxiety as compared to those with less than monthly bingeing

(OR, 3.14; 95% CI, 1.07-9.26).

In short, the existing literature revealed a significant linkage between alcohol

consumption and anxiety. The studies had the strength of using a large-scale sample

size and controlling for potential confounders, however, the lack of consistent

measurement of alcohol consumption inhibited comparison among studies.

62

Alcohol Use and General Mental Health

A six year long study examined a cohort of old women aged from 70 to75

years (N = 12,432) from 1996 to 2002 to investigate the alcohol use effect on

mortality and health- related quality of life (Byles, Young, Furuya, & Parkinson,

2006). Interestingly, the authors found that women consuming a moderate intake of

alcohol (1-2 drinks/day on 3-6 days) had better mental health (measured by the

SF-36-MHI) compared to non-drinkers (78 vs. 75, p < .005), and were less likely to

die as compared to non-drinkers. While in the study by Chan et al. (2009), mental

health (SF-36, MCS) was not significantly related with the number of drinks

consumed in the past week in women (β = .154, p = .19), it was positively related in

men (β = .220, p = .01), after adjustment for age, BMI, current smoking, exercise

and current use of estrogen in women. Although Chan and colleagues did not find a

significant benefit of alcohol consumption on general mental health, their study did

discover that the number of alcoholic drinks was negatively related with depression.

Therefore, it was considered that essentially the study by Byles et al. (2006) and

Chan et al. (2009) provided some evidence of moderate alcohol consumption’s effect

on general mental health, particularly in women. It is important to note that these

findings were drawn from older populations, who are less likely to have drinking

problems; therefore, the ability to apply the findings to a younger population remains

uncertain.

63

Diabetes and Mental Health

Epidemiology of Diabetes

Diabetes is a serious chronic illness with multiple complications and

premature mortality (World Health Organisation, 2006). In Australia, 3.6% of the

whole population has diabetes, which equals approximately 700,000 persons; of all,

83% are Type 2 diabetes (Australian Bureau of Statistics, 2006e). Type 2 diabetes is

likely to develop after 40 years of age, and is strongly associated with obesity,

physical inactivity, and unhealthy diet. For example, the Nurses’ Health Study (N =

78,419), applying a longitudinal analysis, concluded that obesity or weight gain was

a significant risk factor of diabetes (Shai et al., 2006). The study showed that for

non-indigenous Americans, each 5-unit increment in BMI resulted in a twofold

increase in the risk of having diabetes. Additionally, it was found that for each 5 kg

of weight gain, the risk of diabetes was increased by 50% (95% CI, 26%-63%). The

adjusted relative risk (RR) of diabetes ranged from 1.55 to 2.36. Apart from obesity,

physical inactivity was also shown to be a critical risk factor, with an adverse effect

independent from obesity (Sullivan, Morrato, Ghushchyan, Wyatt, & Hill, 2005).

Using a nationally representative sample, Sullivan and colleagues (2005) found that

compared to active people of normal weight , inactive individuals of normal weight

had 50% increased risk of having diabetes (95% CI, 1.25-1.86), while the increased

risk for overweight and inactive ones was 1.65 (95% CI, 1.40-1.96).

Given the strong correlation between obesity and physical inactivity for

diabetes, and the pervasive obesity and physical inactivity, it was projected that the

prevalence of diabetes would be doubled in 2030 from a prevalence rate of 2.8% in

2000, if no urgent action were taken. In fact in Australia, it was estimated that every

64

year, 0.8% of Australian adults were developing diabetes, which equalled 275 new

diabetic patients every day (Barr et al., 2005).

The body of research studies on diabetes complications is expanding fast. For

example, diabetes was found to increase the risk of developing heart diseases (Lafitte

et al., 2010), and furthermore, to correspond to a climbing incidence rate of Type 2

diabetes-related amputations (Vamos, Bottle, Majeed, & Millett, 2010). More

importantly, when mental health problems co-exist with diabetes, it is likely to

further worsen the prognosis. In patients with diabetes, depression has been

associated with worsened glycaemia control (Lustman et al., 2000), non-adherence

to treatment (Ciechanowski, Katon, & Russo, 2000), higher disability bed days

(Egede, 2004) and increased mortality (Ismail, Winkley, Stahl, Chalder, & Edmonds,

2007). Anxiety was significantly associated with poor hyperglycaemia control

(Anderson et al., 2002). A recent large-scale study examined panic episodes among

4385 diabetic patients (95% were Type 2 diabetes) and they reported panic episodes

that were associated with poor glycaemia control, increased diabetic complication,

greater disability and lower self-rated health and social and emotional functioning

(Ludman et al., 2006). Research findings about the correlation between diabetes and

two common mental health problems, depression and anxiety, is detailed in the

following sections.

Diabetes and Depression

A growing number of research articles agree that depression is much more

prevalent in patients with Type 2 diabetes. Evidence from systematic reviews and

meta-analysis (Ali, Stone, Peters, Davies, & Khunti, 2006; Anderson, Freedland,

Clouse & Lustman, 2001) demonstrated that the risk of depression among people

with Type 2 diabetes was 60% (95% CI, 1.2-2.0) as compared to those without

65

diabetes. Despite the advantages of this meta-analysis, which enables a more

objective and accurate estimation of the correlation, it was believed that the true

correlation between diabetes and depression may be skewed. This was because in

most studies, patients with diabetes differed from those without it in many factors

that are associated with depression, such as obesity. This weakness was highlighted

by the authors themselves (Ali et al., 2006). In addition, these systematic reviews

were based on cross-sectional studies, hence providing no clue to the impact of

diabetes on depression over an extended time frame.

In 2008, Mezuk and colleagues conducted a meta-analysis on the longitudinal

correlation between Type 2 diabetes and depression. The estimation of Type 2

diabetes predicting depression was based on seven longitudinal studies. The pooled

relative risk (RR) of depression among those with Type 2 diabetes was shown to be

1.15 (95% CI, 1.02-1.30), which indicated an association of modest magnitude

(Mezuk, Eaton, Albrecht, & Golden, 2008). Meta-analysis is not without limitations.

When extracting the data from the original studies, the authors only obtained the

estimates that were most closely adjusted for sociodemographic factors (Mezuk et al.,

2008).

Although this approach allows effective comparison among studies, it does

artificially alter the correlation between Type 2 diabetes and depression due to the

failure of including fully adjusted estimates. For example, one (De Jonge, Roy, Saz,

Marcos, & Lobo, 2006) of the studies included in this review (De Jonge et al., 2006)

followed an older community population (≥ 55 years) for five years using a standard

interview to assess depression. The age and sex adjusted chance of depression was

1.42 (95% CI, 1.04-1.93); however, after controlling for a full range of confounders

including sociodemographic factors, somatic illnesses and cognitive functioning, the

66

strength of the association was attenuated and became non-significant (OR, 1.26;

95% CI, 0.90-1.77). In contrast, in the meta-analysis (Mezuk et al., 2008), the cruder

estimate was analysed.

In all, the current literature provided some evidence that people with diabetes

are more likely to present and develop depression. However, as mentioned by

authors of systematic reviews, the reliability of the estimated risk of depression

among diabetes may be compromised due to a lack of good control of important

confounding factors including unhealthy lifestyle factors. Given the strong linkage

between unhealthy lifestyle factors and Type 2 diabetes, and their (unhealthy

lifestyle factors’) impact on depression, controlling for lifestyle factors was

considered to be fundamental for the effective comparison of depression between

diabetic and non-diabetic individuals.

Diabetes and Anxiety

The co-occurrence of anxiety and diabetes has also been researched, although

the amount of literature seemed to be less than that on depression. Nevertheless,

previous studies in this area indicated a higher prevalence of anxiety among people

with diabetes than those without. For example, the German National Health

Interview and Examination, which was a cross-sectional study involving 4,169

individuals from the community (Kruse, Schmitz, & Thefeld, 2003), found that

patients with diabetes were much more likely to have anxiety disorder when

compared with the non-diabetic population (OR, 1.93; 95% CI, 1.19-3.14) after

controlling for age, sex, marital status, and socioeconomic status. A modified version

of the CIDI was used in the study to diagnose mental disorders (Kruse et al., 2003).

Systematic reviews were located on the prevalence of anxiety among people

with diabetes (Grigsby, Anderson, Freedland, Clouse, & Lustman, 2002), and it was

67

revealed that 40% of diabetic patients had elevated anxious symptoms, with a higher

prevalence in women compared to men (55.3% vs. 32.9%, p <.001). However, the

increased risk of anxiety among people with diabetes in contrast to the general

population was not able to be obtained as a result of the limited number of controlled

studies (Grigsby et al., 2002). After the publication of this systematic review,

Hermanns et al. (2005) claimed that the frequency of criteria-based anxiety disorders

was lower in diabetic patients compared with a non-diabetic reference sample (5.9%

vs. 9.0%). However in that study, a control group was not clearly identified and no

information on reference sources was available, hence, the conclusion of this study

was questionable. In all, it was considered that the cross-sectional studies provided

useful information on the risk of anxiety among people with diabetes, especially

when there was dearth of longitudinal studies at this stage. Further research

investigating the correlation between anxiety and diabetes may need to take lifestyle

factors into consideration.

Examining the Limitations of the Previous Research

This review of previous studies on the correlations between the four lifestyle

factors including overweight/obesity, physical activity, smoking and alcohol use and

mental health identified several limitations. First, much of the above described

research used large numbers of the general population as study samples with an age

range of 18 years and beyond, and a good number of them had an adequate control

of confounding variables. While intended to provide knowledge about the overall

correlations between lifestyle factors and mental health, the conflicting results

suggested that the correlations might not be homogenate. It is known that the

prevalence of mental health problems varies between different age groups, so do

68

lifestyle factors. Therefore, it is likely that the conflicting results are probably related

with age and gender, which underlines the importance of examining some specific

age and gender groups.

Second, when examining the individual relationships between lifestyle

factors and mental health, there is still a lack of adjustment for other lifestyle factors

in the literature. Given the strong associations among lifestyle factors themselves,

controlling for the interacting effect among lifestyle factors may be an important step

to permit reliable results in this regard.

Third, among studies investigating the impact of diabetes on mental health,

most of the studies were limited in the adjustment of potential confounders that

could interfere with the correlation between these two variables. Therefore, this may

have prevented exploration of unbiased relationships. Also, the literature search

identified a considerable number of studies evaluating the mental health in patients

with diabetes, but their general mental health in comparison with that of the general

population requires further research.

Finally, the lack of a theory-based approach is also noticeable among the

literature. Although the empirical studies in this field underscores the necessity of

developing corresponding strategies to enhance mental health related with lifestyle

factors, knowledge about how to improve mental health remains limited. As

discussed, the social cognitive theory recognises the reciprocal relationships between

personal and behavioural factors; therefore, it permits examination of the

associations between lifestyle factors (behavioural factors) and mental health

(personal factors). Furthermore, the social cognitive theory regards self-efficacy as

the central mediator that initiates behaviour change. The significance of self-efficacy

in predicting behaviour change has also been emphasised in many other theories.

69

Therefore, this study does not only aim to investigate the empirical correlations

between lifestyle factors and mental health, but also to examine the usefulness of

social cognitive theory in explaining any particular associations. If self-efficacy is

proved to be the mediator between lifestyle factors and mental health, corresponding

strategies may be implemented to promote self-efficacy, taking into account

consideration of individuals’ mental health level.

Conceptual Framework

To reiterate, the social cognitive theory was chosen as the foundation of

conceptual framework of the current research. The social cognitive theory is a broad

framework which depicts the reciprocal relationships between personal, behavioural

and environmental factors. The central position of self-efficacy in behavioural

change process is also illustrated clearly.

The current research focused on part of the social cognitive theory, which

included the linkage between behavioural factors (lifestyle factors) and personal ones

(mental health), and the mediating role of self-efficacy in this relationships. The

conceptual framework for the research study is depicted in Figure 2.8. The primary

elements in this conceptual framework are: 1) mental health, 2) self-efficacy, and 3)

four lifestyle factors: overweight/obesity, physical activity, smoking and alcohol use.

70

Figure 2.8. The conceptual framework of the study.

As suggested by the conceptual framework, this current study has two

primary aims. One is to develop a comprehensive understanding of the relationships

between lifestyle factors and mental health, and the other is to examine the mediating

role of self-efficacy in the relationships between lifestyle factors and mental health.

Aims of the Research

Aim 1: The Relationships Between Lifestyle Factors and Mental Health

As indicated by earlier research, both an unhealthy lifestyle and poor mental

health can be costly for both individuals and society. The coexistence of unhealthy

lifestyle factors and mental health problems may further exacerbate people’s health

condition. Previous studies provided some evidence of some gender specific patterns

of the correlations between lifestyle factors and mental health, as well as the

heterogeneity of the associations among age groups. The current study thus aimed to

explore the correlations between lifestyle factors and mental health among midlife

and older women in Australia.

71

To address Aim 1, three individual studies were undertaken. Study 1of the

research investigated the effect of lifestyle factors on mental health within a

population of Australian midlife and older women from the general community.

Study 1 aimed to generate knowledge that can be used in helping lifestyle

modification for the general population. Study 2 examined the difference of mental

health between midlife and older women with and without Type 2 diabetes, and

whether lifestyle factors contribute to the disparities. The findings of Study 2

extended the knowledge on effect of lifestyle factors on mental health and

underscored the importance of integrating mental health component in lifestyle

modification. Finally, Study 3 examined the effect of lifestyle factors on mental

health among midlife and older women with Type 2 diabetes. The profile of the

relationships may be useful in developing further intervention programs for a clinical

population.

Aim 2: Examining the Mediating Role of Self-efficacy

The efficacy of social cognitive theory in facilitating midlife and older

women from the general community to make positive lifestyle changes has been

confirmed (Anderson, Mizzari, Kain, & Webster, 2006). In Study 3 of the current

research, the mediating role of self-efficacy in the correlations between lifestyle

factors and mental health was examined in midlife and older women with a chronic

disease. A cross-sectional design and a clinical sample of midlife and older women

with diabetes were applied. Because of the reciprocal nature of the correlations

between lifestyle factors and mental health, the mediating role of self-efficacy was

examined in two directions. One is whether self-efficacy mediates the effect of

mental health on lifestyle factors, and the other is whether self-efficacy mediates the

72

effect of lifestyle factors on mental health. Seven research questions were created,

and presented as below.

Research Questions

1. What is the effect of lifestyle factors on mental health among midlife and

older women?

2. What is long term effect of lifestyle factors on mental health among midlife

and older women?

3. What is the difference in mental health between midlife and older women

with and without diabetes?

4. What are the contributing factors to the differences in mental health between

midlife and older women with and without diabetes?

5. Does self-efficacy mediate the effect of mental health on lifestyle factors

among midlife and older women with diabetes?

6. Does self-efficacy mediate the effect of lifestyle factors on mental health

among midlife and older women with diabetes?

Chapter Summary

The chapter started with a brief review on women and mental health, and

subsequently described the background of health promotion. Findings from the

literature suggest that women seem to be more vulnerable to mental health problems,

and highlight the importance of research on women at several reproductive stages

including menopause. Furthermore, a greater emphasis was placed on the physical

73

effect of lifestyle factors in the existing literature. Despite their well-known adverse

effects, the prevalence rate of unhealthy lifestyle factors shows no sign of reduction.

This contrast in the research suggests a shift of focus from physical health to mental

health.

To enable the examination of the relationships between lifestyle factors and

mental health, as well as producing knowledge for interventions aimed at improving

mental health, an appropriate behavioural theory is essential. Behavioural theories in

the field of health promotion were reviewed and contrasted. In the end, the rationale

for using social cognitive theory for the study was provided. Social cognitive theory

was chosen because of its ability to recognise the reciprocal correlations between

personal and behavioural factors.

Following this theoretical review, the literatures on the relationships between

the four lifestyle factors and mental health was detailed. This section can be viewed

as two parts: one is the relationship of each lifestyle factor with mental health in the

general population, and the other is the difference in mental health between people

with and without diabetes.

The limitations of previous studies were outlined after the literature review.

Finally, the conceptual framework and research questions of the study were

presented. The details on the methodology for the research study will be introduced

in Chapter 3.

74

75

CHAPTER 3: METHODOLOGY

Introduction

As discussed in Chapter 2, the two main aims of the current research were to

investigate the relationships between lifestyle factors and mental health among

midlife and older women from both community and diabetic clinics, and to explore

the mediating effect of self-efficacy in the relationships between lifestyle factors and

mental health within midlife and older women with diabetes. Both prospective and

cross-sectional designs were implemented and general and clinical samples were

utilised to address the aims. As mentioned, the current research has three individual

studies: Study 1, Study 2 and Study 3. Prior to presenting the methodology of the

studies, a brief review of a series of women’s health studies, on which the current

research was based, is undertaken. In addition, the linkage between the current

research and these women’s health projects is explained. Following this, the

methodology of each study of the research is described, which primarily includes the

design, the sampling, the measurement of variables and statistical analysis.

An Overview of the Women’s Health Studies

The women’s health studies upon which the current research study was built

were: the Healthy Ageing of Women Study (HOW), the Women’s Wellness Program

(WWP) and the chronic disease Women’s Wellness Program (CDWWP). Professor

Debra Anderson is the chief investigator of these projects.

The HOW study is a prospective, cross-cultural study comparing the

menopausal status, health related quality of life, as well as lifestyles between

Australian and Japanese midlife women. The data collection of the study occurred at

76

two sites, which were Queensland in Australia and Nagano in Japan, respectively.

The baseline study of the HOW was conducted in 2001, and its follow up study was

undertaken in 2006. The effect of country site of residence on menopausal status and

symptoms (Anderson, Yoshizawa, Gollschewski, Atogami, & Courtney, 2004), and

the difference in health related quality of life between women from these two

countries have been published earlier (Anderson & Yoshizawa, 2007). These

publications were based on the first wave of data collection. The longitudinal

relationships between lifestyle factors on mental health have not been explored

before this PhD study. For the current research, the Australian branch data from

2001 to 2006 were analysed.

The WWP study was designed following the HOW study. It utilised a

randomised controlled trial (RCT) design to evaluate the efficacy of a lifestyle

intervention for women living in the community. The sample of the WWP was

randomly selected from the HOW; therefore, it could be seen as an extension study

from the HOW. The intervention was of three-month duration, and has been proved

to be effective in improving women’s lifestyle and quality of life (Anderson et al.,

2006). Furthermore, it was also shown to have a sustainable effect in maintaining

positive lifestyle change (Smith-DiJulio & Anderson, 2009).

Extending from the WWP, which examined the treatment effect of a lifestyle

intervention for women from the general population, the CDWWP attempted to

replicate the lifestyle intervention with midlife and older women with a chronic

disease. As well, the CDWWP is a RCT, it has an intervention running over a period

of three months and is led by nurses. Up to the time of writing, the data collection of

the CDWWP was still ongoing; therefore, the efficacy of the intervention was yet to

be evaluated. A comparison of sociodemographic factors, lifestyle factors and

77

chronic conditions between women with and without Type 2 diabetes was presented

in the Women’s Health 2010: The 18th Annual Congress (Anderson, Lang, & Xu,

2010).

From the description, it can be seen that there is continuity among these

women’s health studies themselves in the sense of applying the knowledge obtained

from previous studies into the next. Based on the women’s health studies, the current

research specifically examined the data in terms of the relationships between

lifestyle factors and mental health in this population, which has not been studied in

this population prior to this research. The next section describes the design of the

current study, along with an explanation of the linkage between the current study and

these women’s health studies.

Linkage of the Current Research With HOW, WWP and CDWWP

The correlations between studies in the current research and these women’s

health projects are depicted in Figure 3.1. As can be seen from the graph, Study 1

was based on the HOW study; Study 2 was built on the WWP and CDWWP; and

Study 3 study was undertaken with the CDWWP.

78

Figure 3.1.The research design and its relation to women’s health studies.

Methodology of Study 1

Design

Study 1 used the Australian branch data of the HOW study and consequently

adopted a prospective longitudinal design. The first wave data (baseline) was

collected in 2001, and in 2006, women participating in the study in 2001 were

invited to take part in the second round of the survey (follow up).

Sample

Women who took part in both waves of the HOW study, were selected as the

study sample of Study 1. The sampling procedure of Study 1 (HOW) is described

below. In 2001, six postcode areas in South East Queensland, Australia were

selected as the districts of sampling. A balanced sample from both rural and urban

areas was obtained. Broad inclusion criteria were applied, which were: 1) female; 2)

45 to 60 years of age; and 3) being able to communicate in English. This initial

procedure identified 10,923 eligible women in total, and of those, 1,500 women were

randomly selected and invited to participate in the HOW study. Eventually, 886

79

women were recruited in 2001 (baseline), with a response rate of 59%. The detailed

information of the recruitment procedure was also published by Anderson and her

colleagues (Anderson et al., 2004). In 2006, another round of survey was undertaken

and questionnaires were delivered to women who responded in 2001. 564 women

completed and returned their questionnaires. The final response rate for the HOW

study was 37.6%. For women not replying at in 2006, 2 went overseas, 3 were

deceased, 28 did not wish to participate, 104 had changed addresses, a further 13

could not be traced, and an additional 172 did not reply. The details of the recruiting

process are shown in the flow chart below (Figure 3.2).

Figure 3.2. The flow chart of sample recruitment for Study 1.

As noted earlier, the purpose of Study 1 was to examine the relationships

between four lifestyle factors and mental health among midlife and older women. A

80

multiple linear regression model was chosen as the statistical approach. Lifestyle

factors together with other sociodemographic factors were treated as independent

variables. Categorical independent variables were coded as dummy variables before

being entered into the regression model. The total number of predictors was 25 for

the baseline analysis and 26 for the follow-up analysis, individually.

Based on Cohen’s (1988) calculation, the sample size to detect a medium to

large effect for regression equals to ‘50+8k’ (where k is the number of predictors)

for 80% power and at 0.05 significance level, if testing model is the primary aim; or

equals to ‘104+k’ if examining the effect of individual predictor is the goal. Study 1

aimed to test both model fitness and individual effect; therefore, the larger sample

size generated by the given formulas was chosen as the minimal number of women

required for Study 1 to produce reliable results. Having stated the number of

predictors, the minimal adequate sample size was 250 for baseline analysis and 258

for follow-up analysis, respectively. The real sample size of Study 1 was 564, which

was well over the targeted number.

Ethical Clearance

The Queensland University of Technology Human Research Ethics

Committee approved the Study 1.

Measures

Sociodemographic Factors

The measured sociodemographic factors in Study 1 were: age, marital status,

country of birth, ethnic origin, language spoken at home other than English,

81

education level, employment status and annual household income. All of the

variables were self-reported.

For marital status, women were categorised into three groups: married,

separated and single. With reference to multicultural background, women were also

asked about their countries of birth, whether it was Australia or other countries.

Women were also asked “do you identify yourself as an Aboriginal, Torres Strait or

South Sea Islander”. A ‘yes’ or ‘no’ answer was required. Then a question about

whether they speak another language rather than English at home followed, provided

with ‘yes’ or ‘no’ optional answers as well. With regard to education level, women

were given four options, which were junior school or under junior school, senior

school, university or technological institutes, and other. There were three categories

under employment status in Study 1, which included full-time paid employment,

part-time paid employment, and unemployed. Finally, the annual household income

of women was investigated. Women were offered three options, which were less

than or equal to $40,000, more than $40,000, and do not know.

Lifestyle Factors

As discussed in Chapter 2, four lifestyle factors were examined in the current

research study. The measurement of each lifestyle factor in Study 1 is detailed

below.

Weight status

In this research, weight status was measured with BMI, which is a very

widely applied instrument in the field of obesity research (Atlantis & Baker, 2008;

Cassidy et al., 2005; Flegal, Graubard, Williamson, & Gail, 2007; Gariepy et al.,

82

2009). The threshold of BMI, as proposed by the WHO is classified as follows: less

than or equal to 18.49 is underweight; 18.5 to 24.9 is normal weight, from 25.0 to

29.9 is overweight, and equal to or more than 30 is obese (World Health

Organisation, 2000a). This research study used this classification standard to define

women’s weight status. Despite wide acceptance of the classification standard of

BMI among researchers, it is believed that BMI is a continuous variable by nature. In

addition, analysing BMI as a continuous variable offers a few statistical advantages.

First treating BMI as a continuous variable increases precision of results due to

greater statistical power; second, results will be more informative and easier to

interpret than when using it as a categorical variable; and third, there will be greater

parsimony in the statistical model because of reduced number of indicators (Stanley,

2008).

A self-report method of height and weight was applied in all of the women’s

health studies and thus was chosen as the method for the current research study. In

terms of the reliability of self-reported height and weight, research found people tend

to over report their height, and under report their weight, intentionally or not;

therefore the value of BMI is underestimated. Consequently, the prevalence of

overweight and obesity is under estimated (Elgar & Stewart, 2008; Gorber,

Tremblay, Moher, & Gorber, 2007). Nevertheless, self-reported height and weight is

still a common method seen in much research, as it is relatively easy to obtain (Elgar

& Stewart, 2008). Despite the prevalence of unreliable self-reporting, one women’s

study showed that there was substantial agreement between self-reported and

measured BMI, except for women who are pregnant, older than 75 years or without

visit to a physician, and well-educated (Craig & Adams, 2009).

83

Physical activity

In Study 1, which is based on the HOW project, physical activity was

measured by asking the women a single question, which was “how often do you

exercise (including tai chi, fitness, yoga, walking, swimming) every week to improve

your health?” Four options were given, which included “none”, “1-2 times/week”,

“3-4 times/week” and “5-6 times /week” to indicate these women’s level of exercise.

As can be seen from the question, what was measured was essentially exercise,

rather than physical activity. The definition of physical activity is any bodily

movement produced by skeletal muscles that result in energy expenditure

significantly beyond resting level, particularly involving continuous actions of large

muscles (Briffa et al., 2006). Exercise is only a type of physical activity carried out

for a specific purpose. Therefore, although universally cited as physical activity, it

should be noted that in Study 1, physical activity actually was a reflection of

exercise.

Alcohol use

In Study 1, women’s alcohol use was evaluated by asking women a single

question, which was “have you ever drunk alcohol-containing beverages?” Four

choices were given, which were “never”, “drank in the past”, “occasionally” and

“regularly”.

Smoking

In the current study, smoking status was self-reported, and classified into

“non-smoker”, “past-smoker” and “current smokers”.

84

Confounders

For Study 1, in addition to the sociodemographic factors described above,

women’s menopause status was examined, and was included in the analysis for its

potential confounding effect.

Menopausal status

Menopause status was identified by asking women questions about their

menstruation period in consecutive order. The questions were 1) “Have you had a

hysterectomy, an operation to remove your uterus or womb?” 2) “Have you had both

ovaries removed?” 3) “Have you had a menstrual period in the past 12 months?” 4)

“Have you had a menstrual period in the past 3 months?” and 5) “Compared to a

year ago, has the number of days between menstrual periods become less

predictable?”

It was intended that, based on the response to these questions, women were

going to be classified into the four stages as described below. If a woman answered

yes to Question 1 or 2, then she would be regarded as in surgical menopause, which

means this woman had had a hysterectomy or ovaries moved. If a woman answered

yes to Questions 3 and 4, but no to Question 5, then she was thought to be

premenopausal, which was characterised as having no irregularity in her periods in

the previous 12 months and menstruating in the previous 3 months. Furthermore, if a

woman answered no to Question 3 then she would be considered to be

postmenopausal (naturally), which featured no menses for 12 or more months

without having had a hysterectomy and ovaries removed. Besides the three stages of

surgical menopause, premenopause and postmenopause, if a woman answered yes to

Question 3 she was considered to be in perimenopause. Menopausal status was

85

assessed in all the women’s health studies and hence was a measurable variable for

the current study too.

Mental Health

In the current research, the mental health measured was general mental health

and two common psychological problems which were depression and anxiety. The

measurement for each variable is detailed as below.

General mental health measurement

The measurement of general mental health for this study was derived from

the SF-36 (see Appendix B). SF-36 itself is a well validated and widely used quality

of life measurement tool, and consisted of eight scales entitled physical function (PF),

role physical (RP), bodily pain (BP), general health (GH), vitality (VT), social

function (SF), role emotional (RE) and mental health (MH), which in total were

eight scales. One of the eight scales measuring mental health was the Mental Health

Inventory (MHI), which contains 5 items. The norm values of the eight scales for

various populations were published (Ware, Kosinski, & Gandek, 2000b). SF-36

could also be aggregated into two composite scores based on the values of the eight

scales, which are the physical composite score (PCS) and mental composite score

(MCS). In the current study, both MHI and MCS were used to describe women’s

general mental health.

Mental Health Inventory

The MHI is a short version of the original MHI, which contained 38 items.

MHI was constructed from the 5 items that best predicted the summary score for the

86

38-item MHI. It consisted of one or more items from each of the major mental health

dimensions, which include anxiety, depression, loss of behavioural or emotional

control and psychological well-being (Ware, Kosinski, & Gandek, 2000a). The

correlation between the simple sum of the 5-item MHI and the 38-item MHI was

0.95 (Ware, Kosinski, et al., 2000a). The 5-item MHI is comprised of five questions

about one’s mental health experience in the past four weeks, covering both positive

and negative aspects of mental health. The questions were how much time during the

past four weeks: 1) Have you been a very nervous person? 2) Have you felt so down

in the dumps that nothing could cheer you up? 3) Have you felt calm and peaceful? 4)

Have you felt downhearted and blue? and 5) Have you been a happy person? For

each question, women were asked to rate the frequency of the feeling experienced in

that duration. Six options were given with a coding value. In detail, they were “1 =

all of the time”, “2 = most of the time”, “3 = a good bit of the time”, “4 = some of

the time”, “5 = a little of the time” and “6 = none of the time”. The final score for

each item aligned with their coding values except for Questions 3 and 5, whose score

was the reverse of the coding value. The raw score of MHI was the sum of the final

score of five items. To allow comparison with the norm values of mental health and

previous research, a transformed score of MHI was calculated and presented. The

formula for the calculation of the transformed score was:

Transformed Score = [actual raw score – lowest possible raw score] x 100 /

possible raw range

For MHI, the lowest and highest possible raw scores were 5 and 30,

respectively; therefore, the possible raw score range was 25. This transformation

87

converted the raw score to a 0 to 100 scale, with a higher score indicating better

mental health status.

Mental Composite Scores

As discussed, the MCS is one of two aggregate summary measures of SF-36.

It has been a common measure of mental health for individuals with different

diseases (Leese et al., 2008; Salaffi et al., 2009; Walsh et al., 2006) and general

populations (Harkonmäki, Lahelma, Martikainen, Rahkonen, & Silventoinen, 2006)

as well. The MCS is a weighted sum of eight standardised scales based on the 1998

general U.S population. The three steps of scoring, the standardised mean values of

eight scales, and the loading factors were presented by Ware and colleagues (Ware,

Kosinski, & Dewey, 2000). As compared to the eight scales, the aggregate MCS

score achieved a number of advantages including better precision, and reduced floor

and ceiling effects (Ware, Kosinski, & Dewey, 2000). These advantages were

confirmed by a later study carried out by Gandek and associates within a large study

population (Gandek, Sinclair, Kosinski, & Ware, 2004). The internally consistent

reliability reported in Gandek’s study was 0.89 and a greater elimination of floor and

ceiling effects was seen for composite scores.

Anxiety and Depression

The symptoms of anxiety and depression were measured by the

psychological subscale of the Greene Climacteric Scale (GCS-P), see Appendix C.

The GCS is a self-report questionnaire that measures a total of 21 physical,

psychological and vasomotor symptoms associated with the menopause transition. It

was developed by Greene in 1998, when there was a demand for standardised

88

measurement of climacteric symptoms (Greene, 1998). In application, women were

asked to give immediate response to the items according to their feeling at that

moment. This scale has been validated in a number of different populations, with a

Cronbach’s α of 0.91 (Chen, Davis, Wong, & Lam, 2010; Travers, O'Neill, King,

Battistutta, & Khoo, 2005). It has been used as a tool in studies to detect the efficacy

of intervention on climacteric symptoms (Ziaei, Moghasemi, & Faghihzadeh, 2010)

and to monitor the progress of climacteric symptoms resulting from surgery (Collaris,

Sidhu, & Chan, 2010). With regard to the GCS-P, which was the subscale of interest

of the current study, its two week test-retest reliability was 0.87 (Greene, 1998), and

the internal consistency was from 0.87 to 0.90 (Chen et al., 2010; Travers et al.,

2005). For the anxiety and depression scale, the internal consistency was 0.83 and

0.82, respectively. Scoring of GCS was simple. Each item of the GCS is rated by the

participant according to its severity using a four-point scale ranging from “0 = not at

all” to “3 = extremely”. The final score of each subscale is the sum of all the items

under that subscale. The GCS-P contains eleven items, with items 1 to 6 measuring

anxiety and items 7 to 11 measuring depression. Therefore, the possible score range

was 0-18 for anxiety, and 0-15 for depression, respectively. A higher score means

more severe anxiety and depression symptoms. In addition, the use of a total score of

GCS-P was also applied, and a higher total suggests more psychological symptoms.

It was suggested by Greene that the GCS-P could be used to identify menopausal

women who are severely anxious or have clinical anxiety and/or depression using a

cutoff point of 10, which has been contrasted with the HADS (Greene, n.d.).

89

Statistical Analysis

In Study 1, as well as the following Study 2 and 3, the Statistical Package for

the Social Sciences (SPSS), version 16.0 was used for data entry and analysis.

Imputation of missing values was not attempted. The analysis was

undertaken among participants who did not have missing values on relevant

variables. Assumptions of relevant parametric statistical analysis were examined. A

descriptive analysis of sociodemographic factors, lifestyle factors and mental health

was undertaken first. Women who were analysed in Study 1 were compared to those

who dropped out in 2006. When comparing these two groups, an independent

sample t-test was used to detect the differences of continuous variables, such as BMI

and age, and a Chi-square analysis for categorical variables like country of birth.

Then the relationships among lifestyle factors were investigated. One-way

ANOVA analysis was used for the examination of the relationships between

continuous variables and categorical variables; and Chi-square analysis was applied

for the examination of two categorical variables.

Following this, a cross-sectional analysis of the relationships between

lifestyle factors and mental health was conducted. A multi linear regression analysis

(entered approach) was used to explore the correlations, controlling for

sociodemographic factors and menopause status. Categorical independent variables

were coded as dummy variables before entering the regression models.

Finally, a longitudinal prospective analysis about the long-term effect of

lifestyle factors on mental health was undertaken using multi linear regression

analysis, adjusting for baseline mental health. In this analysis, women’s mental

health in 2006 was the dependent variable, and lifestyle factors at baseline were

independent variables. Other confounders including sociodemographic factors and

90

menopausal status were adjusted. Alpha less than .05 was considered to be

significant.

91


Design

Study 2 adopted a cross-sectional design. The baseline data (pre-intervention)

of women who participated in the WWP and CDWWP projects was utilised. Data

used in Study 2 were drawn from these two pre-existing databases, and the student

did not collect new information.

Sample

The sample of the Study 2 was drawn from the WWP and CDWWP projects.

The sampling procedures of the WWP and CDWWP projects are described,

respectively. Starting with the WWP, among the 886 women who were considered

eligible to participate in the HOW study (also called the Queensland Midlife

Women’s Health Study), 240 women were randomly selected and invited to take part

in the WWP study. These women were mailed the consent form in addition to the

study information and the possibility of being allocated to the case or control group

was explained. In the end, 133 women consented. 39 women did not have complete

baseline data; therefore, for Study 2, 94 women with completed baseline data in

WWP were analysed. The sampling procedure has been reported (Anderson et al.,

2006). The inclusion criteria of the WWP study was the same as the HOW. To

reiterate, it was 1) being a female, 2) 45 to 60 years, and 3) being able to speak, read

and write in English.

Women who participated in the CDWWP study were consecutively recruited

from multiple community health centres in Brisbane, Queensland from October 2008

to March 2010. When women attended the diabetes education group session or when

they were referred to individual consultations with diabetes educators, they were

introduced to the study’s details, risk and benefits, and given the contact number of

92

research project officer for inquiry purposes. Women who met the following criteria

were considered eligible: 1) 45 years and older, 2) being able to communicate in

English, 3) without a diagnosis of breast cancer, 4) not currently receiving palliative

care, and 5) having no other contraindications of participating in this program. If

women were willing to take part in the research, they were asked to post the consent

form back with the provided pre-paid envelope, along with their contact information.

When the consent form was received, the baseline questionnaire was posted to

women with a pre-paid envelope for return purposes. 83 women completed and

returned the baseline questionnaires of the CDWWP, and these were used as part of

the study sample of Study 2.

Cohen’s power calculation (1988) strategy was applied in Study 2. As stated

early, the minimal sample size required to observe a medium to large effect for 80%

power at 0.05 significance level is 50+8k. Based on the findings of Study 1, it was

estimated that the number of predictors in Study 2 would be around 10; therefore, the

minimal sample size was 130. The real sample size of 177 in Study 2 surpassed the

expected number, hence was believed to be adequate to produce reliable study

findings.

Ethical Clearance

All the procedures and interventions of the WWP were approved by the

Queensland University of Technology Ethics Committee prior to the study beginning.

For CDWWP, the Queensland University of Technology Human Research Ethics

Committee and Human Research Ethics Committees at the Prince Charles Hospital

approved the implementation of the study (see Appendix G). Women who

participated in both programs had the right to withdraw from the study after they had

consented. No penalty or consequences were applied.

93

Measures


The sociodemographic factors measured in Study 2 were: age, marital status,

country of birth, ethnic origin, language spoken at home other than English,

education level, employment status and annual household income.

Due to the relatively small sample size, the marital status in Study 2 was

combined and re-categorised into two categories, which were being married and

unmarried. Country of birth was a dichotomous variable, divided into being born in

Australia or other countries. Women were also asked “Do you identify yourself as an

Aboriginal, Torres Strait or South Sea Islander”. A yes/no option was provided. This

question was followed by the one about language. Women were asked whether they

spoke a language other than English at home and provided with ‘yes’ or ‘no’ options

as well. With regard to education level, women were given three options, which were

junior school or less than junior school, senior school, and university or

technological institutes. Employment status was also a dichotomous variable, which

included being in paid employment and unpaid/unemployed. Finally, the annual

household income of women was investigated. Women were offered three options,

which were less than or equal to $40,000, more than $40,000, and do not know.

Lifestyle Factors

In the same way as in Study 1, four lifestyle factors: relative body weight,

physical activity, smoking and alcohol use were included in Study 2. The details of

the measurement of each of these lifestyle factors have been provided in Study 1 (see

p. 81-83).

94

Weight status

BMI was again used as the instrument of measuring women’s weight status.

The standard classification of BMI proposed by the WHO was implemented in Study

2 too: less than or equal to 18.49 was underweight; 18.5 to 24.9 was normal weight,

from 25.0 to 29.9 was overweight, and equal to or more than 30 was obese (World

Health Organisation, 2000a). In addition, BMI was also treated as a continuous

variable in analysis, because of the advantages of doing so (see Study 1, p. 82)

The BMI in Study 2 was self-reported as well. The advantages and

disadvantages of self-reported BMI as compared to objective measurement have

been discussed for Study 1 and therefore are not repeated here (see p. 82).

Physical activity

In Study 2, physical activity was measured by the Seattle Physical Activity

questionnaire (SPA) (Smith-Dijulio, via personal contact, see Appendix F). This

questionnaire was initially developed by the researchers from the Seattle Women’s

Health Study. The SPA questionnaire is a self-report scale consisting of three

questions relating to general daily activity, exercise, and rating of overall level of

physical activity. General daily activities that were asked included, but were not

limited to, activities such as housework, caring for children, shopping and so forth,

but not exercise. Four options were provided to women to indicate their level of

general daily activity, which were: “very active”; “moderately active”; “mildly

active”; and “sedentary”. Following the general daily activity section, the

information about exercise was obtained. The question for exercise was: “How many

times did you exercise for at least 15 minutes at a time in the past month?” Examples

of various types of exercises were given, such as callisthenics, jogging, racquet

95

sports, team sports, dance classes, brisk walking, lifting weights, yoga, tai chi and so

on. Five answers were provided, which were “daily”; “5-6 times a week”; “3-4 times

a week”; “1-2 times a week”; and “none”. Finally, women were requested to rate

their overall activity level out of a continuous scale from 0 (not at all active) to 10

(extremely active).

Additionally, the general daily physical activity item was modified slightly,

due to the relatively small numbers in Study 2. For this question, very active and

moderate levels of daily activity were grouped into one category, which was

renamed as “very active/moderate”. The mildly active and sedentary categories

remained unchanged.

Alcohol use

Alcohol use in the preceding week was investigated. Two questions were

used to describe the women’s drinking habits. The first one was “During the past

week, on how many days did you drink any alcohol/alcohol-containing beverages

such as beer, wine or liquor?” Women were asked to circle the number of days from

0 to 7. The followed question was “During the past week, on the days that you drank

alcoholic beverages, how many standard size drinks did you have on average”.

Along with this question, a brief explanation about one standard drink was provided

to reduce the likelihood of misreport. Women’s weekly consumption of

alcohol-containing beverage could be calculated based on the two questions. The

results were contrasted to the short-term and long-term risk drinking standard

published by the National Health Medical Research Council (2009).

96

Smoking

Smoking status was self-reported, and classified into “non-smoker”,

“past-smoker” and “current smokers”. Nicotine dependence was not examined, as it

is beyond the scope of the research.

Confounders

Menopausal status

The set of questions used in Study 1 were also used in Study 2. As discussed,

according responses to questions, the sample was classified into four categories,

which were premenopause, perimenopause, postmenopause, and surgical

menopause (for detailed information, see p. 84-85).

Number of co-morbidities

The following chronic conditions were investigated in Study 2. There has

been a slight difference in some of the chronic health conditions between the WWP

and the CDWWP. For example, the WWP did not include “endometrial cancer”,

while the CDWWP did. Therefore, only the chronic health conditions investigated in

both projects were examined in Study 2. The investigated health conditions were:

headaches/migraine, stroke, high blood pressure, leaking urine when coughing or

sneezing (stress incontinence), back problems, coronary heart diseases (angina, heart

attack, bypass surgery, angioplasty), other heart diseases (irregular beat, heart

failure), irritable bowel problem, thyroid disorder, arthritis or rheumatism, diabetes,

breast cancer, ovarian cancer, cancer (any type), osteoporosis, bone or joint problems

other than arthritis or osteoporosis, and mental health problems. For each of the

conditions, women were asked to give a yes or no answer. The total number of

97

health conditions that a person has was calculated as an indicator of physical health

status.

Mental Health

Similarly to Study 1, mental health measured in Study 2 also included

general mental health, and two common psychological symptoms including anxiety

and depression. The measurement of these mental health measures followed the

same procedures as Study 1 (see p. 85-88).

Briefly speaking, general mental health was measured by the MHI and MCS

in SF-36, psychological symptoms including anxiety and depression were measured

by the GCS-P. This psychological subscale could be further divided to give measures

of anxiety and depression.


Imputation of missing values in Study 2 was not performed either, given the

small number of missing values. Statistical analysis was performed on variables that

did not possess missing values. The assumptions of relevant statistical analysis were

tested prior to or after analysis. The statistical analysis of Study 2 involved three

steps. The first step was descriptive analysis of the sociodemographic factors,

lifestyle factors, mental health and confounders for the whole study sample, as well

as for women with and without diabetes separately. The difference of the

distributions of sociodemographic factors, lifestyle factors, mental health and

confounders between women with and without diabetes was compared. An

independent t-test was used to compare continuous and normal distributed variables

such as BMI; a Mann-Whitney analysis (non-parametric analysis) was used for

98

continuous but skewed data (e.g. number of co-morbidities); and a Chi-square

analysis was implemented for categorical variables such as employment status. The

aim of this step was to explore the factors that differed between women with and

without a diagnosis of diabetes. Factors differed between women with and without

diabetes were included in the final regression models.

Following that, an examination of the binary relationships between

dependent variables (mental health) and independent variables (sociodemographic

factors, lifestyle factors and confounding factors) was undertaken. A Pearson

correlation analysis was conducted for the examination of continuous independent

variables, and a one-way ANOVA was used to test the difference of mental health in

relation to categorical variables such as smoking and alcohol consumption. The

purpose of this step was to find out the significant factors influencing mental health.

These significant factors then were entered into a regression model together with the

ones indicated at the first step.

The third step was hierarchical regression analysis, with each measure of

mental health being a dependent variable. The predictors were generated from the

previous two steps and entered in the regression model subsequently. Categorical

variables were coded as dummy variables before entering the regression models. An

adjusted R2 was presented to indicate the contribution of different independent

variables to mental health. A significance level of having a p value less than .05 was

adopted.

99


Design

Study 3 was conducted with the CDWWP, and was of cross-sectional design.

Apart from the data collected by the CDWWP, the student introduced another two

measures to the program to allow the exploration of the mediating effect of

self-efficacy. One measure was for depression and anxiety, which was added to

enable effective comparison with other literature, and the other measure was

self-efficacy.

Sample

Study 3 used the same sample of the CDWWP, which gave totally 83 women.

The sampling strategies for Study 3 (CDWWP) have been described for Study 2 (see

p. 91-92) and are not repeated here.

According to the study findings of both Study 1 and Study 2, it was likely

that the number of predictors in regression model in Study 3 was 5. Study 3

primarily aimed to test models with self-efficacy as the potential mediator. Using

this estimation and Cohen’s (1988) formula for 80% power at the significance level

of 0.05, the expected minimal sample size for Study 3 was 90. As can be seen, the

power of Study 3 was slightly inadequate when compared to the expected sample

size, thus the results of Study 3 need to be interpreted with caution.

Ethical Clearance

The CDWWP was approved by the Queensland University of Technology

Human Research Ethics Committee and the Prince Charles Human Research Ethics

Committee (also see p. 92).

100

Measures


The sociodemographic factors measured in the Study 3 were identical to

those of Study 2. To reiterate, the factors included were age, marital status, country

of birth, ethnic origin, language spoken at home rather than English, education level,

employment status and annual house income. Age was self-reported. Marital status

was classified into married and unmarried. Country of birth included two options:

Australia or other countries. Ethnicity was self-defined by women as being of

indigenous origin or not. Language spoken at home rather than English was set as a

‘yes’ or ‘no’ question. Education was grouped into three categories including junior

school or less than junior school, senior school and university or institute of

technology. Employment was divided into paid employment and unemployment.

Finally, annual household income was levelled into ≤ $40,000, > $40,000, and do

not know.

Lifestyle Factors

The four lifestyle factors examined in the previous two studies were also

examined in Study 3. They were relative body weight, physical activity, smoking

and alcohol use. The measurement of these variables is briefly listed in this section.

Weight status

Self reported BMI was used in Study 3. For detailed information, see Study 1

(p. 81-82).

101

Physical activity

The SPA questionnaire was also used in the Study 3. The description of the

scale has been provided in Study 2 (see p. 94-95), and hence is not detailed here. The

modified version of the SPA was applied in the Study 3, with three questions asking

about general daily activity, exercise frequency in the past month and rating of

overall physical activity.

Alcohol use

In Study 3, the same two questions inquiring about women’s alcohol

consumption in the past week were adopted (see p. 95). In addition, a third question

was added. This question was “Is the amount you drank last week more, about the

same or less than you would usually drink?” By asking this question, women’s

alcohol use in the past week as compared to their usual drinking level was able to be

described. In other words, the question provided information on validity of the

previous two questions in measuring accurate alcohol use in women.

Smoking

Smoking status was evaluated. As in the previous two studies, women were

classified into “non-smoker”, “past-smoker” and “current smoker” groups.

Confounders

Menopausal status

The same questions were used in Study 3 to identify women’s menopausal

stage (see p. 84-85). As mentioned before, women were grouped into four categories,

102

which were premenopause, perimenopause, postmenopause and surgical

menopause.

Number of co-morbidities

The investigated health conditions in Study 3 included: headaches/migraine,

stroke, high blood pressure, leaking urine when coughing or sneezing (stress

incontinence), back problem, coronary heart diseases (angina, heart attack, bypass

surgery, angioplasty), other heart diseases (irregular beat, heart failure), irritable

bowel problem, thyroid disorder, arthritis or rheumatism, diabetes, breast cancer,

ovarian cancer, endometrial cancer, cancer (any type), osteoporosis, bone or joint

problem other than arthritis or osteoporosis, clinical depression, anxiety disorder and

other mental health problem as specified. For each of the conditions, a ‘yes’ or ‘no’

answer was given. In addition, the year of diagnosis for each health condition was

requested. The number of chronic health conditions, excluding diabetes, was

calculated.

Duration of Diabetes

The period that women have had diabetes for since diagnosis was calculated

by deducting the year of diagnosis the women provided from the year 2009. The

duration was measured to the nearest year.

Use of Antidepressants

In Study 3, women were also asked to list the medications that they were

taking. Based on the information, the usage of taking antidepressants was identified,

and women were grouped into ‘use’ and ‘not use’ categories.

103

Mental Health

Like Study 1 and Study 2, the mental health of midlife and older women with

diabetes had two meanings: general mental health and psychological symptoms

which included anxiety and depression. General mental health was measured in the

same way as Study 1 and Study 2, therefore not repeated here. Depression and

anxiety were measured by using the HADS in Study 3 (for the scale, see Appendix

D).

Anxiety and depression

Anxiety and depression was measured by the HADS, which was developed

as a self-administered psychological scale to identify and quantify anxiety and

depression in medical patients (Zigmond & Snaith, 1983). It contains two subscales,

which were anxiety (HADS-A) and depression subscales (HADS-D), with each of

them containing seven items. Each of the psychological symptoms in the past week

was assessed using a four-point Likert scale ranging from 0 to 3. Based on the

scoring protocol, the possible total score range was from 0 to 21 for depression and

anxiety, respectively. One distinctive feature in which the HADS differs from other

psychological scales is that it excludes symptoms such as dizziness, headaches,

insomnia, and fatigue, which could give false positive results if they were actually

caused by physical disorders (Herrmann, 1997). In addition, the HADS was

developed to observe milder psychiatric symptoms, and therefore avoided a “floor

effect”.

Because of the above stated advantages of the HADS, it has been used

extensively in studies in examining the relationships between physical illnesses and

psychological symptoms (Alati et al., 2004; Collins, Corcoran, & Perry, 2009; Doyle,

104

McGee, De La Harpe, Shelley, & Conroy, 2006; Dunbar et al., 2008; Fu et al., 2007;

Hildrum, Mykletun, Midthjell, Ismail, & Dahl, 2009). Moreover, its psychometric

properties also have been tested in a variety of medical patients (Honarmand &

Feinstein, 2009; Singer et al., 2009; Untas et al., 2009). The results of the research

showed that HADS is easy to administer, well validated, and sensitive to identifying

anxiety and depression.

There is no single universal cutoff score of the two subscales to identify

anxiety and depression, rather, the cutoff score varied according to the populations

studied (Honarmand & Feinstein, 2009; Singer, et al., 2009). For the current research,

a cutoff score of 8 (8+) was chosen to identify cases of anxiety and depression, as

this threshold has been shown to give an optimal balance between sensitivity and

specificity as contrasted against the International Statistical Classification of

Diseases and Related Health Problems (ICD-9) (Bjelland, Dahl, Haug &

Neckelmann, 2002). The review by Bjelland also indicated that the average

Cronbach’s Alpha was 0.83 for HADS-A, and 0.82 for HADS-D. Moreover, the

specificity and sensitivity was approximately 0.8 when using 8 as a cutoff score.

Although widely used as a screening tool to identify anxiety and depression

cases, it is believed that, by its nature, the HADS score is a continuous variable, with

a higher score suggesting more severe mental symptoms .The norm values of HADS

based on gender and age have been published (Hinz & Schwarz, 2001).

Two reasons were considered when selecting the HADS as the measurement

of anxiety and depression. When the HOW and WWP projects were carried out, they

had a study sample consisting of women exclusively aged from 45 to 60 years, as

this is the age range where menopausal transition occurs. In that circumstance, GCS

was believed to be the most appropriate tool for measuring climacteric symptoms,

105

among which psychological symptoms were included as well. In contrast to the

HOW and WWP, which studied midlife and older women from the general

community, women who participated in the CDWWP were patients seeking

treatment in a diabetic clinic. Moreover, these women were much older, as can be

seen from the inclusion criteria of age (≥ 45 years). Seeking treatment and being of

an older age suggested that this group of women are likely to live with more adverse

physical conditions than their younger counterparts in the HOW and WWP projects.

To avoid the “noise” from physical illnesses, the HADS was added to the survey

because of its performance in this regard.

Self-Efficacy

In Study 3, the Self-Efficacy for Managing Chronic Disease 6-Item Scale

(SEMCD, see Appendix E) was chosen as the measurement tool of self-efficacy for

the current research. This scale was selected because it fitted the context of the

research into the perspective of chronic disease management. The psychometric

properties of the scale are detailed as below.

SEMCD was developed by the researchers from the Chronic Disease

Self-Management Program, Stanford University, to measure people’s confidence in

keeping common health related problems caused by chronic disease from interfering

with their daily life (Lorig, Sobel, Ritter, Laurent, & Hobbs, 2001). These items

cover several domains of diseases management tasks commonly seen in many

chronic diseases: symptom control, role function, emotional functioning and

communicating with physicians. For example, one of the items is “How confident

are you that you can keep the fatigue caused by your disease from interfering with

the things you want to do?” Each item was assessed from a 1 to 10 continuous scale,

106

with 1 meaning “not at all confident” and 10 meaning “totally confident”. The score

for the scale is the mean of the six items.

SEMCD has been tested in 605 patients with various chronic diseases such as

arthritis, diabetes or heart disease, and has been approved to an internal consistency

reliability of 0.91 (Stanford Patient Education Research Centre). The mean efficacy

score for that study sample was 5.17 (SD = 2.22). This scale has been used

frequently in a range of self-management programs designed by Lorig and her team

(Lorig, Ritter, & Plant, 2005; Lorig et al., 2006; Lorig et al., 2001).


Missing data imputation was not performed in Study 3. All the statistical

analysis was based on variables that did not have missing values. Assumptions of

relevant statistics were examined as well. The mediating effect of self-efficacy in the

associations between mental health and lifestyle factors was examined by using

Baron and Kenny’s mediation analysis (Baron & Kenny, 1986). As suggested, four

assumptions need to be held to prove that a variable is the mediator of the

relationship between independent and outcome variables (see Figure 3.3). First, the

independent variable must affect the mediator (path a); second, the independent

variable must be shown to affect the outcome variable (path c); third, the mediator

must affect the dependent variable (path b), fourth, the effect of the independent

variable on the dependent variable is reduced when the mediator is controlled (path

c'). It is considered as a perfect mediation if the effect of the independent variable

disappears.

Based on these principles and the context of the current study, two analytical

models were formed to guide statistical analysis (see Figures 3.4 & 3.5).

107

Figure 3.3. Basic causal chains of mediation model.

Both models attempted to examine the mediating effect of self-efficacy in the

correlations between mental health and lifestyle factors. The difference was: the first

model treated lifestyle factors as outcome variables and mental health as predictors

(Figure 3.4); while the second model considered mental health as outcome variables

and lifestyle factors as predictors (Figure 3.5). In other words, the mediating effect of

self-efficacy in both directions of the relationship was tested.

Self-efficacy

Lifestylefactors

Mentalhealth

a b

c

c'

Figure 3.4.Mediation model using lifestyle factors as outcome variables.


108

Self-efficacy

Mentalhealth

Lifestylefactors

a b

c

c'

Figure 3.5.Mediation model using mental health as outcome variables.

According to the analytical models above, the statistical analysis was carried

out in such an order. First, the correlations between mental health variables (anxiety,

depression, general mental health) and lifestyle factors (BMI, physical activity,

smoking and alcohol use) were analysed first (path c). The purpose of this step was

to confirm the relationships between lifestyle factors and mental health, so there is an

effect to be mediated. Several statistical techniques were used at this step:

independent sample t-test for dichotomous and continuous variables, one-way

ANOVA analysis for categorical and continuous variables, and a Pearson correlation

for two continuous variables. Second, the correlations between mental health and

self-efficacy (path a in Model 1 and path b in Model 2) were explored by using

Pearson correlation. Third, the associations between self-efficacy and lifestyle

factors were tested (path a in Model 2 & path b in Model 1). After these three steps,

the mental health variables and lifestyle factors that fulfilled all of the three

conditions were put into the mediation analysis (multiple linear regression) at the last

step, to examine whether self-efficacy mediates the effect of lifestyle factors on

109

mental health, or the effect of mental health on lifestyle factors. A multiple linear

regression was applied.

Apart from the four elemental principles mentioned by Baron and Kenny,

potential confounders that could interfere with the mediation analysis were examined.

The correlations between self-efficacy and sociodemographic factors were examined.

Any significant factors that emerged from this step of analysis would be included in

the regression model as potential confounders. Again, categorical independent

variables were coded as dummy variables as they were in Study 1 and Study 2. All

the analysis was conducted in the SPSS 16.0. An alpha of 0.05 was used in the study.

Chapter Summary

The current study was built on a range of women’s health studies to address

the research questions. This research examined the pre-existing data of a number of

women’s health studies from a particular angle of the relationships between lifestyle

factors and mental health. Study 1 involved data-mining of the HOW study,

exploring the prospective longitudinal relationships between lifestyle factors and

mental health among midlife and older women. Study 2 compared pre-intervention

data from the WWP and CDWWP, and was aimed at determining the contributing

effect of lifestyle factors to the differences in mental health between women from the

general and clinical population (midlife and older women). Finally, Study 3 was

undertaken along with the CDWWP to examine whether self-efficacy has a

mediating effect on the association between mental health and lifestyle factors. The

measurement of variables for each study of the research was outlined.

The results of each study are displayed in from Chapter 4 to Chapter 6,

preceded with a brief review of the research questions for each study.

110

111

CHAPTER 4 RESULTS OF STUDY 1: THE RELATIONSHIPS BETWEEN

LIFESTYLE FACTORS AND MENTAL HEALTH AMONG AUSTRALIAN

MIDLIFE AND OLDER WOMEN

Introduction

In recent years, the modification of health related lifestyle factors was

considered as a crucial way to promote population health. Despite the widely

recognized adverse effect of unhealthy lifestyle factors on physical health, the

prevalence of unhealthy lifestyles has remained high, and constitutes a considerable

burden of the cost of the health system. For example, 60% of Australians were

identified as overweight or obese, which resulted in a direct cost of $830 million.

Studies have been carried out to examine the relationships between lifestyle factors

and psychological conditions, such as anxiety and depression (see chapter 2). Yet,

there is a lack of research on women in this area, although some research has shown

that women are at higher risk of having psychological problems than men. In

addition, the interactive effect among lifestyle factors has not been considered in

depth and most studies have been of cross-sectional design, which does not allow

exploration of temporal relationships between lifestyle factors and mental health.

Study 1 was designed to address these gaps. The research question of Study 1 is:

what is the relationship between lifestyle factors and mental health among Australian

midlife and older women.

Results of Study 1

Based on the HOW study, the Study 1 utilised 564 women who provided the

completed data for the study. The outcome measure of this phase of study was

112

mental health, and the predicting measures were lifestyle factors. Sociodemographic

factors and menopausal status were examined as confounders.

Characteristics of Sample

Totally, 564 women were analysed in Study 1. The mean age of the women

was 55 (SD = 2.76) years (see Table 4.1). Of the women, 78.1 % were married or in

a de facto relationship; 83.1% were born in Australia; 97.9% were non-aboriginal;

28.4% college educated; about 40.8% not in paid employment; 41.8% had a family

annual income of less than $40,000, and 77.5% were postmenopausal (49.4% natural

& 28.1% surgical). No significant differences were found between women who

completed the study and those who dropped out regarding sociodemographic

characteristics, with the exception of age (see Table 4.1). Data showed that women

who were lost to follow-up (54.36 ± 2.87 years) were slightly younger than those

who remained (54.95 ± 2.76 years) in the study, p = .003. Although the difference of

age was statistically significant, it was not believed that the difference of age would

produce a large effect on the outcome measures, as the absolute difference of age

was less than 1 year. In short, it was concluded that women who completed both

waves of the studies did not differ from those who dropped out, and hence were

representative of the study population.

With regard to women’s lifestyle, it was found that for women who

completed both waves of studies, 16.8% of women did not exercise; 61.2% were

non-smokers and 56.8% occasionally drank alcohol. The frequency distribution was

very similar between women who completed both waves of study and those who

dropped out, with no statistical difference identified (see Table 4.1).

In the analysed sample, the mean score for anxiety was 3.54 (SD = 2.89),

depression 3.03 (SD = 2.58), and psychological symptoms 6.53 (SD = 5.01). The

113

general mental health score obtained from MHI was 76.17 (SD = 17.28), and 46.62

(SD = 9.54) for MCS. Also, there was no difference between two groups.

There is a study that has been undertaken with regard to the norm value of

the GCS scale for women aged from 50 to 59 years. It showed that the norm value of

the GCS psychological subscale was 3.0 for anxiety, 3.0 for depression, and 6.0 for

psychological symptoms (Travers et al., 2005). In order to compare the finding of

Study 1 to Australian norm values, the median values for anxiety, depression and

psychological symptoms were calculated. The median values are as follows: 3.0 for

anxiety, 2.0 for depression and 6.0 for psychological symptoms, which was very

close to the norm.

Table 4.1

Descriptive Analysis of Study Sample

Variables Total

(N=886)

Analysed

(N=564)

Dropouts

(N=322)

p

Age 54.73± 2.82 54.95 ± 2.76 54.36 ± 2.87 .003

BMI 27.04 ± 5.89 27.09 ± 5.66 26.95 ± 6.27 .743

Marital status .226

Married/defacto 670 (76.2) 438 (78.1) 232 (73.0)

Separate 111 (12.6) 65 (11.6) 46 (14.5)

Single 98 (11.1) 58 (10.3) 40 (12.6)

Country of birth .462

Australia 723 (82.3) 466 (83.1) 257 (81.1)

Other 155 (17.7) 95 (16.9) 60 (18.9)

Aboriginal 1.000

Yes 19 (2.1) 12 (2.1) 7 (2.2)

No 867 (97.9) 553 (97.9) 314 (97.8)

Speak English at

home

.585

114

Variables Total

(N=886)

Analysed

(N=564)

Dropouts

(N=322)

p

Yes 62 (7.1) 37 (6.7) 25 (7.8)

No 813 (92.9) 518 (93.3) 295 (92.2)

Education .791

<=junior 462 (52.7) 295 (52.8) 167 (52.5)

Senor high 138 (15.7) 86 (15.4) 52 (16.4)

Uni./Tech 251 (28.6) 159 (28.4) 92 (28.9)

Other 26 (3.0) 19 (3.4) 7 (2.2)

Employment .612

Full-time 268 (32.1) 174 (32.4) 94 (31.6)

Part-time 233 (27.9) 144 (26.8) 89 (30.0)

Unpaid 333 (39.9) 219 (40.8) 114 (38.4)

Income .197

<=$40,000 378 (42.7) 236 (41.8) 142 (44.2)

>$40,000 358 (40.4) 240 (42.5) 118 (36.8)

Don’t know 150 (16.9) 89 (15.8) 61 (19.0)

Menopause status .494

Premenopause 60 (6.8) 41 (7.3) 19 (6.0)

Perimenopause 132 (15.1) 85 (15.2) 47 (14.8)

Postmenopause 423 (48.3) 276 (49.4) 147 (46.4)

Surgical 261 (29.8) 157 (28.1) 104 (32.8)

Exercise .413

None 145 (18.1) 86 (16.8) 59 (20.3)

1-2 times/w 239 (29.8) 154 (30.0) 85 (29.3)

3-4 times/w 240 (29.9) 151 (29.4) 89 (30.7)

5-6 times/w 179 (22.3) 122 (23.8) 57 (19.7)

Smoking .989

None 533 (61.1) 340 (61.2) 193 (61.1)

Past 244 (28.0) 156 (28.1) 88 (27.8)

Current 95 (10.9) 60 (10.8) 35 (11.1)

Alcohol .584

None 186 (21.3) 125 (22.4) 61 (19.2)

115

Variables Total

(N=886)

Analysed

(N=564)

Dropouts

(N=322)

p

Abstainer 40 (4.6) 24 (4.3) 16 (5.0)

Occasional 497 (56.8) 317 (56.8) 180 (56.8)

Regular 152 (17.2) 92 (16.5) 60 (18.9)

Anxiety 3.62 ± 2.88 3.54 ± 2.89 3.78 ± 2.89 .240

Depression 3.10 ± 2.65 3.03 ± 2.58 3.25 ± 2.78 .263

Psychological 6.67 ± 5.08 6.53 ± 5.01 6.92 ± 5.20 .293

MHI 75.69 ± 17.38 76.17 ± 17.28 74.84 ± 17.56 .284

MCS 46.48 ± 9.47 46.62 ± 9.54 46.58 ± 9.51 .853

The Correlations Among Lifestyle Factors

The correlations among the four lifestyle factors were analysed using

different statistical techniques according to the type of data. It can be seen from

Table 4.2 that among midlife women, weight status as measured by BMI was not

related to any of the other lifestyle factors, including physical activity, p = .117,

smoking, p = .608, and alcohol use p = .080, using a one-way ANOVA analysis.

This suggested that women who are obese were not necessarily less physically active,

more likely to smoke or drink alcohol.

Physical activity however, was strongly correlated with smoking, p = .012

and alcohol drinking, p = .009, using a Chi-square test. The data showed that the

proportion of women doing certain levels of physical activity was much lower

among current smokers than that in non-smokers (1-2 times/w: 26.1% vs. 30.9%; 3-4

times/w: 19.6% vs. 29.0%; 5-6 times/w: 17.4% vs. 25.5%), and the percentage of

women doing no physical activity was more than two folds in current smokers as

compared to that in non-smokers (37.0% vs. 14.6%). For drinking alcohol, it was

found that women who never drink alcohol had a higher percentage of non physical

116

activity (23.3% vs. 16.5%), and a lower percentage of exercising (16.4% vs. 29.4%)

than regular drinkers. The results suggested that women who drink alcohol on a

regular basis were more active than non-drinkers.

Smoking was significantly related with drinking alcohol, p < .001. A

Chi-square analysis revealed that there was a much higher proportion of regular

alcohol drinkers among current smokers than in non-smokers (20.3% vs. 9.8%).

Furthermore, the proportions of regular alcohol drinkers among past smokers and

non smokers were 30.0% vs. 9.8%, respectively.

To summarise, when analysing the intercorrelations among lifestyle factors,

BMI was not found to be related to physical activity, smoking and alcohol use.

Physical activity was negatively correlated with smoking, but positively with alcohol

use. In addition, smoking and alcohol use were strongly related, with the highest

proportion of regular alcohol drinkers being found among past-smokers, followed by

current smokers, and lowest in non-smokers.

117

Table 4.2

The Correlation Matrix Table of Lifestyle Factors, the Significance Level of Tests (p

values)

BMI Physical activity Smoking Alcohol

BMI1

.117

(1.973)^

.608

(0.498)^

.080

(2.268) ^

Physical

activity1

.012

(16.254)§.009

(21.865)§

Smoking 1< .001

(57.640)§

Alcohol 1

^: F value; §: Chi-square

Multiple Linear Regressions: Lifestyle Factors Predicting Mental Health at Baseline

To reiterate, this section described the results of the cross-sectional analysis

of baseline data. Mental health measures were the dependent variables and lifestyle

factors were the independent variables. Sociodemographic factors and menopausal

status were entered into regression models as confounders. Adjusted R2 was

presented.

Anxiety at Baseline as Dependent Variable

Lifestyle factors, sociodemographic factors and menopausal status together

explained 8.1% of the variance in anxiety as measured by the psychological subscale

of the GCS, F(25, 408) = 2.518, p < .001.

Two sociodemographic factors including age and language spoken at home

emerged to be significant factors in the regression model. Age was found to be

118

negatively related with anxiety symptoms in the study population, r = -.187, p = .001,

which suggested that older women tended to have less anxiety symptoms. Women

who do not speak English at home was found to have lower anxiety scores than

women who do, p = .006 (see Table 4.3).

Among the examined lifestyle factors, only physical activity was discovered

to be the significant predictor of anxiety. When compared to women who do not

exercise, women who exercise 5-6 times per week reported significantly less anxiety

symptoms, p = .010. In addition, menopausal stage was not found to be an

influencing factor in the model either.

Table 4.3

Multiple Linear Regression, Anxiety at Baseline as Dependent Variable (N = 433)

Variables B SE β p R2

Anxiety at baseline 8.1

Age -.187 3.660 -.187 .001

Marital status Married/de facto Ref

Separate/divorce .553 .447 .063 .216

Single/widow -.114 .431 -.013 .792

Country of birth Australia Ref

Other countries .025 .363 .003 .944

Aboriginality Yes Ref

No .776 .911 .040 .395

Language

spoken at home

English Ref

Others -1.711 .613 -.139 .006

Education ≤Junior school Ref

Senior school -.437 .401 -.055 .276

Uni./technology .043 .313 .007 .890

Others -.852 .711 -.058 .231

Employment Full-time Ref

119


Part-time .082 .347 .013 .813

Not paid .514 .352 .092 .145

Annual family

income

Don’t know Ref

≤ $40, 000 -.055 .564 -.010 .922

> $40, 000 -.634 .567 -.115 .264

BMI .034 .023 .071 .145

Physical

activity

None Ref

1-2 times/w .045 .384 .007 .906

3-4 times/w -.536 .382 -.088 .161

5-6 times/w -1.054 .406 -.161 .010

Smoking Never Ref

Past .296 .305 .049 .332

Current .339 .477 .035 .478

Alcohol Never Ref

Past-drinker .081 .725 .006 .911

Occasionally .188 .329 .034 .569

Regularly .167 .443 .023 .706

Menopausal Perimenopause Ref

Premenopause -1.045 .576 -.099 .071

Postmenopause -.440 .397 -.080 .269

Surgical -.399 .422 -.064 .345

Depression at Baseline as Dependent Variable

As shown in Table 4.4, the same group of independent variables

(sociodemographic factors, lifestyle factor and menopausal stage) explained 13.7%

of the variance in depression at baseline, F(25, 414) = 3.794, p < .001.

At this time, only age turned out to be the significant sociodemographic

factor of depression. Similar to the results about anxiety, age was again found to be

negatively related with depression, r = -.213, p < .001. The data revealed that older

120

women reported less depression symptoms after considering lifestyle factors and

menopausal stages Menopausal stage was found to be related with depression, with

women at postmenopause stage having a lower depression score when compared to

women at the perimenopausal stage.

Two lifestyle factors, including BMI and physical activity, were significant.

The relationship between BMI and depression was positive, r = .154, p = .001,

suggesting women with high BMI had more depression symptoms. Furthermore,

women doing exercise 5-6 times per week had less depressive symptoms when

compared to those doing no exercise (β = .187, p = .002). However, no difference

was found for women exercising at 3-4 times per week, and 1-2 times per week.

Table 4.4

Multiple Linear Regression, Depression at Baseline as Dependent Variable (N =

439)


Depression at baseline 13.7

Age -.202 .049 -.213 < .001



Single/widow .107 .394 .013 .786


Other countries -.139 .328 -.020 .672


No -.429 .829 -.024 .605

Language spoken at

home

English Ref

Others -.979 .527 -.089 .064


Senior school -.358 .362 -.049 .322

Uni./technology .212 .286 .038 .460

121


Others -.528 .628 -.039 .402


Part-time .164 .315 .028 .603

Not paid .284 .315 .054 .369

Annual house

income

Don’t know Ref

≤ $40, 000 .328 .481 .063 .497

> $40, 000 -.284 .487 -.055 .560

BMI .068 .021 .154 .001

Physical activity None Ref

1-2 times/w .281 .344 .048 .415

3-4 times/w -.381 .338 -.067 .260

5-6 times/w -1.156 .364 -.187 .002

Smoking Never Ref

Past .188 .278 .033 .499

Current .442 .424 .050 .298

Alcohol Never Ref

Past-drinker -.795 .682 -.058 .245

Occasionally -.437 .297 -.084 .143

Regularly -.320 .405 -.046 .430


Premenopause -.814 .508 -.085 .110

Postmenopause -.960 .363 -.186 .008

Surgical -.620 .381 -.107 .104

Psychological Symptoms at Baseline as Dependent Variable

The group of independent variables as mentioned previously accounted for

11.7% of the variance in psychological symptoms, F(25, 395) = 3.230, p < .001.

Again, age was showed to be a significant factor in predicting psychological

symptoms, r = -.219, p < .001. This result was consistent with previous findings with

regard to depression and anxiety. Taken together, age was still negatively correlated

122

with psychological symptoms among midlife women. Furthermore, women who do

not speak English at home had significantly lower scores of psychological symptoms

(β = -.102, p = .040). Menopausal stage was also found to be significant, with

women at premenopause having less psychological symptoms than women at the

perimenopausal stage (β = -.114, p = .039).

BMI and physical activity were again found to be significant, but not other

lifestyle factors, including smoking and alcohol consumption. BMI was positively

correlated with scores of psychological symptom, r = .112, p = .020, indicating more

psychological symptoms among women with high BMIs. Physical activity was also

shown to be significant, with less psychological symptoms being found among

women exercising 5-6 times per week and those not exercising, β = -.191, p = .002.

Table 4.5

Multiple Linear Regression, Psychological Symptoms at Baseline as Dependent

Variable (N = 420)


Psychological symptoms at baseline 11.7

Age -.393 .095 -.219 < .001



Single/widow -.016 .763 -.001 .984


Other countries -.160 .641 -.012 .802


No .259 1.594 .008 .871

Language spoken

at home

English Ref

Others -2.247 1.092 -.102 .040


Senior school -.935 .716 -.066 .193

123


Uni./technology .272 .555 .026 .625

Others -1.358 1.243 -.053 .275


Part-time .112 .615 .010 .856

Not paid .653 .625 .065 .296

Annual family

income

Don’t know Ref

≤ $40, 000 .131 .990 .013 .895

> $40, 000 -1.107 .992 -.113 .265

BMI .095 .041 .112 .020

Physical activity None Ref

1-2 times/w .404 .686 .037 .556

3-4 times/w -.916 .675 -.085 .176

5-6 times/w -2.219 .718 -.191 .002

Smoking Never Ref

Past .490 .540 .046 .365

Current .638 .838 .038 .447

Alcohol Never Ref

Past-drinker -1.098 1.315 -.043 .404

Occasionally -.323 .585 -.033 .581

Regularly -.393 .792 -.030 .620


Premenopause -2.099 1.011 -.114 .039

Postmenopause -1.354 .707 -.138 .056

Surgical -.933 .746 -.084 .212

Mental Health Inventory at Baseline as Dependent Variable

All the independent variables explained 7.8% of the variance in MHI score,

F(25, 421) = 2.504, p < .001.

Age was the only demographic factor that was shown to significantly

contribute to the variance in MHI scores, r = .198, p < .001. As revealed by the

124

results, older women’s general mental health status increases as they age.

Menopausal status and other sociodemographic factors were found to be irrelevant to

MHI scores.

The significant lifestyle factors in this model were BMI and physical activity.

BMI was negatively related with MHI scores, r = -.164, p = .001, suggesting that

women with higher BMIs had worse general mental health. The analysis also

showed that both women exercising 3-4 times per week, and 5-6 times per week

were found to have higher MHI scores, when compared to women who did not

exercise, p value equals to .018 and .001, respectively.

Table 4.6

Multiple Linear Regression, the Mental Health Inventory at Baseline as Dependent

Variable (N = 446)

Variables B SE Β p R2

MHI at baseline 7.8

Age 1.248 .336 .198 < .001


Separate/divorce -2.772 2.743 -.050 .313

Single/widow -.588 2.738 -.010 .830


Other countries .931 2.286 .020 .684


No -1.278 5.718 -.010 .823

Language

spoken at home

English Ref

Others 1.374 3.603 .019 .703


Senior school 2.217 2.448 .046 .366

Uni./technology 1.554 1.957 .042 .427

Others .295 4.452 .003 .947


125


Part-time -.126 2.148 -.003 .953

Not paid -.721 2.145 -.020 .737

Annual family

income

Don’t know Ref

≤ $40, 000 .245 3.277 .007 .940

> $40, 000 3.682 3.315 .107 .267

BMI -.488 .142 -.164 .001

Physical

activity

None Ref

1-2 times/w 2.305 2.362 .060 .330

3-4 times/w 5.609 2.353 .147 .018

5-6 times/w 8.572 2.519 .209 .001

Smoking Never Ref

Past -.074 1.924 -.002 .969

Current -.751 2.949 -.012 .799

Alcohol Never Ref

Past-drinker .657 4.545 .007 .885

Occasionally .282 2.018 .008 .889

Regularly -1.402 2.745 -.030 .610


Premenopause 5.906 3.500 .092 .092

Postmenopause 2.737 2.489 .079 .272

Surgical 2.097 2.623 .053 .424

Mental Composite Score at Baseline as Dependent Variable

With the same group of independent variables, when mental health was

measured by MCS, they only explained 4.8% of the variance in MCS scores, F(25,

362) = 1.789, p = .012.

Age remained a significant factor for MCS, with a positive relationship

indicated, r = .224, p < .001. Women at premenopause status had significantly higher

scores of MCS compared to those at perimenopause, β = .136, p = .021.

126

With regard to lifestyle factors, only BMI was found to be significant, r =

-.105, p = .045, which suggested that the MCS scores decreased as their BMIs

increased. Other lifestyle factors including physical activity, smoking and alcohol

use were not of statistical significance.

Table 4.7

Multiple Linear Regression: the Mental Composite Scores at Baseline as Dependent

Variable (N = 387)


MCS at baseline 4.8

Age .764 .198 .224 < .001


Separate/divorce -2.616 1.680 -.084 .120

Single/widow -.922 1.576 -.031 .559


Other countries .831 1.394 .033 .552


No .426 3.368 .006 .899

Language

spoken at home

English Ref

Others .125 2.303 .003 .957


Senior school 1.351 1.478 .050 .361

Uni./technology .500 1.162 .025 .667

Others .585 2.639 .012 .825


Part-time -.209 1.267 -.010 .869

Not paid -1.260 1.293 -.065 .330

Annual family

income

Don’t know Ref

≤ $40, 000 .913 2.025 .048 .652

> $40, 000 1.587 2.035 .084 .436

BMI -.176 .087 -.105 .045

127


Physical

activity

None Ref

1-2 times/w -.976 1.440 -.047 .498

3-4 times/w 1.083 1.422 .052 .447

5-6 times/w 2.845 1.537 .127 .065

Smoking Never Ref

Past .368 1.137 .018 .746

Current -.461 1.917 -.013 .810

Alcohol Never Ref

Past-drinker -1.330 2.751 -.026 .629

Occasionally -.671 1.203 -.035 .577

Regularly -.432 1.639 -.017 .792


Premenopause 5.044 2.169 .136 .021

Postmenopause .547 1.472 .029 .710

Surgical 1.829 1.562 .084 .242

Brief Conclusion of the Cross-sectional Analysis

In short, analysis of the baseline data showed that after adjusting for

sociodemographic factors and menopausal status, physical activity was negatively

correlated with mental health. Furthermore, BMI had a positive relationship with

psychological symptoms including depression and general mental health, suggesting

a lower level of mental health in relation to high BMI. For the other two lifestyle

factors, smoking and alcohol use, no correlation was detected in relation to mental

health. The results suggest that women who are physically active have better mental

health, women with a higher BMI tend to have low levels of mental health, and there

is no difference in mental health in relation to smoking and alcohol drinking.

Additionally, age was showed to be a strong protective factor of women’s

mental health, as evidenced by a negative relationship with psychological symptoms

128

including anxiety and depression, and a positive one with general mental health.

Apart from the effect of age, menopausal status was also indicated to be a significant

factor. The results showed that women at premenopause and postmenopause had

better mental health than those who are at perimenopausal stage.

Multiple Linear Regressions: the Prospective Relationships Between Lifestyle

Factors and Mental Health

This section describes the multiple linear regression models using mental

health variables at follow up as dependent variables. The analysis examines the

prospective effects of baseline lifestyle factors on mental health at follow up after

adjustment for baseline mental health and other confounders.

Anxiety at Follow up as a Dependent Variable

As shown in Table 4.8, when anxiety at baseline was included in regression

model, together with other demographic factors, lifestyle factors and menopausal

stage, totally explained 38.6% of the variance in anxiety at follow up, F(26, 379) =

10.788, p < .001.

In this model, country of birth, as well as women’s education level were

found to be significant predictors of anxiety at follow up. Statistically, women born

overseas has slightly more anxiety symptoms than those born in Australia, β = .086,

p = .040, and women who had senior school education presented less anxiety

symptoms than women with an education level under junior school, β = -.099, p

= .021.

Among lifestyle factors, only alcohol consumption was significant. Using

non-alcohol drinkers as a reference category, past-drinkers had significantly less

anxiety symptoms than non-drinkers at follow up, p = .040. Not surprisingly,

129

baseline anxiety had the highest correlation with anxiety at follow up in the whole

model, β = .623, p < .001. The interaction effect of alcohol consumption on anxiety

over time was also plotted in Figure 4.1.

So, adjusting for the baseline anxiety symptoms, only alcohol use was found

to be correlated with women’s anxiety symptoms at follow up.

Table 4.8

Multiple Linear Regression: Anxiety at Follow up as dependent variable (N = 405)


Anxiety at follow up 38.6

Age .049 .042 .054 .243


Separate/divorce -.324 .349 -.040 .355

Single/widow .107 .328 .014 .744




No -.339 .757 -.018 .654

Language

spoken at home

English Ref

Others .620 .488 .053 .205


Senior school -.727 .313 -.099 .021

Uni./technology -.411 .235 -.079 .081

Others .046 .525 .004 .930


Part-time -.346 .265 -.062 .192

Not paid -.155 .267 -.031 .562

Annual family

income

Don’t know Ref

≤ $40, 000 -.307 .442 -.062 .488

> $40, 000 -.413 .439 -.084 .347

BMI -.012 .018 -.026 .522

130


Physical

activity

None Ref

1-2 times/w .050 .288 .009 .863

3-4 times/w -.151 .292 -.027 .606

5-6 times/w -.213 .311 -.036 .494

Smoking Never Ref

Past .085 .229 .016 .709

Current -.297 .381 -.032 .436

Alcohol Never Ref

Past-drinker -1.121 .547 -.088 .041

Occasionally -.207 .254 -.041 .416

Regularly .044 .337 .007 .897


Premenopause -.493 .443 -.052 .266

Postmenopause -.395 .308 -.080 .200

Surgical -.091 .324 -.016 .779

Baseline anxiety .569 .038 .623 < .001

Depression at Follow up as Dependent Variable

The above described independent variables including baseline depression

explained 37.2% of the variance in depression at follow up, F(26, 391) = 10.491, p

< .001.

As indicated from Table 4.9, non sociodemographic factors were related with

depression at follow up. Similar results were found for lifestyle factors, none of

which was showed to be significantly related with depression at follow up (in 5

years). Baseline depression was the only significant predictor of follow up

depression, r = .605, p < .001.

131

Table 4.9

Multiple Linear Regression: Depression at Follow up as Dependent Variable (N =

417)


Depression at follow up 37.2

Age -.029 .043 -.031 .503



Single/widow .152 .338 .019 .653




No -.356 .692 -.020 .608

Language

spoken at home

English Ref

Others -.824 .479 -.073 .086


Senior school .010 .310 .001 .974

Uni./technology -.111 .243 -.020 .649

Others -.297 .541 -.023 .583


Part-time -.372 .269 -.065 .167

Not paid -.017 .269 -.003 .950

Annual family

income

Don’t know Ref

≤ $40, 000 -.155 .423 -.030 .715

> $40, 000 -.094 .425 -.019 .825

BMI .004 .018 .009 .820

Physical

activity

None Ref

1-2 times/w -.366 .298 -.064 .221

3-4 times/w -.404 .292 -.073 .167

5-6 times/w .006 .317 .001 .986

Smoking Never Ref

Past -.084 .236 -.015 .722

Current .455 .383 .049 .236

132


Alcohol Never Ref

Past-drinker .661 .588 .049 .262

Occasionally .137 .253 .027 .588

Regularly .044 .350 .006 .900


Premenopause -.010 .438 -.001 .982

Postmenopause -.229 .312 -.045 .464

Surgical -.403 .326 -.071 .217

Baseline

depression

.596 .043 .605 < .001

Psychological Symptoms at Follow up as Dependent Variable

The accounted variance of psychological symptoms at follow up was 42.5%,

F(26, 359) = 11.398, p < .001, using the socio-demographic, lifestyle factors,

menopausal stage and baseline psychological symptoms.

In this model, as shown in Table 4.10, there was a difference in psychological

symptoms in five years in relation to country of birth, with women born overseas

having slightly increased psychological symptoms, β = .089, p = .032. Apart from

this, only baseline psychological symptoms was the significant predictor, r = .664, p

< .001. Psychological symptoms at follow up did not differ according to lifestyle

factors and menopausal symptoms.

133

Table 4.10

Multiple Linear Regressions: Psychological Symptoms at Follow up as Dependent

Variable (N = 385)


Psychological symptoms at follow up 42.5

Age .038 .078 .023 .622


Separate/divorce -.231 .648 -.015 .721

Single/widow .200 .599 .014 .738


Other countries 1.081 .503 .089 .032


No -.852 1.358 -.025 .531

Language

spoken at home

English Ref

Others .044 .912 .002 .961


Senior school -.596 .572 -.044 .298

Uni./technology -.539 .432 -.056 .212

Others .052 .969 .002 .957


Part-time -.778 .490 -.075 .113

Not paid -.178 .488 -.019 .715

Annual family

income

Don’t know Ref

≤ $40, 000 -.642 .811 -.070 .429

> $40, 000 -.700 .805 -.077 .385

BMI -.007 .033 -.009 .833

Physical

activity

None Ref

1-2 times/w -.318 .536 -.031 .554

3-4 times/w -.549 .534 -.054 .304

5-6 times/w -.298 .571 -.028 .602

Smoking Never Ref

Past .000 .420 .000 .999

Current .042 .706 .002 .952

134


Alcohol Never Ref

Past-drinker -.284 1.053 -.012 .787

Occasionally .008 .463 .001 .986

Regularly .286 .626 .023 .648


Premenopause -.505 .799 -.030 .527

Postmenopause -.368 .565 -.040 .515

Surgical -.202 .592 -.020 .734

Baseline

psychological.628 .040 .664 < .001

Mental Health Inventory at Follow up as Dependent Variable

First, baseline MHI score with other independent variables explained 37.4%

of the variance in the MHI score at five years’ follow up, F(26, 407) = 10.951, p

< .001. Baseline MHI likewise was the most strong predictor of MHI scores at

follow up, with a positive relationship, r = .602, p < .001.

For general mental health, as measured by the MHI in this model, smoking

emerged to be a significant factor of MHI at follow up. The results showed that

current smokers had as much as 7.225 points lower MHI at follow up compared to

non-smokers, p = .003. No difference was detected between past-smokers and

non-smokers. Interestingly, BMI was found to be a significant and protective factor

of MHI at follow up, r = .107, p = .009. The effect of smoking on MHI measured

mental health over time was also plotted in Figure 4.2.

135

Table 4.11

Multiple Linear Regressions: the Mental Health Inventory Scores at Follow up as

Dependent Variable (N = 433)


MHI at follow up 37.4

Age -.214 .274 -.035 .436


Separate/divorce .563 2.212 .011 .799

Single/widow -3.884 2.179 -.072 .075


Other countries -.836 1.815 -.019 .645


No -1.597 4.503 -.014 .723

Language

spoken at home

English Ref

Others 4.916 2.879 .070 .088


Senior school -.346 1.958 -.007 .860

Uni./technology .976 1.560 .028 .532

Others 1.527 3.508 .017 .664


Part-time 1.641 1.705 .044 .336

Not paid -.794 1.704 -.024 .641

Annual family

income

Don’t know Ref

≤ $40, 000 4.092 2.651 .123 .123

> $40, 000 2.161 2.678 .066 .420

BMI .312 .119 .107 .009

Physical

activity

None Ref

1-2 times/w 1.268 1.892 .034 .503

3-4 times/w 1.543 1.896 .042 .416

5-6 times/w 1.273 2.050 .033 .535

Smoking Never Ref

Past 1.282 1.524 .035 .401

Current -7.225 2.415 -.121 .003

136


Alcohol Never Ref

Past-drinker -1.541 3.582 -.018 .667

Occasionally -1.993 1.616 -.060 .218

Regularly -3.304 2.202 -.075 .134


Premenopause 4.499 2.775 .074 .106

Postmenopause 3.158 1.980 .096 .111

Surgical 4.021 2.087 .106 .055

Baseline MHI .573 .039 .602 < .001

Mental Composite Score at Follow up as Dependent Variable

Regression analysis showed that socio-demographic, lifestyle factors, and

menopausal stages together with MCS at follow up scores explained 25.8% of the

variance in MCS at follow up, F(26, 320) = 5.618, p < .001.

Marital status and language spoken at home were significant predictors of

MCS at follow up. As shown in Table 4.12, over time, women who were single

tended to score 4 points lower in MCS than those who were married, p = .020.

Regarding language spoken at home, women not speaking English at home had 5

more points in MCS at follow up, p = .029.

Smoking continued to be a significant factor for MCS scores at follow up,

with current smokers having a significantly lower score of MCS than non-smokers, β

= -.103, p = .036. Similar to the finding about BMI and MHI scores at follow up,

BMI was again discovered to be a significant factor of general mental health, as

measured by MCS. The r value was .281, p = .004. Both of the results conveyed the

information that women with high BMIs were more likely to have a better general

mental health status. The effect of smoking on MCS over time was also displayed in

Figure 4.3.

137

Table 4.12

Multiple linear regressions: Mental Composite Score at Follow up as Dependent

Variable (N = 346)


MCS at follow up 25.8

Age -.084 .215 -.021 .695


Separate/divorce -.307 1.817 -.008 .866

Single/widow -3.940 1.687 -.115 .020


Other countries .263 1.494 .009 .860


No 2.000 3.638 .026 .583

Language

spoken at home

English Ref

Others 5.712 2.608 .112 .029


Senior school -.643 1.594 -.021 .687

Uni./technology -.486 1.249 -.022 .697

Others 1.857 2.777 .033 .504


Part-time .339 1.366 .014 .804

Not paid -.246 1.386 -.011 .859

Annual family

income

Don’t know Ref

≤ $40, 000 3.216 2.294 .148 .162

> $40, 000 3.166 2.281 .147 .166

BMI .281 .097 .144 .004

Physical

activity

None Ref

1-2 times/w -.326 1.558 -.014 .834

3-4 times/w -1.176 1.568 -.050 .454

5-6 times/w -1.314 1.673 -.052 .433

Smoking Never Ref

Past -.668 1.208 -.028 .581

Current -4.436 2.110 -.103 .036

138


Alcohol Never Ref

Past-drinker -1.013 3.121 -.017 .746

Occasionally -.721 1.322 -.033 .586

Regularly -.419 1.787 -.015 .815


Premenopause 4.351 2.308 .105 .060

Postmenopause 3.293 1.560 .153 .036

Surgical 3.634 1.657 .148 .029

Baseline MCS .574 .056 .499 < .001

Further Examination of the Effect of BMI as Women Age

Study from our research showed a significant and positive relationship

between BMI and general mental health (MHI & MCS) at follow up. Such results

have not been consistent with the findings from previous studies (Kasen et al., 2008;

Roberts, Deleger, Strawbridge, & Kaplan, 2003). To confirm the obtained results,

the difference of the change of mental health variables between baseline and follow

up were compared using a one-way ANOVA analysis. The results shown in Table

4.13 were consistent with what was found in regression models that explored the

prospective effect of lifestyle factors mental health. In brief, BMI had no correlations

with anxiety, depression and psychological symptoms, but it was significantly

related to general mental health measured by MHI (p = .020) and MCS (p = .010).

As indicated by the Post Hoc tests, women who were obese at baseline had a

significantly higher increase in general mental health scores, suggesting that the

general mental health of obese women did improve over time, while the general

mental health of women with normal weight remained much the same.

139

Table 4.13

A Comparison of the Changing Scores of Mental Health Among Women With

Different BMI values (N = 385 ~ 492)

Mental health Weight

categories

N Change scores

(f-b)^

SE F p

Depression Under weight 10 -.40 ± 1.17 .371 1.010 .388

Normal weight 187 -.15 ± 2.09 .153

Overweight 151 -.06 ± 2.21 .180

Obese 125 -.50 ± 2.53 .226

Anxiety Under weight 10 -1.40 ± 2.37 .748 1.416 .237

Normal weight 180 -.14 ± 2.04 .152

Overweight 152 -.43 ± 2.62 .213

Obese 116 -.53 ± 2.43 .226

Psychological Under weight 10 -1.80 ± 2.57 .814 .987 .399

Normal weight 172 -.35 ± 3.53 .270

Overweight 139 -.58 ± 3.92 .333

Obese 115 -1.01 ± 4.43 .413

MHI Under weight 12 -2.67± 12.34 3.56 3.282 .021

Normal weight 196 -.27± 13.54 .968

Overweight 162 .91± 16.23 1.28

Obese 123 4.81± 15.67a, b 1.41

MCS Under weight 7 7.46 ± 10.01 3.78 3.840 .010

Normal weight 164 3.58 ± 9.36 .731

Overweight 128 5.83 ± 10.62c .939

Obese 87 8.04 ± 11.11 1.19

Note. a significantly higher than normal weight; b significantly higher than overweight; c significantly

higher than normal weight. ^ f-b: follow up scores minus baseline.

140

Figure 4.1. The impact of alcohol use on the change of anxiety scores over 5 years.

Figure 4.2. The impact of smoking on the change of MHI scores over 5 years.

141

Figure 4.3. The impact of smoking on the change of MCS scores over 5 years.

Since depression and psychological symptoms were not affected by any of

the measured lifestyle factors, a paired t-test was used to compare the mean. As

shown in Table 4.14, women’s depression and psychological scores decreased over

the follow-up period, with the improvement of psychological symptoms being

statistically significant, p = .001.

Table 4.14

Paired t-Test of Depression and Psychological Symptoms Between Baseline and 5

Years Follow Up (N = 454 ~ 493)

Variables Baseline Follow up Mean difference t p

Depression 3.00 ± 2.58 2.81 ± 2.46 0.19 ± 2.23 1.888 .060

Psychological

symptoms

6.35 ± 4.93 5.75 ± 4.59 0.60 ± 3.90 3.302 .001

142

Brief Conclusion of the Prospective Analysis

The longitudinal prospective analysis showed that: 1) women who were past

drinkers had less anxiety symptoms as they age than non-drinkers; 2) no significant

lifestyle factor was determined in predicting depression and psychological symptoms

at follow up, yet there was a trend toward significance for physical activity; 3)

current smokers had a significantly lower levels of general mental health when

compared to non-smokers, no matter how general mental health was measured; and 4)

women who were obese at baseline experienced a better improvement of general

mental health over time than women with normal weight.

Reflection on Research Questions

The major findings of the study 1 are summarized as below:

1. What are relationships between lifestyle factors and mental health among midlife

and older women?

The research found that BMI was positively related with scores of

depression and psychological symptoms, and negatively related with general

mental health scores (MHI). These results suggested that women with a higher

BMI had a lower level of mental health.

Second, women exercising daily had less psychological symptoms

including anxiety and depression, and better general mental health status than

women who did not exercise. In short, daily physical activity improves the

mental health of midlife and older women.

2) What is the long-term effect of lifestyle factors and mental health among midlife

and older women?

143

First, women who were past-drinkers had less anxiety symptoms as they

age than non-drinkers. No difference was found between alcohol drinkers and

non-drinkers.

Second, there was a clear trend that women who did exercise had less

anxiety, depression and psychological symptoms when compared to women who

did not exercise, although the difference was not statistically significant.

Third, women who were current smokers were found to have lower

general mental health scores overtime compared to women who were

non-smokers.

Fourth, BMI was found to be positively related with general mental

health scores at follow up, suggesting that women with a high BMI were more

likely to have higher levels of mental health as they get older.

3) Unexpected findings

Age was found to be negatively related with psychological symptoms

including anxiety and depression, and positively related with general mental

health. This result indicated that as women get older, their mental health

improves.

144

145

CHAPTER 5 RESULTS OF STUDY 2: THE RELATIONSHIPS BETWEEN

LIFESTYLE FACTORS AND MENTAL HEALTH AMONG AUSTRALIAN

MIDLIFE AND OLDER WOMEN WITH AND WITHOUT DIABETES

Introduction

Study 2 can be seen as an extension of a recent meta-analysis which

examines the prevalence of depression among individuals with diabetes (Ali et al.,

2006). This review paper showed that women with diabetes were more likely to have

depression than those without; however, as highlighted by the authors, these two

groups often differed in factors that are associated with depression, such as obesity in

many studies included in that review. Extending from this review study, Study 2 was

designed to explore the difference of mental health levels between midlife and older

women with and without diabetes and the contributing factors to the difference in

mental health quality. It is hypothesised that lifestyle factors may affect the mental

health levels of women with diabetes. If this hypothesis is proved, the knowledge

generated from the study can be used for health professionals or researchers to

develop programs to reduce the gap.

In total, 177 women were analysed in Study 2, with 94 from the WWP study

and 83 women from the CDWWP. In this study, the outcome measures were mental

health, and the independent variables were lifestyle factors and diabetes. Examined

covariates included sociodemographic factors, menopausal status and number of

co-morbidities. A hierarchical linear regression was used as the main analysis

method for Study 2.

146

Results of Study 2

Description of Sociodemographic Characteristics of the Sample

Age

The mean age of the whole sample was 60 years (SD = 8.08), with the

women having a diagnosis of diabetes being significantly older than those without

diabetes, 65.22 years (SD = 9.02) and 56.38 years (SD = 3.97), respectively.

Marital status

Over 60% of the women were currently married. The distributions of women

who were currently married and not married were similar between the diabetes and

non-diabetes groups, with non statistical significance being found, p = .218.

Education

54% of the women had an education level at senior school or beyond. Of

those, 70% of the women had received a college or technological school education.

When examining the frequency of education levels in relation to diabetes, it was

found that the distribution of three education levels was very close and the

Chi-square test failed to detect any significant difference, p = .243.

Employment

With regard to the employment status, there was a significant difference

between women with and without diabetes, p < .001. Among women with diabetes,

about 75% of them were not in paid employment (either full-time or part-time),

while in women without diabetes, only 45.3% women were in this category. Across

147

the whole sample, about 40% of women were currently employed, with nearly 60%

unemployed.

Annual income

Forty percent (40%) of the women reported an annual household income of

more than $40,000, with more than half reporting less than $40,000. A small number

of the women chose not to answer this question. The data revealed that the annual

income status was higher among women without diabetes than those with diabetes,

with a higher proportion of women having an annual income over $40,000. There

was a trend towards significance for this variable, p = .051.

Country of birth

Approximately 70% women in Study 2 were Australian born, and the

distribution of Australian and non-Australian born women were both similar to the

overall levels. The proportion of overseas born Australians in this sample (23.7%)

was very close to that of the national population (22.0%) (Australian Bureau of

Statistics, 2006a).

Aboriginality

Only 1 woman identified herself as indigenous.

Language spoken at home other than English

Ninety three percent (93%) of the women spoke English at home. The

corresponding proportions in women with and without diabetes were similar to this

value, and no differences were found between groups, p = .253.

148

Menopausal status

The majority (87%) of women were postmenopausal (50.6% naturally and

37.9% surgically). The proportion of women at perimenopause among women

without diabetes was twice as high when compared with women with diabetes. In

addition, a higher percentage of women who were in surgical menopause was seen in

women with diabetes. However, the difference in distribution of menopausal status

was not significant, p = .111.

In summary, women with diabetes were older and less likely to be employed

than those who did not have diabetes.

Table 5.1

The Characteristics of Sociodemographic Factors of Women With and Without

Diabetes (N = 176)

Diabetes

(N = 81)

Non-diabetes

(N = 95)

Overall

Mean / N (%)

Chi-squa

re/tP

Age 65.22 ± 9.02 56.38 ± 3.97 60.45 ± 8.08 8.17 <.001

Marital status 1.515 .218

Married 47 (57.3) 63 (66.3) 110 (62.1)

Not married 35 (42.7) 32 (33.7) 67 (37.9)

Education 2.830 .243

≤Junior school 35 (42.7) 44 (49.4) 79 (46.2)

Senior school 18 (22.0) 11 (12.4) 29 (17.0)

Uni/Technology 29 (35.4) 34 (38.2) 63 (36.8)

Employment 16.796 <.001

Employed 20 (24.4) 52 (54.7) 72 (40.7)

149

Diabetes

(N = 81)

Non-diabetes

(N = 95)

Overall

Mean / N (%)

Chi-squa

re/tP

Unemployed 62 (75.6) 43 (45.3) 105 (59.3)

Annual income 5.937 .051

≤ $ 40,000 50 (64.1) 42 (45.7) 92 (54.1)

> $ 40,000 25 (32.1) 46 (50.0) 71 (41.8)

Don’t know 3 (3.8) 4 (4.3) 7 (4.1)

Country of birth .037 .848

Australia 62 (75.6) 73 (76.8) 135 (76.3)

Other countries 20 (24.4) 22 (23.2) 42 (23.7)

Aboriginality 1.180 .277

Yes 1 (1.2) 0 (0.0) 1 (0.6)

No 80 (98.8) 95 (100.0) 175 (99.4)

Languages 1.306 .253

English 8 (9.8) 5 (5.3) 13 (7.3)

Others 74 (90.2) 90 (94.7) 164 (92.7)

Menopausal status 6.014 .111

Premenopause 1 (1.3) 1 (1.1) 2 (1.1)

Perimenopause 5 (6.3) 13 (13.7) 18 (10.3)

Postmenopause 36 (45.6) 52 (54.7) 88 (50.6)

Surgical 37 (46.8) 29 (30.5) 66 (37.9)

150

Number of Co-morbidities

The distribution of the number of co-morbidities (excluding diabetes) was

positively skewed, with a median value of 2, suggesting that half of women in Study

2 had more than two chronic conditions aside from diabetes. To compare the number

of co-morbidities between women with and without diabetes, a Mann-Whitney was

undertaken to detect the difference because of the properties of distribution of the

variable. As seen from Table 5.2, the median number of chronic diseases for women

without diabetes was 1, while the corresponding value for women with diabetes was

3, which was significantly more, Z score = -5.363, p < .001.

Table 5.2

The Number of Co-morbidities (excluding diabetes) in Women With and Without

Diabetes (N = 177)

N Median value Range Z score p

Women with diabetes 81 3 0-13 -5.363 < .001

Women without diabetes 95 1 0-6

The Description and Comparison of Mental Health and Lifestyle Factors of Women

With and Without Diabetes

An Examination of the Data Distribution

The distribution of mental health variables including anxiety, depression,

psychological symptoms and general mental health (as measured by MHI & MCS)

were firstly examined for the purpose of statistical method selection. As indicated by

the Skewness and Kurtosis values (see Table 5.3), the distribution of all mental

health variables except anxiety, although not perfectly normally distributed, was still

quite close to normal, hence an independent sample t-test was used for comparison

between the diabetes and non-diabetes groups. For anxiety, the distribution deviated

151

more significantly from normality than the other mental health variables. Despite

non-normal distribution of anxiety scores, an independent sample t-test was still

applied for two reasons. First, the transformation technique (square root) failed to

improve the distribution of data, and in addition it increased the difficulty of data

interpretation. Second, when both parametric (independent sample t-test) and

non-parametric analyses (Mann-Whitley) were run, it was found that the significance

levels were the same, with both being non-significant (p = .068 vs. p = .133). In short,

when compared to the crude difference of all of the mental health variables between

women with and without diabetes, an independent sample t-test was applied.

Table 5.3

The Statistics of Normality Examination of Mental Health Variables (N = 177)

Mental health variables Mean SD Skewness (SE) Kurtosis (SE)

Anxiety 3.77 2.81 1.04 (.183) 1.78 (.364)

Depression 3.27 2.51 .795 (.183) .138 (.363)

Psychological symptoms 7.02 2.86 .784 (.183) .321 (.364)

MHI 74.02 16.73 -.785 (.083) .271 (.364)

MCS 48.49 12.80 -.922 (.185) -.097 (.368)

The Difference in Mental Health Between Women With and Without Diabetes

As shown in Table 5.4, women without diabetes tended to have less severe

psychological symptoms including depression, and better general mental health

status. The anxiety score for women with diabetes was 4.19 (SD = 3.17), as

compared to 3.41 (SD = 2.43) for women without diabetes, and no significance was

indicated (p = .068). For depression, the score for women with diabetes was 3.78

(SD = 2.73), and for those without was 2.83 (SD = 2.24), t = 2.506, p = .013. The

scores of psychological symptoms were also significantly higher in women with

152

diabetes, which was mostly contributed to by elevated depression scores. The

psychological scores were 7.94 (SD = 5.39) and 6.24 (SD = 4.22) for women with

diabetes and without, respectively, t = 2.294, p = .023. In relation to measures of

general mental health, it was revealed that women with diabetes had an MHI score of

71.26 (SD = 17.73), which was 5 points lower than the score for women without

diabetes 76.38 (SD = 15.54), t = -2.041, p = .043. Yet, for MCS, no statistically

significant differences were found between the two groups, although women with

diabetes had a lower score of MCS (46.51 ± 12.90 vs. 50.01 ± 12.54), t = -1.839, p

= .068.

In short, women with a diagnosis of diabetes had more psychological

symptoms, including depression, and worse general mental health when compared to

women without diabetes. There was no difference in anxiety symptoms between

these groups.

Table 5.4

The Differences in Mental Health Between Women With and Without Diabetes (N =

177)

Diabetes N Mean SD t p

Anxiety Yes 81 4.19 3.17 1.834 .068

No 95 3.41 2.43

Depression Yes 82 3.78 2.73 2.506 .013

No 95 2.83 2.24

Psychological Yes 81 7.94 5.39 2.294 .023

No 95 6.24 4.22

MHI Yes 81 71.26 17.73 -2.041 .043

No 95 76.38 15.54

MCS Yes 77 46.51 12.90 -1.839 .068

No 95 50.01 12.54

153

The Difference in Lifestyle Factors Between Women With and Without Diabetes

Weight status

The mean BMI of women with diabetes was 33.05 (SD = 7.18), suggesting

that the majority of the women with diabetes were obese. In contrast, the mean BMI

of women without diabetes was 27.90 (SD = 6.41), which was significantly lower

than their counterparts, t = 4.968, p < .001. The WHO standard was applied to

classify women based on their BMI. It was found that only one woman was

classified as underweight (without diabetes). Therefore, when comparing the

proportions of different weight categories between women with and without diabetes,

underweight was combined into normal weight level. As seen in Table 5.5, the

proportion of women having a normal range BMI in the non-diabetes group was

36.8%, which was 5 times higher than 7.3% in the diabetes group. While for obesity,

the proportion for women with diabetes was twice as much as that of women without

diabetes, 62.2% and 33.7%, respectively. The proportions of women being

overweight were quite similar between diabetic and non-diabetic groups (30.5% vs.

29.5%). In brief, a comparative analysis showed that obesity was a more severe

problem for women with diabetes, χ² = 24.207, p < .001.

Physical activity

Physical activity was measured by three dimensions, which included general

daily activity, exercise and rating of overall physical activity. For general daily

activity, there were four categories in the questionnaire, yet due to the small number

in “very active” and “moderately active” levels, they were combined into one

category before analysis. The percentages of women doing moderate to very active

levels of daily activity was 40.7% and 66.7% for women with and without diabetes,

http://en.wikipedia.org/wiki/Chi_(letter)

154

respectively. On the other hand, the percentage of women having a sedentary

lifestyle was 19.8% for the group with diabetes, as compared to 5.4% for the

non-diabetic group. The level of general daily activity differed significantly between

the two groups, χ² = 14.476, p = .001. In terms of the frequency of exercising for at

least 15 minutes, it was found that half of the women with diabetes did not take any

kind of exercise in the past month, whereas only a quarter of women without

diabetes reported not to. Moreover, the percentage of women doing daily exercise in

the non-diabetic group was 10.6%, which was twice as many as that in women with

diabetes (5.0%). The proportions of the rest of the exercise categories were

consistently higher in women without diabetes than those with diabetes. Similar to

general daily activity, the difference of exercise levels between the two groups also

achieved a level of statistical significance, χ² = 13.142, p = .011. Furthermore,

women’s self-reported rating of overall physical activity was 3.57 (SD = 2.10) for

women without diabetes, which was about 2 points lower than 5.16 for women

without diabetes (SD = 2.15), t = -4.855, p < .001. In summary, women with diabetes

had a significantly lower level of physical activity than women without diabetes.

Smoking

Referring to Table 5.5, it was found that there was a higher prevalence of

non-smokers among women without diabetes than those with diabetes: 82.1% vs.

69.5%, respectively. The proportion of past smokers among women with diabetes

was 26.8%, which was substantially higher than 2.1% in women without diabetes. In

contrast, the prevalence of current smokers was much higher in the non-diabetic

group, 15.8% vs. 3.7%, respectively. Statistically, there was a significant difference

in smoking habits between women with and without diabetes, χ² = 27.125, p < .001.




155

The data showed that a reasonable number of women with diabetes used to be

smokers, but have quit smoking when they entered the study. Smoking is more of an

issue for women without diabetes.

Alcohol use

The days of alcohol drinking in the past week were positively skewed,

consequently, a median value was reported and a non-parametric test of two

independent samples (Mann-Whitney analysis) was undertaken to compare the

difference. The median value of the number of days drinking alcohol was 0 for

women with diabetes, and 1 for women without diabetes, Z score = -2.570, p = .010.

The number of standard drinks (daily and weekly) was calculated for each woman

and contrasted with the Australian Alcohol Guidelines to determine the women’s

short-term and long-term risk in relation to alcohol consumption (National Health

and Medical Council (NHMRC), 2001). According to this guideline for women,

short-term risk categories included low risk: “up to 4 drinks on any day, no more

than 3 days per week”, risky: “5 to 6 drinks on any one day” and high risk: “7 or

more drinks on any one day”. The long-term risk categories included low risk: “up to

14 drinks per week”, risky: “15 to 28 drinks per week”, and high risk: “29 or more

per week”. Using these standards, about 2% of women in the whole sample were

drinking at a short-term risk level (risky/high risk) and 4.5% at a long-term risk level

(risky/high risk). There was no difference in the prevalence of short-term and

long-term risk between women with and without diabetes, as suggested by Fisher’s

exact test, p = .338 and p = .071, respectively.

156

Table 5.5

The Differences in Lifestyle Factors Between Women With and Without Diabetes (N

= 177)

Diabetes Test of

significance

p

Yes (N = 82) No (N = 95)

BMI (continuous) 33.05 ± 7.18 27.90 ± 6.41 4.968 < .001

BMI (categorical) 24.207 < .001

18.5-24.9 6 (7.3) 35 (36.8)

25.0-29.9 25 (30.5) 28 (29.5)

≥30.0 51(62.2) 32 (33.7)

General daily activity 14.476 .001

Very/moderate 33 (40.7) 62 (66.7)

Mild 32 (39.5) 26 (28.0)

Sedentary 16 (19.8) 5 (5.4)

Exercise 13.142 .011

Daily 4 (5.0) 10 (10.6)

5-6 times/week 8 (10.0) 12 (12.8)

3-4 times/week 15 (18.8) 22 (23.4)

1-2 times/week 13 (16.2) 27 (28.7)

None 40 (50.0) 23 (24.5)

Overall activity 3.57 ± 2.10 5.16 ± 2.15 -4.855 < .001

Smoking 27.125 < .001

Non-smokers 57 (69.5) 78 (82.1)

Past smokers 22 (26.8) 2 (2.1)

Current smoker 3 (3.7) 15 (15.8)

Alcohol -2.570 .010

Days of drinking 0 1

Short-term .338

Low risk 79 (96.3) 94 (98.8)

Risky & high risk 3 (3.7) 1 (1.1)

Long-term .071

Low risk 80 (98.8) 88 (92.6)

Risky & high risky 1 (1.2) 7 (7.4)

157

The Difference of Eight Scales and Composite Score in SF-36 Between Women With

and Without Diabetes

The transformed scores of eight SF-36 scales, as well as the two composite

scores were calculated and presented for women with and without diabetes,

separately. To ensure the accuracy of MCS scores, correlations between two

composite scores and eight scale scores were analysed as an approach to examine

potential errors (Ware, Kosinski, & Dewey, 2000). The results were displayed in

Table 5.6. The results showed that PF, RP, BP scales correlated highest with PCS

and lowest with MCS; and the MH, RE and SF correlated highest with MCS and

lowest with PCS, which was consistent with correlations described by Ware (Ware,

Kosinski, & Dewey, 2000). Therefore, it was believed that the MCS score was

measured accurately.

Table 5.6

Scoring Check: the Correlations Between Eight Scales and Composite Scores of

SF-36 (N = 172 ~ 175)

PF RP BP GH VT SF RE MH

PCS .850** .811** .820** .712** .480** .448** .096 .062

MCS .116 .300** .127 .325** .600** .622** .864** .845**

**p < .001

The difference of general mental health varied with the measurement

including MHI and MCS. As shown in Table 5.7, the MHI scores for women with

diabetes were 71.26 (SD = 17.73), which were significantly lower than 76.38 for

women without diabetes (SD = 5.54), t = -2.041, p = .043. When measured by MCS,

the mental health score for women with diabetes was 46.51 (SD = 12.90), and the

corresponding scores for women without diabetes were 50.09 (SD = 12.54), yet no

158

statistical significance was achieved, t = -1.839, p = .068. For the rest of the seven

scales and PCS, women with diabetes consistently showed a lower score than

women without diabetes. All the differences were of statistical significance.

159

Table 5.7

The Differences in Quality of Life Between Women With and Without Diabetes (N = 176)

SF-36 scales Diabetic (N = 81) Non-diabetic (N = 95) Test of

significance

p

Mean SD Mean SDEight scales

Physical function (PF) 58.44 27.94 77.95 19.59 -5.252 < .001

Role physical (RP) 55.56 42.20 77.89 35.52 -3.762 < .001

Bodily pain (BP) 57.94 27.27 69.35 22.49 -3.007 .003

General health (GH) 54.13 20.26 72.28 19.40 -6.045 < .001

Vitality (VT) 48.29 21.44 58.37 20.62 -3.182 .002

Social function (SF) 71.19 27.83 83.42 23.73 -3.156 .002

Role emotional (RE) 67.92 39.85 82.46 33.63 -2.618 .010

Mental health (MH) 71.26 17.73 76.38 15.54 -2.041 .043

Composite scores

PCS 39.91 13.53 48.59 10.56 -4.605 < .001

MCS 46.51 12.90 50.09 12.54 -1.839 .068

160

The Binary Analysis Between Independent Variables and Dependent Variables

Before conducting regression analysis of the relationships between lifestyle

factors and mental health, binary correlations between independent variables

including lifestyle factors and diabetes and dependent variables, which were five

mental health measurements, were undertaken. The correlations between

confounding variables including sociodemographic factors, number of

co-morbidities and menopausal status, and the dependent variables related to mental

health status were also analysed. This step provided a crude description of the

associations between dependent variables and independent and confounding

variables. Significant variables were further examined in hierarchical regression

models.

Sociodemographic Factors and Mental Health

The associations between sociodemographic factors and five mental health

measurements were analysed first. A Pearson correlation was carried out to assess

the relationships between age and mental health variables, and results showed no

correlation between age and any of the five mental health variables. For other

dichotomous sociodemographic factors including marital status, employment,

country of birth and language spoken at home, an independent sample t-test was run.

Similarly to the results for age, no difference in mental health was seen in relation to

these sociodemographic factors. Lastly, a one-way ANOVA analysis was undertaken

for variables of annual income and education, which both had more than two

categories. Again, no statistical difference was found in relation to mental health.

The significant levels of the above tests were displayed in Table 5.8. It was found

that there was no correlation between sociodemographic factors and mental health

among midlife and older women with diabetes.

161

Table 5.8

Level of Significance (p values) of the Relationships Between Sociodemographic

Factors and Mental Health (N = 176)

Anxiety Depression Psychological MHI MCS

Age a .996 .537 .761 .604 .693

Marital status b .676 .675 .622 .292 .129

Employment b .079 .784 .262 .181 .490

Country of birth b .731 .281 .401 .454 .417

Language other

than English b

.511 .773 .489 .289 .209

Annual income c .337 .543 .714 .289 .411

Education c .389 .317 .312 .575 .984aPearson correlation; bIndependent sample t-test; cOne-way ANOVA

Lifestyle Factors and Mental Health

Following the analysis of the relationships between sociodemographic factors

and mental health, a parallel analysis was undertaken for the relationships between

lifestyle factors and mental health variables. For BMI, mild and positive correlations

were found between BMI and psychological symptoms including anxiety and

depression, but not for general mental health (MHI & MCS). The correlation

coefficient values for anxiety, depression and psychological symptoms were .175 (p

= .022), .186 (p = .015) and .202 (p = .008), respectively.

Regarding physical activity, the analysis showed that general daily activity

was only negatively related with depression, F(2, 171) = 3.170, p = .044, with active

women having less depression symptoms. There was a similar trend for anxiety,

psychological symptoms and general mental health, but no significant differences

were found. Exercise was analysed as a continuous variable in the analysis using a

Pearson correlation analysis. It was shown that exercise was significantly associated

162

with all of the mental health measures, except anxiety. The strength of the

correlations was mild, suggesting that a high level of exercise is related to better

mental health (see Table 5.9 for correlation coefficients). In terms of the rating of

overall physical activity, significant results were revealed in the relationships with all

mental health measures. The rating of overall physical activity was negatively related

with psychological symptoms including anxiety (r = -.252, p = .001) and depression

(r = -.306, p < .001). Also, it was positively correlated with MHI (r = .209, p = .006)

and MCS (r = .196, p = .011) measurements.

Significant results were found between smoking status and psychological

symptoms including depression (see Table 5.9). Interestingly, past smokers had the

highest scores of depression, F(2, 174) = 6.047, p = .003, and psychological

symptoms, F(2, 173) = 4.300, p = .015. No difference was found for anxiety and

general mental health (MHI & MCS).

For alcohol drinking, both long-term and short-term risk categories were

examined. But as described in the previous section, the prevalence of both long-term

and short-term drinking in this sample was quite low (2% to 4%), which caused an

inadequate sample size in these categories. As a result, although an independent

sample t test was still run, it was believed that the results were unlikely to be reliable

due to insufficient sample sizes in risky drinking groups.

Based on the above results, it was decided that BMI, physical activity and

smoking would be included in the subsequent hierarchical regressions models.

Alcohol use was accordingly excluded from regression.

163

Table 5.9

The Relationships Between Lifestyle Factors and Mental Health Variables (N = 170 ~ 177)

GCS-Anxiety GCS-Depression GCS-P MHI MCS

BMI .175* .186 * .202 ** -.128 -.061

General physical activity

Very & moderate 3.46 ± 2.72 2.80 ± 2.40* 6.26 ± 4.72 76.00 ± 16.99 50.02 ± 12.87

Mild 3.98 ± 3.09 3.78 ± 2.57 7.72 ± 5.05 72.55 ± 16.96 47.07 ± 12.99

Sedentary 4.38 ± 2.46 3.67 ± 2.56 8.04 ± 4.61 69.20 ± 15.31 45.10 ± 12.25

Exercise .118 .234** .187 * -.205 ** -.194 *

Overall rating -.252*** -.306*** -.304*** .209** .196*

Smoking

Non-smokers 3.67 ± 2.88 2.96 ± 2.45** 6.63 ± 4.84* 74.55 ± 17.13 49.09 ± 12.89

Past 4.96 ± 2.80 4.83 ± 2.78 9.74 ± 4.74 71.50± 16.86 44.36 ± 12.34

Current 3.00 ± 1.85 3.50 ± 2.64 6.50 ± 4.16 73.41 ± 13.63 49.15 ± 12.30

Alcohol short-term

Low risk 3.71 ± 2.78 3.34 ± 2.52 6.93 ± 4.82 74.33 ± 16.67 48.71 ± 12.75

Risky & high risk 6.25 ± 3.30 4.75 ± 2.22 11.00 ± 5.29 61.00 ± 16.12 39.33 ± 12.94

Alcohol short-term

Low risk 3.74 ± 2.80 3.28 ± 2.53 6.98 ± 4.85 74.48 ± 16.80 48.50 ± 12.85

Risky & high risk 4.13 ± 3.36 3.00 ± 2.20 7.13 ± 5.25 73.50 ± 15.26 49.63 ± 12.60

*p < .05; **p < .01; ***p < .001

164

Table 5.10

The Correlations Between Number of Co-morbidities, Menopausal Status and Mental Health Variables (N = 168 ~ 176)

Anxiety Depression Psychological MHI MCS

# of co-morbidities .383*** .443*** .452*** -.276*** -.193*

Menopausal status

Premenopause 3.50 ± 3.54 4.50 ± 3.54 8.00 ± 7.07 70.00 ± 31.11 44.83 ± 15.75

Perimenopause 3.72 ± 2.54 2.89 ± 1.97 6.61 ± 3.97 77.33± 14.58 51.71 ± 10.84

Postmenopause 3.61 ± 2.86 3.01± 2.39 6.63 ± 4.81 73.79± 17.16 47.86 ± 13.50

Surgical menopause 4.03 ± 2.89 3.68 ± 2.79 7.68 ± 5.18 73.76± 16.46 48.87 ± 12.70

*p < .05; ***p < .001

165

As shown in Table 5.10, the number of co-morbidities was found to be

strongly correlated with all the mental health variables. The magnitudes of the

correlations have been mild to moderate, suggesting that women living with more

chronic conditions had poorer mental health.

No significant result was found for menopausal status in relation to mental

health. Therefore, the number of co-morbidities would be included as a confounding

variable in regression models.

Predicting Mental Health: Hierarchical Regression Analysis

Hierarchical regression analysis was conducted to examine the relative

contribution to anxiety of sociodemographic factors, diabetes, lifestyle factors and

number of co-morbidities. Parallel analysis was undertaken for other mental health

measures including depression, psychological symptoms and two constructs of

general mental health including MHI and MCS. It was decided that diabetes would

be entered into the regression as the first step. Age and employment status were

entered into the regression as the second step, as they differed significantly between

women with and without diabetes. Lifestyle factors including BMI, rating of overall

physical activity and smoking status were entered at the third step. The number of

co-morbidities was entered at the last step.

It needs to be noted that although general activity and exercise were also

found to be significantly related with mental health, these two items were not

included in the regression models for a couple of reasons. First, rating of overall

physical activity, to some extent, is a summary of one’s general activity and exercise

levels. It can be seen that there is a significant overlap between them. Entering all of

the three items of physical activity together into the regression model would result in

166

substantial multicolinearity. So, there are two options: using two items: general

physical activity and exercise, or using one item: rating of overall physical activity.

Second, a comparison of the two approaches was carried out. When the two physical

activity items of general physical activity and exercise were entered into regression

models with other lifestyle factors, using depression as the dependent variable, they

totally explained about 7% of the variance. When the rating of overall physical

activity replaced general activity and exercise, and were entered into regression

models with other lifestyle factors, this explained 10% of the variance of depression.

Similar findings were seen in the regression models using anxiety, psychological

symptoms, MHI and MCS as dependent variables. Thus, it is concluded that using a

rating of overall physical activity has two advantages in terms of statistical

modelling. First, it introduces one less item into the regression model; therefore the

model is more concise. Second, using the rating of overall physical activity

explained more variance of mental health measures, indicating a better fit to the

model.

Anxiety

As indicated in Table 5.11, at Step 1, diabetes explained very little of the

variance in anxiety (less than 1%), F(1, 162) = 2.213, p > .05. This result was

expected and consistent what was shown in Table 5.4.

At Step 2, when age and employment status were entered into regression

together with diabetes, the three variables explained 3% of the variance in anxiety,

F(3, 160) = 2.243, p > .05. At this time, age was the only significant predictor of

anxiety. Yet the 2% increase of the variance explained for anxiety was not

significant.

167

When lifestyle factors were introduced into the model, 6% of the variance in

anxiety was accounted, F(7, 156) = 2.726, p = .011. The additional 3% increase was

shown to be significant. However, when diabetes, sociodemographic factors and

lifestyle factors were put together, only the rating of overall of physical activity was

a significant predictor of anxiety.

At Step 4, where the number of co-morbidities was entered into the model,

the explained percentage of the variance in anxiety increased significantly to 17%,

F(8, 155) = 5.220, p < .001, which was an additional 10% beyond lifestyle factors.

With all the variables entered, age, rating of overall physical activity and number of

co-morbidities were significant predictors of anxiety.

Table 5.11

Hierarchical Regression Analysis to Predict Anxiety (N = 163)

Step Predicting variables B SE β R2 R2 change

Dependent variable: anxiety

1 Diabetes -.651 .438 -.116 .007 .013

2 Diabetes -.965 .522 -.172 .022 .027

Age -.080 .040 -.213*

Employment status .157 .101 .150

3 Diabetes .024 .600 .004 .069 .069*

Age -.058 .040 -.156


BMI .020 .034 .053

Overall rating of physical

activity-.261 .107 -.214*

Past smokers .979 .708 .115

Current smokers -.813 .725 -.089

4 Diabetes .457 .574 .081 .172 .103***

Age -.068 .038 -.183


168


BMI -.005 .032 -.013


activity-.251 .101 -.205*

Past smokers .914 .668 .108

Current smokers -.972 .685 -.107

# of co-morbidities .452 .100 .358***

*p < .05; **p < .01; ***p < .001

Depression

For depression, diabetes itself explained 2% of the variance in depression,

F(1, 163) = 4.640, p = .033 (see Table 5.12).

At Step 2, when age and employment were entered into regression, diabetes

remained significant, but age and employment status were not significant predictors.

The three variables explained 2.7% of the variance in depression, F(3, 161) = 2.574,

p = .056. There was less than a 1% increase in the percentage of variance explained,

and this was not statistically significant.

Lifestyle factors were brought in at Step 3. Together with the variables at

Step 2, all the variables explained about 10% of the variance in depression, F(7, 157)

= 4.086, p < .001. This was about an additional 10% beyond what was accounted for

by diabetes, age and employment status. However, diabetes was not a significant

predictor, only the rating of overall physical activity and the category of past

smokers being the significant predictors of depression.

At the last step, when the number of co-morbidities was added, 22% of the

variance in depression was accounted for, F(8, 156) = 6.868, p < .001. The number

of co-morbidities greatly increased the variance in depression that could be

explained, and was a significant predictor of depression. Moreover, both the rating of

169

overall physical activity and being a past smoker remained significant predictors in

the model.

Table 5.12

Hierarchical Regression Analysis to Predict Depression (N = 164)


Dependent variable: depression

1 Diabetes -.822 .381 -.166* .022 .028*

2 Diabetes -1.267 .458 -.257** .028 .018

Age -.045 .035 -.136

Employment status -.036 .089 -.039

3 Diabetes -.316 .515 -.064 .116 .108***

Age -.023 .034 -.068


BMI .010 .029 .030


activity-.252 .092 -.233**

Past smoker 1.592 .596 .217**

Current smoker .467 .622 .058

4 Diabetes .069 .490 .014 .223 .106***

Age -.032 .032 -.096


BMI -.013 .027 -.038


activity-.243 .086 -.225**

Past smokers 1.552 .559 .211**

Current smokers .324 .585 .040


*p < .05; **p < .01; ***p < .001

170

Psychological Symptoms

Similar to anxiety, the results showed that diabetes explained little variance

in psychological symptoms (1.7%), F(1,162) = 3.760, p = .054.

When age and employment status were entered, diabetes became a significant

factor. These three variables explained 2.5% of the variance in psychological

symptoms, F(3, 160) = 2.395, p = .070, but the increase was not statistically

significant, as indicated by Table 5.13.

At Step 3, lifestyle factors were added into the model and all the variables

explained 8% of the variance in psychological symptoms, F(7, 156) = 3.604, p

= .001. This was a considerable increase from Step 2. In addition, the significance of

diabetes disappeared at this step, but rating of overall physical activity and past

smokers became significant predictors.

When the number of co-morbidities were put into the model, all the variables

explained 23% of the variance in psychological symptoms, F(8, 155) = 7.089, p

< .001. There was an additional 10% increase on the basis of lifestyle factors. When

all the variables were combined together, the rating of overall physical activity, past

smokers and number of co-morbidities remained significant predictors of

psychological symptoms.

171

Table 5.13

Hierarchical Regression Analysis to Predict Psychological Symptoms (N = 163)

Step Predicting variables B SE Β R2 R2 change

Dependent variable: psychological symptoms

1 Diabetes -1.439 .742 -.151 .017 .023

2 Diabetes -2.199 .889 -.230* .025 .020

Age -.123 .067 -.193


3 Diabetes -.280 1.005 -.029 .101 .096**

Age -.080 .066 -.126


BMI .032 .056 .050


activity-.512 .179 -.246**

Past smoker 2.494 1.186 .172*

Current smoker -.343 1.214 -.022

4 Diabetes .543 .943 .057 .230 .129***

Age -.099 .062 -.155


BMI -.015 .053 -.023


activity-.493 .166 -.237**

Past smokers 2.370 1.097 .164*

Current smokers -.646 1.125 -.042


*p < .05; **p < .01; ***p < .001

General Mental Health: the Mental Health Inventory

As shown in Table 5.14, at Step 1 of the regression, diabetes was not

associated with MHI scores, and only explained 1% of variance in MHI scores, F(1,

162) = 2.822, p = .095.

172

When age and employment status were entered, diabetes became a significant

factor with a p value of .05. There was not much increase with regard to the

percentage of variance being explained, F(3, 160) = 1.374, p = .253.

At Step 3, when lifestyle factors were put into the regression model, all the

variables only explained about 3% of the variance of MHI scores, F(7, 156) = 1.363,

p = .225, with none of the variables being significant.

The last model with the additional independent variable of number of

co-morbidities explained 5% of the variance in MHI scores, F(8, 155) = 2.074, p

= .041. At this step, the number of co-morbidities was the only significant factor in

the model, with a positive relationship being revealed.

Table 5.14

Hierarchical Regression Analysis to Predict the Mental Health Inventory Scores (N

= 163)


Dependent variable: MHI

1 Diabetes 4.371 2.602 .131 .011 .017

2 Diabetes 6.181 3.130 .185* .007 .008

Age .259 .238 .116


3 Diabetes 3.759 3.682 .113 .015 .033

Age .209 .243 .094


BMI -.123 .205 -.055


activity1.075 .656 .147

Past smokers -.662 4.256 -.013

Current smokers -3.365 4.562 -.060

4 Diabetes 2.131 3.671 .064 .050 .039*

173


Age .247 .239 .110


BMI -.030 .204 -.013


activity1.031 .645 .141

Past smokers -.539 4.181 -.011

Current smokers -2.588 4.491 -.046

# of co-morbidities -1.658 .640 -.220*

*p < .05; **p < .01; ***p < .001

General Mental Health: the Mental Composite Scores

The MCS score was explained poorly by the variables in Study 2. As can be

seen from Table 5.15, all of the variables explained about 1% of the variance of

general mental health as measured by the MCS. None of the lifestyle factors was of

statistical significance in relation to MCS. Even the overall rating of physical activity

and the number of co-morbidities became non-significant.

174

Table 5.15

Hierarchical Regression Analysis to Predict the Mental Composite Scores (N = 159)


Dependent variable: MCS

1 Diabetes 2.362 2.024 .092 .002 .009

2 Diabetes 3.592 2.424 .141 -.005 .006

Age .107 .185 .062


3 Diabetes 1.649 2.845 .065 .001 .031

Age .078 .188 .045


BMI .063 .160 .037


activity.945 .507 .170

Past smokers -2.919 3.407 -.075


4 Diabetes .797 2.883 .031 .011 .015

Age .098 .188 .057


BMI .105 .161 .061


activity.929 .505 .167

Past smokers -2.946 3.390 -.076


# of co-morbidities -.799 .507 -.140

*p < .05; **p < .01; ***p < .001

175

Reflection on Research Questions

1) What is the difference in mental health between women with and without

diabetes, without consideration of lifestyle factors?

Women with diabetes presented more psychological symptoms

including depression and a lower level of general mental health when

compared to women without diabetes. In other words, women with diabetes

have lower levels of mental health than those without.

2) What are the contributing factors to the difference in mental health between

women with and without diabetes?

After controlling for lifestyle factors, there was no difference in

mental health between women with and without diabetes. The previously

identified difference in mental health between women with and without

diabetes groups was not related with diabetes itself, but with a lower level of

physical activity, being a past-smoker and the number of co-morbidities those

women with diabetes present.

176

177

CHAPTER 6 RESULTS OF STUDY 3: ANALYSIS OF THE MEDIATION OF

SELF-EFFICACY IN THE RELATIONSHIPS BETWEEN LIEFSTYLE

FACTORS AND MENTAL HEALTH IN MIDLIFE AND OLDER WOMEN

WITH DIABETES

Introduction

Self-efficacy was the central concept in social cognitive theory, and believed

to be essential in guiding behaviour changes. The multimodal intervention used in

the WWP study is based on social cognitive theory, and has been shown to be

effective in improving the lifestyles for midlife and older women from the general

community. The purpose of Study 3 was to examine the mediating effect of

self-efficacy in the relationships between lifestyle factors and mental health. In the

context of the current research study, if self-efficacy mediates the relationships,

concerted effect should be devoted to enhance individuals’ self-efficacy and lead to

desired outcomes. Study 3 has the potential to generate knowledge that might be

useful for developing effective interventions for mental health improvement.

The baseline data of women who participated in the CDWWP project (N =

83) was utilised in Study 3. A cross-sectional design was adopted. The variables in

Study 3 were sociodemographic factors, lifestyle factors, mental health, self-efficacy

and other confounders including duration of diabetes, use of antidepressants, number

of co-morbidities and menopausal status. The sampling procedure and results of

recruitment were depicted in Figure 6.1.

178

Figure 6.1. Sample recruitment procedure of Study 3.

During the period from October 2008 to March 2010, 1604 patients in total

were seen in diabetes clinics. Of those, 875 patients were men, 48 patients were

women younger than 45 years, and 82 women were missed out. This resulted in 599

women who were 45 years or older, who were potentially eligible for the study.

When women were introduced to the research study, 251 women refused to take

consent packs which contained study information. A further 143 women were

excluded due to a failure to meet inclusion criteria. Finally, 205 consent packs were

given out, 83 women filled out the consent forms, and completed baseline

questionnaires later on. The response rate was 40%.

Because the study sample of the third study had significant overlap with

Study 2, the characteristics of the study population (CDWWP) were not repeated

here to for the purpose of conciseness (see p. 146-159).

179

Results of Study 3

The Correlations Between Mental Health and Lifestyle Factors

This step of analysis was to find out the correlations between mental health

and lifestyle factors (Path c). Based on Baron and Kenny’s principles, this initial step

was undertaken to determine the significant correlations between mental health and

lifestyle factors, in other words, to establish whether there is a correlation that can be

mediated.

As shown in Table 6.1, BMI was found to be significantly and positively

correlated with depression, r = .275, p = .014, suggesting that higher BMI values

were related with higher depression scores. Apart from depression, no correlations

were identified between BMI and anxiety and general mental health. The rating of

overall physical activity was negatively and strongly correlated with depression, r =

-.284, p = .013, but not with anxiety and general mental health. The independent

sample t-test failed to detect any significant difference in mental health between

women who drank and did not drink in the preceding week. For smoking status, a

significant difference in general mental health (MCS) was found. Interestingly,

women who were current smokers had higher MCS scores (58.10 ± 3.16) than those

who were not current smokers (46.34 ± 13.00), p = .003. However, it needs to be

noted that only three women were current smokers, thus, the results may not be

reliable due to the low statistical power.

To sum up, this step of analysis found two significant correlations between

mental health and lifestyle factors, that is a correlation of depression with BMI and

the rating of overall physical activity. Thus, the mediating effect of self-efficacy was

tested in these two correlations.

180

Table 6.1

The Correlations Between Lifestyle Factors and Mental Health (N = 83)

Anxiety Depression MHI MCS

BMI§ .185 .275* -.123 -.088Overall physical -.172 -.284* .147 .019

alcohol use^

Yes 5.52 ± 4.19 4.61 ± 3.55 73.39 ± 17.91 50.12 ± 12.07

No 5.50 ± 4.59 4.98 ± 3.54 71.31 ± 17.75 46.00 ± 13.06

Smoking^

Current 4.00 ± 2.00 4.67 ± 4.04 82.67 ± 10.67 58.10 ± 3.16**

Non-current 5.71 ± 4.56 4.86 ± 3.50 71.00 ± 17.88 46.34 ± 13.00§Pearson correlation; ^ t-test; *p < .05, **p < .01

The Correlations Between Mental Health and Self-Efficacy in Managing Chronic

Disease

This section was comprised of three components, which were a description of

self-efficacy in managing chronic diseases, followed by a description of mental

health, and finished with the correlation analysis between self-efficacy and mental

health measures.

Description of Self-Efficacy in Managing Chronic Diseases

Table 6.2 described women’s self-efficacy levels in managing chronic

diseases. The six items focused on women’s confidence in their ability to perform

these activities in order to minimise the negative effects of chronic diseases. The

results generally indicated that women’s self-efficacy was relatively high, with a

mean score of each item ranging from 6.76 for fatigue management, to 7.28 for

medication compliance. The mean score of self-efficacy in all of the six items was 7.

181

Table 6.2

Mean Score of Self-Efficacy in Managing Chronic Diseases (N = 80)

Items of SEMCD Mean ± SD

Confidence in keeping fatigue from interfering with things you

want to do

6.76 ± 2.00

Confidence in keeping physical discomfort or pain from

interfering with things you want to do

6.87 ± 2.07

Confidence in keeping emotional distress from interfering with

things you want to do

7.28 ± 2.02

Confidence in keeping other symptoms/health problems from

interfering with things you want to do

6.94 ± 1.96

Confidence in doing different tasks/activities needed to manage

your health condition

7.03 ± 2.04

Confidence in doing things other than taking medication to

reduce the effect of illness on your life

7.28 ± 1.99

Mean score 7.03 ± 1.67

Description of Mental Health

Anxiety and depression

Anxiety and depression in Study 3 was measured by the HADS. As displayed

in Table 6.3, the mean anxiety score was 5.65 (SD = 4.50) and the mean depression

score was 4.86 (SD = 3.50), respectively. To develop a better understanding of the

anxiety and depression levels for women with diabetes, the prevalence of anxiety

and depression was compared with that of a general population using a cutoff score

of 8. As seen in Table 6.3, the percentage of anxiety and depression in this sample

was 34.9% and 20.5%, respectively, which was nearly 4 to 5 times as much as that of

the general population (9.6% for anxiety and 4.9% for depression).

182

Table 6.3

The Description of Anxiety and Depression as Measured by the Hospital Anxiety and

Depression Scale (N = 80)

Type Variables Mean / N SD / %

Continuous Anxiety 5.65 4.50

Depression 4.86 3.50

Categorical Anxiety 29 34.9

Depression 17 20.5

General mental health

The frequency distributions of eight individual scales of SF-36 were

described. As described by the mental health component of SF-36 (see Table 6.4), it

was found that in the past four weeks, over 50.0% of women identified themselves as

nervous people for different amounts of time. 44.6% of women had “felt down in

dumps where nothing could cheer them up”, and 63.9% of women “felt downhearted

and blue” in the previous month, with most of them reporting these feelings to have

short duration. Half of the women reported that they have “felt calm and peaceful for

most or a good bit of the time”, with a small percentage of women “never feeling

calm and peaceful” within the investigated time. Almost 70% of women regarded

themselves as “a happy person for a reasonable amount of time”, and 6.0% of them

“have not been happy in the previous four weeks”.

183

Table 6.4

The Frequency Distribution of Mental Health Domain in SF-36 (N = 80)

All of the

time

N (%)

Most of

the time

N (%)

A good

bit of

time

N (%)

Some of

the time

N (%)

A little of

the time

N (%)

None of

the time

N (%)

Been a

nervous person

4 (4.8) 2 (2.4) 6 (7.2) 12 (14.5) 22 (26.5) 37 (44.6)

Felt down in

the dumps

2 (2.4) 1 (1.2) 5 (6.0) 10 (12.0) 19 (22.9) 46 (55.4)

Felt calm and

peaceful

6 (7.2) 23 (27.7) 18 (21.7) 17 (20.5) 14 (16.9) 5 (6.0)

Felt

downhearted

1 (1.2) 2 (2.4) 4 (4.8) 19 (22.9) 27 (32.5) 30 (36.1)

Been a happy

person

14 (16.9) 29 (34.9) 12 (14.5) 18 (21.7) 5 (6.0) 5 (6.0)

In terms of vitality, 44.6% and 38.6% of the women found themselves “full

of pep” and felt that they “had lots of energy” for a good bit of time to all of the time,

respectively. While on the other hand, also about 33.7% of women felt “worn out”

for a good bit of time or more in the past 4 weeks, and half of the women felt “tired”

for more than good bit of time (see Table 6.5).

Table 6.5

The Frequency Distribution of Vitality Domain of SF-36 (N =80)

All of the

time

N (%)

Most of

the time

N (%)

A good

bit of time

N (%)

Some of

the time

N (%)

A little of

the time

N (%)

None of

the time

N (%)

Feel full of pep 3 (3.6) 17 (20.5) 17 (20.5) 24 (28.9) 13 (15.7) 9 (10.8)

Have a lot energy 0 (0.0) 12 (14.5) 20 (24.1) 18 (21.7) 20 (24.1) 13 (15.7)

Feel worn out 2 (2.4) 10 (12.0) 16 (19.3) 21 (25.3) 24 (28.9) 10 (12.0)

Feel tired 10 (12.0) 15 (18.1) 16 (19.3) 21 (25.3) 18 (21.7) 3 (3.6)

184

Two items were applied to investigate the effect of physical and emotional

health on women’s social activity. As indicated in Table 6.6, more than half of the

women found their social activities were interfered with, with 21.7% being slightly

impacted, 18.1% moderately, 15.7% quite a bit, and 1.2% extremely. Likewise, more

than 50% of women reported social activities were affected for various time

durations, with 16.9% reporting a little of the time, 21.7% some of the time, 14.5%

most of the time, and 4.8% for all of the time.

Table 6.6

The Frequency Distribution of Social Function of SF-36 (N = 83)

Not at all Slightly Moderately Quite a

bit

Extremely

To what extent

social activities

were affected

36 (43.4) 18 (21.7) 15 (18.1) 13 (15.7) 1 (1.2)

All of the

time

Most of

the time

Some of

the time

A little of

the time

None of

the time

How much time

social activities

were affected

4 (4.8) 12 (14.5) 18 (21.7) 14 (16.9) 35 (42.2)

The extent to which women were affected by emotional problem was

described in Table 6.7. Specifically, 25.9% of women had to cut down time on work

or activities due emotional distress, 37.0% of women accomplished less than they

liked, and 29.3% of women did not work as carefully as usual.

185

Table 6.7

The Frequency Distribution of Role Emotional of SF-36 (N = 83)

Yes No

Cut down time on work or activities 21 (25.9) 60 (74.1)

Accomplish less than you would like 30 (37.0) 51 (63.0)

Did not work as carefully as usual 24 (29.3) 58 (70.7)

For physical functioning, it was found that 60% of the women reported that

their vigorous activity was limited a lot because of health conditions; while only

about 20% of the women claimed that moderate activity was affected so. Half of the

women had no difficulty in carrying groceries due to their health, and the percentage

was similar regarding to climbing one flight of stairs. About one third of the women

reported that their health condition significantly limited bending/kneeling, walking

more than a mile or several blocks. In addition, 66.7% to 88.9% of women were not

affected in relation to self-bathing or dressing, and walking one block (see Table

6.8).

Table 6.8

The Frequency Distribution of Physical Function of SF-36 (N = 83)

Yes, limited a

lot

Yes, limited

a little

No, not

limited at all

Limits imposed on vigorous activity 50 (61.7) 25 (30.9) 6 (7.4)

Limits imposed on moderate activity 16 (19.8) 34 (42.0) 31 (38.3)

Limits imposed on carrying groceries 13 (16.0) 33 (40.7) 35 (43.2)

Limits imposed on several flights of

stairs

26 (32.1) 29 (35.8) 26 (32.1)

Limits imposed on one flight of stairs 12 (14.8) 26 (32.1) 43 (53.1)

Limits imposed on bending or

kneeling

23 (28.4) 38 (46.9) 20 (24.7)

186

Yes, limited a

lot

Yes, limited

a little

No, not

limited at all

Limits imposed on walking more than

a mile

26 (31.7) 27 (32.9) 29 (35.4)

Limits imposed on walking several

blocks

23 (28.4) 26 (32.1) 32 (39.5)

Limits imposed on walking one block 9 (11.1) 18 (22.2) 54 (66.7)

Limits imposed on bathing or dressing 2 (2.5) 7 (8.6) 72 (88.9)

The effects of physical health on women’s work and other activities were

presented in Table 6.9. It can be seen that 28% of women had to cut down the

amount of time that they spent on work and other activities in the past four weeks.

Moreover, approximately half of the women accomplished less, had been limited in

the kind of work they do, and/or had difficulty in performing the work or other

activities because of health issues.

Table 6.9

The Frequency Distribution of Role Physical of SF-36 (N = 83)

Yes No

Cut down the time you spent on work or other activities 23 (28.0) 59 (72.0)

Accomplished less than you would like 43 (52.4) 39 (47.6)

Were limited in the kind of work or other activities 37 (45.1) 45 (54.9)

Had difficulty performing the work or other activities 37 (44.6) 45 (54.9)

Regarding the severity of pain women had during the past four weeks, about

half of the women experienced pain at moderate to very severe levels, with 30% of

them reporting having moderate pain. The effect of pain on women’s work was not

as bad as the pain they experienced. It was found that over 80% of women reported

be non-affected or mildly affected (see Table 6.10).

187

Table 6.10

The Frequency Distribution of Bodily Pain of SF-36 (N = 83)

None Very mild Mild Moderate Severe Very severe

How much bodily pain have you had 13 (15.7) 12 (14.5) 22 (26.5) 24 (28.9) 10 (12.0) 2 (2.4)

Not at all A little bit Moderately Quite a bit Extremely

How much did the pain interfere 26 (31.3) 28 (33.7) 14 (16.9) 13 (15.7) 2 (2.4)

Table 6.11

The Frequency Distribution of General Health of SF-36 (N = 83)

Excellent Very good Good Fair Poor

General health 1 (1.2) 7 (8.4) 34 (41.0) 38 (45.8) 3 (3.6)

Definitely true Mostly true Don’t know Mostly false Definitely false

Get sick more easily 1 (1.2) 13 (16.0) 12 (14.8) 25 (30.9) 30 (37.0)

As healthy as anybody 11 (13.6) 34 (42.0) 15 (18.5) 14 (17.3) 7 (8.4)

Expect my health to get worse 7 (8.6) 15 (18.5) 32 (39.5) 17 (21.0) 10 (12.3)

My health is excellent 1 (1.2) 35 (43.2) 10 (12.3) 14 (17.3) 21 (25.9)

188

When women were asked to rate their general heath, half of them reported

their general health condition was good to excellent, and 45.8% of women rated their

health as fair (see Table 6.11). Four additional items were asked to assess women’s

general health. The results showed that 17% of women thought they got sick more

easily than other people, and about 70% of the women did not agree with this

statement. In addition, 25% of women believed that their health was deteriorating.

Last, about half of the women agreed that their health was excellent, and about half

of them disagreed.

After the presentation of frequency tables for eight scales in SF-36, the two

composite scores of SF-36 were calculated. Furthermore, a scoring check suggested

by Ware and colleagues (2000) was undertaken. The correlations between the eight

scales of SF-36, and two composite scores were displayed in Table 6.12. Similar to

the results in Study 2, it showed that PCS had the highest correlations with PF, RP

and BP, but lowest correlations with SF, RE and MH. In contrast, MCS correlated

highest with SF, RE and MH, but lowest with PF, RP and BP. The PCS and MCS

correlations with GH and VT were both moderate. These results provided evidence

that the scoring of the composite scores have been accurate.

189

Table 6.12

The Correlations Between Eight Scales of SF-36 and Two Composite Scores (N = 83)

PF RP BP GH VT SF RE MH

PCS .854** .827** .778** .657** .566** .476** .038 -.034

MCS .056 .225* .232* .285* .451** .510** .855** .833**

*p < .05; ** p < .001; PF = physical function; RP = role physical; BP = bodily pain; GH = general health; VT = vitality; SF = social function;

MH = mental health; PCS = physical component summary; MCS = mental component summary

190

Correlations Between Mental Health and Self-Efficacy in Managing ChronicDisease

The cross-sectional relationships between SEMCD and different measures of

mental health were analysed by using a binary Pearson correlation analysis. It can be

seen from Table 6.13 that SEMCD was significantly related with all of the mental

health variables, with the strongest correlation being found with depression, r = -.632,

p < .001. These results suggested that women with high self-efficacy tended to have

less anxiety and depression symptoms, and better general mental health. Apart from

this, it was noticed that all the other mental health measures were strongly

interrelated with each other, p < .001.

Table 6.13

Associations Between Self-Efficacy in Managing Chronic Diseases and Mental

Health (N = 83)

SEMCD Anxiety Depression MHI MCS

SEMCD 1 -.422*** -.632*** .452*** .367***

Anxiety 1 .676*** -.649*** -.650***

Depression 1 -.513*** -.502***

MHI 1 .833***

MCS 1

*** p < .001

Associations Between Self-Efficacy in Managing Chronic Disease and Lifestyle

Factors

When examining the relationships between self-efficacy in managing chronic

diseases and lifestyle factors, an independent sample t-test and One-way ANOVA

191

analysis were applied. The characteristics of lifestyle factors for women with

diabetes were displayed in Table 6.14 for clarification, but not described in detail.

There was a significant negative relationship between BMI and SEMCD, r =

-.258, p = .025, suggesting that women with a higher BMI tended to have a lower

level of SEMCD. The magnitude of relationship was mild. However, when women

were classified into normal, overweight and obese categories according to their

BMIs, no statistical difference was found between groups in terms of their

self-efficacy score. This could be related with the small number of women in the

normal weight category (N = 5), which suggested limited statistical power to detect

differences. Women’s rating of overall physical activity was positively related with

SEMCD, r = .281, p = .018. This showed that women with more physical activity

had a high level of self-efficacy in managing chronic diseases. Apart from the

findings for BMI and physical activity, no difference was found in self-efficacy in

relation to other lifestyle factors including smoking and alcohol use.

Table 6.14

Associations Between Self-Efficacy in Managing Chronic Diseases and Lifestyle

Factors (N = 80)

Variables SEMCD

(Mean ± SD)

Test of

significance

p

BMI (continuous) 33.51 ± 7.28 7.03 ± 1.67 -.258 .025

BMI (categorical) 1.392 .255

Normal 6 (7.6) 7.37 ± 1.91

Overweight 22 (27.8) 7.52 ± 1.38

Obese 51 (64.6) 6.83 ± 1.77

Physical activity

General activity 1.151 .322

Very/moderate 30 (37.5) 7.28 ± 1.76

192

Variables SEMCD

(Mean ± SD)

Test of

significance

p

Mild 34 (42.5) 7.13 ± 1.61

Sedentary 16 (20.0) 6.51 ± 1.59

Exercise -1.970 .053

Yes 39 (49.4) 7.46 ± 1.79

No 40 (50.6) 6.72 ± 1.47

Rating of overall

physical activity

3.55 ± 2.00 .263 .025

Alcohol consumption -1.510 .135

Yes 23 (28.0) 7.46 ± 1.66

No 58 (70.7) 6.84 ± 1.67

Smoking habit -1.575 .119

Smoker 3 (3.6) 8.50 ± 1.26

Non-smoker 80 (96.4) 6.97 ± 1.66

The Associations of Self-Efficacy With Sociodemographic Factors and Other

Confounders

The purpose of this section was to examine the correlations between SEMCD

and sociodemographic factors, as well as other confounders to determine the

significant covariates to be included in further mediation analysis.

The analysis showed that none of the sociodemographic factors was

correlated with SEMCD. Other variables including the duration of diabetes, use of

antidepressants and number of co-morbidities were significantly correlated with

SEMCD, menopausal status was not (Table 6.15). In detail, the number of

co-morbidities was negatively related with SEMCD, and the strength of this

association was moderate, r = -.543, p < .001. Women who have had diabetes for

more than two years had lower SEMCD scores than those who have had diabetes for

less than two years (7.77 ± 1.13 vs. 6.55 ± 1.79), t = 3.716, p = .001. Lastly, women

193

who were using antidepressants reported a SEMCD score of 5.83 (SD = 2.24) while

the corresponding scores for those who did not were 7.26 (SD = 1.45), t = 2.946, p

= .004. Based on the analysis, the confounders of the duration of diabetes, use of

antidepressants and the number of co-morbidities were to be included in the

mediation analysis.

Table 6.15

The Differences of Self-Efficacy in Managing Chronic Diseases in Relation to

Sociodemographic Factors and Other Confounders (N = 77 ~ 83)

Mean ± SD / N (%)SEMCD

(Mean ± SD)

Test of

Significance

p

Age 65.43 ± 9.26 7.03 ± 1.67 -.070 .539

# of co-morbidities§ 3.00 7.03 ± 1.67 -.543 < .001

Marital status .678 .500

Married 47 (56.6) 7.14 ± 1.68

Not married 36 (43.4) 6.88 ± 1.67

Country of origin .531 .328

Australia 59 (71.1) 6.90 ± 1.75

Others 24 (28.9) 7.31 ± 1.45

Aboriginality -1.328 .188

Yes 1 (1.2) 4.83

No 81 (98.8) 7.05 ± 1.66

Speaking another

language than English

-1.030 .306

Yes 10 (12.0) 6.52 ± 1.62

No 73 (88.0) 7.10 ± 1.67

Education 2.598 .081

≤Junior school 34 (42.5) 6.57 ± 1.97

Senior school 16 (20.0) 7.08 ± 1.47

Uni./Technology 30 (37.5) 7.51 ± 1.26

Employment .157 .675

194

Mean ± SD / N (%)SEMCD

(Mean ± SD)

Test of

Significance

p

Paid employment 19 (22.9) 7.17 ± 1.37

Unemployed 64 (77.1) 6.98 ± 1.76

Annual income .059 .943

≤$ 40,000 47 (61.0) 7.09 ± 1.68

>$ 40,000 27 (35.1) 6.96 ± 1.70

Don’t know 3 (3.89) 6.89 ± 1.78

Duration of diabetes 3.716 .001

≤ 2 years 31 (38.8) 7.77 ± 1.13

> 2 years 49 (61.2) 6.55 ± 1.79

Use of antidepressant 2.946 .004

Yes 13 (15.7) 5.83 ± 2.24

No 70 (84.3) 7.26 ± 1.45

Menopausal status .504 .681

Pre- 1 (1.3) 8.33

Peri- 7 (8.8) 7.55 ± .64

Post 32 (40.0) 7.03 ± 1.50

Surgical 40 (50.0) 6.89 ± 1.92§Median

Mediation Analysis

Based on the above analysis, the mediating effect of SEMCD could be tested

in the correlations between depression and BMI, and depression and rating of overall

physical activity. The confounders in the mediation analysis were duration of

diabetes, use of antidepressants, and number of co-morbidities. The hypothesised

mediation models and the correlation coefficients between variables were depicted in

from Figures 6.2 to 6.5. As the relationships between lifestyle factors and mental

health are reciprocal, four models were examined. The first model used physical

activity as the outcome measure and depression as the predictor to examine whether

self-efficacy mediates the effect of depression on physical activity. The second

195

model used BMI as the outcome measure and depression as the predictor to examine

whether self-efficacy mediates the effect of depression on BMI. The third model

used depression as the outcome measure and BMI as the predictor to examine the

mediating effect of self-efficacy. Finally, the fourth model used depression as the

outcome measure and physical activity as the predictor and examined the mediating

effect of self-efficacy.

Figure 6.2.Mediating model l: using depression to predict physical activity.

196

Figure 6.3.Mediation model 2: using depression to predict BMI.

Figure 6.4.Mediation model 3: using BMI to predict depression.

Figure 6.5.Mediation model 4: using physical activity to predict depression.

197

Model 1: Using Depression to Predict Physical Activity

As shown in Table 6.16, when physical activity was regressed on depression

only, depression was a significant predictor, β = -.298, p = .012. And depression

alone explained 7.6% of the variance in rating of overall physical activity, F(1, 69) =

6.726 , p = .012.

At Step 2, when SEMCD was entered into the regression model, the

significance of depression vanished, yet SEMCD was non-significant. The model

with depression and SEMCD totally explained 7.7% of the variance in the rating of

physical activity, F(2, 68) = 3.913, p = .025, with little improvement from Step 1.

This suggested that SEMCD had some influence on the correlation between

depression and physical activity, but it did not work as a mediator. Entering other

confounders did not change the significance level of the correlation between

depression and physical activity.

Table 6.16

Multiple Linear Regression, Using Depression to Predict Physical Activity (N = 70)

Step Variables B SE Β R2 R2 change

1 Depression -.161 .062 -.298* .076* .089*

2 Depression -.108 .080 -.201 .077* .014

SEMCD .183 .175 .155

3 Depression -.130 .084 -.240 .066 .029

SEMCD .247 .191 .209

# of co-morbidities .008 .106 .011

Duration of diabetes .730 .493 .186

Antidepressants .188 .651 .037

*p < .05

198

Model 2: Using Depression to Predict BMI

When BMI was regressed on depression (Step 1), depression explained 5%

of the variance in BMI, F(1, 71) = 4.811, p = .032. Depression itself was a

significant predictor of BMI, β = .252, p = .032.

Similar to physical activity, once SEMCD was entered into regression,

depression became non-significant, but neither was SEMCD. These two variables

explained 5.6% of the variance in BMI, F(2, 70) = 3.126, p = .050. The change of R2

was not significant (see Table 6.17).

At Step 3, other confounders were entered into the regression model too, yet

none of the variables was significant. In addition, all the variables as a group

explained less variance in BMI as compared to Step 1 and Step 2, F(5, 67) = 1.559 ,

p = .184. Therefore, it was concluded that the results did not support the hypothesis

of SEMCD being the mediator of the effect of depression on BMI.

Table 6.17

Multiple Linear Regression, Using Depression to Predict BMI (N = 70)

Step Variables B SE Β R2 R2 change

1 Depression .479 .218 .252* .050* .063*

2 Depression .277 .277 .146 .056* .019

SEMCD -.682 .574 -.173

3 Depression .230 .297 .121 .037 .022

SEMCD -.573 .636 -.145


Duration of diabetes -.747 1.801 -.054

Antidepressants 2.255 2.268 .130

*p < .05

199

Model 3: Using BMI to Predict Depression

Step 1 used BMI as the independent variable, and depression as the

dependent variable. As shown in Table 6.18, BMI had a significant relationship with

depression, β = .252, p = .032. BMI, by itself, explained 5.0% of the variance in

depression, F(1, 71) = 4.811, p = .032.

When SEMCD was entered into the regression model with BMI, the

significance of BMI vanished, and SEMCD was shown to be a strong factor

influencing depression, β = -.590, p < .001. Moreover, the variance accounted by

BMI and SEMCD was 37.1%, F(2, 70) = 22.194, p < .001, which was a substantial

increase beyond what was explained by BMI.

The effect of confounders was controlled at Step 3. As revealed, the strength

of the correlation between SEMCD and depression was attenuated somewhat,

remained statistically significant, β = -.388, p = .002. None of the remaining

variables were found to be significant. All the variables explained 41.4% of the

variance in depression, which was shown to be a significant improvement, ΔR2 =

6.7%, p = .050.

In brief, SEMCD was a mediator of the relationship between BMI (predictor)

and depression (outcome). In other words, the finding that women with high BMI are

more likely to have depressive symptoms is because they have low self-efficacy in

managing chronic disease.

200

Table 6.18

Multiple Linear Regression, Using BMI to Predict Depression (N = 72)

Step Variables B SE β R2 R2 change

1 BMI .132 .060 .252* .050* .063*

2 BMI .051 .051 .097 .371*** .325***

SEMCD -1.225 .201 -.590***

3 BMI .039 .050 .073 .414*** .067*

SEMCD -.806 .243 -.388**


Duration of diabetes 1.204 .724 .166


*p < .05; **p < 0.01; ***p < .001

Model 4: Using Physical Activity to Predict Depression

The effect of the rating of overall physical activity on depression was first

analysed. At Step 1, physical activity was a significant predictor of depression, β =

-.298, p =.012. Physical activity alone explained 7.6% of the variance in depression,

F(1, 69) = 6.726, p = .012.

Then, SEMCD was entered into regression analysis with physical activity.

The results showed that physical activity became a non-significant factor in

predicting depression, while SEMCD became a significant one, β = -.594, p < .001.

These two variables explained 39.6% of the variance in depression, F(2, 68) =

23.949, p < .001.

Finally, when other confounders were entered into the model, the magnitude

of the correlation between depression and self-efficacy decreased, but remained

strongly significant, β = -.412, p < .001. In addition, no other factors were found to

be significantly impacting on depression. The whole group of variables accounted

201

for 43.0% of the variance in depression, but this was not a significant improvement,

ΔR2 = 5.7%, p = .082.

In short, the analysis showed that in the relationship between physical

activity (predictor) and depression (outcome), SEMCD was a mediator. The reason

for more depression symptoms among women with low physical activity was

actually low self-efficacy in this group of women.

Table 6.19

Multiple Linear Regressions, Using Physical Activity to Predict Depression (N = 70)

Step Variables B SE β R2 R2 change

1 Physical activity -.552 .213 -.298* .076* .089*

2 Physical activity -.243 .179 -.131 .396*** .324***

SEMCD -1.300 .212 -.594***

3 Physical activity -.272 .176 -.147 .430*** .057

SEMCD -.902 .256 -.412***


Duration of diabetes 1.108 .712 .152


*p < .05; ***p < .001

Reflection on the Research Questions

1) What is the relationship between mental health and lifestyle factors among

midlife and older women with diabetes?

Among this population, depression was positively correlated with

BMI and negatively related with rating of overall physical activity. Apart from

depression, no other correlations were identified between other mental health

and lifestyle factors.

202

2) Does self-efficacy in managing chronic diseases mediate the correlations

between depression and BMI, and depression and physical activity among

midlife and older women with diabetes?

When examining the effect of lifestyle factors (BMI & physical activity)

on depression, self-efficacy was a mediator. However, when the effect of

depression on lifestyle factors (BMI & physical activity) was evaluated,

self-efficacy was not a mediator.

3) Additional findings

Women who had diabetes for more than two years had a lower level of

self-efficacy in managing chronic disease than women who have had diabetes for

less than two years. Furthermore, women using antidepressants had a lower level

of self-efficacy in managing chronic diseases than those who do not.

203

CHAPTER 7: DISCUSSION

Introduction

This chapter summarises the major findings of each study, compares the

findings of the current research with previous studies, and discusses the strengths

and limitations of the study. Prior to each section of the discussion, the

corresponding major finding is briefly outlined. This chapter is comprised of five

sections: the first lists the primary findings of each individual study of the current

research; the second discusses the relationship of each lifestyle factor with mental

health in midlife and older women from the general population; the third illustrates

the relationship between lifestyle and mental health in midlife and older women with

chronic diseases; the fourth demonstrates the mediating effect of self-efficacy in the

associations between lifestyle factors and mental health among midlife and older

women with a chronic disease, and the fifth presents the strengths and limitations.

The Lifestyles of Australian Midlife and Older Women

Research question:

What are the inter-relationships among lifestyle factors among Australian

midlife and older women?

The current study examined four common lifestyle factors including BMI,

physical activity, smoking and alcohol use in Australian midlife and older women. It

is important to gain an understanding of the characteristics of lifestyle factors among

Australian midlife and older women.

204

The current research found that there is no difference in other lifestyle factors

including physical activity, smoking and alcohol use in relation to BMI, suggesting

that overweight and obese women do not necessarily have a greater number of other

unhealthy lifestyle factors. When examining the relationships between physical

activity, smoking and alcohol use, the research found women who are current

smokers had lower level of physical activity than non-smokers. In addition, women

who reported to be regular alcohol drinkers are more likely to perform physical

activity. Finally, women who are current smokers are also more likely to be regular

alcohol drinkers.

The absence of the relationships between BMI and other three lifestyle

factors differs from the literature in this area. Studies have reported that people who

are overweight or obese have lower levels of physical activity than those with

normal weight (Sanchez et al., 2008; Stamatakis et al., 2009), and that people who

are current smokers are less likely to be overweight (John, Hanke, Rumpf, & Thyrian,

2005; Munafo, Tilling, & Ben-Shlomo, 2009). In contrast, the current study found

results otherwise. 81% of obese women in the current study report doing different

levels of physical activity, suggesting that these women may have initiated physical

activity and attempted to place it into practice when they participated in the study.

Therefore, effort should be made towards facilitating women to stabilise the

behaviour of doing physical activity. Furthermore, in this research, no relationship

between smoking and BMI was indicated, although the small number of current

smokers in the current study may have limited the power of detecting a significant

result.

The relationship between smoking status and physical activity collaborates

with a review study by Kaczynski et al. (2008), who claim that current smokers have

205

lower level of physical activity as compared to non-smokers. There are two

hypotheses for this relationship. One suggests that smokers are less health conscious

than those who have never smoked. When this attitude towards health is reflected in

physical activity, this group pays less attention to such activity and consequently

approaches it with less effort. The other hypothesis proposes that physiologically,

smokers tend to have an impaired lung capacity, which in turn prevents them from

carrying out adequate physical activity. In reality, it is likely that these two

mechanisms work jointly rather than separately. Nevertheless, current smokers are

more likely to be physically inactive compared to non-smokers, therefore; physical

activity promotion should be a strong focus for current smokers, particular.

Regarding the correlation between physical activity and alcohol use, the

current study found that regular alcohol drinkers are more likely to engage in a high

level of physical activity, and less likely to lead a sedentary lifestyle than

non-drinkers. It needs to be noted that the majority of women drink alcohol in

moderation in this study. This finding corresponds with a German study, whose

results revealed that non-alcohol use is a risk factor for exercising less than two

hours per week (Haenle et al., 2006), despite of the difference in physical activity

measurement. This relationship may be explained by the socioeconomic factors

(Fine et al., 2004), for example, education. Among women who drink alcohol on a

regular basis, over 40% of the women received tertiary education (university or

technological college), while among women who never drink, the corresponding

percentage was 22%. Higher level of education may indicate that these women are

more conscious and knowledgeable about their own health, which could be a

possible explanation to the relationship between alcohol drinking and physical

activity.

206

The Effect of Lifestyle Factors on Mental Health in Midlife and Older Women

Overweight and Obesity and Mental Health

Research question:

What is the relationship between BMI and mental health among Australian

midlife and older women?

The prevalence rates of overweight and obesity in Australian midlife and

older women in the current study are 33.0% and 25.6%, respectively. These figures

are very similar to those of the Australian female population, among which 30% of

women are classified as overweight and 22% as obese (Thorburn, 2005). In contrast,

among midlife and older women with diabetes, the prevalence of obesity is much

higher than that of the general population (62.2% vs. 25.6%).

The current study found that Australian midlife and older women who are

overweight or obese had lower levels of mental health than women with normal

weight. This finding collaborates with a number of previous studies, which clearly

state a positive relationship between BMI and depression (Barry et al., 2008;

Bruffaerts et al., 2008; Eunkyung, 2009; Zhao et al., 2009).

However, the relationship between BMI and mental health changes as

women age. It is indicated that over time, a higher BMI seemed to be beneficial to

women’s mental health. To our best knowledge, this is the first study showing that as

women age, BMI can actually change from a negative factor for mental health to a

positive one. Some literature does state that being overweight or obese has no

relationship with individuals’ mental health. For example, Vogelzangs and

colleagues studied a cohort of elderly people aged from 70 to 79 years in Netherland,

and found that BMI has no impact on depression as women age (Vogelzangs et al.,

207

2010). One Norwegian study examined anxiety, and also failed to find a relationship

between BMI and anxiety (Bjerkeset et al., 2008). Moreover, other studies comment

that despite the limited physical function that is commonly seen among overweight

and obese people, their mental health remains unaffected (Huang, Frangakis, & Wu,

2006; Mond & Baune, 2009; Renzaho et al., 2010; Vasiljevic et al., 2008; Wee et al.,

2010). In comparison, the current study shows that as midlife and older women age,

those who are overweight or obese have better mental health than those with normal

weight. Relating to the underlying mechanisms between obesity and mental health

may be helpful to explain why the effect of obesity and overweight on mental health

changes as women age.

Two primary pathways of the linkage between mental health (depression) and

obesity are proposed, one is the biophysical mechanism and the other is the

psychological pathway. The biological mechanism refers to the dysregulation of the

hypothalamic-pituitary-adrenocortical (HPA) system which is seen in both obese and

depressed patients (Bornstein, Schuppenies, Wong, & Licinio, 2006; Chrousos,

2000). It is evidenced by the facts that several neuropeptidergic and neurotransmitter

systems involving molecules and norepinephrine are involved in the regulation of

mood as well as body weight. This hypothesis is further supported by the evidence

that antidepressant treatment can result in a side effect of weight gain, while

treatment of obesity often causes depression (McElroy et al., 2004). The other

explanation is psychologically associated, which is strongly related with the social

construction of how obesity is perceived. In the present day, stigma and

discrimination toward obese persons are pervasive and pose numerous consequences

for their psychological and physical health (Puhl & Heuer, 2010). People with

obesity are often blamed with having a weak personality, a lack of self-discipline and

208

even being intelligently inferior (Rogge et al., 2004). This probably explains why

women with a higher BMI have more psychological symptoms.

However, despite the unfavourable societal attitude toward obesity, how an

individual would be influenced mentally heavily depends on his or her

self-perspective about his or her own weight. The psychological symptoms of obese

people may have been rooted in the inability to reduce the gap between a desired

body weight and the real one. Therefore concern about one’s appearance becomes

one of the important pathways via which heavy people experience psychological

problems (Hrabosky & Thomas, 2008; Markowitz, Friedman, & Arent, 2008;

Muennig, Jia, Lee, & Lubetkin, 2008). In relating to the current study (Morrow,

2001), women become more accepting of their bodies as they age because of the

psychosocial changes occurring at this stage of life (Keel, Baxter, Heatherton, &

Joiner, 2007). Going from midlife and beyond, women actually start to have more

space for themselves, both financial and time wise. They are less driven by the

socio-cultural norms of success and achievement. In addition to that, women at

midlife tend to have a more stable income, be released from child raising and have

better marriage relationships (Hunter, Sundel, & Sundel, 2002; Pudrovska, 2009).

These positive factors in combination can be an enormous source of satisfaction and

freedom, which can distract women from their appearance, and still result in

improved mental health.

An alternative explanation is based on the context of the study, and is related

with women’s physical activity levels. Some earlier studies find that physical

inactivity is more common among overweight or obese populations (Sanchez et al.,

2008; Stamatakis et al., 2009), whereas in the current study, women who are obese

have been as equally active as women with normal weight (see Appendix A). It is

209

indicated that the majority of obese women carry out exercise routinely. Considering

these women’s activity level, it is regarded that overweight or obese women may

have benefited from doing regular exercise and may have developed a strong sense

of well-being. Previous study has also shown that midlife women who are engaged

in health promotion activities (e.g. physical activity) report greater extent of life

satisfaction (Degges-White & Myers, 2006).

Physical Activity and Mental health

Research question:

What is the relationship between physical activity and mental health among

Australian midlife and older women?

It was revealed that 16.7% of the women did not perform any exercise on a

weekly basis, while more than 50% of them exercised more than three times a day.

Compared to the Australian population data, which indicates 30% of Australians lead

a sedentary lifestyle, the prevalence of a sedentary lifestyle in the current study is

less than half of the norm (Bauman & Owen, 1999). These data suggests that

Australian midlife and older women are more physically active than the general

population.

Regarding the effect of physical activity on mental health, the current study

revealed that women exercising daily had much better mental health including less

anxiety, depression and better general mental health. This positive effect of physical

activity remained unchanged as women age. This finding of the study is widely

supported by a substantial amount of research in this field (Aoyagi et al., 2010; Bhui

& Fletcher, 2000; Brown et al., 2005; Fukukawa et al., 2004; Galper et al., 2006;

210

Goodwin, 2003; Rakovac et al., 2007; Vallance et al., 2010; Van Gool et al., 2007;

Wyshak, 2001). The current study’s findings strongly suggest that physical activity

is a useful means of improving mental health for Australian midlife and older

women or midlife and older women in other western countries.

Despite the widely accepted notion that physical activity improves mental

health and promotes a sense of well-being, there is a lack of agreement about how to

measure physical activity. The physical activity measured in one study may not be

comparable to that examined in another. For example, the study by Goodwin et al.

(2003) measured physical activity by asking “how often do you get physical exercise

in your job and recreational activities?”, which essentially evaluated the frequency of

doing exercise. However, Bhui and his associates investigated how many times a

week participants took part in activities and for how long on each occasion (Bhui &

Fletcher, 2000), considering both frequency and duration. Furthermore, a study

undertaken in Japan measured physical activity objectively by using a pedometer

(Aoyagi et al., 2010). As can be seen from these examples, apparently, there is a lack

of agreement on how physical activity should be evaluated, which in turn increases

the difficulty of identifying the minimal effective level of physical activity in

reducing depression and anxiety.

Several mechanisms of physical activity (exercise in this case) reducing

depression and anxiety have been discussed, with some of them being primarily

physiologically based, while others stem from a psychological perspective. In fact,

the beneficial effect of physical activity on mental health is likely to be produced by

the combination of both mechanisms. These mechanisms have been reviewed by Lox

et al. (Lox, Martine Ginis, & Petruzzello, 2006). The existing biophysical hypotheses

of depression include the endorphin hypothesis and the monoamine hypothesis. The

211

former hypothesis explains that the body produces endorphin that is a natural

painkiller during exercise; which makes people feel happy. The latter suggested that

exercise changes mood by accelerating the communication between brain

neurotransmitters that are related in emotion regulation, thus reducing depressive

symptoms. The theory explaining the anxiety reduction effect of exercise is called

thermogenic theory. It argues that exercise elevates the body temperature, which is

sequentially sensed by the brain, therefore triggering a muscle relaxation response

and resulting in a feeling of anxiety reduction.

The psychological hypotheses for depression and anxiety are quite similar.

The master hypothesis assumes that alleviated depression is associated with a sense

of accomplishment from doing exercise. Socially, it is an opportunity for individuals

to obtain support if exercising with others (Donaghy, 2007). Regarding anxiety,

research states that exercise distracts people from their daily routine, and creates an

opportunity to leave worries behind, therefore, resulting in a decreased level of

anxiety. The applicability of these theories is not yet well explored, but it is agreed

that exercise decreases negative psychological effects via biophysical and

psychological pathways in conjunction with each other.

Smoking and Mental Health

Research question:

What is the relationship between smoking and mental health among


The study identified 10.8% of current smokers among a large community

sample of Australian midlife and older women. This prevalence of current smokers

212

is approximately half of the national proportion, 18.4% of the Australian population

(Social Research Centre, 2006). The detrimental effect of smoking on mental health

was not shown at an earlier stage, but did become significant as women age.

The negative effect of smoking on mental health has been well established by

earlier research studies (Massak & Graham, 2008; Mino et al., 2001; Sarna et al.,

2008; Tavafian et al., 2009; Tselebis et al., 2001). For example, the Nurses’ Health

Study in America examined 158, 736 nurses from 1992/1993 to 2000/2001, during

which nurses’ mental health appeared to improve as they age, but the disparities of

mental health across smoking status remained (Sarna et al., 2008). In the current

study, when women first entered the study at 55 years of age, not much difference in

mental health was seen between women who smoked and those did not. While as

women age, the difference of mental health between those with different levels of

smoking enlarged. There was a steady improvement for non-smoking women and a

clear deterioration for women who smoked. The improvement of mental health as

women age has been demonstrated by the Nurses’ Health Study. In addition, the

current study suggests that smoking has actually prevented some women from

enjoying optimal mental health.

Two mechanisms have been proposed to explain the adverse effect of

smoking on mental health. The first hypothesis is that smoking per se leads to

depression by neurochemical changes brought on by smoking, which is primarily

supported by the studies examining the incidence of psychological illnesses among

adolescents (Johnson et al., 2000; Wu & Anthony, 1999). The second illustrates that

the co-morbidity of smoking and psychological illnesses is a result of shared

predisposition to genetic and environmental factors (Dierker, Avenevoli, Stolar, &

Merikangas, 2002; Duncan & Rees, 2005; Goodwin & Hamilton, 2002).

213

Given the strong impact of smoking on mental health, it is interesting that no

relationship between the two variables was found when women were younger, for

which explanations are sought. Re-examination of the mental health level of women

finds that when compared to women of younger ages i.e. 30 years, the mental health

of midlife and older women in this study is actually better (Australian Bureau of

Statistics, 2007). When a cutoff score of 10 was used to categorise women into

clinical depression/non-depression and clinical anxiety/non-anxiety groups, the

prevalence rates of depression and anxiety were both found to be 2.3%. These

percentages are much lower than the lifetime prevalence of mental disorders (14.4%)

in the Australian sample (Australian Bureau of Statistics, 2007). Considering the

much lower proportion of smokers in the study population, as mentioned in the

beginning of this section, the non-significant relationship between smoking and

mental health among midlife and older women at their 55 years of age may be

attributed to an inadequate number of women who smoke or have mental health

problems. Nevertheless, it is regarded that smoking produces detrimental

psychological effects on midlife and older women. Continued effort should be made

to encourage women to quit smoking and improve their mental health to the largest

extent as possible.

Alcohol Use and Mental Health

Research question:

What is the relationship between alcohol use and mental health among


214

The study showed that the prevalence of women drinking alcohol on a

regular basis was 16.5%, half of the women (50%) drank alcohol occasionally, and

the other 22.4% of the women never drank. When alcohol drinking was categorised

according to the National Health and Medical Council’s guidelines, the study found

that the prevalence of short-term risky and high risk drinking was 2.26%, and

long-term risky and high risk drinking was 4.52%, respectively. Compared to the

national data for Australian women, which revealed that the proportion of risky and

high risk drinking (including short- and long-term) level has increased from 6.2% in

1995 to 11.7% for the period between 2004 and 2005 (Australian Bureau of

Statistics, 2006c), risky and high risk drinking seems less of a problem for midlife

and older women.

In addition, the study suggests that the relationship between alcohol use and

mental health changes as women age. No relationship between the two variables is

found initially, yet as time goes by, women who were past alcohol drinkers present

less anxiety symptoms than women who were non-drinkers. No difference in mental

health was found between current drinkers and non-drinkers.

The non-significant relationship between current alcohol drinking and mental

health is in contrast to some earlier studies, which claim that moderate alcohol

drinking is related to improved quality of life (mentally and physically) and healthier

ageing for older women (Byles et al., 2006; Chan et al., 2009). The evaluation of

alcohol drinking in the current study is compared to that of earlier studies as an

attempt to explain the differences.

Regarding the assessment of alcohol, the study obtained information by

asking women “how often do you drink alcohol?”. It can be seen from this question

that what was measured is the frequency of alcohol consumption. Graham et al. have

215

discussed the effect of alcohol use measurement on its associations with depression,

and concluded that depression is unrelated with frequency of alcohol use, but more

linked with a larger number of standard drinks per occasion (Graham, Massak,

Demers, & Rehm, 2007). This suggests that frequency-based categorisation may

result in different classification levels from what is obtained by using a quantity

based method. However, as mentioned earlier in this section, the percentage of

women drinking at risky and high risk levels is quite low (2.26% to 4.52%) in this

sample. Therefore, the likelihood of missing the capture of harmful alcohol drinking

is rare.

It is seen that as women age, the anxiety levels of women who were past

drinkers were lower than non-drinkers. Literature has demonstrated that people who

quit alcohol drinking are more likely to have mental health problems such as

depression and anxiety as part of their withdrawal syndrome (Ducci et al., 2007). It

is considered that there may be a few reasons behind this phenomenon. First, the

past-drinkers in the current study may not have been heavy problem drinkers before

they quit. In this circumstance, even though they refrain from drinking alcohol,

psychological problems are less likely to occur. Second, reduced anxiety levels may

be related with physical activity. In the current study, the prevalence of women

doing daily exercise among past-drinkers is 43%, which is nearly two times higher

than that of non-drinkers (16.4%). The beneficial psychological effect of physical

activity has been widely acknowledged, therefore, it is considered that women who

drank in the past benefit psychologically from doing adequate activity in the long

term.

216

Age and Mental Health

The mental health of Australian midlife and older women improves as they

age. This is evidenced by reduced psychological symptoms including depression.

This finding collaborates with the results of two large women’s studies: the

Australian Longitudinal Study of Women’s Health (ALSWH) and the Nurses’

Health Study. Both found that generally women’s mental health improves as they

age, although some negative factors (e.g. smoking) could prevent this trend from

occurring in a subgroup of women. The participants in the current study are middle

aged or older, around which time of life a considerable number of changes, both

biological (e.g. menopausal symptoms) and social (e.g.empty nest), often occur.

Traditional views tend to highlight these factors as the sources of stress for women,

but overlook the positive aspects of midlife. Actually, research (Hunter et al., 2002)

has shown most women experience a high level of well-being, optimism and

satisfaction about life, power and personal achievement at middle-age, although the

deterioration of physical health is an inevitable factor that can confront some

women.

The Effects of Lifestyle Factors on the Mental Health of Women With and Without

Diabetes

Research questions:

a. What is the difference in mental health between women with and without

diabetes?

b. What are the contributing factors of the difference in mental health between

women with and without diabetes?

217

Based on ten controlled studies estimating the prevalence of depression

among diabetic populations, Ali et al. concluded that the prevalence of depression

among people with Type 2 diabetes was nearly twice that of those without (Ali, et al.,

2006).

Although such meta-analysis has the advantages of using a larger sample size

and a greater ability to generalise the results to other populations, it is not without its

limitations. As commented by the authors, diabetic and non-diabetic groups included

in many studies differ from each other in variables known to be associated with

depression. For example, confounders like socioeconomic factors, obesity and

co-morbid conditions were not always adjusted for when comparing the depression

levels between the two groups. Therefore, the resulting analysis may not accurately

reflect the true differences. Again, for anxiety, the German National Study, which

included 4,169 individuals from the community (Kruse et al., 2003), found that

patients with diabetes are 90% (95%CI, 1.19-3.14) more likely to have anxiety

disorder when compared to those without. However, this study also has a limitation

of only controlling for sociodemographic factors. The current study, thus, has the

advantage of considering the confounding effect of lifestyle factors and

co-morbidities that could have an effect on the relationship between diabetes and

mental health.

The current study showed that before adjusting for sociodemographic factors,

lifestyle factors and number of chronic diseases, women with diabetes were found to

have more depressive symptoms and a lower level of general mental health, but not

anxiety. After controlling for confounders, the difference in depression and general

mental health vanished to being non-significant. Therefore, this means that diabetes

itself is not related with lower levels of mental health.

218

Rather, lifestyle factors including physical inactivity and the number of

co-morbidities impact on the mental health of women. The relationship between

physical activity and mental health again confirms the existing literature (Brown et

al., 2005; Dunn, Trivedi, & O'Neal, 2001), and lends further support to the beneficial

psychological effect of physical activity and adverse effect of physical inactivity.

The important implication of the result is that there is a potential to improve the

mental health of women with diabetes, as physical inactivity can be modified with

effort.

What needs to be considered is that women with diabetes often have other

chronic conditions that can strongly affect their ability and motivation to do physical

activity. This could be the reason why women with diabetes reported lower levels of

physical activity. Therefore, research aiming to increase physical activity for women

with diabetes must take this into consideration to enable a tailored and effective

program that is suitable for the population.

Overweight and obesity did not impose an adverse effect on women’s mental

health. Some of the research studies on obesity and health-related quality of life

indicated that although overweight and obesity can sometimes limit people’s

physical function and compromise their level of physical function, their mental

health remains less disturbed (Goins, Spencer, & Krummel, 2003; Sturm &Wells,

2001). Two other lifestyle factors, smoking and alcohol use, were not related with

mental health, which is probably due to the small numbers of women who smoke

and drink in this sample.

In conclusion, the current study found the levels of mental health of women

with diabetes are lower than women without. The disparity in mental health results

from lower levels of physical activity and other co-morbidities, but not the condition

219

of having diabetes. Given that physical inactivity is modifiable, there is a potential

for the mental health of women with diabetes to be improved.

The Mediating Effect of Self-Efficacy, Mental Health and Lifestyle Factors

Research question:

Does self-efficacy mediate the relationships between lifestyle factors and

mental health among midlife and older women with diabetes?

The study suggests that self-efficacy mediates the relationships between

lifestyle factors, especially BMI and physical activity and depression. Specifically,

when examining the effect of lifestyle factors including BMI and physical inactivity

on depression, self-efficacy fully mediates the relationship. However, it was not

shown to be the one when examining the effect of depression on these lifestyle

factors.

From a statistical perspective, it can be explained why self-efficacy was not

shown to be a mediator when using depression to predict the lifestyle factors of BMI

and physical activity. It is noticed that, in particular, the correlation between

depression and self-efficacy in the study was strong. This strong correlation is

supported by social cognitive theory, in which mood is described as a source of

self-efficacy (Bandura, 1997). Based on Baron and Kenny’s theory, this means that

there is a strong correlation between the predicting variable (depression) and the

mediator (self-efficacy).

These researchers have stated that when a strong relationship exists between

a predicting variable and mediator, multicollinearity is likely to occur in regression

models. The consequence of multicollinearity is larger standard errors for

coefficients of each independent variable. In addition, finding a statistically

220

significant coefficient becomes more difficult due to lowered statistical power

(Kenny, 2009; Marill, 2004).

In the study, when lifestyle factors were regressed on depression and

self-efficacy, the observed powers of depression and self-efficacy were both low

(approximately 20%), and explained why self-efficacy is not the mediator of the

effect of depression on the lifestyle factors of BMI and physical activity. While BMI

and physical activity were analysed as initial variables (predictors) that cause

depression, self-efficacy was proved to be the mediator of the relationship. In this

circumstance, the correlations between the initial variables of BMI and physical

activity and the mediator (self-efficacy) were mild; therefore, the likelihood of

creating multicollinearity in regression models is low. Also, it was found that the

observed power of predictors was found to be adequate (about 80%), which

supported the reliability of the results.

Taken all together, the study showed that self-efficacy is the mediator of the

effect of the lifestyle factors of BMI and physical activity on depression, but not the

other way around. Although self-efficacy in managing chronic diseases is strongly

indicated as a full mediator of the effect of lifestyle factors on depression, it is

acknowledged that this study did not exam other potential mediators that could

possibly alternate the results. Nethertheless, the result that depression is related with

a lower self-efficacy level is supported by a number of other researchers (Aberle et

al., 2009; Bowser, Utz, Glick, Harmon, & Rovnyak, 2009; Sacco et al., 2005;

Wagner, Tennen, & Osborn, 2010). In addition, research with a larger sample size

may be warranted to test the mediating effect of self-efficacy with the effect of

depression on lifestyle factors.

221

Regarding the mediating effect of self-efficacy, the current study validated

the conclusion of the research by Sacco and colleagues (2005) that self-efficacy fully

mediates the relationship between BMI and self-efficacy among people with diabetes.

In addition, the current research also shows that self-efficacy fully mediates the

relationship between physical activity and depression. It is known that women with

diabetes also live with other chronic conditions that could limit their ability to

exercise. The resulting lower level of physical activity then may cause or exacerbate

depression because of isolation, lower energy and increased body weight. Moreover,

depression can reversely impact on people’s motivation to do physical activity, thus

a vicious circle is formed. Based on the findings of the current study, it can be seen

that self-efficacy is the central aspect in behaviour change; therefore, researchers or

clinicians should work tirelessly to enhance patients’ self-efficacy in order to initiate

and stabilise the effect of behavioural change. How to enhance patients’ self-efficacy

is discussed in the section on theoretical reflection below.

Self-Efficacy, Duration of Diabetes and Use of Antidepressants

Research question:

How does the duration of diabetes and use of antidepressants affect

self-efficacy in managing chronic disease?

The finding that a longer duration of diabetes is related with lower

self-efficacy in managing chronic disease is unexpected. Initially, it was

hypothesised that women with a longer duration of diabetes would have more

opportunities to build up their skills of diabetes management and thus should have

less psychological problems related to diabetes management. Age and the number of

222

co-morbidities were checked to explore the underlying reasons, because it was

thought that age may affect women’s ability to understand information and the

number of co-morbidities indicates the burden of disease management.

Further analysis of the differences in age and number of co-morbidities in

relation to diabetes duration was undertaken. No relationship between age and

duration of diabetes was found (65.18 ± 9.31 vs. 65.59 ± 9.32, t = -.196, p = .845);

therefore, age is not the factor that influences the correlation between duration of

diabetes and self-efficacy. Similarly, the number of co-morbidities was not shown to

be related the diabetes duration. Thus, complicated disease management tasks do not

explain the relationship between lower self-efficacy and a longer duration.

Finally, it is regarded that although women have had diabetes for a long time

(e.g. over two years), they do not necessarily get opportunities to learn and enhance

their chronic disease management skills. When the symptoms deteriorate over time,

the management tasks do get complicated and mastering these skills becomes more

difficult than when the diabetes started. Yet for women with a shorter duration of

diagnosis, the management tasks are relatively easier, so gaining confidence appears

to be a quicker and more efficient process when they are taught how to manage

diabetes.

Theoretical Reflection

The findings from the current study partially support the hypothesised

conceptual framework proposed in Chapter 2 (also see Figure 7.1). Based on the

social cognitive theory, the initial conceptual framework has two primary

assumptions. The first assumption is that that lifestyle factors including overweight

and obesity, physical activity, smoking and alcohol use have impacts on mental

223

health, and second is that self-efficacy mediates the effect of mental health on

lifestyle factors.

As indicated by the study results, the first assumption about the effect of

lifestyle factors on mental health is met, and in addition, the effect of each lifestyle

factor on mental health varies. The second assumption of the framework is not

supported. Instead, the mediating effect of self-efficacy on the effect of lifestyle

factors on mental health is strongly indicated.

In brief, the new conceptual framework clearly shows that lifestyle factors do

affect mental health, and self-efficacy is the mediator in the relationship among

women with Type 2 diabetes. Hence, it indicates that individuals’ mental health

needs to be considered when implementing lifestyle modification programs. Also, it

is suggested that concerted effort is required to promote people’s self-efficacy, which

may be effective to facilitate lifestyle change as well as mental health improvement.

The detailed information of the conceptual framework is delineated below.

All of the four lifestyle factors have impacts on mental health, but the

patterns of their relationships with mental health are different. First, physical activity

has the most robust relationship with mental health. It reduces psychological

symptoms, including depression and anxiety, and promotes general mental health.

From the other aspect, people who have inadequate level of physical activity

experience more frequent and severe mental health problems. Second, the effect of

BMI on mental health varies with time. In Australian midlife and older women,

those with a high BMI value have better improvement of mental health as they age,

while in women with diabetes, those with higher BMI values have more depression

symptoms. The variation of the impact of BMI indicates that women who are

overweight or obese do not necessarily have lower level of mental health. The effect

224

of BMI on mental health may be influenced by other factors. Third, smoking

produces an adverse effect on mental health of midlife and older women. Although

the prevalence rate of smoking among midlife and older women is lower than the

population data, the psychological adverse effect on individuals who smoke is not

any less. Women who are non-smokers have experienced a steady improvement of

mental health over time, yet women who are smokers display a downward trend. The

adverse psychological effect of smoking may be difficult to detect in a short period,

but if women continue to smoke, their mental health will eventually be influenced.

Thus, smoking should be largely discouraged. Finally, women who were formerly

alcohol drinkers experience less anxiety symptoms as they age. This result may be

taken with caution as the number of abstainers in the current research is small, which

may limit the ability of generalising the results to a broader population.

Lifestyle modification programs for midlife and older women may be more

effective and cost-effective, if they consider the features of lifestyle profiles of this

population. The understanding of the relationships between lifestyle factors and

mental health may be more profound knowing the interactions among lifestyle

factors themselves. Among midlife and older women with diabetes, lower levels of

physical activity and a high BMI is related to depression. Self-efficacy fully

mediates the effect of these two lifestyle factors on depression, but the effect of

depression on these two factors. Clinically, this implies that women living with

unhealthy lifestyle factors including physical inactivity and obesity are not confident

in managing chronic diseases, suggesting additional support for these women in

improving lifestyle and disease management. The enhancement of self-efficacy for

these women will help reducing depression symptoms, better managing disease tasks,

and promoting health related quality of life.

225

It is highly recommended that clinical strategies aiming at improving

self-efficacy should heavily rely on the four major sources of self-efficacy. The four

major sources are: 1) mastery experience, 2) vicarious experience, 3) verbal

persuasion, and 4) self-appraisal. The most important source of self-efficacy is

mastery experience, which means people gain the most confidence from successful

completion of a certain task. For example, if health professionals are able to work

with women in developing new skills of controlling blood sugar by appropriate

physical activity, women’s self-efficacy in this regard will be largely enhanced. At

the same time, the depression symptoms may also be relieved effectively by a

rewarding feeling from having completed the task. The other three strategies should

also be implemented in combination to maximise the positive outcomes.

The theoretical framework generated from the current research provides

some guidance for developing lifestyle modification programs that could be tailored

for midlife and older women. Firstly, lifestyle modification needs to incorporate

mental health as an integrated component of the program because of their

relationships. Second, among all of the four factors, physical activity should be the

first priority due to its robust protective impact. Finally, the framework indicates for

women with diabetes, health professionals should work hard to enhance self-efficacy.

226

Figure 7.1. A conceptual framework for the relationships between lifestyle factors and mental health among Australian midlife and

older women.

Positive relationship Negative relationship

227

Strengths and Limitations

The strengths of the current research include the following: the study utilises

a large number and representative sample of midlife and older women. A

longitudinal prospective approach is applied to examine the temporal relationships

between lifestyle factors and mental health. In addition, the research study includes a

clinical population of women with diabetes and is able to conduct a comparative

study between general and clinical populations. The comparative study extends the

knowledge obtained from examining general populations, overcomes a shortage of

previous studies, and further enhances the understanding of the relationships

between lifestyle factors and mental health. More importantly, a theoretical

framework based on the social cognitive theory is used to explain the relationships

between mental health and lifestyle factors, which contribute knowledge about

behavioural change mechanisms in order to assist in the further development of

useful intervention. This is one of the few studies using an exclusive female sample

aged 45 years and older, which is important as the knowledge generated from the

study may facilitate healthy aging for women.

Based on the HOW project, the research has investigated the temporal

relationships between lifestyle factors and mental health over a period of five years.

The advantage of having an extended time frame is that it permits relatively reliable

observation of the improvement or deterioration of mental health problems, which

are chronic conditions that may not be easily observable in a shorter period (Pasco et

al., 2008). Furthermore, the study also acknowledges the potential interactions

among lifestyle factors themselves, and examines their relationships with each other

before analysing their individual effects on mental health. By doing this, the

interpretation of the individual temporal relationships between lifestyle factors and

228

mental health can be partly linked to the characteristics of other lifestyle factors at a

population level, thus enabling a more comprehensive understanding of this research

question in this female population. The results offer valuable information to health

promotion for women of this age group and give suggestions about some areas

worthy of attention from the psychological point of view.

The research study also clarifies the differences in mental health between

women with and without diabetes, and furthermore, finds out the factors that

contribute to this difference. As highlighted in a systematic review on the prevalence

of depression among people with diabetes (Ali et al., 2006), the diabetic and

non-diabetic populations included in many studies differed on variables that are

associated with depression, such as lifestyle factors. The current study addressed this

issue by controlling for a number of lifestyle factors and co-morbidities as well, thus

the results may be more reliable. It should be noticed that when comparing women

with and without diabetes, the age of the two groups did not match, which may have

introduced some bias into the results. But statistical analysis was controlled to

minimise the bias. In addition, although the study clearly shows that a low level of

physical activity is one of the important factors in relation to lower levels of mental

health among women with diabetes, no causal relationship is indicated due to a

cross-sectional design.

The reliance on self-reporting measures for all of the variables in the study

may introduce some information bias to the study. For example, when using a

self-report strategy to obtain people’s height and weight, the BMI is often

underestimated. This is because, intentionally or not, people tend underestimate

weight and overestimate height (Gorber et al., 2007). But given the large sample size

of the study, obtaining an objective measurement of height and weight is financially

229

very difficult. Hence, a self-reported BMI has been applied. Despite the disadvantage

of using a self-reported BMI, there is evidence showing that in women, a substantial

strong agreement exists between measured and self-reported BMIs, except for

women who are pregnant, older than 75 years or without physician visits (Craig &

Adams, 2009). Apart from BMI, the assessment of alcohol drinking in the study is

frequency focused; therefore, the actual quantity of alcohol consumed per occasion

cannot be calculated. In this case, abnormal drinking patterns (e.g. binge drinking)

are not examined. However, as discussed, the prevalence of risky and high risk

drinking among this particular population is quite low, so, the possibility of failing to

capture abnormal drinking is low.

The literature indicates two ways of assessing mental health, which are

clinical interviews or self-report psychometric scales. In the current study, mental

health is examined by using self-report psychological scales. Some argue that

clinical interviews are objective methods, which are considered to be more reliable

and valid than psychometric scales. However, the interest of the current research is

not clinical mental health disorders, such as major depression, but more on women’s

mental health symptoms, which can be much more minor. In addition, the

measurement tools of mental health utilised in the study are all well validated and

appropriate to the study population.

As highlighted by Bandura, the measurement of self-efficacy should be

circumstance specific, indicating a high level of self-efficacy in one domain does not

necessarily imply the same level of confidence in the other. For example, someone

with high self-efficacy in mathematics does not necessarily feel confident in

literature. In the current study, self-efficacy in managing chronic disease is evaluated,

while the examined behaviours are health related lifestyle such as physical activity.

230

Managing chronic diseases is a multitask which involves other activities beyond

lifestyle modification, for example, adhering medications. From this point of view,

self-efficacy in managing chronic diseases may not be the most accurate measure of

women’s self-efficacy in improving their lifestyles. However, it needs to be noticed

that self-efficacy also has the property of generality (Bandura, 1997), which means

that self-efficacy in one area can actually be extended to another. Given that lifestyle

modification is one of the integrated aspects of chronic diseases management,

women’s confidence in their abilities to manage chronic diseases may still be a good

reflection of the self-efficacy in exercise or diet. In addition, this scale of

self-efficacy in managing chronic disease has other advantages, too. First, it suits the

population among whom multiple chronic conditions are not uncommon, despite the

focus on diabetes. Second, it is a brief version containing six items only, which is

unlikely to burden participants, and the possibility of inaccurate responses resulting

from tiredness from filling out the questionnaires is low.

This research study examines the relationships between mental health and

lifestyle factors by using both longitudinal and cross-sectional designs, and is able to

compare the general population with a clinical cohort. The usefulness of social

cognitive theory in explaining these relationships is examined as well. It is expected

that knowledge generated from the study will contribute to the study of women’s

health, especially in the areas of healthy ageing and health promotion.

231

CHAPTER 8: CONCLUSIONS

Introduction

This chapter starts with a brief summary of the major findings of the research,

followed by the implications generated from the research for education, research and

policy development, respectively, and finally finishes with the conclusions for the

study. The current research attempts to develop a comprehensive understanding of

the effect of lifestyle factors on mental health in Australian midlife and older women.

It has the potential to enhance the knowledge of health professionals, researchers and

policy makers with regard to lifestyle factors from a psychological perspective. It

may also be beneficial to the development of health promotion programs that aim to

improve the lifestyles and mental health for women in Australia.

Summary of Major Findings

Effect of Lifestyle Factors on Mental Health in Midlife and Older Women

The four lifestyle factors of overweight and obesity, physical activity,

smoking and alcohol use have a different impact on mental health among midlife and

older women.

First, women with a higher BMI have lower levels of mental health

(depression and general mental health) than women with normal weight. However,

as women age, the mental health of women who are overweight and obese becomes

better than that of women with normal weight. Second, women who are physically

active have higher levels of mental health than those who are not. Third, smoking

adversely impacts on their mental health. This negative effect may not exhibit when

women were younger, yet the adverse effect of smoking is clear over the long term.

232

Fourth, those who were formerly drinkers of alcohol had fewer anxiety symptoms

than women who were always non-drinkers, as both groups age.

Although the patterns of the relationship of each individual lifestyle factor

with mental health varies, all the evidence supports women living with a healthier

lifestyle have a better status of mental health. Among the four modifiable lifestyle

factors, physical activity is deemed to be particularly important for midlife and older

women due to its strong association with mental health and the effects produced on

other lifestyle factors.

Effect of Lifestyle Factors on Mental Health in Midlife and Older Women With and

Without Diabetes

Women with diabetes do have lower levels of mental health than women

without. However, the difference is not related with diabetes per se, rather, it is

contributed to by the low level of physical activity and the higher number of

co-morbidities that women with diabetes commonly present.

This finding extends and verifies the knowledge about the relationship

between physical activity and mental health. More importantly, it sheds light on the

development of strategies that could be used in improving the mental health of

women with diabetes. Given the effect of physical activity on mental health, and its

modifiable nature, it is believed that physical activity promotion will benefit midlife

and older women both physically and psychologically.

233

The Mediating Role of Self-Efficacy in the Associations Between Lifestyle Factors

and Mental Health Among Midlife and Older Women With Diabetes

Self-efficacy fully mediates the effect of the lifestyle factors of BMI and

physical activity on depression. Women with longer durations of diabetes have lower

levels of self-efficacy in managing chronic diseases, suggesting a strong need for

interventions facilitating women to master chronic disease management skills. In

addition, women using antidepressants have lower levels of self-efficacy than

women who do not.

Based on the findings, self-efficacy enhancement strategies may be

particularly important for mental health improvement. Previous studies consistently

show that people with higher levels of self-efficacy are more likely to assume

successful behaviour changes, while the current studies suggest that individuals are

also likely to benefit psychologically if their self-efficacy is enhanced. Therefore,

future intervention programs should continue to place a heavy emphasis on

facilitating people’s self-efficacy by providing for the four major sources of

self-efficacy in an effective manner.

Implications

Implications of the Study

The profile of lifestyle factors among Australian midlife and older women

illustrates that concerted effort should be made towards promoting and stabilising

physical activity. The impact of alcohol use on mental health seems minor among

women. Considering the low prevalence of risky and high risk drinking, alcohol use

may not be the major target of health promotion for this population, but this finding

234

may not apply to other populations such as males (Australian Bureau of Statistics,

2006c). A great number of studies indicate overweight and obesity is related with

more severe mental health problems (Carroll et al., 2010; Kasen et al., 2008;

Luppino et al., 2010), but the current study shows the mental health of women who

are overweight and obese improves better as women age. Although the traditional

view of middle age consistently links middle age with stress, decline of health and so

forth, middle age is also characterised by career success, release from child-bearing

responsibilities, financial stability (Australian Bureau of Statistics, 2006d) and more

personal space (Hunter et al., 2002).

Overall, in health promotion for midlife and older women, physical activity is

the priority. It is a robust protector of mental health for women with and without

diabetes. In addition, physical activity is also one of the essential strategies for losing

weight (Hamman et al., 2006); therefore, focusing on physical activity also facilitates

weight management at the same time. Emphasis on physical activity rather than BMI

value may also help to avoid potential distress resulting from failure or fluctuation of

weight loss (Teixeira et al., 2002).

In terms of how to motivate women to increase their levels of physical

activity, self-efficacy can never be overemphasised. Women with high self-efficacy

have better mental health and healthier lifestyle factors, and self-efficacy is the

mediator in the relationship between mental health and lifestyle factors. Thus, there

is solid evidence that self-efficacy should be centrally targeted in behaviour change

programs, which is likely to benefit women both psychologically and physically.

235

Implications for Health Education

The percentages of elderly people in Australia continue to expand (Australian

Bureau of Statistics, 2008b), so does the proportion of chronic diseases within this

population (Australian Bureau of Statistics, 2006f). Leading a healthy lifestyle at

midlife is considered to be vitally important for preventing and improving clinical

outcomes of chronic diseases amongst older population. Therefore, motivating

clients to make lifestyle changes will continue to be a major task of health

professionals. When conducting health education regarding lifestyle change for

clients, health providers need to bear in mind that women who live with unhealthy

lifestyles also have lower levels of mental health, which may affect their ability to

make successful lifestyle changes (Wott & Carels, 2010). It is recommended that

before consultation, health professionals conduct a systematic assessment of clients’

lifestyle and mental health status. The assessment of lifestyle factors allows health

professionals to individualise their care (Clark & Hampson, 2001) and possibly

effective use of limited clinical resources. Some brief mental health assessment tools

like HADS may be helpful in identify potential mental health problems in a timely

manner, and provide useful information for further consultation. If a client is found

to have mental issues, health professionals may want to help them solve these

problems first. If a client is mentally well, but still reporting difficulty in changing

lifestyle, health professionals may need to consider whether any of the

socio-environmental factors is preventing him/her from to do so (Strayhorn, 2009).

Moreover, it is worthwhile for health professionals to work towards

self-efficacy enhancement by providing clients with the four sources of self-efficacy

proposed by Bandura (1997). Task accomplishment is the strongest source of

gaining confidence. Setting up reasonable goals for changes within set time periods

236

for clients and implementing individualised care are important strategies to realise

optimal behavioural change at each stage. After a certain period of practice, it is

important to review the progress of behavioural change. This is a great opportunity

to evaluate the previously developed plan and bring up with solutions to problems if

there is any. In brief, health professionals shall spare no effort to facilitate clients to

make positive lifestyle change. It is a long journey and it is not easy, but it is

certainly worth all the effort.

Implications for Policy Making

The importance of a healthy lifestyle is clearly illustrated in the current

research which has been informed by a psychological perspective. Considering the

existing literature on the impacts of lifestyle factors on physical health, it is

concluded that a healthy lifestyle can benefit individuals both physically (Breslow et

al., 2001; Ma et al., 2008; Parker et al., 2007; Rothenbacher et al., 2006) and

psychologically. Therefore, lifestyle promotion or modification should continue to

be the focus of policy development. At the policy making level, governments should

work very closely and intensively with researchers and health professionals in this

field. They should utilise the knowledge generated from the clinical practice as well

as research field to develop targeted policies to facilitate behavioural changes. For

example, if a lack of safe environment and instruments prevents a large number of

people exercising, building such infrastructures would be a very appropriate and

effective approach (American Heart Association, 2006; Strayhorn, 2009). Or, if the

public have little awareness of access to health care or the potential consequences of

living unhealthy lifestyle, mass media can be used to disseminate information.

237

In the meantime, government or policy makers should also strive to create a

supportive environment for researchers and clinicians, as they are the driving force

of improving lifestyle for the population. Adequate support will motivate clinicians

to work more efficiently with clients and also allow researchers to conduct large

scale and rigorous studies to inform policy making in turn. In addition, government

should aim to largely promote collaboration between clinicians and researchers. It is

important to translate knowledge from the research to the first line of practice;

therefore, clinicians can be informed with the most updated information and

researchers can apply their research into practice. Not only government, but also

employers are able to contribute to health promotion for their staffs. For example,

organisations can encourage employees to use stairs as means of increasing physical

activity.

In brief, government and policy makers are at vital position in promoting

health for the society. If they make tireless effort to create a supportive environment

for both research and clinical practice, achieving the goal of health promotion may

become easier.

Implications for Future Research

The longitudinal analysis in the current research offers valuable information

about the temporal effect of lifestyle factors on mental health among midlife and

older women from the general population over an extended time frame. However,

one’s lifestyle can change overtime (Sternfeld et al., 2004). A woman who has a

sedentary lifestyle can initiate regular physical activity, and in contrast, a slim and

healthy woman can also put on weight. In other words, there is a fair chance that

women’s lifestyle patterns alternate over time. In this case, capturing the changes of

238

lifestyle patterns of participants becomes essential. Future research may need to

follow the participants more frequently when research resources permit, in order to

develop a better understanding of the effect of lifestyle factors on mental health.

The sample size in the study examining the mediating role of self-efficacy

may not be adequate; therefore, when considering depression as predictor and

lifestyle factors as outcomes, the mediating effect of self-efficacy failed to be

detected. A larger sample size is needed to overcome the collinearity effect between

self-efficacy and depression. In addition, further studies may be warranted to include

other potential mediators to validate the mediating role of self-efficacy. Due to the

cross-sectional design, the causal relationship between self-efficacy, lifestyle factors

and mental health cannot be determined. It is recommended that the psychological

effect of interventions based on self-efficacy should be evaluated in future research.

Conclusions

The current research investigates the longitudinal impact of lifestyle factors

on mental health among a large and representative sample of midlife and older

women from the general community; examines the difference in mental health

between midlife and older women with and without diabetes as well as the

contributing effect of lifestyle factors to the disparities; and tests the mediating effect

of self-efficacy in explaining the relationships between lifestyle factors and mental

health among midlife and older women with diabetes.

Women living with unhealthy lifestyles generally have lower levels of mental

health. Diabetes does have an impact on mental health. To be precise, low levels of

physical activity and other co-morbidities, which are related to diabetes, produce a

negative effect on mental health. This finding highlights that the mental health of

women with diabetes can be improved by increasing their physical activity levels.

239

The social cognitive theory explains part of the relationships between mental health

and lifestyle factors, and there is strong evidence to suggest that self-efficacy is the

mediator of the effect of lifestyle factors on mental health. It is recommended that

future research should focus on enhancing the self-efficacy of participants to

improve their mental health.

240

REFERENCES

Aberle, I., Scholz, U., Bach-Kliegel, B., Fischer, C., Gorny, M., Langer, K., et al.

(2009). Psychological aspects in continuous subcutaneous insulin infusion: A

retrospective study. Journal of Psychology, 143(2), 147-160.

Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and

Human Decision Processes, 50(2), 179-211.

Ajzen, I., & Fishbein, M. (1980). Understanding attitudes and predicting social

behavior. Englewood Cliffs: Prentice-Hall.

Al-Delaimy, W. K., Willett, W. C., Manson, J. E., Speizer, F. E., & Hu, F. B. (2001).

Smoking and mortality among women with type 2 diabetes: The Nurses'

Health Study cohort. Diabetes Care, 24(12), 2043-2048.

Alati, R., Kinner, S., Najman, J. M., Fowler, G., Watt, K., & Green, D. (2004).

Gender differences in the relationships between alcohol, tobacco and mental

health in patients attending an emergency department. Alcohol and

Alcoholism, 39(5), 463-469.

Alati, R., Lawlor, D. A., Najman, J. M., Bor, W., O'Callaghan, M., & Williams, G.

M. (2005). Is there really a 'J-shaped' curve in the association between

alcohol consumption and symptoms of depression and anxiety? Findings

from the Mater- University Study of Pregnancy and its outcomes. Addiction,

100(5), 643-651.

Ali, S., Stone, M. A., Peters, J. L., Davies, M. J., & Khunti, K. (2006). The

prevalence of co-morbid depression in adults with Type 2 diabetes: a

systematic review and meta-analysis. Diabetic Medicine, 23(11), 1165-1173.

American Heart Association. (2010). Women and cardiovascular disease:statistics.

Retrieved 22 November 2009, from

241

http://www.americanheart.org/downloadable/heart/1260905040318FS10WM

10.pdf

American Heart Association. (2006). Physical activity intervention studies: what we

know and what we need to know: a scientific statement from the American

Heart Association Council on Nutrition, Physical Activity, and Metabolism

(subcommittee on physical activity); Council on Cardiovascular Disease in

the Young; and the Interdisciplinary Working Group on Quality of Care and

Outcomes Research. Circulation, 114, 2379-2752.

Anderson, D., Lang, C. P., & Xu, Q. (2010). Is lifestyle intervention for mature

women in the general population suitable for patients with diabetes?

Preliminary results comparing two populations. Paper presented at the

Women's Health: The 18th Annual Congress. Washington D.C., U.S.A.

Anderson, D., Mizzari, K., Kain, V., & Webster, J. (2006). The effects of a

multimodal intervention trial to promote lifestyle factors associated with the

prevention of cardiovascular disease in menopausal and postmenopausal

Australian women. Health Care for Women International, 27(3), 238-253.

Anderson, D., Yoshizawa, T., Gollschewski, S., Atogami, F., & Courtney, M. (2004).

Menopause in Australia and Japan: effects of country of residence on

menopausal status and menopausal symptoms. Climacteric, 7(2), 165-174.

Anderson, D. J., & Yoshizawa, T. (2007). Cross-cultural comparisons of

health-related quality of life in Australian and Japanese midlife women: the

Australian and Japanese Midlife Women's Health Study. Menopause, 14(4),

697-707.

Anderson, R. J., Grigsby, A. B., Freedland, K. E., de Groot, M., McGill, J. B.,

Clouse, R. E., et al. (2002). Anxiety and poor glycemic control: a

242

meta-analytic review of the literature. International Journal of Psychiatry in

Medicine, 32(3), 235-247.

Aoyagi, Y., Park, H., Park, S., & Shephard, R. J. (2010). Habitual physical activity

and health-related quality of life in older adults: Interactions between the

amount and intensity of activity (the Nakanojo Study). Quality of Life

Research, 19(3), 333-338.

Araya, R., Gaete, J., Rojas, G., Fritsch, R., & Lewis, G. (2007). Smoking and

common mental disorders: a population-based survey in Santiago, Chile.

Social Psychiatry and Psychiatric Epidemiology, 42(11), 874-880.

Atlantis, E., & Baker, M. (2008). Obesity effects on depression: systematic review of

epidemiological studies. International Journal of Obesity, 32(6), 881-891.

Australian Bureau of Statistics. (1997). Mental health and Wellbeing: profile of

adults, Australia (No. 4326.0). Canberra: Australian Bureau of Statistics.

Australian Bureau of Statistics. (2006a). Census of population and housing: media

releases and fact sheets (2914.0.55.002). Canberra: Australian Bureau of

Statistics.

Australian Bureau of Statistics. (2006b). Tobacco smoking in Australia: a snapshot,

2004-2005 (4831.0.55.001). Canberra: Australian Bureau of Statistics.

Australian Bureau of Statistics. (2006c). Alcohol consumption in Australia: a

snapshot, 2004-05 (4832.0.55.001). Canberra: Australian Bureau of

Statistics.

Australian Bureau of Statistics. (2006d). Australian social trends 2006 (4102.0).

Canberra: Australian Bureau of Statistics.

243

Australian Bureau of Statistics. (2006e). Diabetes in Australia: a snapshot, 2004-05

(4820.0.55.001). Canberra: Australian Bureau of Statistics.

Australian Bureau of Statistics. (2006f). National health survey: summary of results,

2004-2005 (4364.0). Canberra: Australian Bureau of Statistics.

Australian Bureau of Statistics. (2006g). Physical activity in Australia: a snapshot,

2004-2005 (4835.0.55.001). Canberra: Australian Bureau of Statistics.

Australian Bureau of Statistics. (2007). National survey of mental health and

wellbeing: summary of results (4326.0). Canberra: Australian Bureau of

Statistics.

Australian Bureau of Statistics. (2008a). Overweight and obesity in adults


Australian Bureau of Statistics. (2008b). Population projections, Australia (cat no.

3222.0). Canberra: Australian Bureau of Statistics.

Australian Bureau of Statistics. (2009). National health survey: summary of results,


Australian Institute of Health and Welfare. (2004). Lonitudinal studies of ageing:

implications for future studies. Retrieved 22 November, from

http://www.aihw.gov.au/publications/age/lsa/lsa.pdf

Australian Institute of Health and Welfare. (2008). Diabetes Australian facts 2008

(No. CVD 40). Canberra: AIHW.

Avis, N. E., Stellato, R., Crawford, S., Bromberger, J., Ganz, P., Cain, V., et al.

(2001). Is there a menopausal syndrome? Menopausal status and symptoms

across racial/ethnic groups. Social Science & Medicine, 52(3), 345-356.

244

Ballard, K. D., Kuh, D. J., & Wadsworth, M. E. J. (2001). The role of the menopause

in women's experiences of the 'change of life'. Sociology of Health and

Illness, 23(4), 397-424.

Bandura, A. (1977). Self-efficacy: toward a unifying theory of behavioral change.

Psychol Rev, 84(2), 191-215.

Bandura, A. (1986). Social foundations of though and action: a social cognitive

theory. Englewood Cliffs, N.J.: Prentice-Hall.

Bandura, A. (2004). Health promotion by social cognitive means. Health Education

& Behavior, 31(2), 143-164.

Bandura., A. (1997). Self-efficacy: the exercise of control. New York: W.H.

Freeman.

Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction

in social psychological research: Conceptual, strategic, and statistical

considerations. Journal of Personality and Social Psychology, 51(6),

1173-1182.

Barr, E. L. M., Magliano, D. J., Zimmet, P. Z., Polkinghorne, K. R., Atkins, R. C.,

Dunstan, D. W., et al. (2005). AusDiab 2005-the Australian diabetes,obesity

and lifestyle study. Retrieved 22 November 2009, from

http://www.diabetes.com.au/pdf/AUSDIAB_Report_Final.pdf

Barry, D., Pietrzak, R. H., & Petry, N. M. (2008). Gender differences in associations

between body mass index and DSM-IV mood and anxiety disorders: results

from the National Epidemiologic Survey on Alcohol and Related Conditions.

Annals of Epidemiology, 18(6), 458-466.

245

Bauman, A., & Owen, N. (1999). Physical activity of adult Australians:

epidemiological evidence and potential strategies for health gain. Journal of

Science and Medicine in Sport / Sports Medicine Australia, 2(1), 30-41.

Begg, S., Vos, T., Barker, B., Stevenson, C., Stanley, L., & Lopez, A. D. (2007). The

burden of disease and injury in Australia 2003. Canberra:AHIW.

Bender, T., Nagy, G., Barna, I., Tefner, I., Kadas, E., & Geher, P. (2007). The effect

of physical therapy on beta-endorphin levels. European Journal of Applied

Physiology, 100(4), 371-382.

Bhui, K., & Fletcher, A. (2000). Common mood and anxiety states: gender

differences in the protective effect of physical activity. Social Psychiatry and

Psychiatric Epidemiology, 35(1), 28-35.

Bjerkeset, O., Romundstad, P., Evans, J., & Gunnell, D. (2008). Association of adult

body mass index and height with anxiety, depression, and suicide in the

general population: the HUNT study. American Journal of Epidemiology,

167(2), 193-202.

Bonnet, F., Irving, K., Terra, J. L., Nony, P., Berthezene, F., & Moulin, P. (2005).

Anxiety and depression are associated with unhealthy lifestyle in patients at

risk of cardiovascular disease. Atherosclerosis, 178(2), 339-344.

Bornstein, S. R., Schuppenies, A., Wong, M. L., & Licinio, J. (2006). Approaching

the shared biology of obesity and depression: the stress axis as the locus of

gene-environment interactions. Molecular Psychiatry, 11(10), 892.

Bowser, D. M., Utz, S., Glick, D., Harmon, R., & Rovnyak, V. (2009). The

relationship between diabetes mellitus, depression, and missed appointments

in a low-income uninsured population. Diabetes Educator, 35(6), 966-977.

246

Breslow, R. A., Ballard-Barbash, R., Munoz, K., & Graubard, B. I. (2001).

Long-term recreational physical activity and breast cancer in the National

Health and Nutrition Examination Survey I epidemiologic follow-up study.

Cancer Epidemiology, Biomarkers & Prevention, 10(7), 805-808.

Briffa, T. G., Maiorana, A., Sheerin, N. J., Stubbs, A. G., Oldenburg, B. F., Sammel,

N. L., et al. (2006). Physical activity for people with cardiovascular disease:

recommendations of the National Heart Foundation of Australia. The Medical

Journal of Australia, 184(2), 71-75.

Brown, W. J., Ford, J. H., Burton, N. W., Marshall, A. L., & Dobson, A. J. (2005).

Prospective study of physical activity and depressive symptoms in

middle-aged women. American Journal of Preventive Medicine, 29(4),

265-272.

Bruffaerts, R., Demyttenaere, K., Vilagut, G., Martinez, M., Bonnewyn, A., De

Graaf, R., et al. (2008). The relation between body mass index, mental health,

and functional disability: a European population perspective. Canadian

Journal of Psychiatry, 53(10), 679-688.

Byles, J., Young, A., Furuya, H., & Parkinson, L. (2006). A drink to healthy aging:

The association between older women's use of alcohol and their

health-related quality of life. Journal of the American Geriatrics Society,

54(9), 1341-1347.

Cai, L., Lubitz, J., Flegal, K. M., & Pamuk, E. R. (2010). The predicted effects of

chronic obesity in middle age on medicare costs and mortality. Med Care,

48(6), 510-517.

Caldwell, T. M., Rodgers, B., Jorm, A. F., Christensen, H., Jacomb, P. A., Korten, A.

E., et al. (2002). Patterns of association between alcohol consumption and

247

symptoms of depression and anxiety in young adults. Addiction, 97(5),

583-594.

Cameron, A. J., Dunstan, D. W., Owen, N., Zimmet, P. Z., Barr, E. L., Tonkin, A.

M., et al. (2009). Health and mortality consequences of abdominal obesity:

evidence from the AusDiab study. Medical Journal of Australia, 191(4),

202-208.

Capewell, S., Ford, E. S., Croft, J. B., Critchley, J. A., Greenlund, K. J., & Labarthe,

D. R. (2009). Cardiovascular risk factor trends and potential for reducing

coronary heart disease mortality in the United States of America. Bulletin of

the World Health Organisation, 88(2), 120-130.

Carpenter, K. M., Hasin, D. S., Allison, D. B., & Faith, M. S. (2000). Relationships

between obesity and DSM-IV major depressive disorder, suicide ideation,

and suicide ideation and suicide attempts: results from a general population

study. American Journal of Public Health, 90(2), 251-257.

Carr, D., & Friedman, M. A. (2005). Is obesity stigmatizing? Body weight,

perceived discrimination, and psychological well-being in the United States.

Journal of Health and Social Behavior, 46(3), 244-259.

Carroll, D., Blanck, H., Serdula, M., & Brown, D. (2010). Obesity, physical activity,

and depressive symptoms in a cohort of adults aged 51 to 61. Journal of

Aging and Health, 22(3), 384-398.

Cassidy, A. E., Bielak, L. F., Zhou, Y., Sheedy, P. F., 2nd, Turner, S. T., Breen, J. F.,

et al. (2005). Progression of subclinical coronary atherosclerosis: does

obesity make a difference? Circulation, 111(15), 1877-1882.

Chan, A. M., von Muhlen, D., Kritz-Silverstein, D., & Barrett-Connor, E. (2009).

Regular alcohol consumption is associated with increasing quality of life and

248

mood in older men and women: the Rancho Bernardo Study. Maturitas,

62(3), 294-300.

Chen, R. Q., Davis, S. R., Wong, C. M., & Lam, T. H. (2010). Validity and cultural

equivalence of the standard Greene Climacteric Scale in Hong Kong.

Menopause, 17(3), 630-635.

Chen, Y., Jiang, Y., & Mao, Y. (2009). Association between obesity and depression

in Canadians. Journal of Women's Health, 18(10), 1687-1692.

Chrousos, G. P. (2000). The role of stress and the hypothalamic-pituitary-adrenal

axis in the pathogenesis of the metabolic syndrome: neuro-endocrine and

target tissue-related causes. International Journal of Obesity and Related

Disorders, 24(S2), S50.

Chu, S. Y., Callaghan, W. M., Kim, S. Y., Schmid, C. H., Lau, J., England, L. J., et

al. (2007). Maternal obesity and risk of gestational diabetes mellitus.

Diabetes Care, 30(8), 2070-2076.

Ciechanowski, P. S., Katon, W. J., & Russo, J. E. (2000). Depression and diabetes:

impact of depressive symptoms on adherence, function, and costs. Archives

of Internal Medicine, 160(21), 3278-3285.

Clark, M., & Hampson, S. E. (2001). Implementing a psychological intervention to

improve lifestyle self-management in patients with Type 2 diabetes. Patient

Education and Counselling, 42(3), 247-256.

Cohen, J. W. (1988). Statistical power analysis for the behavioural sciences (2nd ed.).

Hillsdale, NJ: Lawrence Erlbaum Associates.

Cohen, L. S., Soares, C. N., Vitonis, A. F., Otto, M. W., & Harlow, B. L. (2006).

Risk for new onset of depression during the menopausal transition: the

249

Harvard study of moods and cycles. Archives of General Psychiatry, 63(4),

385-390.

Colagiuri, S., Lee, C. M. Y., Colagiuri, R., Magliano, D., Shaw, J. E., Zimmet, P. Z.,

et al. (2010). The cost of overweight and obesity in Australia. The Medical

Journal of Australia, 192(5), 260-264.

Collaris, R., Sidhu, K., & Chan, J. M. (2010). Prospective follow-up of changes in

menopausal complaints and hormone status after surgical menopause in a

Malaysian population. Menopause, 17(2), 351-358.

Collins, M. M., Corcoran, P., & Perry, I. J. (2009). Anxiety and depression

symptoms in patients with diabetes. Diabetic Medicine, 26(2), 153-161.

Craig, B. M., & Adams, A. K. (2009). Accuracy of body mass index categories

based on self-reported height and weight among women in the United States.

Maternal & Child Health Journal, 13(4), 489-496.

Cuijpers, P., Smit, F., Ten Have, M., & De Graaf, R. (2007). Smoking is associated

with first-ever incidence of mental disorders: a prospective population-based

study. Addiction, 102(8), 1303-1309.

Cummings, K. M., Jette, A. M., & Rosenstock, I. M. (1978). Construct validation of

the health belief model. Health Education Monographs, 6(4), 394-405.

De Jonge, P., Roy, J. F., Saz, P., Marcos, G., & Lobo, A. (2006). Prevalent and

incident depression in community-dwelling elderly persons with diabetes

mellitus: results from the ZARADEMP project. Diabetologia, 49(11),

2627-2633.

De Moor, M. H. M., Beem, A. L., Stubbe, J. H., Boomsma, D. I., & De Geus, E. J. C.

(2006). Regular exercise, anxiety, depression and personality: A

population-based study. Preventive Medicine, 42(4), 273-279.

250

Deecher, D., Andree, T. H., Sloan, D., & Schechter, L. E. (2008). From menarche to

menopause: Exploring the underlying biology of depression in women

experiencing hormonal changes. Psychoneuroendocrinology, 33(1), 3-17.

Degenhardt, L., & Hall, W. (2003). Patterns of co-morbidity between alcohol use

and other substance use in the Australian population. Drug and Alcohol

Review, 22(1), 7-13.

Degges-White, S., & Myers, J. E. (2006). Women at midlife: an exploration of

chronological age, subjective age, wellness, and life satisfaction. Adultspan

Journal, 5(2), 67-80.

Department of Health and Ageing. (2008). Lifescripts. Retrieved 22 November 2009,

from

http://www.health.gov.au/internet/main/publishing.nsf/Content/health-pubhlt

h-strateg-lifescripts-index.htm

Dhaliwal, S. S., & Welborn, T. A. (2009). Central obesity and cigarette smoking are

key determinants of cardiovascular disease deaths in Australia: a public

health perspective. Prev Med, 49(2-3), 153-157.

DiClemente, R. J., Crosby, R. A., & Kegler, M. C. (2009). Emerging theories in

health promotion practice and research (2nd ed.). San Francisco:

Jossey-Bass.

Dierker, L. C., Avenevoli, S., Stolar, M., & Merikangas, K. R. (2002). Smoking and

depression: an examination of mechanisms of comorbidity. The American


Donaghy, M. E. (2007). Exercise can seriously improve your mental health: fact or

fiction? Advances in Physiotherapy, 9(2), 76-88.

251

Dong, C., Sanchez, L. E., & Price, R. A. (2004). Relationship of obesity to

depression: a family-based study. International Journal of Obesity & Related

Metabolic Disorders, 28(6), 790-795.

Doyle, F., McGee, H. M., De La Harpe, D., Shelley, E., & Conroy, R. N. (2006).

The Hospital Anxiety and Depression Scale depression subscale, but not the

Beck Depression Inventory-Fast Scale, identifies patients with acute coronary

syndrome at elevated risk of 1-year mortality. Journal of Psychosomatic

Research, 60(5), 461-467.

Ducci, F., Enoch, M.-A., Funt, S., Virkkunen, M., Albaugh, B., & Goldman, D.

(2007). Increased anxiety and other similarities in temperament of alcoholics

with and without antisocial personality disorder across three diverse

populations. Alcohol, 41(1), 3-12.

Dunbar, J. A., Reddy, P., Davis-Lameloise, N., Philpot, B., Laatikainen, T.,

Kilkkinen, A., et al. (2008). Depression: an important comorbidity with

metabolic syndrome in a general population. Diabetes Care, 31(12),

2368-2373.

Duncan, B., & Rees, D. I. (2005). Effect of smoking on depressive symptomatology:

a reexamination of data from the National Longitudinal Study of Adolescent

Health. American Journal of Epidemiology, 162(5), 461-470.

Dunn, A. L., Trivedi, M. H., & O'Neal, H. A. (2001). Physical activity dose-response

effects on outcomes of depression and anxiety. Medicine and Science in

Sports Exercise, 33(Suppl. 6), S587-597.

Dzewaltowski, D. A., Noble, J. M., & Shaw, J. M. (1990). Physical activity

participation: Social cognitive theory versus the theories of reasoned action

252

and planned behavior. Journal of Sport & Exercise Psychology, 12(4),

388-405.

Ebrahim, S., Beswick, A., Burke, M., & Davey Smith, G. (2009). Multiple risk

factor interventions for primary prevention of coronary heart disease.

Cochrane Database of Systematic Reviews, 4, Art. No.: CD001561. DOI:

10.1002/14651858.CD001561.pub2

Egede, L. E. (2004). Effects of depression on work loss and disability bed days in

individuals with diabetes. Diabetes Care, 27(7), 1751-1753.

Elgar, F. J., & Stewart, J. M. (2008). Validity of self-report screening for overweight

and obesity. Evidence from the Canadian Community Health Survey.

Canadian Journal of Public Health, 99(5), 423-427.

Emmanuel, G., & Jatkin, D. K. (2007). Risk of dementia with increased body mass

index. Age and Ageing, 36(1), 23-29.

Eunkyung, P. (2009). Gender as a moderator in the association of body weight to

smoking and mental health. American Journal of Public Health, 99(1),

146-151.

Ezzati, M., Henley, S. J., Lopez, A. D., & Thun, M. J. (2005). Role of smoking in

global and regional cancer epidemiology: current patterns and data needs.

International Journal of Cancer, 116(6), 963-971.

Ezzati, M., & Lopez, A. D. (2004). Regional, disease specific patterns of

smoking-attributable mortality in 2000. Tobacco Control, 13(4), 388-395.

Fine, L. J., Philogene, G. S., Gramling, R., Coups, E. J., & Sinha, S. (2004).

Prevalence of multiple chronic disease risk factors: 2001 National Health

Interview Survey. American Journal of Preventive Medicine, 27(Suppl.1),

18-24.

253

Fishbein., M., & Ajzen., I. (1975). Belief, attitude, intention and behavior: an

introduction to theory an research. Reading, Mass, London:

Addison-Wesley.

Flegal, K. M., Graubard, B. I., Williamson, D. F., & Gail, M. H. (2007).

Cause-specific excess deaths associated with underweight, overweight, and

obesity. Journal of the American Medical Association, 298(17), 2028-2037.

Fox, C. S., Massaro, J. M., Hoffmann, U., Pou, K. M., Maurovich-Horvat, P., Liu, C.

Y., et al. (2007). Abdominal visceral and subcutaneous adipose tissue

compartments: association with metabolic risk factors in the Framingham

Heart Study. Circulation, 116(1), 39-48.

Freeman, E. W., Sammel, M. D., Lin, H., Gracia, C. R., & Kapoor, S. (2008).

Symptoms in the menopausal transition: hormone and behavioral correlates.

Obstetrics & Gynecology, 111(1), 127-136.

Freeman, E. W., Sammel, M. D., Liu, L., Gracia, C. R., Nelson, D. B., & Hollander,

L. (2004). Hormones and menopausal status as predictors of depression in

women in transition to menopause. Archives of General Psychiatry, 61(1),

62-70.

Fu, C.-W., Tan, A. W., Sheng, F., Luan, R.-S., Zhan, S.-Y., Chen, W.-Q., et al.

(2007). The prevalence of anxiety symptoms and depressive symptoms in

patients with somatic disorders in urban China: a multi-center cross-sectional

study. International Journal of Psychiatry in Medicine, 37(2), 185-199.

Fukukawa, Y., Nakashima, C., Tsuboi, S., Kozakai, R., Doyo, W., Niino, N., et al.

(2004). Age Differences in the Effect of Physical Activity on Depressive

Symptoms. Psychology and Aging, 19(2), 346-351.

254

Galán, I., Rodríguez-Artalejo, F., Díez-Gañán, L., Tobías, A., Zorrilla, B., &

Gandarillas, A. (2006). Clustering of behavioural risk factors and compliance

with clinical preventive recommendations in Spain. Preventive Medicine,

42(5), 343-347.

Gallicchio, L., Schilling, C., Miller, S. R., Zacur, H., & Flaws, J. A. (2007).

Correlates of depressive symptoms among women undergoing the

menopausal transition. Journal of Psychosomatic Research, 63(3), 263-268.

Galper, D. I., Trivedi, M. H., Barlow, C. E., Dunn, A. L., & Kampert, J. B. (2006).

Inverse association between physical inactivity and mental health in men and

women. Medicine and Science in Sports and Exercise, 38(1), 173-178.

Gandek, B., Sinclair, S. J., Kosinski, M., & Ware, J. E., Jr. (2004). Psychometric

evaluation of the SF-36 health survey in Medicare managed care. Health

Care Financing Review, 25(4), 5-25.

Gariepy, G., Wang, J., Lesage, A. D., & Schmitz, N. (2010). The longitudinal

association from obesity to depression: results from the 12-year National

Population Health Survey. Obesity, 18(5), 1033-1038.

Gariepy, G., Nitka, D., & Schmitz, N. (2009). The association between obesity and

anxiety disorders in the population: a systematic review and meta-analysis.

International Journal of Obesity, 34(3), 407-419.

Gaylord, G., Gruener, R., Rodgers, J., & Zalice, K. K. (Eds.). (2008).Mental health

nursing. Upper Saddle River, N.J.: Pearson/Prentice Hall.

Goins, R. T., Spencer, S. M., & Krummel, D. A. (2003). Effect of obesity on

health-related quality of life among Appalachian elderly. Southern Medical

Journal, 96(6), 552-557.

255

Goldney, R. D., Dunn, K. I., Air, T. M., Dal Grande, E., & Taylor, A. W. (2009).

Relationships between body mass index, mental health, and suicidal ideation:

population perspective using two methods. The Australian and New Zealand


Goodwin, R. D. (2003). Association between physical activity and mental disorders

among adults in the United States. Preventive Medicine, 36(6), 698-703.

Goodwin, R. E., & Hamilton, S. P. (2002). Cigarette smoking and panic: the role of

neuroticism. American Journal of Psychiatry, 159(7), 1208-1213.

Gorber, S. C., Tremblay, M., Moher, D., & Gorber, B. (2007). A comparison of

direct vs. self-report measures for assessing height, weight and body mass

index: a systematic review. Obesity Reviews, 8(4), 307-326.

Graham, K., Massak, A., Ders, A., & Rehm, J. (2007). Does the association between

alcohol consumption and depression depend on how they are measured?

Alcoholism, Clinical and Experimental Research, 31(1), 78-88.

Greene, J. G. (n.d.) The greene climacteric scale. Retrieved 6 January 2010, from

http://www.menopausematters.co.uk/greenescale.php, Retrieved on 6 January

2010

Greene, J. G. (1998). Constructing a standard climacteric scale. Maturitas, 29(1),

25-31.

Grigsby, A. B., Anderson, R. J., Freedland, K. E., Clouse, R. E., & Lustman, P. J.

(2002). Prevalence of anxiety in adults with diabetes: a systematic review.

Journal of Psychosomatic Research, 53(6), 1053-1060.

Guh, D. P., Zhang, W., Bansback, N., Amarsi, Z., Birmingham, C. L., & Anis, A. H.

(2009). The incidence of co-morbidities related to obesity and overweight: a

systematic review and meta-analysis. BMC Public Health, 9, 88.

256

Hämäläinen, J., Kaprio, J., Isometsa, E., Heikkinen, M., Poikolainen, K., Lindeman,

S., et al. (2001). Cigarette smoking, alcohol intoxication and major

depressive episode in a representative population sample. Journal of

Epidemiology and Community Health, 55(8), 573-576.

Hach, I., Ruhl, U. E., Klose, M., Klotsche, J., Kirch, W., & Jacobi, F. (2007).

Obesity and the risk for mental disorders in a representative German adult

sample. European Journal of Public Health, 17(3), 297-305.

Hach, I., Ruhl, U. E., Klotsche, J., Klose, M., & Jacobi, F. (2006). Associations

between waist circumference and depressive disorders. Journal of Affective

Disorders, 92(2-3), 305-308.

Haenle, M. M., Brockmann, S. O., Kron, M., Bertling, U., Mason, R. A., Steinbach,

G., et al. (2006). Overweight, physical activity, tobacco and alcohol

consumption in a cross-sectional random sample of German adults. BMC

Public Health, 6, 233-245.

Hamman, R. F., Wing, R. R., Edelstein, S. L., Lachin, J. M., Bray, G. A., Delahanty,

L., et al. (2006). Effect of weight loss with lifestyle intervention on risk of

diabetes. Diabetes Care, 29(9), 2102-2107.

Hardeman, W., Johnston, M., Johnston, D. W., Bonetti, D., Wareham, N. J., &

Kinmonth, A. L. (2002). Application of the Theory of Planned Behaviour in

behaviour change interventions: a systematic review. Psychology & Health,

17(2), 123-158.

Harkonmäki, K., Lahelma, E., Martikainen, P., Rahkonen, O., & Silventoinen, K.

(2006). Mental health functioning (SF-36) and intentions to retire early

among ageing municipal employees: the Helsinki Health Study.

Scandinavian Journal of Public Health, 34(2), 190-198.

257

Hart, C. L., & Smith, G. D. (2009). Alcohol consumption and use of acute and

mental health hospital services in the West of Scotland Collaborative

prospective cohort study. Journal of Epidemiology and Community Health,

63(9), 703-707.

Hasin, D. S., Goodwin, R. D., Stinson, F. S., & Grant, B. F. (2005). Epidemiology of

major depressive disorder: results from the National Epidemiologic Survey

on Alcoholism and Related Conditions. Archives of General Psychiatry,

62(10), 1097-1106.

Hassmén, P., Koivula, N., & Uutela, A. (2000). Physical exercise and psychological

well-being: a population study in Finland. Preventive Medicine, 30(1), 17-25.

Haynes, J. C., Farrell, M., Singleton, N., Meltzer, H., Araya, R., Lewis, G., et al.

(2005). Alcohol consumption as a risk factor for anxiety and depression:

results from the longitudinal follow-up of the National Psychiatric Morbidity

Survey. The British Journal of Psychiatry, 187(6), 544-551.

Healy, J. (2004). The benefits of an ageing population. Retrieved 5 February 2010,

from https://www.tai.org.au/documents/dp_fulltext/DP63.pdf

Heo, M., Pietrobelli, A., Fontaine, K. R., Sirey, J. A., & Faith, M. S. (2006).

Depressive mood and obesity in US adults: comparison and moderation by

sex, age, and race. International Journal of Obesity, 30(3), 513-519.

Hermanns, N., Kulzer, B., Krichbaum, M., Kubiak, T., & Haak, T. (2005). Affective

and anxiety disorders in a German sample of diabetic patients: prevalence,

comorbidity and risk factors. Diabetic Medicine, 22(3), 293-300.

Herrmann, C. (1997). International experiences with the Hospital Anxiety and

Depression Scale-A review of validation data and clinical results. Journal of

Psychosomatic Research, 42(1), 17-41.

258

Hildrum, B., Mykletun, A., Midthjell, K., Ismail, K., & Dahl, A. A. (2009). No

association of depression and anxiety with the metabolic syndrome: the

Norwegian HUNT study. Acta Psychiatrica Scandinavica, 120(1), 14-22.

Hinz, A., & Schwarz, R. (2001). Anxiety and depression in the general population:

normal values in the Hospital Anxiety and Depression Scale. Psychotherapie,

Psychosomatik, Medizinische Psychologie, 51(5), 193-200.

Honarmand, K., & Feinstein, A. (2009). Validation of the Hospital Anxiety and

Depression Scale for use with multiple sclerosis patients. Multiple Sclerosis,

15(12), 1518-1524.

Hrabosky, J. I., & Thomas, J. J. (2008). Elucidating the relationship between obesity

and depression: Recommendations for future research. Clinical Psychology:

Science and Practice, 15(1), 28-34.

Hu, G., Jousilahti, P., Barengo, N. C., Qiao, Q., Lakka, T. A., & Tuomilehto, J.

(2005). Physical activity, cardiovascular risk factors, and mortality among

Finnish adults with diabetes. Diabetes Care, 28(4), 799-805.

Huang, I. C., Frangakis, C., & Wu, A. W. (2006). The relationship of excess body

weight and health-related quality of life: evidence from a population study in

Taiwan. International Journal of Obesity, 30(8), 1250-1259.

Hunter, S., Sundel, S. S., & Sundel, M. (2002).Women at midlife: Life Experiences

and Implications for the Helping Professions.Washington, D.C. : National

Association of Social Workers.

Husky, M. M., Mazure, C. M., Paliwal, P., & McKee, S. A. (2008). Gender

differences in the comorbidity of smoking behavior and major depression.

Drug and Alcohol Dependence, 93(1-2), 176-179.

259

Ismail, K., Winkley, K., Stahl, D., Chalder, T., & Edmonds, M. (2007). A cohort

study of people with diabetes and their first foot ulcer: the role of depression

on mortality. Diabetes Care, 30(6), 1473-1479.

Jackson, T. (2006). Relationships between perceived close social support and health

practices within community samples of American women and men. Journal

of Psychology, 140(3), 229-246.

Janssen, I., Katzmarzyk, P. T., & Ross, R. (2002). Body mass index, waist

circumference, and health risk: evidence in support of current National

Institutes of Health Guidelines. Archives of Internal Medicine, 162(18), 2074.

Jee, S. H., Sull, J. W., Park, J., Lee, S.-Y., Ohrr, H., Guallar, E., et al. (2006).

Body-mass index and mortality in Korean men and women. Journal of the

American Medical Association, 355(8), 779-787.

Jeon, C. Y., Lokken, R. P., Hu, F. B., & Van Dam, R. M. (2007). Physical activity of

moderate intensity and risk of type 2 diabetes: a systematic review. Diabetes

Care, 30(3), 744-752.

John, U., Hanke, M., Rumpf, H. J., & Thyrian, J. R. (2005). Smoking status,

cigarettes per day, and their relationship to overweight and obesity among

former and current smokers in a national adult general population sample.


Johnson, J. G., Cohen, P., Pine, D. S., Klein, D. F., Kasen, S., & Brook, J. S. (2000).

Association between cigarette smoking and anxiety disorders during

adolescence and early adulthood. Journal of the American Medical

Association, 284(18), 2348-2351.

260

Johnson, S. S., Driskell, M., Johnson, J. L., Prochaska, J. M., Zwick, W., &

Prochaska, J. O. (2006). Efficacy of a transtheoretical model-based expert

system for antihypertensive adherence. Disease Management, 9(5), 291-301.

Jorm, A. F., Korten, A. E., Christensen, H., Jacomb, P. A., Rodgers, B., & Parslow,

R. A. (2003). Association of obesity with anxiety, depression and emotional

well-being: a community survey. Australian and New Zealand Journal of

Public Health, 27(4), 434-440.

Kaczynski, A. T., Manske, S. R., Mannell, R. C., & Grewal, K. (2008). Smoking and

physical activity: a systematic review. American Journal of Health Behavior,

32(1), 93-110.

Kasen, S., Cohen, P., Chen, H., & Must, A. (2008). Obesity and psychopathology in

women: a three decade prospective study. International Journal of Obesity,

32(3), 558-566.

Kaufert, P. A., Gilbert, P., & Tate, R. (2008). The Manitoba Project: a

re-examination of the link between menopause and depression. Maturitas,

61(1-2), 54-66.

Keel, P. K., Baxter, M. G., Heatherton, T. F., & Joiner, T. E., Jr. (2007). A 20-year

longitudinal study of body weight, dieting, and eating disorder symptoms.

Journal of Abnormal Psychology, 116(2), 422-432.

Keizer, I., & Eytan, A. (2005). Variations in smoking during hospitalization in

psychiatric in-patient units and smoking prevalence in patients and

health-care staff. International Journal of Social Psychiatry, 51(4), 317-328.

Kenny, D. A. (2009). Mediation. Retrieved 12 May 2010, from

http://davidakenny.net/cm/mediate.htm

261

Kessler, R. C., Berglund, P., Demler, O., Jin, R., Merikangas, K. R., & Walters, E. E.

(2005). Lifetime prevalence and age-of-onset distributions of DSM-IV

disorders in the National Comorbidity Survey Replication. Archives of

Geneneral Psychiatry, 62(6), 593-602.

Khaled, S. M., Bulloch, A., Exner, D. V., & Patten, S. B. (2009). Cigarette smoking,

stages of change, and major depression in the Canadian population.

Canadian Journal of Psychiatry, 54(3), 204-208.

Klose, M., & Jacobi, F. (2004). Can gender differences in the prevalence of mental

disorders be explained by sociodemographic factors? Archives of Women's

Mental Health, 7(2), 133-148.

Klungsoyr, O., Nygard, J. F., Sorensen, T., & Sandanger, I. (2006). Cigarette

smoking and incidence of first depressive episode: an 11-year,

population-based follow-up study. American Journal of Epidemiology,

163(5), 421-432.

Knowler, W. C., Barrett-Connor, E., Fowler, S. E., Hamman, R. F., Lachin, J. M.,

Walker, E. A., et al. (2002). Reduction in the incidence of type 2 diabetes

with lifestyle intervention or metformin. New England Journal of Medicine,

346(6), 393-403.

Korhonen, T., Broms, U., Varjonen, J., Romanov, K., Koskenvuo, M., Kinnunen, T.,

et al. (2007). Smoking behaviour as a predictor of depression among Finnish

men and women: a prospective cohort study of adult twins. Psychological

Medicine, 37(5), 705-715.

Krishnan, S., Rosenberg, L., & Palmer, J. R. (2009). Physical activity and television

watching in relation to risk of type 2 diabetes: the Black Women's Health

Study. American Journal of Epidemiology, 169(4), 428-434.

262

Kruse, J., Schmitz, N., & Thefeld, W. (2003). On the association between diabetes

and mental disorders in a community sample: results from the German

National Health Interview and Examination Survey. Diabetes Care, 26(6),

1841-1846.

Lafitte, M., Barandon, L., Pucheu, Y., Pillois, X., Gin, H., Bonnet, J., et al. (2010).

After acute coronary syndrome, diabetic patients with peripheral vascular

disease remain at high risk of cardiovascular events despite secondary

prevention measures. Archives of Cardiovascular Diseases, 103(2), 97-105.

Lawrence, D., Mitrou, F., & Zubrick, S. R. (2009). Smoking and mental illness:

results from population surveys in Australia and the United States. BMC

Public Health, 9, 285.

Lee, I. M., & Skerrett, P. J. (2001). Physical activity and all-cause mortality: what is

the dose-response relation? Medicine and Science in Sports and Exercise,

33(Suppl.6), S459-471.

Lee, J. S., Kritchevsky, S. B., Tylavsky, F. A., Harris, T., Everhart, J., Simonsick, E.

M., et al. (2004). Weight-loss intention in the well-functioning,

community-dwelling elderly: associations with diet quality, physical activity,

and weight change. American Journal of Clinical Nutrition, 80(2), 466-474.

Leese, M., Schene, A., Koeter, M., Meijer, K., Bindman, J., Mazzi, M., et al. (2008).

SF-36 scales, and simple sums of scales, were reliable quality-of-life

summaries for patients with schizophrenia. Journal of Clinical Epidemiology,

61(6), 588-596.

Lewinsohn, P. M., Seeley, J. R., Roberts, R. E., & Allen, N. B. (1997). Center for

Epidemiologic Studies Depression Scale (CES-D) as a screening instrument

263

for depression among community-residing older adults. Psychology and

Aging, 12(2), 277-287.

Li, T. Y., Rana, J. S., Manson, J. E., Willett, W. C., Stampfer, M. J., Colditz, G. A.,

et al. (2006). Obesity as compared with physical activity in predicting risk of

coronary heart disease in women. Circulation, 113(4), 499-506.

Lorig, K., Doak, C., Doak, L., & Giloth, B. E. (2001). Patient education, a practical

approach (3rd ed. P.22). Lodon: Sage Publications.

Lorig, K., Ritter, P. L., & Plant, K. (2005). A disease-specific self-help program

compared with a generalized chronic disease self-help program for arthritis

patients. Arthritis and Rheumatism, 53(6), 950-957.

Lorig, K., Ritter, P. L., Villa, F., & Piette, J. D. (2008). Spanish diabetes

self-management with and without automated telephone reinforcement: two

randomized trials. Diabetes Care, 31(3), 408-414.

Lorig, K., Ritter, P. L., Laurent, D. D., & Plant, K. (2006). Internet-based chronic

disease self-management: a randomized trial. Medical Care, 44(11), 964-971.

Lorig, K., Sobel, D. S., Ritter, P. L., Laurent, D., & Hobbs, M. (2001). Effect of a

self-management program on patients with chronic disease. Effective Clinical

Practice, 4(6), 256-262.

Lox, C. L., Martine Ginis, K. A., & Petruzzello, S. J. (2006). The psychology of

exercise : integrating theory and practice Scottsdale, Ariz: Holcomb

Hathaway Publishers.

Lu, S., Tseng, H., Lin, L., Luh, W., & Shu, B. (2009). Factors related to depression

during menopause: a study in Southern Taiwan. Journal of Nursing Research,

17(2), 128-135.

264

Ludman, E., Katon, W., Russo, J., Simon, G., Von Korff, M., Lin, E., et al. (2006).

Panic episodes among patients with diabetes. General Hospital Psychiatry,

28(6), 475-481.

Luppino, F. S., De Wit, L. M., Bouvy, P. F., Stijnen, T., Cuijpers, P., Penninx, B. W.

J. H., et al. (2010). Overweight, obesity, and depression: a systematic review

and meta-analysis of longitudinal studies. Archives of General Psychiatry,

67(3), 220-229.

Lustman, P. J., Anderson, R. J., Freedland, K. E., de Groot, M., Carney, R. M., &

Clouse, R. E. (2000). Depression and poor glycemic control: a meta-analytic

review of the literature. Diabetes Care, 23(7), 934-942.

Ma, G., Luan, D., Li, Y., Liu, A., Hu, X., Cui, Z., et al. (2008). Physical activity

level and its association with metabolic syndrome among an employed

population in China. Obesity Reviews, 9 (Suppl. 1), 113-118.

Maki, P. M. (2008). Menopause and anxiety: immediate and long-term effects.

Menopause, 15(6), 1033-1035.

Mancuso, C. A., Rincon, M., McCulloch, C. E., & Charlson, M. E. (2001).

Self-efficacy, depressive symptoms, and patients' expectations predict

outcomes in asthma. Medical Care, 39(12), 1326-1338.

Marill, K. A. (2004). Advanced statistics: linear regression, part II: multiple linear

regression. Academic Emergency Medicine, 11(1), 94-102.

Markowitz, S., Friedman, M. A., & Arent, S. M. (2008). Understanding the relation

between obesity and depression: Causal mechanisms and implications for

treatment. Clinical Psychology, 15(1), 1-20.

Martinsen, E. W. (2008). Physical activity in the prevention and treatment of anxiety

and depression. Nordic Journal of Psychiatry, 62 (Suppl. 47), 25-29.

265

Massak, A., & Graham, K. (2008). Is the smoking-depression relationship

confounded by alcohol consumption? An analysis by gender. Nicotine &

Tobacco Research, 10(7), 1231-1243.

Mather, A. A., Cox, B. J., Enns, M. W., & Sareen, J. (2009). Associations of obesity

with psychiatric disorders and suicidal behaviors in a nationally

representative sample. Journal of Psychosomatic Research, 66(4), 277-285.

Mathers C, Vos T, & C, S. (1999). The burden of disease and injury in Australia

(PHE 17). Canberra: AIHIW.

Mathers, C. D., Vos, E. T., Stevenson, C. E., & Begg, S. J. (2000). The Australian

Burden of Disease Study: measuring the loss of health from diseases, injuries

and risk factors. The Medical Journal of Australia, 172(12), 592-596.

McClure, J. B., Divine, G., Alexander, G., Tolsma, D., Rolnick, S. J., Stopponi, M.,

et al. (2009). A comparison of smokers' and nonsmokers' fruit and vegetable

intake and relevant psychosocial factors. Behavioral Medicine, 35(1), 14-22.

McElroy, S. L., Kotwal, R., Malhotra, S., Nelson, E. B., Keck, P. E., & Nemeroff, C.

B. (2004). Are mood disorders and obesity related? A review for the mental

health professional. The Journal of Clinical Psychiatry, 65(5), 634-651.

Medina-Shepherd, R. B. (2008). The efficacy of the health belief model in predicting

Spanish-speaking Hispanic women's behavior regarding mammography

screening. US: ProQuest Information & Learning.

Mezuk, B., Eaton, W. W., Albrecht, S., & Golden, S. H. (2008). Depression and type

2 diabetes over the lifespan: a meta-analysis. Diabetes Care, 31(12),

2383-2390.

266

Mino, Y., Shigemi, J., Otsu, T., Ohta, A., Tsuda, T., Yasuda, N., et al. (2001).

Smoking and mental health: cross-sectional and cohort studies in an

occupational setting in Japan. Preventive Medicine, 32(4), 371-375.

Mond, J. M., & Baune, B. T. (2009). Overweight, medical comorbidity and

health-related quality of life in a community sample of women and men.

Obesity, 17(8), 1627-1634.

Morrow, C. (2001). Psychosocial health of well women through the life cycle. In J.

A. Rosenfeld (Ed.), Handbook of women's health : an evidence-based

approach (pp. 123). New York Cambridge University Press.

Motl, R. W., & McAuley, E. (2010). Physical activity, disability, and quality of life

in older adults. Physical Medicine and Rehabilitation Clinics of North

America, 21(2), 299-308.

Muennig, P., Jia, H., Lee, R., & Lubetkin, E. (2008). I think therefore I am:

perceived ideal weight as a determinant of health. American Journal of


Munafo, M. R., Tilling, K., & Ben-Shlomo, Y. (2009). Smoking status and body

mass index: a longitudinal study. Nicotine & Tobacco Research, 11(6),

765-771.

Mykletun, A., Overland, S., Aarø, L. E., Liabø, H.-M., & Stewart, R. (2008).

Smoking in relation to anxiety and depression: Evidence from a large

population survey: The HUNT study. European Psychiatry, 23(2), 77-84.

Nakata, A., Takahashi, M., Ikeda, T., Hojou, M., Nigam, J. A., & Swanson, N. G.

(2008). Active and passive smoking and depression among Japanese workers.

Preventive Medicine, 46(5), 451-456.

267

National Health and Medical Council (NHMRC). (2001). Australian alcohol

guidelines: health risks and benefits. Retrieved 18 April 2010, from

http://www.nhmrc.gov.au/publications/synopses/ds9syn.htm

National Health and Medical Research Council (NHMRC). (2009). Australian

guidelines to reduce health risks from drinking alcohol. Retrieved 22 April

2010, from

http://www.nhmrc.gov.au/_files_nhmrc/file/publications/synopses/ds10-alco

hol.pdf

National Heart Lung and blood Institutes. (1998). Clinical guidelines of

identification, evaluation and treatment of overweight and obesity in adults

(No. 98-4083).Maryland: NIH Publication.

Orozco, L. J., Buchleitner, A. M., Gimenez-Perez, G., Roqué i Figuls, M., Richter,

B., & Mauricio, D. (2008). Exercise or exercise and diet for preventing type 2

diabetes mellitus. Cochrane Database Systematic Reviews, (3). Art. No.:

CD003054. DOI: 10.1002/14651858.CD003054.pub3.

Parker, E. D., Jacobs Jr, D. R., Schreiner, P. J., Schmitz, K. H., & Dengel, D. R.

(2007). Physical activity in young adults and incident hypertension over 15

years of follow-up: The CARDIA Study. American Journal of Public Health,

97(4), 703-709.

Parry, B. L. (2007). Sleep disturbances at menopause are related to sleep disorders

and anxiety symptoms. Menopause, 14(5), 812-814.

Pasco, J. A., Williams, L. J., Jacka, F. N., Ng, F., Henry, M. J., Nicholson, G. C., et

al. (2008). Tobacco smoking as a risk factor for major depressive disorder:

population-based study. The British Journal of Psychiatry, 193(4), 322-326.

268

Patja, K., Jousilahti, P., Hu, G., Valle, T., Qiao, Q., & Tuomilehto, J. (2005). Effects

of smoking, obesity and physical activity on the risk of type 2 diabetes in

middle-aged Finnish men and women. Journal of Internal Medicine, 258(4),

356-362.

Patten, S. B., Brandon-Christie, J., Devji, J., & Sedmak, B. (2000). Performance of

the composite international diagnostic interview short form for major

depression in a community sample. Chronic Diseases in Canada, 21(2),

68-72.

Pelican, J. M. (2007). Understanding differentiation of health in late modernity - by

use of sociological systems theory. In D. V. McQueen, I. Kickbusch & L.

Potvin (Eds.), Health and modernity : the role of theory in health promotion.

New York: Springer.

Pisinger, C., Toft, U., Aadahl, M., Glümer, C., & Jørgensen, T. (2009). The

relationship between lifestyle and self-reported health in a general population:

the Inter99 study. Preventive Medicine, 49(5), 418-423.

Plotnikoff, R. C., Hotz, S. B., Johnson, S. T., Hansen, J. S., Birkett, N. J., Leonard, L.

E., et al. (2009). Readiness to shop for low-fat foods: a population study.

Journal Of The American Dietetic Association, 109(8), 1392-1397.

Pomerleau, C. S., Zucker, A. N., & Stewart, A. J. (2003). Patterns of depressive

symptomatology in women smokers, ex-smokers, and never-smokers.

Addictive Behaviors, 28(3), 575-582.

Poortinga, W. (2007). The prevalence and clustering of four major lifestyle risk

factors in an English adult population. Preventive Medicine, 44(2), 124-128.

Prochaska, J. O., & Velicer, W. F. (1997). The transtheoretical model of health

behavior change. American Journal of Health Promotion, 12(1), 38-48.

269

Prochaska, J. O., Velicer, W. F., Rossi, J. S., Goldstein, M. G., Marcus, B. H.,

Rakowski, W., et al. (1994). Stages of change and decisional balance for 12

problem behaviors. Health Psychology, 13(1), 39-46.

Pudrovska, T., & A. (2009). Psychological implications of parenthood in midlife:

Evidence from sibling models. US: ProQuest Information & Learning.

Puhl, R. M., & Heuer, C. A. (2010). Obesity stigma: important considerations for

public health. American Journal of Public Health, 100(6), 1019-1028.

Rakovac, M., Baric, R., & Heimer, S. (2007). Prevalence of certain chronic diseases

sensitive to kinesiological treatment in physically active vs inactive

postmenopausal women. Kinesiology, 39(2), 124-130.

Renzaho, A., Wooden, M., & Houng, B. (2010). Associations between body mass

index and health-related quality of life among Australian adults. Quality of

Life Research, 19(4), 515-520.

Resnicow, K., McCarty, F., & Baranowski, T. (2003). Are precontemplators less

likely to change their dietary behavior? A prospective analysis. Health

Education Research, 18(6), 693-705.

Riva, M., Bambra, C., Curtis, S., & Gauvin, L. (2010). Collective resources or local

social inequalities? Examining the social determinants of mental health in

rural areas. European Journal of Public Health, doi: 10.1093/eurpub/ckq064

Roach, J. B., Yadrick, M. K., Johnson, J. T., Boudreaux, L. J., Forsythe Iii, W. A., &

Billon, W. (2003). Using self-efficacy to predict weight loss among young

adults. Journal of the American Dietetic Association, 103(10), 1357-1359.

Roberts, R. E., Deleger, S., Strawbridge, W. J., & Kaplan, G. A. (2003). Prospective

association between obesity and depression: evidence from the Alameda

270

County Study. International Journal of Obesity and Related Metabolic

Disorders, 27(4), 514-521.

Rodgers, B., Korten, A. E., Jorm, A. F., Christensen, H., Henderson, S., & Jacomb, P.

A. (2000). Risk factors for depression and anxiety in abstainers, moderate

drinkers and heavy drinkers. Addiction, 95(12), 1833-1845.

Rogge, M. M., Greenwald, M., & Golden, A. (2004). Obesity, stigma, and civilized

oppression. Advances in Nursing Science, 27(4), 301-315.

Rohrer, J. E., Rush Pierce, J. J., & Blackburn, C. (2005). Lifestyle and mental health.


Rosenstock, I. (1974). Historical origins of the health belief model. Health

Education Monographs, 2, 328-335.

Rothenbacher, D., Koenig, W., & Brenner, H. (2006). Lifetime physical activity

patterns and risk of coronary heart disease. Heart, 92(9), 1319-1320.

Sacco, W. P., Wells, K. J., Vaughan, C. A., Friedman, A., Perez, S., & Matthew, R.

(2005). Depression in adults with type 2 diabetes: the role of adherence, body

mass index, and self-efficacy. Health Psychology, 24(6), 630-634.

Salaffi, F., Sarzi-Puttini, P., Girolimetti, R., Atzeni, F., Gasparini, S., & Grassi, W.

(2009). Health-related quality of life in fibromyalgia patients: a comparison

with rheumatoid arthritis patients and the general population using the SF-36

health survey. Clinical and Experimental Rheumatology, 27(5 Suppl. 56),

S67-S74.

Sanchez-Villegas, A., Henriquez, P., Bes-Rastrollo, M., & Doreste, J. (2006).

Mediterranean diet and depression. Public Health Nutrition, 9(8A),

1104-1109.

271

Sanchez, A., Norman, G. J., Sallis, J. F., Calfas, K. J., Rock, C., & Patrick, K. (2008).

Patterns and correlates of multiple risk behaviors in overweight women.


Sarna, L., Bialous, S. A., Cooley, M. E., Jun, H.-J., & Feskanich, D. (2008). Impact

of smoking and smoking cessation on health-related quality of life in women

in the Nurses' Health Study. Quality of Life Research, 17(10), 1217-1227.

Scott, K. M., Bruffaerts, R., Simon, G. E., Alonso, J., Angermeyer, M., Girolamo, G.

D., et al. (2008). Obesity and mental disorders in the general population:

results from the world mental health surveys. International Journal of

Obesity, 32(1), 192-200.

Shai, I., Jiang, R., Manson, J. E., Stampfer, M. J., Willett, W. C., Colditz, G. A., et al.

(2006). Ethnicity, obesity, and risk of type 2 diabetes in women: a 20-year

follow-up study. Diabetes Care, 29(7), 1585-1590.

Shortridge-Baggett, L. M. (2001). Self-efficacy: measurement and intervention in

nursing. Scholarly Inquiry for Nursing Practice, 15(3), 183-188.

Singer, S., Kuhnt, S., Götze, H., Hauss, J., Hinz, A., Liebmann, A., et al. (2009).

Hospital anxiety and depression scale cutoff scores for cancer patients in

acute care. British Journal of Cancer, 100(6), 908-912.

Skogen, J. C., Harvey, S. B., Henderson, M., Stordal, E., & Mykletun, A. (2009).

Anxiety and depression among abstainers and low-level alcohol consumers.

The Nord-TrØndelag Health Study. Addiction, 104(9), 1519-1529.

Smith-DiJulio, K., & Anderson, D. (2009). Sustainability of a multimodal

intervention to promote lifestyle factors associated with the prevention of

cardiovascular disease in midlife Australian women: a 5-year follow-up.

Health Care for Women Internationalt, 30(12), 1111-1130.

272

Smith-DiJulio, K., Woods, N. F., & Mitchell, E. S. (2008). Well-being during the

menopausal transition and early postmenopause: a longitudinal analysis.

Menopause, 15(6), 1095-1102.

Social Research Center. (2006). National tobacco survey: smoking prevalence and

consumption, 1997-2005. Retrieved 11 May 2010, from

http://www.quitnow.info.au/internet/quitnow/publishing.nsf/Content/hotspots

-lp

Somers, J. M., Goldner, E. M., Waraich, P., & Hsu, L. (2006). Prevalence and

incidence studies of anxiety disorders: a systematic review of the literature.

Canadian Journal of Psychiatry, 51(2), 100-113.

Stamatakis, E., Hirani, V., & Rennie, K. (2009). Moderate-to-vigorous physical

activity and sedentary behaviours in relation to body mass index-defined and

waist circumference-defined obesity. British Journal of Nutrition, 101(5),

765-773.

Stanford Patient Education Research Center. Self-Efficacy for Managing Chronic

Disease 6-Item Scale. CA: Stanford University.

Stanley, E. L. (2008). Why we should use simpler models if the data allow this:

relevance for ANOVA designs in experimental biology. BMI Physiology,

8(16), doi:10.1186/1472-6793-1116.

Sternfeld, B., Wang, H., Quesenberry, C. P., Jr., Abrams, B., Everson-Rose, S. A.,

Greendale, G. A., et al. (2004). Physical activity and changes in weight and

waist circumference in midlife women: findings from the Study of Women's

Health Across the Nation. American Journal of Epidemiology, 160(9),

912-922.

273

Storr, C. L., Cheng, H., Alonso, J., Angermeyer, M., Bruffaerts, R., de Girolamo, G.,

et al. (2010). Smoking estimates from around the world: data from the first

17 participating countries in the World Mental Health Survey Consortium.

Tobacco Control, 19(1), 65-74.

Strawbridge, W. J., Deleger, S. p., Roberts, R. E., & Kaplan, G. A. (2002). Physical

activity reduces the risk of subsequent depression for older adults. American

Journal of Epidemiology, 156(4), 328-334.

Strayhorn, G. (2009). Health disparities: the case for diabetes. In R. L. Braithwaite, S.

E. Taylor & H. M. Treadwell (Eds.), Health issues in the Black community

(3rd ed.). (pp. 279-289). San Francisco: Jossey-Bass.

Sturm, R., & Wells, K. B. (2001). Does obesity contribute as much to morbidity as

poverty or smoking? Public Health, 115(3), 229-235.

Sullivan, P. W., Morrato, E. H., Ghushchyan, V., Wyatt, H. R., & Hill, J. O. (2005).

Obesity, inactivity, and the prevalence of diabetes and diabetes-related

cardiovascular comorbidities in the U.S., 2000-2002. Diabetes Care, 28(7),

1599-1603.

Takeuchi, T., Nakao, M., & Yano, E. (2004). Relationship between smoking and

major depression in a Japanese workplace. Journal of Occupational Health,

46(6), 489-492.

Tanasescu, M., Leitzmann, M. F., Rimm, E. B., & Hu, F. B. (2003). Physical activity

in relation to cardiovascular disease and total mortality among men with type

2 diabetes. Circulation, 107(19), 2435-2439.

Tavafian, S.-S., Aghamolaei, T., & Zare, S. (2009). Water pipe smoking and

health-related quality of life: a population-based study. Archives of Iranian

Medicine, 12(3), 232-237.

274

Taylor-Piliae, R. E., Fair, J. M., Haskell, W. L., Varady, A. N., Iribarren, C., Hlatky,

M. A., et al. (2010). Validation of the Stanford Brief Activity Survey:

examining psychological factors and physical activity levels in older Adults.

Journal of Physical Activity & Health, 7(1), 87-94.

Teixeira, P., J. , Going, S., B., Houtkooper, L., B., Cussler, E., C. , Martin, C., J.,

Metcalfe, L., L., et al. (2002). Weight loss readiness in middle-aged women:

psychosocial predictors of success for behavioral weight reduction. Journal

of Behavioral Medicine, 25(6), 499-523.

Thorburn, A. W. (2005). Prevalence of obesity in Australia. Obesity Reviews, 6(3),

187-189.

Tobias, M., Templeton, R., & Collings, S. (2008). How much do mental disorders

contribute to New Zealand's tobacco epidemic? Tobacco Control, 17(5),

347-350.

Travers, C., O'Neill, S. M., King, R., Battistutta, D., & Khoo, S. K. (2005). Greene

Climacteric Scale: norms in an Australian population in relation to age and

menopausal status. Climacteric, 8(1), 56-62.

Trosclair, A., & Dube, S. R. (2010). Smoking among adults reporting lifetime

depression, anxiety, anxiety with depression, and major depressive episode,

United States, 2005-2006. Addictive Behaviors, 35(5), 438-443.

Tselebis, A., Panaghiotou, A., Theotoka, I., & Ilias, I. (2001). Nursing staff anxiety

versus smoking habits. International Journal of Nursing Practice, 7(3),

221-223.

Turley, M., Tobias, M., & Paul, S. (2006). Non-fatal disease burden associated with

excess body mass index and waist circumference in New Zealand adults.

Australian & New Zealand Journal of Public Health, 30(3), 231-237.

275

Untas, A. l., Aguirrezabal, M., Chauveau, P., Leguen, E., Combe, C., & Rascle, N.

(2009). Anxiety and depression in hemodialysis: validation of the Hospital

Anxiety and Depression Scale (HADS). Néphrologie & Thérapeutique 5(3),

193-200.

Vallance, J. K., Murray, T. C., Johnson, S. T., & Elavsky, S. (2010). Quality of life

and psychosocial health in postmenopausal women achieving public health

guidelines for physical activity. Menopause, 17(1), 64-71.

Vamos, E. P., Bottle, A., Majeed, A., & Millett, C. (2010). Trends in lower extremity

amputations in people with and without diabetes in England, 1996-2005.

Diabetes Research & Clinical Practice, 87(2), 275-282.

Van Gool, C. H., Kempen, G. I. J. M., Bosma, H., van Eijk, J. T. M., van Boxtel, M.

P. J., & Jolles, J. (2007). Associations between lifestyle and depressed mood:

longitudinal results from the Maastricht Aging Study. American Journal of


Vasiljevic, N., Ralevic, S., Marinkovic, J., Kocev, N., Maksimovic, M., Milosevic, G.

S., et al. (2008). The assessment of health-related quality of life in relation to

the body mass index value in the urban population of Belgrade. Health and

Quality of Life Outcomes, 6, 106-106.

Vogelzangs, N., Kritchevsky, S. B., Beekman, A. T. F., Brenes, G. A., Newman, A.

B., Satterfield, S., et al. (2010). Obesity and onset of significant depressive

symptoms: results from a prospective community-based cohort study of older

men and women. The Journal of Clinical Psychiatry, 71(4), 391-399.

Wagner, J. A., Tennen, H., & Osborn, C. Y. (2010). Lifetime depression and

diabetes self-management in women with Type 2 diabetes: a case-control

study. Diabetic Medicine, 27(6), 713-717.

276

Walsh, T. L., Homa, K., Hanscom, B., Lurie, J., Sepulveda, M. G., & Abdu, W.

(2006). Screening for depressive symptoms in patients with chronic spinal

pain using the SF-36 Health Survey. The Spine Journal, 6(3), 316-320.

Wang, J., & Patten, S. B. (2001). Alcohol consumption and major depression:

findings from a follow-up study. The Canadian Journal of Psychiatry, 46(7),

632-638.

Wannamethee, S. G., & Shaper, A. G. (2003). Alcohol, body weight, and weight

gain in middle-aged men. The American Journal of Clinical Nutrition, 77(5),

1312-1317.

Ware, J. E., Jr., Kosinski, M., & Dewey, J. E. (2000). Chapter 7 Scoring SF-36v2

physical and mental summary measures. In How to score version 2 of the

SF-36® health survey. Lincoln: RI: QualityMetric Incorporated.

Ware, J. E., Jr., Kosinski, M., & Gandek, B. (2000a). Development of the SF-36

SF-36 Health Survey: Manual & Interpretation Guide. Lincoln, RI:

QualityMetric Incorporated.

Ware, J. E., Jr., Kosinski, M., & Gandek, B. (2000b). Validity: norm-based

interpretation SF-36 Health Survey: Manual & Interpretation Guide. Lincoln,

RI: QualityMetric Incorporated.

Warren, C. W., Jones, N. R., Eriksen, M. P., & Asma, S. (2006). Patterns of global

tobacco use in young people and implications for future chronic disease

burden in adults. Lancet, 367(9512), 749-753.

Wee, H., Wu, Y., Thumboo, J., Lee, J., & Tai, E. (2010). Association of body mass

index with Short-Form 36 physical and mental component summary scores in

a multiethnic Asian population. International Journal of Obesity, 34(6),

1034-1043.

277

Wen, L. M., Thomas, M., Jones, H., Orr, N., Moreton, R., King, L., et al. (2002).

Promoting physical activity in women: evaluation of a 2-year

community-based intervention in Sydney, Australia. Health Promotion

International, 17(2), 127-137.

Widome, R., Linde, J. A., Rohde, P., Ludman, E. J., Jeffery, R. W., & Simon, G. E.

(2009). Does the association between depression and smoking vary by body

mass index (BMI) category? Preventive Medicine, 49(5), 380-383.

Willi, C., Bodenmann, P., Ghali, W. A., Faris, P. D., & Cornuz, J. (2007). Active

smoking and the risk of type 2 diabetes: a systematic review and

meta-analysis. Journal of American Medical Association, 298(22),

2654-2664.

Wolitzky-Taylor, K. B., Castriotta, N., Lenze, E. J., Stanley, M. A., & Craske, M. G.

(2010). Anxiety disorders in older adults: a comprehensive review.

Depression and Anxiety, 27(2), 190-211.

World Health Organisation. (2000a). Obesity: preventing and managing the global

epidemic. WHO Technical Report Series 894. Geneva. Retrieved 25 June

2010, from http://whqlibdoc.who.int/trs/WHO_TRS_894_(part1).pdf

World Health Organisation. (2000b). Technical report series 894: "obesity:

preventing and managing the global epidemic." Geneva: WHO.

World Health Organisation. (2003). WHO launches annual Move for Health day as

global initiative to promote benefits of physical activity, Retrieved 07 August

2009, from

http://www.who.int/mediacentre/news/releases/2003/pr15/en/index.html

World Health Organisation. (2005a).Mental Health. Retrieved on 02 June 2010,

from http://www.who.int/mental_health/en/

278

World Health Organisation. (2005b). Preventing chronic disease: a vital investment.

Retrieved 02 June 2009, from

http://www.who.int/chp/chronic_disease_report/contents/en/index.html

World Health Organisation. (2006). Diabetes. Retrieved 07 June 2009, from

http://www.who.int/mediacentre/factsheets/fs312/en/index.html

World Health Organisation. (2008). Management of substance abuse-Alcohol.

Retrieved 14 July 2010, from

http://www.who.int/substance_abuse/facts/alcohol/en/index.html

World Health Organisation Regional Office for Europe. (1986). Ottawa chater for

health promotion. Geneva: WHO.

Wott, C. B., & Carels, R. A. (2010). Overt weight stigma, psychological distress and

weight loss treatment outcomes. Journal of Health Psychology, 15(4),

608-614.

Wu, L. T., & Anthony, J. C. (1999). Tobacco smoking and depressed mood in late

childhood and early adolescence. American Journal of Public Health, 89(12),

1837-1840.

Wyshak, G. (2001). Women's college physical activity and self-reports of

physician-diagnosed depression and of current symptoms of psychiatric

distress. Journal of Women's Health & Gender-Based Medicine, 10(4),

363-370.

Yusuf, S., Hawken, S., Ounpuu, S., Dans, T., Avezum, A., Lanas, F., et al. (2004).

Effect of potentially modifiable risk factors associated with myocardial

infarction in 52 countries (the INTERHEART study): case-control study.

Lancet, 364(9438), 937-952.

279

Zender, R., & Olshansky, E. (2009). Women's mental health: depression and anxiety.

The Nursing Clinics of North America, 44(3), 355-364.

Zhao, G., Ford, E. S., Dhingra, S., Li, C., Strine, T. W., & Mokdad, A. H. (2009).

Depression and anxiety among US adults: associations with body mass index.


Ziaei, S., Moghasemi, M., & Faghihzadeh, S. (2010). Comparative effects of

conventional hormone replacement therapy and tibolone on climacteric

symptoms and sexual dysfunction in postmenopausal women. Climacteric,

13(2), 147-156.

Zigmond, A. S., & Snaith, R. P. (1983). The Hospital Anxiety and Depression Scale.

Acta Psychiatrica Scandinavica, 67(6), 361-370.

280

APPENDICES

Appendix A: A Comparison of Physical Activity Among Different BMI Categories

Table A1

The Levels of Physical Activity Across BMI Categories (N = 485)

BMI

Physical Activity

None 1-2 times /week

3-4 times /week

5-6 times /week

Underweight 0 (0) 2 (33.3) 1 (16.7) 3 (50.0)

Normal weight 28 (14.6) 54 (28.1) 58 (30.2) 52 (27.1)

Overweight 28 (17.2) 47 (28.8) 49 (30.1) 39 (23.9)

Obese 25 (20.2) 42 (33.9) 33 (26.6) 24 (19.4)

χ² 7.582

p .557

281

Appendix B: Medical Outcomes Study Short Form (SF-36)

Please answer each of the following questions. Some questions may look like others, but

each one is different. Please take the time to read and answer each question carefully, and

mark an in the one box that best describes your answer.

1. In general, would you say your health is:

Excellent Very good Good Fair Poor

1 2 3 4 5

2. Compared to one year ago, how would you rate your health in general now?

Much betternow than oneyear ago

Somewhat betternow than oneyear ago

About thesame as one year

ago

Somewhat worsenow than oneyear ago

Much worsenow than oneyear ago

1 2 3 4 5

3. The following items are about activities you might do during a typical day.Does your health now limit you in these activities? If so, how much?

Yes,limiteda lot

Yes,limiteda little

No, notlimitedat all

a.a.a.a. Vigorous activities, such as running, lifting heavyobjects, participating in strenuous sports 1 2 3

b.b.b.b. Moderate activities, such as moving a table, pushing avacuum cleaner, bowling, or playing golf 1 2 3

c.c.c.c. Lifting or carrying groceries 1 2 3

d.d.d.d. Climbing several flights of stairs 1 2 3

e.e.e.e. Climbing one flight of stairs 1 2 3

f.f.f.f. Bending, kneeling, or stooping 1 2 3

g.g.g.g. Walking more than a kilometre 1 2 3

h.h.h.h. Walking several blocks 1 2 3

i.i.i.i. Walking one block 1 2 3

j.j.j.j. Bathing or dressing yourself 1 2 3

282

4. During the past 4 weeks, have you had any of the following problems with your work orother regular daily activities as a result of your physical health?

Yes No

a.a.a.a. Cut down on the amount of time you spent on work or otheractivities 1 2

b.b.b.b. Accomplished less than you would like 1 2

c.c.c.c. Were limited in the kind of work or other activities 1 2

d.d.d.d. Had difficulty performing the work or other activities (forexample, it took extra effort) 1 2

5. During the past 4 weeks, have you had any of the following problems with your work orother regular daily activities as a result of any emotional problems (such as feelingdepressed or anxious)?

Yes No

a.a.a.a. Cut down on the amount of time you spent on work or otheractivities 1 2

b.b.b.b. Accomplished less than you would like 1 2

c.c.c.c. Did work or other activities less carefully than usual 1 2

6. During the past 4 weeks, to what extent has your physical health or emotional problemsinterfered with your normal social activities with family, friends, neighbours, or groups?

Not at all Slightly Moderately Quite a bit Extremely

1 2 3 4 5

7. How much bodily pain have you had during the past 4 weeks?

None Very mild Mild Moderate Severe Very Severe

1 2 3 4 5 6

8. During the past 4 weeks, how much did pain interfere with your normal work(including both work outside the home and housework)?

Not at all A little bit Moderately Quite a bit Extremely

1 2 3 4 5

283

9. These questions are about how you feel and how things have been with you during thepast 4 weeks. For each question, please give the one answer that comes closest to theway you have been feeling. How much of the time during the past 4 weeks?

All ofthetime

Mostof thetime

A goodbit of

the time

Someof thetime

A littleof thetime

Noneof thetime

a.a.a.a. Did you feel full of life? 1 2 3 4 5 6

b.b.b.b. Have you been a very nervous person? 1 2 3 4 5 6

c.c.c.c. Have you felt so down in the dumps thatnothing could cheer you up? 1 2 3 4 5 6

d.d.d.d. Have you felt calm and peaceful? 1 2 3 4 5 6

e.e.e.e. Did you have a lot of energy? 1 2 3 4 5 6

f.f.f.f. Have you felt downhearted and blue? 1 2 3 4 5 6

g.g.g.g. Did you feel worn out? 1 2 3 4 5 6

h.h.h.h. Have you been a happy person? 1 2 3 4 5 6

i.i.i.i. Did you feel tired? 1 2 3 4 5 6

10.During the past 4 weeks, how much of the time has your physical health or emotionalproblems interfered with your social activities (like visiting friends, relatives, etc.)?

All of thetime

Most of thetime

Some of thetime

A little of thetime

None of thetime

1 2 3 4 5

11.How TRUETRUETRUETRUE or FALSEFALSEFALSEFALSE is each of the following statements for you?

Definitely true

Mostlytrue

Don'tknow

Mostlyfalse

Definitely false

a.a.a.a. I seem to get sick a little easier thanother people 1 2 3 4 5

b.b.b.b. I am as healthy as anybody I know 1 2 3 4 5

c.c.c.c. I expect my health to get worse 1 2 3 4 5

d.d.d.d. My health is excellent 1 2 3 4 5

284

Appendix C: Greene’s Climacteric Scale

Please indicate the extent to which you are bothered at the moment by any of these

symptoms by placing a tick in the appropriate box.

Not at all A little Quite abit Extremely

a. Heart beating quickly or strongly

b. Feeling tense or nervous

c. Difficulty in sleeping

d. Excitable

e. Attacks of panic

f. Difficulty in concentrating

g. Feeling tired or lacking in energy

h. Loss of interest in most things

i. Feeling unhappy or depressed

j. Crying spells

k. Irritability

l. Feeling dizzy or faint

m. Pressure or tightness in head or body

n. Parts of body feel numb or tingling

o. Headaches

p. Muscle and joint pains

q. Loss of feeling hands or feet

r. Breathing difficulties

s. Hot Flushes

t. Sweating at night

u. Loss of interest in sex

285

Appendix D: The Hospital Anxiety and Depression Scale

Please read each item and circle the number next to the reply which comes closest to

how you have been feeling in the past week. Don't take too long over your replies;

your immediate reaction to each item will probably be more accurate than a long

thought out response.

a. I feel tense or ‘wound up’

0. Most of the time1. A lot of the time2. From time to time, occasionally3. Not at all

b. I still enjoy the things I used to enjoy

0. Definitely as much1. Not quite as much2. Only a little3. Hardly at all

c. I get a sort of frightened feeling as if something awful is about to happen

0. Very definitely and quite badly1. Yes, but not too badly2. A little, but it doesn’t worry me3. Not at all

d. I can laugh and see the funny side of things

0. As much as I always could1. Not quite so much now2. Definitely not so much now3. Not at all

e. Worrying thoughts go through my mind

0. A great deal of the time1. A lot of the time2. From time to time but not too often3. Only occasionally

f. I feel cheerful

0. Not at all1. Not often2. Sometimes3. Most of time

g. I can sit at ease and feel relaxed

0. Definitely1. Usually

286

2. Not often3. Not at all

h. I feel as if I am slowed down:

0. Nearly all the time1. Very often2. Sometimes3. Not at all

i. I get a sort of frightened feeling like ‘butterflies’ in the stomach

0. Not at all1. Occasionally2. Quite often3. Very often

j. I have lost interest in my appearance

0. Definitely1. I don’t take so much care as I should2. I may not take quite as much care3. I take just as much care as ever

k. I feel restless as if I have to be on the move

0. Very much indeed1. Quite a lot2. Not very much3. Not at all

l. I look forward with enjoyment to things

0. As much as I ever did1. Rather less than I used to2. Definitely less than I used to3. Hardly at all

m. I get sudden feelings of panic

0. Very often indeed1. Quite often2. Not very often3. Not at all

n. I can enjoy a good book or radio or TV program

0. Often1. Sometimes2. Not often3. Very seldom

287

Appendix E: Self-efficacy in Managing Chronic Diseases

We would like to know how confident you are in doing certain activities. For each of the

following questions, please choose the number that corresponds to your confidence that you

can do the tasks regularly at the present time.

1 2 3 4 5 6 7 8 9 10Not at allconfident

Totallyconfident

Confidence(1-10)

a) How confident are you that you can keep the fatigue caused by your diseasefrom interfering with the things you want to do? ________

b) How confident are you that you can keep the physical discomfort or pain ofyour disease from interfering with the things you want to do? ________

c) How confident are you that you can keep the emotional distress caused byyour disease from interfering with the things you want to do? ________

d) How confident are you that you can keep any other symptoms or healthproblems you have from interfering with the things you want to do? ________

e) How confident are you that you can do the different tasks and activities neededto manage your health condition so as to reduce you need to see a doctor?

________f) How confident are you that you can do things other than just taking medication

to reduce how much your illness affects your everyday life? ________

Scoring: the mean score of the six items is calculated.

288

Appendix F: The Seattle Physical Activity Questionnaire

1. General daily activity includes activities such as housework, caring for children,shopping, gardening or activity at work. It does not include exercising. How do you

describe your current general daily activity level?

Very active (involves strenuous labour)

Moderately active

Mildly active (some walking/stair climbing)

Sedentary (mostly sitting)

2. During the past month, how many times did you exercise for at least 15 minutes at a

time? (Exercise includes activities such as callisthenics, jogging, racquet sports,

team sports, dance classes, brisk walking, lifting weights, yoga, Tai Chi etc)

Daily

5-6 times a week

3-4 times a week

1-2 times a week

None

3. Overall, how do you rate your current level of physical activity (general daily activityplus exercise)? Rate from (00) not at all active to (10) extremely active.

00 01 02 03 04 05 06 07 08 09 10NOT AT ALL ACTIVE EXTREMELY ACTIVE

289

Appendix G: Ethical Approval

290

291

292

293

294

295

QUEENSLAND UNIVERSITY OF TECHNOLOGY SCHOOL OF NURSINGeprints.qut.edu.au/43715/1/Qunyan_Xu_Thesis.pdf · 2011. 7. 28. · Xu, Q., Anderson, D., & Lurie-Beck, J. The effect of central

Documents