Sociodemographic, physical and psychosocial factors associated with depressive behaviour in family caregivers of stroke survivors in the acute care phase

BRAIN INJURY, VOL. 18, NO. 8 (AUGUST 2004), 797–809

Sociodemographic, physical and psychosocial

factors associated with depressive behaviour

in family caregivers of stroke survivors in the

acute care phase

JOAN S. GRANT, MICHAEL WEAVER,

TIMOTHY R. ELLIOTT,

ALFRED R. BARTOLUCCI and

JOYCE NEWMAN GIGER

University of Alabama at Birmingham, Birmingham, AL, USA

(Received 27 March 2003; accepted 20 January 2004)

Primary objective: This study examined sociodemographic, physical and psychosocial characteristics offamily caregivers of stroke survivors at risk for depression.Research design: A correlational study.Methods and procedures: Data were collected from 52 family caregivers and care recipients with ischemicstroke. Data were collected 1–2 days before discharge from a rehabilitative facility and at 5, 9 and 13weeks post-discharge.Main outcomes and results: Compared to African Americans, Caucasians had �3.7 times higher oddsof being at risk. The odds of being at risk for depressive behaviour decreased by 30% for each unitincrease in belonging and increased by 5% for each unit increase in burden.Conclusions: Family caregivers who are isolated from others may benefit from therapeutic interven-tions that facilitate social interaction and decrease caregiver burden. Ethnic differences in caregiverdepression merit further research.

Introduction

Stroke is the nation’s third leading cause of death, killing almost 160 000 Americanseach year. Because at least one-third of these individuals have residual deficitsthat require assistance with activities of daily living after a stroke, they are caredfor most frequently by a spouse or other family member [1]. Unfortunately, thisdisease takes a physical and emotional toll on both stroke survivors and their familymembers [2].

Many studies report that these family caregivers experience significant levels ofdepressive symptomology as a result of the caregiving process [3–6]. For example,Kotila et al. [7] found that depressive symptomology was common amongstroke survivors (n¼ 321) and their caregivers (n¼ 195) in the acute care phase

Brain Injury ISSN 0269–9052 print/ISSN 1362–301X online # 2004 Taylor & Francis Ltd

http://www.tandf.co.uk/journalsDOI: 10.1080/02699050410001671766

Correspondence to: J. S. Grant, DSN, RN, CS, University of Alabama at Birmingham, Universityof Alabama School of Nursing, NB #407, 1530 3rd Avenue South, Birmingham, AL 35294-0113,USA. e-mail: [email protected]

(i.e. 3 months post-stroke). In a review of studies to examine the prevalence ofstroke caregiver depressive behaviour, Han and Haley [4] used seven studies toestimate rates that ranged from 34–52%.

Caregiver depressive behaviour has negative consequences for both the strokesurvivor and family caregiver. For example, in a large longitudinal cohort study,Schulz and Beach [8] found that spousal caregivers who experienced mentalor emotional strain related to caregiving had mortality risks 63% higher thannon-caregiving controls. Caregiving literature also supports that caregiver depressivebehaviour may be associated with less than optimal care of stroke survivors in thehome [9] and overwhelmed caregivers are more likely to institutionalize dependentfamily members [10]. In fact, caregiver burnout appears to be the primaryreason why individuals are admitted to nursing homes instead of the care recipient’sphysical condition worsening [5].

Although a variety of factors are linked to depressive behaviour in family care-givers of other populations, such as individuals with Alzheimer’s disease or dementia[11, 12], there is little research examining factors that contribute to depressivebehaviours in the acute care phase of stroke family caregiving. Research to identifythose factors in caregivers of stroke survivors is important in designing interventionsto assist caregivers in effectively managing caregiving problems.

Therefore, the purpose of this study was to identify a set of variables that identifyfamily caregivers of stroke survivors at risk for developing depressive symptomsin the acute care phase of the caregiver role. In this study, the acute care phasewas defined as up to 3 months post-discharge from a rehabilitative facility.The independent contributions of selected stroke survivor and family caregiversociodemographic characteristics, caregiver general and physical functioning,satisfaction with healthcare services, social support, preparedness and burden towarddepressive symptoms were examined.

Sociodemographic characteristics and depressive behaviour

Age [6, 13–15], income [16] and ethnicity [17, 18] are inconsistent predictorsof depressive behaviour in stroke caregiver populations. For example, while someinvestigators report that White Americans have greater depressive symptomatologythan African Americans [18], Grant [17] found that younger African Americancaregivers had a greater risk for depression than White caregivers in a study com-paring home vs telephone contacts to deliver social problem-solving training(n¼ 30).

Some investigators report a significant relationship between gender anddepressive symptomatology in caregivers. For example, daughter caregivers appearto have less depressive behaviour than wife caregivers [19] and caregiving husbandsexperience fewer depressive symptoms than their female counterparts [13, 20].However, studies that control for the care recipient’s physical and cognitivefunctioning commonly find no differences associated with gender [21]. Otherinvestigators also suggest that hours spent in caregiving affect caregiver depressivesymptomatology [6, 22].

Stroke survivor characteristics such as age and the type, location, extent andprior incidence of stroke may influence functioning of the stroke survivor andsubsequent caregiver outcomes [23, 24]. While many studies suggest that stroke

798 J. S. Grant et al.

survivor physical disabilities and caregiver depressive behaviour are unrelated[7, 17, 25], cognitive impairment appears to be a significant factor [26].

Physical and psychological characteristics anddepressive behaviour

Empirical data suggest that caregiver physical and psychosocial characteristics,such as general health, physical functioning, satisfaction with healthcare services,social support, preparedness and burden may influence depressive behaviour infamily caregivers of stroke survivors. Literature findings support the role of physicalfunctioning and general health in predicting depressive behaviours in caregivers ofindividuals who experienced a stroke. For example, Grant et al. [3] reportedthat physical functioning was an important factor in predicting early depressivesymptoms in caregivers of individuals who experienced a stroke (n¼ 52).Similarly, Hodgson et al. [27] indicated that physical functioning was importantin family caregivers’ emotional well-being. Other investigators also support theimportance of caregiver health in predicting depressive symptoms [28, 29].

Grant et al. [30] examined the best predictors of caregiver depressive symptomsat onset of the caregiver role among 74 persons providing care to stroke survivors.Caregivers at risk for problems with depressive behaviour were best predicted bya negative problem orientation, inadequate caregiver preparedness and impairedsocial functioning. Within the model, only negative problem orientation andsocial functioning were significant independent predictors of depressive behaviour.Other investigators have also reported the importance of perceived caregiverpreparedness in the caregiver role [31].

Investigators also concur that social support is an important factor in caregiverdepressive behaviour [7, 32]. In family members caring for individuals withdementia, MaloneBeach and Zarit [33] and Grant et al. [3] found that instrumentalsupport, or provision of tangible services or aid, was more important than eitherinformational or emotional support in reducing depression in caregiving women.Li et al. [19], however, found that emotional support was a key factor in reducingdepressive symptoms in wives (n¼ 103) and daughters (n¼ 149) who cared forelderly parents with behavioural problems, including stroke survivors.

In examining the effect of caregiver burden on depressive behaviour, someinvestigators have found a significant relationship between caregiver burden anddepressive behaviour [28, 34] while others have found no relationship betweenthe variables [3]. These studies used different populations and the caregiver rolespanned from onset to 1 year, perhaps explaining the variability in burden as asignificant predictor of depressive behaviour.

Methods

Participants

The sample consisted of 52 individuals who had an admitting diagnosis of ischemicstroke (includes brain infarction) caused by either thrombi or emboli. Familymembers who were primarily responsible for assisting the stroke survivors withactivities of daily living (ADLs) and instrumental ADLs were also included in thesample. Only 15% who met inclusion criteria declined to be in the study.

Depressive behaviour in family caregivers 799

Functional deficits were measured by the Functional Independence Measure(FIMSM) [35] and scores ranged from 36–103, with an average of 68. The strokesurvivors commonly experienced deficits such as hemiplegia, dysphasia and memoryloss. There were more stroke survivors who experienced left-sided (n¼ 31) hemi-spheric strokes than those who experienced right-sided (n¼ 17) hemisphericstrokes, while two had bilateral strokes and two were unknown. The strokesurvivors ranged in age from 37–92 years old, with an average age of 74 years,and were almost equally mixed by gender (n¼ 28 females; 24 males).

Primary family caregivers were selected because one individual typically isresponsible for caregiving, usually the wife, daughter or daughter-in-law [36].Family caregivers (n¼ 46 women and 6 men) ranged in age from 25–74 yearsold, with an average age of 56 years. They were primarily wives (n¼ 17) or daugh-ters (n¼ 18). The majority completed either high school (n¼ 28) or college(n¼ 17). There were more Caucasians (n¼ 37) than African Americans (n¼ 15).

Procedure

A trained research nurse made a minimum of two contacts with the family care-givers and stroke survivors in the rehabilitative facility to verify eligibility criteria,explain the study and secure informed consent. Data collection on all instrumentsoccurred 1–2 days before discharge of the stroke survivor and at 5, 9 and 13 weekspost-discharge in the home of either the stroke survivor or primary family caregiver.The measures were administered in a random order. Before beginning the pilot, thereliability of all instruments was examined using a similar sample of family caregiversof stroke survivors (n¼ 19), which yielded acceptable reliability coefficients.

Measures

Sociodemographic data formFamily caregiver sociodemographic variables were age, gender, ethnicity, religion,relationship to the individual recovering from the stroke, educational level, socio-economic status, participation in therapy in the rehabilitation setting (i.e. physical,occupational and speech), number of hours spent working outside the home andin caregiving activities, number of individuals who assisted the caregiver in theircaregiver role and number of hours they spent in caregiving role. Sociodemographicvariables for the individual recovering from the stroke were age; gender; ethnicity;and the type, location and total number of strokes. These data were recorded ona sociodemographic data form.

Functional deficitsFunctional deficits of the stroke survivor were measured with the FunctionalIndependence Measure (FIMSM) [35]. The FIMSM has been used as a measure offunctional ability in numerous studies that span various populations and hasadequate validity and reliability for stroke survivors [37–40]. The instrument con-tains 13 items that address motor function (eating, grooming, bathing, dressing,toileting, bowel and bladder control, transfers and locomotion) and five itemsthat measure cognitive function (communication and social cognition). Each itemon the scale ranges from 1 (total assistance) to 7 (complete independence), withhigher numbers indicating fewer functional deficits.

800 J. S. Grant et al.

General health and physical functioningGeneral health and physical functioning were measured with Medical OutcomesStudy Short Form Health Survey (SF-36) [41] items measuring those dimensions(physical functioning¼ 10 items; general health ¼ five items). The two sub-scaleshave acceptable internal consistency (�¼ 0.88–0.94 for physical functioning;0.80–0.95 for general health) and test–re-test reliability coefficients (r¼ 0.81–0.90for physical functioning; 0.80–0.83 for general health) reported in research par-ticipants having various sociodemographic characteristics, diagnoses and diseaseseverity. Test–re-test coefficients for physical functioning and general healthwere comparable to preliminary studies (0.93 and 0.91, respectively). Higherscores indicate better general health and physical functioning [42, 43].

Satisfaction with health careSatisfaction with health care services was measured with the Client SatisfactionQuestionnaire (CSQ), an 8-item, 4-point, Likert-type scale [44]. This measure ofglobal satisfaction with health care has internal consistency reliability coefficientsof 0.80 with older adults [45] and the pilot data yielded internal consistency andtest–re-test reliability coefficients of 0.71 and 0.89, respectively. Higher scoresindicate greater satisfaction with health care.

Social supportThe 30-item Interpersonal Support Evaluation List (ISEL) was used to measuresocial support [46]. The instrument measures perceived availability of someone totalk about problems (appraisal), people to do things with (belonging) and materialaid (tangible support). The scale has been used with older populations and familycaregivers of stroke survivors and empirical evidence supports its validity andreliability [47, 48]. Lower scores indicate greater social support.

Caregiver preparednessCaregiver preparedness was measured by the Preparedness for Caregiving Scale(PCS), a five-item Likert-type scale that appraises the caregiver’s perceptionsabout their preparedness to give care and to handle the stresses of caregiving [49].Estimates of reliability indicate it is internally consistent (�¼ 0.67–0.92) and stableover time (r¼ 0.81–0.92) for primary family caregivers of different patient popula-tions [50, 51]. The pilot data with family caregivers of stroke survivors yieldedsimilar psychometric data (�¼ 0.77–0.88; r¼ 0.89). Higher scores indicate greatercaregiver preparedness.

Caregiving burdenCaregiving burden perceptions were measured with the difficulty sub-scale of theCaregiving Burden Scale (CBS). The CBS is a 14-item rating scale that measuresdemands and difficulty associated with doing direct, instrumental and interpersonalcare tasks by family caregivers. The difficulty sub-scale was used because it is a betterpredictor of caregiving burden than demand alone or the total score [52, 53]. Thereis support for content and construct validity of the measure and acceptable internalconsistency reliability coefficients for the difficulty scale of 0.91 [52, 53]. Pilot workwith family caregivers of stroke survivors yielded comparable alphas of 0.87–0.91.Higher scores indicate greater caregiver burden.

Depressive behaviour in family caregivers 801

Depressive behaviourDepressive behaviour was measured with the 20-item Centre for EpidemiologicStudies Depression Scale (CES-D) [54]. This valid and reliable instrument hasbeen used across numerous populations, including family caregivers of strokesurvivors [16, 55].

Higher scores indicate greater depressive behaviour. The CES-D is often used infield research to classify persons who may have a greater likelihood of meetingclinical criteria for a depressive syndrome. Generally, a cut-off score of 20 or greateris used to differentiate older persons who may be at risk for a depressive disorder[56]. For this studies purposes, this cut-off score was used to classify caregivers aseither at-risk or not at-risk for depression in the analyses.

Statistical analyses

Because sample size was not sufficient to accommodate all potential predictors inan initial model, bivariate relationships between each potential predictor with thedichotomized depression score were initially examined in order to eliminate thosevariables that did not exhibit a bivariate relationship. Using a repeated measures(generalized estimating equations) Generalized Linear Mixed Model (GLMM),variables were selected for inclusion in the initial model if the bivariate relationshipp-value was �0.20. If an instrument had both total and sub-scale scores meetingthe p� 0.20 inclusion criterion, the scale having the largest correlation with depres-sive behaviour (i.e. either total or sub-scale score) was selected for inclusion in theinitial model to avoid colinearity.

The initial full GMM consisted of the following predictor variables: time;caregiver ethnicity, belonging (social support), burden, physical function, generalhealth and satisfaction with healthcare; stroke survivor age, ethnicity, cognition(component of FIMSM); and number of individuals assisting with care. However,number of family members assisting with care was found to be collinear with timeand stroke survivor ethnicity was collinear with caregiver ethnicity. As a result, thenumber of family members assisting with care and caregiver ethnicity were removedfrom the initial full model. Non-significant ( p>0.05) variables were removedone at a time, as long as the removal did not significantly degrade model fit.Once removed, variables were not allowed to re-enter. Interactions were notincluded in the models because of the relatively limited (relative to the numberof potential predictor variables) sample size. This is a limitation of the study, as astatistically significant interaction may obscure the main effects for one or more ofthe interacting variables.

Results

At baseline, 19 caregivers were classified as at-risk for depressive syndrome basedon CES-D scores that exceeded the cut-off score of 19 [56]. The remaining 33caregivers were classified as low-risk. Table 1 provides the frequency and percentageof at-risk and low-risk individuals over time.

Table 2 presents the means and standard deviations on the major sociodemo-graphic and self-report variables for at-risk depressed and non-depressed caregivers.The most parsimonious set of predictors of risk for depressive behaviour in family

802 J. S. Grant et al.

caregivers of stroke survivors in the acute care phase of the caregiver role includedcaregiver burden, ethnicity and people to do things with, or belonging, acomponent of social support (table 3).

Compared to African Americans, Caucasians had �3.7 times higher odds ofbeing at-risk for depression. For each unit increase in belonging scores, the oddsof being at-risk for depression decreased by �30%. Clinically, a 2-point increase inbelonging scores (e.g. on a scale ranging from 0–10) would reduce the odds of beingat-risk for depression by 55%. For each unit increase in caregiver burden, the oddsof being at-risk for depression increased by �5%. The Caregiver Burden scoresrange from 14–70, with values of 14 indicating no burden, 28 a small amount ofburden, 42 moderate burden, 56 a large amount and 70 indicating a great deal ofburden in caregiving tasks. Clinically, an 8-point increase in burden scores wouldincrease the odds of being at-risk for depression by 50%. Time and general healthwere non-significant, but retained in the model because removing them reducedgoodness-of-fit, as measured by the –2 Log Likelihood.

Discussion

Other studies show depressive symptomatology rates as high as 44% [57], while 37%experienced depressive symptomotology in this study; identical to another studyusing a similar sample of stroke family members at onset of the caregiver role [3].The present study revealed that several demographic and psychosocial variablesare related to caregiver distress in the first 3 moths after discharge from strokerehabilitation. Considerable research has documented relations between socialsupport and caregiver adjustment, but the reasons for these associations have notbeen well understood. Specific elements of support that often are identified ascrucial include the amount of available support, the adequacy of the support andthe overall satisfaction with the support received [58]. However, unmet social andinter-personal needs are also associated with caregiver distress [59, 60]. Additionally,there is some indication that elements of social support may differentially affectadjustment at different times in the adjustment process. For example, tangiblesupport emerged in the work as one of the best predictors of depressive sympto-matology among caregivers 1–2 days prior to their return to the community fromthe rehabilitation facility [3].

The results suggest that caregivers who experience decreases in their sense ofbelonging with others may be particularly vulnerable to distress in the weeksfollowing their return home. Perhaps these caregivers had unmet personal needsthat were previously addressed in relationships that were disrupted by the stroke, andin turn experienced a social isolation that compromized their adjustment.

Table 1. Frequency and percentage of caregivers at risk and those at low risk for problems with depressivebehaviour over time

Time

Variable Baseline 5 Weeks 9 Weeks 13 Weeks

At risk for problems with depressive behaviour 19 (37%) 19 (40%) 9 (21%) 14 (34%)Low risk for problems with depressive behaviour 33 (63%) 29 (60%) 34 (79%) 27 (66%)

Depressive behaviour in family caregivers 803

Table 2. Means and (standard deviations) on self-report variables for caregivers at risk and those at low risk for depressive behaviour over time

Depressive status of family caregivers

At risk Low risk

Variables T1 T2 T3 T4 T1 T2 T3 T4

General health (SF-36)Physical functioning 67.9 (20.5) 67.4 (25.2) 48.3 (21.1) 60.7 (31.4) 79.2 (23.6) 78.8 (25.0) 78.4 (23.4) 73.9 (22.4)General health 58.9 (18.5) 56.3 (19.9) 45.9 (18.0) 53.4 (17.6) 74.5 (21.1) 70.3 (21.1) 67.1 (21.2) 67.6 (23.0)

Social supportAppraisal 6.2 (2.3) 6.4 (2.6) 5.0 (3.0) 6.4 (2.7) 8.1 (1.6) 8.2 (1.6) 8.6 (1.4) 8.6 (1.4)Belonging 7.3 (3.1) 7.2 (2.8) 7.0 (2.7) 6.6 (3.0) 9.2 (1.1) 9.1 (1.5) 9.1 (1.7) 9.3 (1.3)Tangible support 8.1 (2.5) 8.5 (2.4) 8.4 (2.6) 8.3 (2.8) 9.4 (0.90) 9.6 (0.73) 9.4 (1.2) 9.6 (1.0)

Caregiver burdenDirect care 7.5 (2.6) 7.3 (3.4) 7.7 (3.4) 7.1 (3.2) 6.6 (2.7) 5.0 (2.1) 5.6 (2.6) 4.8 (2.0)Instrumental 18.6 (6.4) 20.1 (7.0) 18.9 (9.0) 20.1 (5.8) 15.8 (5.6) 13.9 (4.9) 13.7 (4.7) 12.4 (4.6)Interpersonal 10.7 (3.2) 11.6 (3.7) 11.4 (4.1) 13.1 (3.6) 8.9 (3.6) 7.6 (2.4) 8.1 (3.3) 7.2 (2.3)

Preparedness for caregiving scale 13.2 (4.0) 11.3 (3.6) 13.1 (4.7) 12.2 (4.0) 15.2 (3.1) 15.5 (2.8) 15.6 (3.0) 15.8 (3.2)Satisfaction with health care 26.6 (4.7) 23.2 (6.4) 20.9 (7.2) 21.7 (8.1) 26.7 (4.2) 27.9 (3.4) 26.3 (4.4) 26.9 (3.6)

804

J.S.Gran

tet

al.

These caregivers may also have experienced a sense of relational deprivation if thecare recipient was the primary source of this type of support and the deficits incurredby the stroke impaired their role in the relationship.

Despite the significant contributions of burden to the prediction of caregiverdepressive behaviour, the concept of burden has been scrutinized in recent work.Some critics observe that most measures of burden may be best construed as meas-ures of caregiver stress [61]. From this perspective, then, the results may indicatethat increases in caregiver stress—as assessed in part by the burden scale—areassociated with an increased likelihood of developing symptoms often characteristicof depression.

The significant relations between caregiver ethnicity and depressive behaviourare consistent with other studies that have reported similar findings [7, 18, 62].African-American caregivers appear to report fewer symptoms of depression thanCaucasian caregivers, but both groups display increases in ill health and physicalsymptoms over time [63]. The reasons for differences in emotional adjustmentremain unknown. Caution should be exercised in basing interpretations anddeveloping interventions on matters of biological race, as stereotypes may beinadvertently used to limit or justify services [64]. Further work is neededto explicate the social-cultural mechanisms that may account for these observeddifferences in self-reported levels of depressive behaviour.

Findings must be viewed cautiously because of the relatively small sample size,especially in terms of the absolute number of at-risk individuals. For example,Harrell et al. [65] recommend a minimum of 10 events per predictor variable inorder to obtain adequate coefficient estimates. While use of a GLMM approach torepeated measures helps somewhat, the fact remains that modelling was based ononly 51 non-randomly selected individuals, thereby limiting the degree to whichthese results may be generalized. While an important first step in developing infor-mation in this population, these results need to be validated in independent samples.

A note about the overall non-significant Time effect may be in order at thispoint. An examination of changes in risk for depressive behaviour between adjacentmeasurement points indicated that risk may not have been constant across the timeperiod. The sudden point decrease in risk for depressive behaviour at week 9 mayhave been a result of a higher dropout rate by those caregivers at particular risk fordepressive behaviour by that point, resulting in a bias toward lower depressive riskscores in the remaining sample.

Table 3. Final model predicting risk for depressive behaviour

Variables DF Chi-squared p-value Odds ratio

Time (weeks): Overall 3 6.23 0.10*0 vs 5 1 0.02 0.88 1.055 vs 9 1 6.11 0.01 0.329 vs 13 1 1.60 0.21 1.73

Belonging (social support) 1 8.30 <0.01 0.70Caregiver burden 1 5.98 0.01 1.05General health 1 2.98 0.08* 0.98Caregiver ethnicity 1 4.25 0.04 3.7

*Not statistically significant; retained in model as removal significantly degraded fit.

Depressive behaviour in family caregivers 805

This study, however, does identify potential factors that may be predictiveof potential depressed and non-depressed family caregivers of stroke survivors.Future research examining the effectiveness of interventions designed to increasebelongingness and decrease caregiver burden should be explored. Empiricaldata suggest that telephone follow-up by clinicians who assist caregivers to viewcaregiving problems less negatively and to use systematic problem-solving strategiesin dealing with them are effective in reducing caregiver burden and risk fordepressive behaviour [66, 67]. In their role as caregivers, family members shouldbe encouraged to set goals for doing enjoyable things with the stroke survivor andother people. These activities with other people to increase a sense of belongingnessare potentially very important in improving a positive mood and decreasingdepressive behavior [30].

Acknowledgements

This study was supported by a grant to the first author from the National Institute ofNursing Research, the University of Alabama at Birmingham and NeuroscienceNursing Foundation of the American Association of Neuroscience Nurses.

References

1. AGENCY FOR HEALTH CARE POLICY AND RESEARCH: Post-stroke Rehabilitation (AHCPR PublicationNo. 95-0662) (Rockville, MD: US Department of Health and Human Services), 1995.

2. NATIONAL STROKE ASSOCIATION: The Cost of Stroke to all Americans. Available online at:http://www.stroke.org/brain_stat.cfm (accessed 6 November 2001).

3. GRANT, J. S., BARTOLUCCI, A., ELLIOTT, T. et al.: Sociodemographic, physical, and psychosocialcharacteristics of depressed and non-depressed family caregivers of stroke survivors. Brain Injury,14: 1089–1100, 2000.

4. HAN, B. and HALEY, W. E.: Family caregiving for patients with stroke: review and analysis. Stroke,30: 1478–1485, 1999.

5. NATIONAL FAMILY CAREGIVERS ASSOCIATION: Family caregiving statistics. Available online at:http://www.nfcacares.org/NFC1998_stats.html (accessed 30 June 2002).

6. NIEBOER, A. P., SCHULZ, R., MATTHEWS, K. A. et al.: Spousal caregivers’ activity restriction anddepression: a model for changes over time. Social Science Medicine, 47: 1361–1371, 1998.

7. KOTILA, M., NUMMINEN, H., WALTIMO, O. et al.: Depression after stroke: results of theFINNSTROKE study. Stroke, 29: 368–372, 1998.

8. SCHULZ, R. and BEACH, S. R.: Caregiving as a risk factor for mortality: the caregiver health effectsstudy. JAMA, 282: 2215–2219, 1999.

9. EVANS, R. L., BISHOP, D. S. and HASELKORN, J. K.: Factors predicting satisfactory home care afterstroke. Archives of Physical Medicine & Rehabilitation, 72: 144–147, 1991.

10. ANESHENSEL, C., PEARLIN, L., MULLAN, J. et al.: Profiles in caregiving: the unexpected career (San Diego:Academic Press), 1995.

11. ALSPAUGH, M. E., STEPHENS, M. A., TOWNSEND, A. L. et al.: Longitudinal patterns of risk fordepression in dementia caregivers: objective and subjective primary stress as predictors. Psychologyand Aging, 14: 34–43, 1999.

12. MESHEFEDJIAN, G., MCCUSKER, J., BELLAVANCE, F. et al.: Factors associated with symptoms ofdepression among informal caregivers of demented elders in the community. Gerontologist, 38:247–253, 1998.

13. DENNIS, M., O’ROURKE, S., LEWIS, S. et al.: A quantitative study of the emotional outcome ofpeople caring for stroke survivors. Stroke, 29: 1867–1872, 1998.

14. GIVEN, B. and GIVEN, C.: Family caregiving for the elderly. In: J. J. Fitzpatrick, R. L. Taunton andA. K. Jacox (editors) Annual Review of Nursing Research (New York: Springer), pp. 77–101, 1991.

15. DRAPER, B. M., POULOS, C. J., COLE, A. et al.: A comparison of care-givers for elderly stroke anddementia victims. Journal of American Geriatric Society, 40: 896–901, 1992.

806 J. S. Grant et al.

16. SCHULZ, R., TOMPKINS, C. and RAU, M.: A longitudinal study of the psychological impact ofstroke on primary support persons. Psychology and Aging, 3: 131–141, 1988.

17. GRANT, J. S.: Social problem-solving partnerships with family caregivers. Rehabilitation Nursing,24: 254–260, 1999.

18. WRIGHT, L. K., HICKEY, J. V., BUCKWALTER, K. C. et al.: Emotional and physical health of spousecaregivers of persons with Alzheimer’s disease and stroke. Journal of Advanced Nursing, 30:552–563, 1999.

19. LI, L. W., SELTZER, M. M. and GREENBERG, J. S.: Social support and depressive syndromes:differential patterns in wife and daughter caregivers. Journal of Gerontology Series B—PsychologicalSciences & Social Sciences, 52B: S200–S211, 1997.

20. BOOKWALA, J. and SCHULZ, R.: A comparison of primary stressors, secondary stressors, and depres-sive symptoms between elderly caregiving husbands and wives: the caregiver health effects study.Psychology and Aging, 15: 607–616, 2000.

21. PUCHNO, R. and RESCH, N.: Husbands and wives as caregivers: Antecedents of depressionand burden. Gerontologist, 29: 159–165, 1989.

22. BEACH, S. R., SCHULZ, R., YEE, J. L. et al.: Negative and positive health effects of caring for adisabled spouse: longitudinal findings from the caregiver health effects study. Psychology and Aging,15: 259–271, 2000.

23. GRANGER, C. V., HAMILTON, B. B. and FIEDLER, R. C.: Discharge outcome after strokerehabilitation. Stroke, 23: 978–982, 1992.

24. LOEWEN, S. C. and ANDERSON, B. A.: Reliability of the modified motor assessment scale andthe Barthel Index. Physical Therapy, 68: 1077–1081, 1988.

25. ANDERSON, C. S., LINTO, J. and STEWART-WYNNE, E.: A population-based assessmentof the impact and burden of caregiving for long-term stroke survivors. Stroke, 26:843–849, 1995.

26. SCHWARZ, K. A. and BLIXEN, C. E.: Does home health care affect strain and depressive sympto-matology for caregivers of impaired older adults. Journal of Community Health Nursing, 14: 39–48,1997.

27. HODGSON, S. P., WOOD, V. A. and LANGTON-HEWER, R.: Identification of stroke carers at risk:a preliminary study of predictors of carers’ psychological well-being at one year post stroke.Clinical Rehabilitation, 10: 337–346, 1996.

28. WHITLATCH, C. J., FEINBERG, L. F. and SEBESTA, D. S.: Depression and health in family caregivers:adaptation over time. Journal of Aging and Health, 9: 222–243, 1997.

29. LEE, S. and WU, S.: Determinants of burden and depression among family caregivers. NursingResearch (China), 6: 57–68, 1998.

30. GRANT, J. S., WEAVER, M., ELLIOTT, T. R., BARTOLUCCI, A. A. and NEWMAN GIGER, J. Familycaregivers of stroke survivors: characteristics of caregivers at-risk for depression. RehabilitationPsychology, 49: 172–179, 2004.

31. SMITH, C. E., MAYER, L. S., PERKINS, S. B. et al.: Caregiver learning needs and reactions tomanaging home mechanical ventilation. Heart & Lung, 23: 157–163, 1994.

32. CARNWATH, C. M. and JOHNSON, D. A. W.: Psychiatric morbidity among spouses of patientswith stroke. British Medical Journal, 294: 409–411, 1987.

33. MALONEBEACH, E. E. and ZARIT, S. H.: Dimensions of social support and social conflict aspredictors of caregiver depression. International Psychogeriatrics, 7: 25–38, 1995.

34. THOMMESSEN, B., AASLAND, D., BRAEKHUS, A. et al.: The psychosocial burden on spouses ofthe elderly with stroke, dementia and Parkinson’s disease. International Journal of GeriatricPsychiatry, 17: 78–84, 2002.

35. UNIFORM DATA SYSTEM FOR MEDICAL REHABILITATION: Guide for the Uniform Data Set for MedicalRehabilitation, Version 5.0 (Buffalo: State University of New York at Buffalo ResearchFoundation), 1996.

36. POLLOCK, A.: Carers’ literature review. Nursing Times, 90: 31–33, 1994.37. CHAU, N., DALER, S., ANDRE, J. M. et al.: Inter-rater agreement of two functional independence

scales: the Functional Independence Measure (FIMSM) and a subjective uniform continuousscale. Disability Rehabilitation, 16: 63–71, 1994.

38. DODDS, T. A., MARTIN, D. P., STOLOV, W. et al.: A validation of the functional independencemeasure and its performance among rehabilitation inpatients. Archives of Physical Medicine &Rehabilitation, 74: 531–536, 1993.

Depressive behaviour in family caregivers 807

39. GRANGER, C., COTTER, A. C., HAMILTON, B., et al.: Functional assessment scales: a study ofpersons after stroke. Archives of Physical Medicine & Rehabilitation, 74: 133–138, 1993.

40. HAMILTON, B. B. and GRANGER, C. V.: A rehabilitation uniform data system (State University ofNew York at Buffalo: Functional Assessment Study Center, Department of RehabilitationMedicine), 1992.

41. MEDICAL OUTCOMES TRUST: SF-36 health survey update: July 1993 (Boston: Author), 1993.42. MCHORNEY, C. A., WARE, J. E., LU, J. F. R. et al.: The MOS 36-Item Short-Form Health Survey

(SF-36): III. Tests of data quality, scaling assumptions, and reliability across diverse patient groups.Medical Care, 32: 40–66, 1994.

43. WARE, J., SNOW, K., KOSINSKI, M. et al.: SF-36 Health Survey: Manual and interpretation guide(Boston: The Health Institute, New England Medical Center), 1993.

44. LARSEN, D. L., ATTKISSON, C. C., HARGREAVES, W. A. et al.: Assessment of client/patient satisfac-tion. Evaluation Program Planning, 2: 197–207, 1979.

45. CRYNS, A., NICHOLS, R., KATZ, L. et al.: The hierarchical structure of geriatric patient satisfaction.Medical Care, 27: 802–816, 1989.

46. COHEN, S., MERMELSTEIN, R., KAMARCK, T. et al.: Measuring the functional components ofsocial support. In: I. G. Sarason and B. R. Sarason (editors) Social support: Theory, research andapplications (Boston: Martinus Nijhoff ), pp. 73–94, 1985.

47. HEITZMANN, C. A. and KAPLAN, R. M.: Assessment of methods for measuring social support.Health Psychology, 7: 75–109, 1988.

48. ZIMMER, J. G., GROTH-JUNCKER, A. and MCCUSKER, J.: A randomized study of a home healthcare team. American Journal of Public Health, 75: 134–141, 1985.

49. ARCHBOLD, P. G., STEWART, B. J., GREENLICK, M. R. et al.: Mutuality and preparednessas predictors of caregiver role strain. Research in Nursing & Health, 13: 375–384, 1990.

50. ARCHBOLD, P. G. and STEWART, B. J.: Family Cargiving Inventory (Available from Patricia G.Archbold, Department of Family Nursing, School of Nursing, Oregon Health SciencesUniversity, Portland, OR 97201-3098), 1986.

51. ARCHBOLD, P. G., STEWART, B. J. and GREENLICK, M. R.: Caregiver Relief Project (Grant #5; R01NR01140) (Washington, DC: National Center for Nursing Research), 1983.

52. CAREY, P., OBERST, M. T., MCCUBBIN, M. A. et al.: Appraisal and caregiving burden in familymembers caring for patients receiving chemotherapy. Oncology Nursing Forum, 18: 1341–1348,1991.

53. OBERST, M. T., THOMAS, S. E., GASS, K. A. et al.: Caregiving demands and appraisal of stressamong family caregivers. Cancer Nursing, 12: 209–215, 1989.

54. RADLOFF, L. S.: The CES-D Scale: a new self-report depression scale for research in the generalpopulation. Applied Psychological Measures, 1: 385–401, 1977.

55. MCLEAN, J., ROPER-HALL, A., MAYER, P. et al.: Service needs of stroke survivors and theirinformal carers: a pilot study. Journal of Advances in Nursing, 16: 559–564, 1991.

56. HIMMELFARB, S. and MURRELL, S.: Reliability and validity of five mental health scales in olderpersons. Journal of Gerontology, 38: 333–339, 1983.

57. STEIN, P. N., WAYNE, A. G. and HIBBARD, M.: An examination of depression in the spouses ofstroke patients. Rehabilitation Psychology, 37: 121–130, 1992.

58. MITTLEMAN, M., FERRIS, S., SHULMAN, E. et al.: Efficacy of multicomponent individualizedtreatment to improve the well-being of Alzheimer’s caregivers. In: E. Light, G. Niederehe andB. D. Lebowitz (editors) Stress effects on family caregivers of Alzheimer’s patients: Research and inter-ventions (New York: Springer), pp. 156–184, 1994.

59. MALONEBEACH, E. and ZARIT, S.: Dimensions of social support and social conflicts as predictors ofcaregiver depression. International Psychogeriatrics, 7: 25–38, 1995.

60. RIVERA, P. A., ROSE, J. M., FUTTERMAN, A. et al.: Dimensions of perceived social supportin clinically depressed and non-depressed female caregivers. Psychology and Aging, 6: 232–237,1991.

61. CHWALISZ, K. and KASLER, V.: Perceived stress: a better measure of caregive burden. Measurementand Evaluation in Counseling and Development, 28: 88–98, 1995.

62. HALEY, W., WEST, C., WADLEY, V. et al.: Psychological, social, and health impact of caregiving:a comparison of black and white dementia family caregivers and noncaregivers. Psychology andAging, 10: 540–552, 1995.

808 J. S. Grant et al.

63. ROTH, D. L., HALEY, W., OWEN, J. et al.: Latent growth models of the longitudinal effects ofdementia caregiving: a comparison of African American and white family caregivers. Psychologyand Aging, 16: 427–436, 2001.

64. ELLIOTT, T. and RIVERA, P.: The experience of families and their carers in healthcare. In:S. Llewelyn and P. Kennedy (editors) Handbook of Clinical Health Psychology (Oxford: Wiley &Sons), in press.

65. HARRELL, F., LEE, K. and MARK, D.: Tutorial in biostatistics. Multivariable prognostic models:issues in developing models, evaluation assumptions and adequacy, and measuring and reducingerrors. Statistics in Medicine, 15: 361–387, 1996.

66. GRANT, J. S., ELLIOTT, T. R., WEAVER, M. et al.: A telephone intervention with family caregiversof stroke survivors after rehabilitation. Stroke, 33: 2060–2065, 2002.

67. STEWART, M. J., HART, G., MANN, K. et al.: Telephone support group intervention for personswith hemophilia and HIV/AIDS and family caregivers. International Journal of Nursing Studies, 38:209–225, 2001.

Depressive behaviour in family caregivers 809