1*1 Library and Archives - University of Guelph Atrium

THE BURDEN OF ACUTE GASTROINTESTINAL ILLNESS IN GALVEZ, ARGENTINA

AND THE METROPOLITAN REGION, CHILE

A Thesis

Presented to

The Faculty of Graduate Studies

of

The University of Guelph

by

M. KATHLEEN THOMAS

In partial fulfillment of requirements

for the degree of

Doctor of Philosophy

May, 2010

© M. Kathleen Thomas, 2010

1*1 Library and Archives Canada

Published Heritage Branch

395 Wellington Street Ottawa ON K1A 0N4 Canada

Bibliotheque et Archives Canada

Direction du Patrimoine de I'edition

395, rue Wellington Ottawa ON K1A 0N4 Canada

Your file Votre reference ISBN: 978-0-494-67821-3 Our We Notre r6f6nence ISBN: 978-0-494-67821-3

NOTICE: AVIS:

The author has granted a nonexclusive license allowing Library and Archives Canada to reproduce, publish, archive, preserve, conserve, communicate to the public by telecommunication or on the Internet, loan, distribute and sell theses worldwide, for commercial or noncommercial purposes, in microform, paper, electronic and/or any other formats.

L'auteur a accorde une licence non exclusive permettant a la Bibliotheque et Archives Canada de reproduce, publier, archiver, sauvegarder, conserver, transmettre au public par telecommunication ou par Nnternet, prefer, distribuer et vendre des theses partout dans le monde, a des fins commerciales ou autres, sur support microforme, papier, electronique et/ou autres formats.

The author retains copyright ownership and moral rights in this thesis. Neither the thesis nor substantial extracts from it may be printed or otherwise reproduced without the author's permission.

L'auteur conserve la propriete du droit d'auteur et des droits moraux qui protege cette these. Ni la these ni des extraits substantiels de celle-ci ne doivent etre imprimes ou autrement reproduits sans son autorisation.

In compliance with the Canadian Privacy Act some supporting forms may have been removed from this thesis.

Conformement a la loi canadienne sur la protection de la vie privee, quelques formulaires secondaires ont ete enleves de cette these.

While these forms may be included in the document page count, their removal does not represent any loss of content from the thesis.

Bien que ces formulaires aient inclus dans la pagination, il n'y aura aucun contenu manquant.

1*1

Canada

ABSTRACT

THE BURDEN OF ACUTE GASTROINTESTINAL ILLNESS IN GALVEZ, ARGENTINA AND THE METROPOLITAN REGION, CHILE

M. Kathleen Thomas Advisor:

University of Guelph, 2010 Dr. Scott A. McEwen

This thesis is an investigation of the burden of acute gastrointestinal

illness (Gl) in Galvez, Argentina and the Metropolitan region, Chile. A cross-

sectional survey was administered door-to-door to 2915 and 6047 randomly

selected residents of Galvez and the Metropolitan region, respectively. Two time

periods, coinciding with high and low Gl seasons, were selected for survey

administration. Shorter (i.e., 7 and 15 day) and longer (i.e., 30 day) recall

periods were incorporated in the surveys. The surveys were designed to

determine the distribution and population-level burden of Gl, identify risk factors

including food and water consumption and food safety behaviours, estimate

pathogen-specific community-level incidence rates, and evaluate the effect of

different recall periods on Gl incidence rates in population-level burden of Gl

studies. High response rates of 61% and 76% were obtained in Galvez and the

Metropolitan region, respectively. The annual incidence rate of Gl in the

Metropolitan region ranged from 0.98 to 2.3 episodes per person-year, for a 30-

day and a 7-day recall period, respectively, and the annual incidence rate of

diarrhea in Galvez, ranged from 0.46 to 1.68 episodes per person-year for a 30-

day and a 7-day recall period, respectively. In general, little difference in the rate

of Gl was seen between men and women, while children 0-4 and youths 10-19

had higher rates of Gl in both Galvez and the Metropolitan region. In the

Metropolitan region, Gl was associated with consumption of undercooked

poultry, undercooked beef, cheese made from unpasteurized milk and higher

water consumption, though causality cannot be inferred due to the different recall

periods used for risk factors versus Gl symptoms and the cross-sectional study

methodology. The estimated community-level, pathogen-specific incidence rates

for Salmonella, Campylobacter and Shigella infections were greater than the

laboratory confirmed incidence rates, likely due to ill residents not seeking

medical care, not submitting a stool sample and some laboratories not routinely

testing for certain enteric pathogens. This research has demonstrated the

significant burden of Gl in these South American communities that is not

captured by traditional surveillance and outbreak registries, and has highlighted

risk factors that can inform public health policy and prevention activities.

ACKNOWLEDGEMENTS

"It takes a village to raise a child"

- African proverb

This thesis would not have been possible without the help, guidance and

support from the many people that make up my 'village'. My thanks and

gratitude to the entire faculty and staff of the Department of Population Medicine

for their assistance, support and mentorship throughout these past 3 V2 years; in

particular to William Sears for his statistical expertise and colourful

conversations. Thanks to the Public Health Agency of Canada, the International

Development Research Centre, and the Pan American Health Organization for

supporting this research. Thanks to my study partners and participants in

Argentina and Chile. Thanks to my fellow graduate students for their friendship,

their time for sharing ideas and commiserating about the common struggles of

graduate school. Thanks to my friends, for their thoughtful patience and kind

encouragement. Thanks to my advisory committee, without them none of this

would have been possible. Specifically, thanks to Dr. Richard Reid-Smith for his

calm guidance, intelligence and resourcefulness throughout this process; to Dr.

Enrique Perez for his dedication, constant demonstration that there is always a

way to get things done and his pursuit of equity in public health; to Dr. Shannon

Majowicz for her continued mentorship, positive attitude and incredible content

knowledge; and to Dr. Scott McEwen for his wisdom and advice, epidemiology

expertise and kindness. I was truly blessed to have four complementary

individuals that provided me with the necessary support, expertise and

mentorship, I could not have asked for a better team. Thanks to my family, for

their love and support, in particular my parents for their faith in me and for

exposing me to other parts of the world early on. And finally, thanks to my fiance

Fernando, for his love and support, for his patience and willingness to travel and

for his encouragement to pursue my goals. It truly 'takes a village to make a

Doctoral student'.

a

TABLE OF CONTENTS

CHAPTER ONE: Introduction and Literature Review 1

INTRODUCTION 1

LITERATURE REVIEW 4

RATIONALE 36

OBJECTIVES 39

REFERENCES 41

TABLES 46

CHAPTER TWO: Burden of acute gastrointestinal illness in Galvez,

Argentina, 2007 53

ABSTRACT 53

INTRODUCTION 54

MATERIALS AND METHODS 55

RESULTS 60

DISCUSSION 62

REFERENCES 68

TABLES 71

FIGURES 79

CHAPTER THREE: Burden of acute gastrointestinal illness in the

Metropolitan region, Chile, 2008 80

SUMMARY 80

INTRODUCTION 80

in

METHODS 82

RESULTS 86

DISCUSSION 89

REFERENCES 95

TABLES 98

FIGURES 105

CHAPTER FOUR: Risk factors for acute gastrointestinal illness related to food consumption trends, food purchasing and hygiene habits among residents of the Metropolitan Region, Chile, 2008 106

ABSTRACT 106

INTRODUCTION 107

METHODS 108

RESULTS 113

DISCUSSION 117

CONCLUSIONS 123

REFERENCES 125

TABLES 129

FIGURES 138

CHAPTER FIVE: Water consumption trends and associations with acute gastrointestinal illness in the Metropolitan Region, Chile, 2008 139

ABSTRACT 139

BACKGROUND 140


IV

RESULTS 145

DISCUSSION 149

CONCLUSIONS 156

REFERENCES 157

TABLES 161

FIGURES 166

CHAPTER SIX: Estimated numbers of human infections due to Salmonella, Campylobacter and Shigella, Metropolitan region, Chile, 2008 171

ABSTRACT 171

BACKGROUND 172


RESULTS 176

DISCUSSION 177

CONCLUSIONS 182

REFERENCES 184

TABLES 189

CHAPTER SEVEN: Conclusions 193

SUMMARY, DISCUSSION AND CONCLUSIONS 193

APPENDICES 205

APPENDIX I: Survey tool - Argentina (7 day recall period) 206

APPENDIX II: Survey tool - Argentina (30 day recall period) 210

APPENDIX III: Formulas for calculating prevalence, incidence rate and

V

incidence proportion 214

APPENDIX IV: Survey tool - Chile (7, 15 and 30 day recall periods) 215

APPENDIX V: Expected probability formulas - Chile 222

APPENDIX VI: Survey tool - Chile (Laboratory survey) 223

APPENDIX VII: Survey tool - Chile (Sentinel clinic case-control study).231

VI

LIST OF TABLES

Table Title Page 1.1 Summary of published population-based burden of acute 46

gastrointestinal illness (Gl) studies 1991 - 2006, by region, in chronological order.

1.2 Summary of food and water consumption data from 50 population-based studies, in various countries 1992 - 2006, in chronological order.

2.1 Respondent representativeness, demographic distribution and 71 the prevalence of acute gastrointestinal illness per study phase and recall period in Galvez Argentina, 2007.

2.2 Symptoms and their duration for both study phases and recall 73 periods combined, Galvez, Argentina, 2007.

2.3 Days of missed work and school by cases and care-givers for 74 both study phases and recall periods combined, Galvez, Argentina, 2007.

2.4 Final multivariate model of risk factors associated with acute 75 gastrointestinal illness in Galvez, Argentina, 2007.

2.5 Medications and access to medical care, for both study 76 phases and recall periods combined, Galvez, Argentina, 2007.

2.6 Number and mean, minimum, and maximum percentage of 77 cases that sought medical attention and the estimated underreporting, for both study phases and recall periods, Galvez, Argentina, 2007.

2.7 Minimum set of results proposed for studies of acute 78 gastrointestinal illness (25) for both study phases and recall periods, Galvez, Argentina, 2007.

3.1 Socio-demographic distribution of Metropolitan region 98 residents, survey respondents and monthly prevalence of acute gastrointestinal illness by category, Chile 2008.

3.2 Number of cases by gastrointestinal illness symptom, 99 prevalence, annual incidence rate and annual incidence proportion, by recall period and phase, Metropolitan region, Chile 2008.

vn

3.3 Number and percent of cases (n=467) by secondary 100 symptoms, duration of gastrointestinal symptoms and duration of missed activities due to gastrointestinal illness, Metropolitan region, Chile, 2008.

3.4 Number and percent of cases (n=467) by treatments, use of 101 medical care and reasons for not seeking medical care by Gl cases, Metropolitan region, Chile, 2008.

3.5 Univarable analysis results of association with acute 102 gastrointestinal illness, Metropolitan region, Chile. 2008.

3.6 Final multivariable model of risk factors associated with acute 103 gastrointestinal illness, Metropolitan region, Chile, 2008.

3.7 Descriptive statistics of acute gastrointestinal illness based on 104 30-day recall period following the proposed standard case definition of gastrointestinal illness, Metropolitan region, Chile, 2008.

4.1 Age, gender and education distribution of Metropolitan region 129 residents and survey respondents, Chile 2008.

4.2 Distribution of food purchasing habits in the seven days prior 130 to interview, monthly prevalence of acute gastrointestinal illness, and unconditional associations between location of food item purchase Gl in the 30 days prior to interview, Metropolitan region, Chile, 2008.

4.3 Frequency of hand washing and hygiene behaviours, monthly 132 acute gastrointestinal illness prevalence and unconditional associations between hand washing and hygiene behaviours and Gl in the 30 days prior to interview, Metropolitan region, Chile, 2008.

4.4 Proportion of respondents that consumed high risk food at 134 least once in the seven days prior to interview, monthly prevalence of acute gastrointestinal illness and unconditional associations between consumption of individual high risk food and Gl, Metropolitan region, Chile, 2008 (n=6047).

4.5 Associations between reported consumption of high risk food 135 items and gender in the seven days prior to interview, Metropolitan region Chile, 2008.

4.6 Associations of reported consumption of high risk food items 136

vm

and age group in the seven days prior to interview compared with referent group of individuals 20-59 years of age, Metropolitan region, Chile, 2008.

4.7 Final multivariable model of socio-demographic, food 137 consumption and hygiene behaviour factors associated with Gl, Metropolitan region, Chile, 2008.

5.1 Age, gender and education distribution of Metropolitan region 161 residents and survey respondents, Chile 2008.

5.2 Mean number of drinking water servings consumed, by 162 occupation, adjusted for other variables in the multivariable model, Metropolitan region, Chile, 2008 (n=5765).

5.3 Final multivariable logistic regression model of risk factors for 163 Gl and proportion of respondents with Gl by risk factor, Metropolitan region, Chile, 2008 (n=5709).

6.1 Data sources and input distributions used to estimate under- 189 reporting of gastro-intestinal infections (Salmonella, Campylobacter, and Shigella) in the Metropolitan region, Chile, 2008.

6.2 The annual reported number of infections and incidence by 190 Instituto de Salud Publica (ISP) and estimated annual number of cases, incidence and under-reporting multipliers for Salmonella, Campylobacter, and Shigella infections, Metropolitan region, Chile, 2008.

6.3 Sensitivity analysis of correlation of input distributions and 191 overall under-reporting multiplier by pathogen for the Metropolitan region, Chile, 2008.

6.4 Summary of international pathogen-specific incidence rates 192 for comparison with results from the Metropolitan region, Chile, 2008.

IX

LIST OF FIGURES

Figure Title Page 2.1 Theoretical burden of illness pyramid for Galvez, Argentina, 79

2007.

3.1 Monthly prevalence of Gl by symptoms and age group, 105 Metropolitan region, Chile, 2008.

4.1 Unconditional odds of acute gastrointestinal illness (Gl) and 138 proportion of meals consumed in different locations, Metropolitan region, Chile, 2008.

5.1 Estimated mean number of servings of water consumed in the 166 24 hours prior to interview by gender and study phase from multivariable model with 95% confidence interval error bars, Metropolitan region, Chile, 2008 (n=5765).

5.2 Estimated mean number of servings of water consumed in the 167 24 hours prior to interview by study phase and education level with from multivariable model with 95% confidence interval error bars, Metropolitan region, Chile, 2008 (n=5765).

5.3 Estimated mean number of servings of water consumed in the 168 24 hours prior to interview by gender and age category from multivariable model with 95% confidence interval error bars, Metropolitan region, Chile, 2008 (n=5765).

5.4 Estimated mean number of servings of water consumed in the 169 24 hours prior to interview by 3-way interaction of socioeconomic level and study phase and being a bottled water user from multivariable model with 95% confidence interval error bars, Metropolitan region, Chile, 2008 (n=5765).

5.5 Probability of being a case of Gl by socio-economic level and 170 gender from multivariable model with 95% confidence interval error bars, Metropolitan region, Chile, 2008 (5709).

x

CHAPTER ONE

Introduction and Literature Review

Introduction

Acute gastrointestinal illness (Gl) causes significant morbidity, mortality

and socio-economic burden around the world (1-4). Developed countries tend to

experience lower rates of mortality, but high rates of morbidity and socio

economic loss. Developing countries typically have a higher mortality burden,

though deaths caused by diarrhea appear to be declining (5,6). Clean water,

sanitation and safe food are key components to preventing and controlling Gl at

the population-level (7). These public health areas are at the forefront of

international public health organizations' objectives and priorities as well as local

public health workers concerns (8-11).

Understanding the magnitude, distribution and specific risk factors

associated with Gl is a good first step in mitigation (12). With this baseline

information, interventions, targeted surveillance and research activities can be

developed and their success assessed. Likewise, this fundamental knowledge

provides insight into the impacts of broader worldwide trends, such as

globalization, climate change and international travel and trade, on the rates,

distributions and risk factors of disease. It also informs international estimates

such as the World Health Organization Global Burden of Disease Assessments

specific to foodbome diseases (3).

1

Existing surveillance based on disease notification and outbreak

investigation reporting activities around the world substantially underestimates

the true burden of Gl in the population (13). Many cases of Gl are not captured

by local and national surveillance systems, and outbreak reports can also be

biased due to a lack of source identification, incomplete case capture and

unidentified outbreaks not being included (14,15). In order to have a clearer

understanding of disease status within the population, targeted research

activities, including burden of illness studies, need to be undertaken (16,17).

These studies focus on capturing the magnitude and impact of a particular

illness. Studies can be designed to capture a variety of qualitative and

quantitative burden indicators, for example case counts, direct costs associated

with treatment, and indirect costs associated with lost productivity (18).

To date, the majority of research pertaining to population-level burden of

Gl has occurred in developed countries (12,16,17,19-27), with few studies

occurring within developing countries (28-30). Common results from studies in

developed populations include higher rates of Gl in females (16,17,24), young

children (16,17,20,22,24,26), and those with higher levels of education

(20,22,26).

In developed countries, methods used for burden of illness studies

commonly include surveys of the population, laboratories, physicians, and

analysis of reportable disease surveillance system data. With the information

gained from these surveys it is possible to generate multipliers and to determine

estimates of disease in the population and the under-reporting rate (31-33).

2

Among the scientists currently using the existing methodologies for

estimation of burden of Gl, there is debate regarding the advantages and

disadvantages of prospective cohort designs versus retrospective approaches

using cross-sectional surveys that seek information about symptoms in the

recent past (27). The prospective approach is more costly and time consuming

but provides more accurate estimates of disease incidence (22,27). The

retrospective approach, though less expensive and time consuming, is thought to

suffer from a type of recall bias known as 'telescoping' (i.e., the tendency to recall

illness events from the more distant past into the recent past) which results in

potential overestimation of the burden of Gl (27).

Methods utilized in developed countries may not be suitable for use in

developing countries. Studies spanning many months are expensive; they

require substantial personnel and financial investments that are prohibitive in

countries with limited public health resources and extensive competing public

health priorities. Trained telephone survey companies are used in some

developed countries for burden of illness studies but these companies are less

common or non-existent in developing countries. Residents with fixed telephone

lines may not be representative of the population in some developing countries,

especially with the popularity of cellular phones1. In addition, literacy rates tend

to be lower in developing countries, limiting the utility of mail-in surveys.

To develop intervention and policy, understanding the source of the illness

is of great importance (34,35). Disease attribution involves crediting cases of

' International Telecommunications Union, http://www.itu.int/ITU-D/ICTEYE/lndicators/lndicators.aspx#. accessed February 23, 2010.

3

http://www.itu.int/ITU-

illness to their source or cause; in terms of foodborne disease, it is the attribution

of cases of illness to the responsible food vehicle (34). This is currently a

developing area of research which can play a key role in informing food safety

initiatives (36). With an increased understanding of what foods are contaminated

and what levels of exposure to these foods exist, a more complete food safety

risk profile can be created. This can be used to generate interventions or

monitoring strategies. An initial step of this work is determining what foods

people are consuming on a regular basis and to identify what food-based risks

are prevalent in the population. This enables targeted food safety initiatives to be

developed and tested for effectiveness and economic viability.

Literature Review

This literature review focuses on three main types of studies:

1. Burden of acute gastrointestinal illness population surveys. The studies are

presented by region and in chronological order, including key results and

estimates of the burden of acute gastrointestinal illness by age, gender and

socio-economic level. Study designs and response rates are presented in brief

for general comparison; however, detailed descriptions of the methodologies and

study discussions are not included.

2. Food and water consumption surveys. Patterns in food and water

consumption by socio-demographic factors, and proportions of bottled water and

water filtration use results are described in a population-level burden of Gl study

4

context. Nutrition-focused, water intervention trials or other non-population,

burden of illness based studies are not included in this review.

3. Community estimates for pathogens. Brief descriptions of analytic

approaches, data sources and under-reporting estimates for specific pathogens

are presented. Literature outside of the population-based, burden of Gl context

is not reviewed.

1. Burden of acute gastrointestinal illness population surveys

1.1 Europe

The Netherlands-1991:

Hoogenboom-Verdegaal et al. (21) conducted a community-based study

in four regions of the Netherlands that provided information on the incidence and

severity of gastroenteritis in single households or families. A cohort study was

conducted from March to July 1991 in ten geographically representative

municipalities located within four regional health services. Each of the four

health services contacted approximately 1,500 people by letter from the

municipal registers and one person from each household was asked to complete

the survey. Participants were asked about the presence and persistence of

gastrointestinal symptoms, if they sought medical care and what type. Weekly

questionnaires were to be returned over the 17 consecutive weeks of the study.

Those who developed symptoms during the study period were asked to submit a

stool sample immediately. Participants were asked to report symptoms of

diarrhea (two or more stools in a day), fever, vomiting, nausea, abdominal pain

5

and or cramps, and the presence of blood and/or mucus in stools. Grade 1

gastroenteritis was defined as diarrhea or vomiting and at least two additional

symptoms of nausea, abdominal pain, cramps, blood or mucus in stools within a

1 week period. Grade 2 gastroenteritis was defined as diarrhea or vomiting and

at least two additional symptoms as in grade 1 but both on the same day and

lasting at least 2 days within a 1 week period. In total, there were 2,257

participants (36% response rate) of which 425 and 115 had symptoms consistent

with grade 1 and 2 gastroenteritis, respectively. A doctor was consulted by 18%

and 22% of the cases for grade 1 and 2 gastroenteritis, respectively, with half of

these consultations being in person and the remainder by telephone. Women

and children 0-18 years of age had higher risks of grade 1 gastroenteritis

compared to males and adults 19-35 years of age.

Infectious Intestinal Disease (IIP) Study- England 1993 - 1996:

Wheeler et al. (27) studied the incidence and etiology of cases of

infectious intestinal disease (IID) (a) in the community and (b) presenting to

general practitioners, and their relationship with national surveillance (laboratory)

reports. Prospective cohort and cross-sectional study designs were used.

General practices were selected from the Medical Research Council's general

practice research framework and 70, serving approximately 460 000 people,

volunteered to participate in the study. The sample was representative of

national general practices in terms of location, urban and rural characteristics,

and social deprivation index. From each practice, 200 people of all ages were

6

selected at random and invited to participate by letter and telephone. A case of

I ID was defined as anyone experiencing loose stools or significant vomiting

(more than once in 24 hours, incapacitating, or accompanied by cramps or fever)

lasting less than 2 weeks, in the absence of a known non-infectious cause and

preceded by a symptom-free period of three weeks. Participants returned

weekly postcards for six months indicating absence of symptoms. Those with

symptoms sent a stool specimen to the Leeds Public Health Laboratory. A

second cohort was selected for an additional six months. Data were collected

from 1993 to 1996. In total, 9,776 people were recruited (40% response rate) for

a total of 4,026 person-years and there were 781 cases of IID; an incidence rate

of 19.4/100 person years. The retrospective estimate of reported diarrhea in the

month prior to recruitment from the cross-sectional survey was 564/8,674, which

extrapolates to an incidence rate of 55/100 person years. The ratio of community

incidence to general practice presentation rates was 5.8. For every isolate

reported to the national surveillance there were 1.4 positive laboratory results,

6.2 stools submitted, 23 cases presented to the practitioner and 136 cases in the

community.

Sensor (A population based cohort study on Gastroenteritis) - Netherlands 1998

-1999:

De Wit et al. (22) studied the incidence of gastroenteritis in the general

population and associated causative pathogens in different age groups.

Prospective population-based cohort and nested case-control study designs

7

were used. Two 6-month cohorts were recruited and data were collected from

December 1998 - December 1999. An age-stratified random sample was

selected from all people registered at the 44 participating sentinel general

practices from the Netherlands Institute of Primary Health Care, 27 practices

participated in the first cohort and 31 in the second. Selected individuals were

invited by mail with a follow-up letter 3 weeks later to non-respondents. A

baseline questionnaire was completed at the beginning of the study and

participants returned a card weekly, reporting presence or absence of symptoms.

Participants were instructed to contact the study coordinator by telephone if they

developed symptoms of vomiting or diarrhea, and a stool sample was to be

collected immediately. The case definition was at least three loose stools within

24 hours or vomiting at least three times within 24 hours or diarrhea with two or

more additional symptoms in 24 hours; additional symptoms were diarrhea,

vomiting, abdominal cramps, abdominal pain, fever, nausea, blood in stool, or

mucus in stool. If there was an obvious non-infectious cause the case was

excluded. For each case, a control was selected matched on age, degree of

urbanization and region, and requested to submit stool samples. In total, 4,860

people participated (42% response rate) with a total of 1,050 case episodes. The

overall incidence rate standardized by age, sex and cohort was 283 per 1,000

person-years. Incidence was slightly higher in the first cohort, much higher in

children 0-11 years of age and increased with educational level. Seasonal peaks

were seen in January and June. Antibiotic use in the week prior to illness was

reported by 2% of cases. Norwalk-like virus (NLV) was the most commonly

8

identified pathogen for all age groups except those 18-64 where Staphylococcus

aureus toxins were most common.

Norway 1999-2000:

Kuusi et al. (24) conducted a national survey to estimate the incidence

and burden of gastroenteritis in Norway. A cross-sectional survey design was

used. Data were collected from June 1999 - June 2000. A total of 3,000 people

were selected at random from the Norwegian population registry and were

mailed the survey. A reminder letter was mailed one week after initial mail-out

and a second reminder and an additional copy of the survey were mailed out to

non-respondents within 3 weeks. Surveys of children under 15 years of age

were completed by their parents. A case was defined as a person who had

diarrhea (three or more loose stools in 24 hours) or at least three of the following

symptoms: vomiting, nausea, abdominal cramps or fever > 38 °C in the 4 weeks

prior to completion of survey. Those with chronic diarrheal illness were excluded.

In total, 1,843 respondents (61% response rate) completed the survey,

symptoms were reported by 265 (14%) of respondents of which 171 met the

case definition. Women and those between ages 25 and 39 years were most

likely to be cases. Approximately 17% of cases visited their physician, 8% of

cases submitted a stool sample and 4% of cases were hospitalized for their

illness. For those 0-14 years of age, consumption of water from a private supply

(well or surface source) was identified as a significant risk factor while consuming

water from a chlorinated water source was found to be a protective factor in this

9

age group. For those 21-40 years of age, travel outside of Norway was identified

as a risk factor. For all other age groups, no significant differences were found

for any of the risk factors.

Ireland 2000-2001:

Scallan et al. (16) studied the extent of acute gastroenteritis in the

population in Northern Ireland and the Republic of Ireland as well as the

associated health-seeking behaviour and impact on missed work and school. A

cross-sectional survey was administered by telephone. Data were collected from

December 2000 to November 2001. Random digit dialling was used to survey

private households with a fixed telephone line. The individual with the next

birthday was selected to respond to the survey, with proxy respondents for

children < 12 years of age and at the discretion of the parent for children 12-16

years of age. A case was defined as anyone experiencing diarrhea with three or

more loose stools in 24 hours or bloody diarrhea or vomiting together with at

least one other symptom (diarrhea, abdominal pain/cramps, or fever) in the 4

weeks prior to interview and in the absence of a known non-infectious cause. In

total, 9,903 interviews were completed (64.8% response rate) with 4.5% of

respondents reporting at least one episode of acute gastroenteritis in the 4

weeks prior to interview, an incidence rate of 0.60 episodes per person-year.

Women, children < 14 years of age, adults 25-44 years of age and those

reporting a professional or non-manual occupation were more likely to report

being a case. A higher frequency of reporting of acute gastroenteritis was found

10

in Northern Ireland compared to the Republic of Ireland. From multivariable

analysis sex (being female), age (<5 years old), occupation of main earner

(professional/non-manual) and jurisdiction (Northern Ireland) were significantly

associated with reporting acute gastroenteritis. Seasonal peaks were found in

Northern Ireland in December and April, and in April and September in the

Republic of Ireland. A physician was consulted by 29.2% of cases and 9.1% of

cases were asked to submit a stool sample. For 17.4% of cases, they

themselves or a family member were required to take time off work, while for

19.0% of cases someone in their household had to take time off from school.

1.2 North America

FoodNet - USA 1996 - 1997:

Herikstad et al. (20) conducted a survey of the Foodborne Diseases Active

Surveillance Network (FoodNet) to better understand and more precisely quantify

the amount and burden of illness caused by acute diarrhea in the United States.

A cross-sectional survey was administered by telephone. Data were collected

between July 1996 and June 1997. Random digit dialling was used to identify

households within the FoodNet sites. Respondents were asked about symptoms

of diarrhea, defined as three or more loose stools or bowel movements in any 24

hour period within the four weeks prior to interview. Those respondents with a

chronic illness in which diarrhea is a major symptom (e.g., colitis, irritable bowel

syndrome) or who had had surgery to remove part of their stomach or intestine

were excluded from the analysis. Diarrheal illness was defined as diarrhea

11

lasting longer than one day or which resulted in significant impairment of daily

activities. Overall, 9,003 surveys were completed (71% response rate); 8624

surveys were completed by eligible respondents of which 977 had experienced

symptoms of diarrhea in the 4 weeks prior to interview, resulting in a rate of 1.4

episodes of diarrhea per person-year, and 492 had experienced diarrheal illness.

Prevalence of diarrheal illness was highest among children <5 years of age and

lowest among adults > 65 years of age. Prevalence of diarrheal illness increased

with increasing education and was more common in respondents living in urban

areas. Of those with diarrheal illness, 7% reported taking antibiotics to treat their

illness and 12% reported visiting a healthcare provider. Of those who visited a

healthcare provider, 21% reported being requested to provide a stool sample of

which 89% complied and 8% of those who visited a healthcare provider were

hospitalized.

FoodNet - USA 1998 - 1999:

Imhoff et al. (23) conducted the second survey of the FoodNet sites to

continue tracking the burden of self-reported diarrheal illness. A cross-sectional

survey was administered by telephone between July 1998 and June 1999.

Random digit dialling was used to identify households within the FoodNet sites.

One individual per household was randomly selected from a roster of household

members. Parents or guardians responded for children <12 years of age.

Respondents were asked if they had experienced diarrhea, defined as three or

more loose stools in any 24 hour period in the 4 weeks prior to interview. People

12

with a chronic illness (e.g., Crohn's disease or irritable bowel syndrome) and

people with part of their stomach or intestine removed were excluded from the

analysis. Diarrheal illness was defined as diarrhea that lasted more than 1 day.

Overall 12,755 surveys were completed (41% response rate); 12,075 were

eligible respondents of which 1,192 (10%) reported having symptoms of diarrhea

in the 4 weeks prior to interview resulting in a rate of 1.3 episodes of acute

diarrhea per person-year, of which 645 had diarrheal illness. Children < 5 years

of age and adults 25 - 44 years of age had the highest prevalences of acute

diarrheal illness, while adults > 65 years of age had the lowest. There were no

differences in prevalence between males and females, or between urban and

rural dwellers. Of those with diarrheal illness, 12% reported taking antibiotics to

treat their illness and 21% sought medical care of which 16% were asked to

submit a stool sample - 97% complied. Overall, 2% of those with diarrheal illness

were hospitalized.

FoodNet - USA 1996 - 2003:

Jones et al. (12) summarized four population-based surveys that

assessed acute diarrheal illness within FoodNet sites. In addition to the two

studies already mentioned by Herikstad et al. (20) and Imhoff et al. (23), two

additional similar cross-sectional surveys were conducted by telephone in March

2000 - February 2001 and March 2002 - February 2003. Diarrhea was defined

as three or more loose stools in 24 hours and acute diarrheal illness was defined

as diarrhea with duration greater than one day or impairment of daily activities.

13

Those with chronic illness with diarrhea as a major symptom were excluded from

analysis. The 2000/2001 survey had a total of 14,139 eligible respondents of

which 7.3% experienced symptoms of diarrhea and 5.0% experienced diarrheal

illness. The 2002/2003 survey had a total of 15,578 eligible respondents of which

7.2% and 5.2% experienced symptoms of diarrhea and diarrheal illness,

respectively. In both surveys, children < 5 years of age and adults 18-35 years of

age had the highest prevalences of diarrheal illness. Females in the 2002/2003

survey had a significantly greater prevalence of diarrheal illness compared to

males, 6.0% compared to 4.5%. When all four studies were combined, the

overall rate of diarrheal illness among blacks and those with less than a high

school education were lower than the comparison groups of whites and high

school graduates. The rate of diarrheal illness was higher among rural residents

and medically uninsured compared to urban residents and medically insured

respondents, respectively. Sex, race, age and insurance status were

significantly associated with acute diarrheal illness in multivariable modeling.

Overall 19.5% of those suffering diarrheal illness sought medical care, 3.9% were

asked to submit a stool sample and 3.7% of respondents complied with this

request, additionally 1.9% of all respondents with diarrheal illness were

hospitalized.

National Studies on Acute Gastrointestinal Illness (NSAGI) - Canada 2001-2002:

Majowicz et al. (17) estimated the magnitude and distribution of acute

gastrointestinal illness in a Canadian community. A cross-sectional survey was

14

administered by telephone from February 2001 to February 2002. Population

sampling was conducted using a randomized list of residential telephone

numbers, and one individual from each household was identified randomly by

selecting the individual with the next birthday. Respondents were asked whether

they experienced any symptoms of diarrhea or vomiting in the previous 28 days.

In total 3,496 surveys were completed (36.6% response rate); 351 respondents

were identified to be cases of acute gastrointestinal illness; an additional 77

identified their symptoms as due to a pre-existing condition and were included in

the non-case group for analysis. Prevalence was significantly higher in children

0-9 years of age and young adults 20-24 years of age and there were

significantly more female than male cases. Overall, 1% of cases reported being

hospitalized for their illness.

NSAGI - Canada 2002-2003:

Thomas et al. (26) described the frequency, magnitude, distribution and

clinical burden of acute, self-reported gastrointestinal illness in British Columbia.

A cross-sectional telephone survey was administered from June 2002 to June

2003. A randomized list of residential telephone numbers was used and one

individual from each household was identified randomly by selecting the

individual with the next birthday. Respondents were asked about experiencing

any symptoms of diarrhea or vomiting in the previous 28 days. In total 4,612

surveys were completed (44.3% response rate); 451 respondents were identified

as acute cases of gastrointestinal illness; an additional 131 cases were identified

15

as chronic cases and were included in the non-case group for analysis.

Prevalence was highest in those 0-9 and 10-14 years of age, higher in females

than in males and higher in those with a university, graduate or professional

degree. Overall, 11.6% of cases visited a healthcare provider, of which 23.1 %

were asked to submit a stool sample; 83.3% complied. Of cases 18 years of age

or older, 32.5% took time away from work because of their illness of which 43.5%

lost income as a result of their absence.

NSAGI - Canada (2005 - 2006):

Sargeant et al. (25) estimated the burden, severity and demographic

distribution of acute gastrointestinal illness in Ontario. A cross-sectional

telephone survey was administered from May 2005 to April 2006. Ontario

residential telephone numbers were randomly selected and one individual from

each household was identified at random by selecting the person with the next

birthday. The survey elicited information on symptoms of vomiting or diarrhea in

the previous 28 days as well as medical history and medical system or

medication use to treat illness. In total, 2,090 surveys were completed (36.6%

response rate); 178 respondents were identified as acute cases of Gl; an

additional 34 were identified as chronic cases and were included in the non-case

group for analysis. Prevalence was higher among rural residents (compared to

urban residents), females (compared to males) and in the months of February

and April (compared to December). Overall, 22% of cases sought medical care,

16

of which 33% were asked to submit a stool sample and 100% complied with this

request.

1.3 Australia

OzFoodNet - Australia 2001 - 2002:

Hall et al. (19) conducted a national community survey to estimate the

number of cases of gastroenteritis and to identify any regional, seasonal,

demographic and socioeconomic risk factors for gastroenteritis in Australia. A

cross-sectional survey was administered via telephone from September 2001 to

August 2002. Random digit dialling was used to select households and

interviewers asked to speak to the person with the next birthday. Parents or

guardians responded on behalf of children <15 years of age. Respondents were

asked about symptoms of vomiting and diarrhea within the 4 weeks prior to

interview. Infectious gastroenteritis was defined as at least three loose stools or

two vomits in 24 hours, or at least four loose stools or three vomits in 24 hours if

respiratory symptoms were present. In total, 6,987 surveys were completed

(67% response rate); 11.2% of respondents reported having any diarrhea or

vomiting and 7.4% met the case definition for symptoms in the 4 weeks prior to

interview. Prevalence was higher among females than males and among young

children (0-4 years of age). From multivariable analysis, region, season, age and

sex were found to be associated with having gastroenteritis. There were greater

odds of gastroenteritis in Northern Territory compared to Queensland, in summer

compared to spring, for children 0-4 years of age and for females compared to

17

males. For those with less than year 10 of education, higher levels of

gastroenteritis were associated with lower and higher income levels, and lower

levels of gastroenteritis were associated with middle levels of income. For those

with higher levels of education, income had little effect. Those without medical

insurance were more likely to report gastroenteritis.

1.4 Jordan

Jordan 2003 - 2004:

Gargouri et al. (29,30) estimated the burden of disease due to Salmonella,

Shigella and Brucella infections in Jordan. As part of this work, national cross-

sectional population surveys were conducted in September 2003 and May 2004

to estimate the numbers of people experiencing symptoms consistent with

Salmonella, Shigella and Brucella infections (i.e., diarrhea for salmonellosis and

shigellosis, and persistent fever for brucellosis and salmonellosis caused by

Salmonella Typhi infection). Respondents were selected at random and surveys

were administered face-to-face. Those less than one year of age were excluded

from the survey. Questions pertained to diarrhea and persistent fever in the 30

days prior to interview as well as whether the ill person sought medical care, and

if so whether they went to a Ministry of Health (MOH) facility and submitted a

sample. Diarrhea was defined as three or more loose stools that took the shape

of a container in a 24 hour period, and persistent fever was defined as a fever

lasting more than 48 hours. In September 2003, 759 surveys were completed

(91% response rate); 7.8% of respondents experienced diarrhea and 2.3%

18

experienced persistent fever. Of the diarrhea cases, 19.5% sought medical

attention and 8.9% went to a Ministry of Health (MOH) facility. In May 2004, 819

surveys were completed (98% response rate); 6.1% of respondents experienced

diarrhea and 1.1 % experienced persistent fever. Of the diarrhea cases, 40.8%

sought medical care and 71.1 % went to a MOH facility. Overall the rate of

diarrhea in the community was 0.8 episodes per person-year.

1.5 Latin America

Cuba 2005 - 2006:

Prieto et al. (28) determined the temporal and demographic distribution

and burden of self-reported acute gastrointestinal illness in Cuba. A cross-

sectional survey was administered door-to-door within three sentinel sites during

two time periods, June - July 2005 (rainy season) and November 2005 -

January 2006 (dry season). Households were randomly selected from a list

maintained by the medical office at the site. One individual per household was

selected to complete the interview. Though attempts were made to select this

person at random by asking for the person with the next birthday, in most

interviews the person who answered the door was the respondent. Acute

gastrointestinal illness was defined as three or more bouts of loose stools in a 24

hour period in the 30 days prior to interview. In total, 6,399 interviews were

completed (97.3% response rate); 680 respondents had experienced symptoms

of acute gastrointestinal illness. Prevalence varied by site and season. From

multivariable analysis, gastrointestinal illness was higher in the rainy season, in

19

children 0-12 years of age and teens 13-17 years of age compared to adults 25-

54 years of age, and in males. Proportions of cases that visited a physician

ranged from 17.1 % to 38.1%, of which 33.3% to 53.9% were asked to submit a

stool sample and 72.7% -100.0% complied. Of those who sought treatment,

0.0% - 31.6% received an antibiotic for treatment. Hospitalization was low with

1.2% of cases being hospitalized for their illness.

Table 1.1 provides a summary of the above described burden of acute

gastrointestinal illness surveys. These studies were the most recent and relevant

to this thesis, however additional studies do exist (37). Three studies used

prospective cohort methodology while the remaining studies were conducted as

cross-sectional studies. Only two studies were conducted in developing

countries (Cuba and Jordan) and both were administered face-to-face, which

likely contributed to the high response rates achieved. Additionally, these

surveys were conducted during two time periods rather than over an entire year,

as was the case with the studies conducted in developed countries (excepting

the first study in the Netherlands (21)). In many of the studies, a greater

occurrence of Gl was found in women and children. Additionally, some studies

found GI to be associated with higher education and certain months or seasons.

Incidence of Gl ranged from 0.18 to 1.4 episodes per person-year, with the three

studies conducted as prospective cohorts having the lowest incidence estimates.

It is likely that variation in incidence estimates reflects not only potential true

differences in disease burden in these locations, but also differences in

20

methodology and case definition employed in the study, thus making direct

comparisons more difficult.

2. Food and water consumption community surveys

2.1 Food consumption surveys

The Food and Drug Administration (FDA) National Survey - USA 1992 - 1993:

Klontz et al. (38) assessed the prevalence of selected food consumption

and preparation behaviours associated with increased risks of food-borne illness

and demographic characteristics related to such behaviours. The FDA

completed a telephone survey of 1,620 individuals (65% response rate) with

questions pertaining to consuming raw animal protein foods and handling of

cutting boards. More than 50% of respondents reported eating foods with raw

eggs, 23% usually ate hamburgers undercooked and 17% reported eating raw

shellfish. After cutting raw meat or chicken, 26% of respondents did not wash

their cutting board with soap or bleach or replace with a different cutting board to

continue preparing food. Males, adults 18-39 years of age and those with more

than a high school education were more likely to consume raw animal protein

foods, compared to females, adults >39 years of age and those with less than

high school education, respectively. Similarly, males and individuals 18-39 years

of age were more likely to use a dirty cutting board.

21

The Behavioral Risk Factor Surveillance System (BRFSS) - USA 1995 - 1996:

Yang et al. (39) reported the results of standard food safety questions that

were added to the BRFSS survey in 8 states in 1995 and 1996. The BRFSS

survey is an on-going, telephone health survey system that tracks health

conditions and risk behaviours in the United States.2 Surveys were administered

to non-institutionalized adults (>17 years of age) by telephone in 1995 and 1996.

The standard food safety questions pertained to handling of raw meat or chicken,

consumption of high risk food items, knowledge of safe food handling practices

and occurrence of diarrhea. In total, 19,356 interviews were completed and

responses pertained to the 12 months prior to interview. Approximately 50% of

respondents had eaten undercooked eggs in the previous 12 months.

Consumption of undercooked hamburgers was more common among males than

females, decreased with increasing age and increased with increasing salary and

education. Nearly 20% of respondents did not wash their hands or the cutting

board with soap after handling raw meat or chicken.

FoodNet - USA 1998 - 1999:

Samuel et al. (40) reported results from food consumption questions of the

FoodNet survey administered to 12,755 Americans > 17 years of age. In total,

10,209 responses were included in the analysis. Approximately 18% of

respondents consumed runny eggs in the 7 days prior to interview, followed by

alfalfa sprouts (8%) and pink hamburgers (7%). Overall, 62% of respondents

2 Center for Disease Control, Behavioral Risk Factor Surveillance System, http://www.cdc.qov/brfss/.

22

http://www.cdc.qov/brfss/

reported not eating any of the risky foods in the 7 days prior to interview. Males

were more likely than females to consume more than 1 risky food. Adults >64

years of age were less likely to consume risky foods while immuno-compromised

respondents were more likely to consume risky foods compared to their healthy

counterparts.

C-EnterNet - Canada 2005 - 2006:

Nesbitt et al. (36) evaluated the food consumption patterns of the general

population in the Waterloo Region of Ontario and described, from a food safety

perspective, demographic factors that related to the consumption of specific food

items. A cross-sectional survey was administered by telephone from November

2005 to March 2006 to randomly selected residents. Households were randomly

sampled from a list of residential telephone numbers and the individual with the

next birthday was selected for interview. Questions pertained to location of meal

consumption, types of food consumed, water consumption, hygiene and food

safety practices, acute gastrointestinal illness in the previous week and previous

four weeks, demographics, grocery purchasing habits and food preparation

habits. Questions about food consumption pertained to the seven days prior to

interview. Where appropriate, questions pertaining to grocery purchasing were

asked of the person identified as most familiar with these practices for the

household. In total, 2,332 surveys were completed (32.7% response rate). In

general, males were more likely than females to consume foods considered high

risk for the transmission of enteric pathogens. Respondents >65 years old were

23

more likely to consume eggs, including undercooked eggs, than other age

groups. Consumption of chicken nuggets, hamburgers (not made at home) and

deli meats was highest among children, teens and young adults suggesting

elevated interest in convenience foods for these age groups. Home was the

most commonly reported place for meal preparation with an average of 25.8

meals per week prepared at home. The percentage of respondents whose food

consumption patterns in the week prior to survey were typical of a normal week

ranged from 80.3% to 95.5% depending on general food category.

2.2 Water consumption surveys

FoodNet - USA 1998 - 1999:

Lee et al. (41) reported results of water consumption-related questions

that were incorporated into FoodNet surveys administered to 7 sites from 1998-

1999. In total 12,755 people participated and 63.8% identified municipal water

as their primary water source, followed by bottled water (17.8%), and private well

water (15.0%). Of tap water drinkers, 30% treated their water, with filtration

being the most common method (76.0%). Reasons for drinking bottled water

included improved taste and odour (49.1%), avoiding chemicals (28.0%) and

avoiding germs (16.5%). Experiencing diarrheal illness was not associated with

any of the water exposure variables.

NSAGI - Canada 2001 - 2002:

24

Jones et al. (42) described the drinking water consumption patterns and

associations with various demographic characteristics of the residents of

Hamilton, Ontario. A cross-sectional telephone survey was administered

September 2001 - March 2002. A residential telephone listing was used to

identify households and the individual with the next birthday was selected to

respond to the survey. Proxy respondents were used for subjects > 12 years of

age, and at the discretion of the parent or guardian for those 1 2 - 1 8 years of

age. Questions pertained to the amount of water consumed in the previous 24

hour period including all plain water or water consumed from beverages made

from water combined with frozen juices or flavoured crystals as well as the

amount of water consumed from bottled water. Other questions pertained to use

of water treatment devices as well as demographic variables. A total of 1,757

surveys were completed (37.4% response rate). Total daily water intake ranged

from zero to thirty-two 250 ml servings, with a median of four servings. From

multivariable analysis, water consumption decreased with increasing age and

was higher for respondents with higher levels of education than 'less than high-

school', and residents using in-home water treatment devices consumed more

water than those who did not use water treatment devices. In total, 27.3% of

respondents were classified as bottled water users (i.e., those who consumed

75% or more of their total water as bottled water) while 59.7% of respondents did

not consume any bottled water. From multivariable analysis, the probability of

being a bottled water user increased with increasing age until the age of 31 and

then decreased as age increased; in-home water treatment users were less likely

25

to be bottled water users, and the odds of bottled water use was less for

weekdays compared to weekends. Use of water treatment devices was reported

by 49% of respondents, with jug filters being used by 66.2% of these

respondents.

NSAGI - Canada 2002 - 2003:

Jones et al. (43) described the drinking water consumption patterns and

associations with demographic characteristics and acute gastrointestinal illness

in three communities in British Columbia. A cross-sectional telephone survey

was administered June 2002 - June 2003. Questions pertained to the amount of

water consumed as tap water and as bottled water in the 24 hours prior to

interview, as well as use of water treatment devices in the home, source of water

(municipal, private source, both, or other source), demographic factors and

recent symptoms of vomiting and diarrhea. In total, 4,612 surveys were

completed (44% response rate). Total daily water consumption ranged from zero

to thirty-six 250ml servings with a median of four servings in the 24 hours prior to

interview. From multivariable analysis, bottled water users consumed more

water, water consumption was higher on weekdays than on weekends,

respondents with higher levels of education consumed more water, and females

over the age of 25 years consumed more water than males. In total, 23% of

respondents were identified as bottled water users (i.e., those who consumed

75% or more of their total water as bottled water) while 67.4% of respondents

consumed no bottled water in the 24 hours prior to interview. From multivariable

26

analysis, bottled water use increased with increasing age until approximately 26

years of age and then began to decline with increasing age. An increased

probability of being a bottled water user was associated with municipal water

source compared to private water source, a higher level of income and with not

using an in-home water treatment device. In-home water treatment devices were

used by 47% of the respondents, with jug filters being the most common choice

(53%). Odds of reporting symptoms of acute gastrointestinal illness increased

with increasing amount of water consumed, controlling for age and sex.

Sweden 1999-2003:

Westrell et al. (44) estimated the drinking water consumption and

evaluated potential demographic differences that could impact water intake in

Sweden using three data sources: 1. a national environmental health survey

conducted in 1999; 2. a waterborne disease outbreak investigation in 2002; and,

3. a small water consumption study from 2003. The national survey was

administered to 15,496 residents of which 11,233 responded (73% response

rate) and 10,957 provided answers regarding water intake within the home. The

waterborne disease outbreak investigation reported results of a cohort study

questionnaire administered to all 605 permanent residents in the area of the

outbreak, of which 387 responded (64% response rate) and 157 people provided

information on water consumption. Heated tap water and bottled water

consumption was the focus of the small water consumption study with 75

respondents (63% response rate). Daily consumption of cold tap water ranged

27

from 0.2 to 2.4 litres with an average of 0.86 litres. Women drank more tap water

than men, people >70 years of age consumed the most water while those aged

40-49 years consumed the least. Average daily bottled water intake was low with

0.06 litres/day being consumed; however, increasing bottled water consumption

was associated with increasing income.

C-EnterNet - Canada 2005 - 2006:

Pintar et al. (45) examined the daily amount of water consumed and

different factors related to water consumption. A cross-sectional survey was

administered by telephone from November 2005 to March 2006 to randomly

selected residents of the Waterloo Region of Ontario. Questions pertained to

amount of water consumed in the 24 hours prior to interview, use of water

treatment devices and amount of bottled water consumed along with

demographic variables. Interviews were completed by 2,332 respondents

(32.7% response rate). Answers regarding amount of water consumed were

obtained from 2,189 respondents, among which total water consumption ranged

from 0 to 6.25L per day, mean of 1.39L. From multivariate modeling, men

consumed less water than women and this volume decreased with age of

respondent. Frequent bottled water drinkers (i.e., those that consumed 75% or

more of their total water as bottled water) consumed less water. A higher level of

education was associated with higher water consumption, except for those with

an advanced post-graduate degree who consumed less than the referent group

with some trade, college or university training. Respondents who had an

28

advanced water treatment system consumed more water if they also consumed

bottled water, compared to those who consumed bottled water and had either no

water treatment system or a carbon filter system. In total, 34% of respondents

were identified as bottled water users and 48% consumed no bottled water in the

24 hours prior to interview. From multivariable analysis of the non-bottled water

users, older respondents consumed less water and this effect was more

pronounced in males than in females. Overall, males consumed less water than

females, a higher level of education was associated with higher water

consumption (except for those with an advanced degree where it was associated

with lower water consumption), and respondents that used an advanced

treatment device consumed less water than those that used a carbon filter.

Table 1.2 summarizes the food and water consumption studies reviewed

above. All of these studies were conducted in developed countries and all but the

Swedish water consumption study solely used cross-sectional methodology. The

English IID study attempted to identify foods associated with I ID; however, due to

their methodology (prospective cohort with a nested case-control) they were

unable to achieve this goal (46). Foods considered to be higher risk for Gl were

regularly consumed in these study populations. Several studies report that more

water is consumed by women than men and those with higher levels of

education. A wealth of related literature on nutrition and nutrition related

diseases (e.g., diabetes, obesity, etc.), water intervention trials, outbreak

29

investigations and other non-population, burden of acute gastrointestinal illness

based studies exists but is outside the scope of this review.

3. Community estimates for pathogens

IIP Study-England:

Wheeler et al. (27) studied the incidence and etiology of infectious

intestinal disease (IID) in England in 1993 - 1996 as described earlier.

Pathogen-specific incidence rates in the community were calculated based on

the number of incident cases occurring in the population cohort divided by the

number of person-weeks of follow-up. Pathogen-specific general practice rates

were calculated based on the number of incident cases presenting at a general

practice divided by the practice population. Community and general practice

pathogen-specific incidence rates were combined with rates of positive laboratory

testing and reporting to national surveillance to construct reporting pyramids for

all IID, Campylobacter, Salmonella, Rotavirus, and Small round structured

viruses. Ratios of the number of community cases to those reported to the

national surveillance for Salmonella, Campylobacter, Rotavirus and Small round

structured viruses were 3.2:1, 7.6:1, 35.0:1 and 1562:1, respectively.

USA:

Mead et al. (35) used data from numerous sources including FoodNet

surveys to estimate the number of food-related illnesses and deaths in the United

States. Total cases reported, by pathogen, was obtained from passive and

30

active surveillance systems, outbreak reports, and individual studies. Under-

ascertainment by pathogen was taken into account using multipliers generated

from FoodNet surveys, literature and expert opinion. A factor of 38 was used for

pathogens that cause primarily non-bloody diarrhea. A factor of 20 was used for

pathogens that typically cause bloody diarrhea. A factor of 2 was used for

pathogens that typically cause very severe illness. Estimated total annual cases

were generated for a number of bacteria, parasites and viruses, including

approximately 1.4 million cases of salmonellosis, 2.5 million cases of

campylobacteriosis, 0.5 million cases of shigellosis and 73,000 infections due to

Escherichia coli 0157:H7 in the entire population.

FoodNet-USA:

Voetsch et al. (31) used data from FoodNet surveys and other sources to

estimate the number of non-Typhoidal Salmonella infections and resultant

physician visits, hospitalizations and deaths in the community that occurred

annually in the USA from 1993 - 1996. Multipliers were generated for those with

both bloody and non-bloody symptoms, with the relevant proportions determined

from a Salmonella case-control study. Data from FoodNet active surveillance

provided age-specific incidence rates of salmonellosis for the FoodNet sites,

which were extrapolated to the entire USA population to determine an estimated

annual total number of laboratory confirmed cases of salmonellosis. The

FoodNet active surveillance system provided information on the rate of

hospitalization and death. Data from a laboratory survey were used to estimate

31

the frequency of laboratory testing for Salmonella and the sensitivity of laboratory

tests was estimated from the literature. Data from population surveys were used

to estimate the proportion of cases that seek medical attention, have a stool

sample requested and submit a stool sample. For each culture-confirmed case

of Salmonella it was estimated that there were 38.6 cases in the community and

that Salmonella caused 14,860 hospitalizations and 415 deaths annually.

NSAGI-Canada:

Thomas et al. (47) used data from the NSAGI surveys and other literature

to estimate the number of community cases of illness due to Salmonella,

Campylobacter and verotoxigenic Escherichia coli (VTEC) in Canada circa 2000.

Multipliers were determined for both bloody and non-bloody diarrhea and the

proportion of pathogen-specific cases in each category was based on information

from the international literature. The Canadian National Notifiable Disease

registry (NND) provided data on the annual number of laboratory confirmed

cases. Data from the NSAGI public health reporting survey were used to

determine the frequency of case reporting from local to provincial health

authorities. Data from the NSAGI laboratory survey were used to estimate the

frequency of laboratory reporting to the local health authority and the frequency

of sample testing for each pathogen. Sensitivities of laboratory tests were taken

from international literature. The NSAGI population surveys provided estimates

for the frequency of cases seeking medical attention, being requested to submit a

sample and complying with the sample request. Conservative and liberal

32

estimates were calculated for each pathogen to account for uncertainty in the

estimates. It was estimated that for every case of VTEC, Salmonella and

Campylobacter infection reported to NND, there were 10-47, 13-37 and 23-49

infections annually in the Canadian population, respectively. This corresponds to

an annual rate of 0.7 - 3.3, 2.5 - 6.9 and 9.1 - 19.3 per 1,000 Canadians, of

VTEC, Salmonella and Campylobacter infections, respectively.

QzFoodNet - Australia:

Hall et al. (48) estimated multipliers to be applied to the annual number of

cases of salmonellosis, campylobacteriosis and Shiga toxin-producing

Escherichia coli (STEC) infections reported to the Australian Notifiable Diseases

Surveillance System, as well as the community incidence of these infections in

Australia. Severity of symptoms (bloody vs. non-bloody, and duration (1-2 days,

3-4 days and 5 or more days)) was used to categorize cases of gastroenteritis,

and multipliers were calculated according to these categories. Infections by the

three pathogens of interest were classified by these same severity categories.

To generate the case categories, data from the Australian National

Gastroenteritis (NGS) survey and unpublished reports on practices for treatment

and management of gastroenteritis by general practitioners were used. Data

from the Royal College of Pathologists Australasia Quality Assurance Programs

were used to determine the probability of correctly identifying Salmonella and

Campylobacter in stool samples by laboratories. Expert opinion was used to

determine the probability of a positive laboratory result being reported to health

33

authorities. Symptom profiles of salmonellosis and campylobacteriosis were

taken from unpublished case-control studies in Australia. Unpublished data from

OzFoodNet provided information on reported cases of STEC and laboratory

sensitivity of detecting STEC. The numbers of annual community Salmonella,

Campylobacter and STEC infections in Australia were estimated to be 49,843,

224,972 and 4,420, respectively, circa 2000-2004. This corresponds to an

annual rate of 262, 1184 and 23 Salmonella, Campylobacter and STEC

infections per 100,000 population.

Jordan:

Gargouri et al. (30) estimated the burden of disease due to Salmonella,

Shigella and Brucella infections in Jordan. National cross-sectional population

surveys were conducted in September 2003 and May 2004 to estimate the

number of people experiencing symptoms consistent with Salmonella, Shigella

and Brucella infections (i.e., diarrhea for non-typhoidal salmonellosis and

shigellosis and persistent fever for brucellosis and salmonellosis caused by

Salmonella Typhi infection). During September 2003 and May 2004 a survey of

all laboratories that received blood or stool samples to test for Salmonella,

Shigella or Brucella was conducted to determine the number of stool cultures,

blood cultures, Brucella tube agglutination tests, and the number of laboratory

confirmed cases of Salmonella, Shigella or Brucella. National surveillance data

from September 2003 to August 2004 were reviewed to compare the number of

reported Salmonella, Shigella or Brucella infections in Jordan to the number

34

reported in the Ministry of Health (MOH) laboratory survey. The annual average

number of cases was estimated by summing the two month estimates and

multiplying by six. Multipliers were calculated using the proportion of ill people >1

year of age who sought care only and those who sought care at MOH facilities

and submitted a clinical stool or blood specimen to an MOH laboratory. These

multipliers were applied to the numbers of laboratory confirmed non-typhi

Salmonella, Shigella or Brucella infection obtained from the MOH laboratory

survey to estimate the burden of disease. No laboratory samples were positive

for Salmonella Typhi or Salmonella Paratyphi A, B or C. The total numbers of

national non-typhi Salmonella and Shigella infections were estimated to be

16,266 and 6,606 per year, respectively, which corresponds to rates of 306 and

124 per 100,000 people. The total number of national cases of Brucella infection

was estimated to be 6912 per year, corresponding to a rate of 130 cases per

100,000 people.

The six above-reviewed studies provide community pathogen-specific

burden estimates; only one was conducted in a developing country. Some work

has been done to better capture the uncertainty of these estimates with

confidence intervals and credible intervals calculated for estimates by the English

and Australian studies, respectively. Several other key studies (49,50) have

incorporated and made refinements to these sorts of calculations; however, their

focus was specific to the burden attributable to foodborne disease, (i.e., number

of cases, hospitalizations and deaths due to foodborne disease, by pathogen)

35

and thus are not included in this review. Other studies describe pathogen-

specific burden but were conducted in specific demographic groups (i.e., children

<5 years old) typically using prospective cohort methodology rather than

population study-based multiplier estimations, and, thus, are not included in this

review.

Rationale

Although the reviewed studies have established some accepted

approaches to burden of illness research, several important methodologic

questions remain. For example: what is the impact of using different recall

periods in a population-level burden of acute gastrointestinal illness study?; how

can population-level burden of acute gastrointestinal illness study methods be

adapted for use in developing countries or countries with limited resources? In

addition, while the burden of Gl is relatively well understood in developed

countries, it remains largely undescribed in developing or under-resourced

countries, including those of South America. Burden of Gl questions specific to

South American include: what is the population-level burden of Gl in South

America?; what population-level food and water consumption-related risk factors

for Gl exist in South America?; and what are the community-level enteric

pathogen-specific burden estimates in South America? Answering these

questions is the basis for the rationale of this thesis.

The concern around recall bias is central to study design methodology for

data collection in a population study. Recall bias, a form of information bias, can

36

result in misclassification of study participants (51). This misclassification can

contribute to over- or under-estimation of the true burden of illness as well as

potentially distort relationships with factors and behaviours of interest (e.g.,

socio-demographic factors, medical care seeking behaviours etc.). These errors

can be exacerbated, for example, when these data are extrapolated to generate

estimates for entire populations or regions, used to inform risk assessments or to

generate pathogen-specific community estimates.

The prospective cohort approach is less likely to suffer from recall bias,

but is costly and more difficult to complete than the survey approach. In addition,

this approach can include the requirement that participants who experience Gl

symptoms to submit a stool sample, as was the case in England (27) and the

Netherlands (21,22). This may deter respondents from reporting symptoms in

order to avoid having to collect a stool sample, contributing to under-estimation.

The cross-sectional study design can be subject to recall bias in the form of

'telescoping' or forgetting, which can result in over- or under-estimates, but is

less costly and time consuming and does not typically require stool samples from

those with symptoms.

Route of survey administration can affect survey cost as well as response

rate and study participation. Face-to-face surveys are more labour-intensive,

time-consuming and costly compared to telephone or mail-in survey

administration; however, response and participation rates may be compromised.

Maintaining anonymity is a challenge for face-to-face interviews which may make

potential responders hesitant to participate. A lack of literacy may generate

37

selection bias in a mail-in survey where respondents would be required to read

and understand each question in order to participate. Similarly, the use of

cellular phones or lack of land-line telephones may generate selection bias in a

telephone survey if participants are selected from a land-line directory or are

unwilling to participate in a survey using their cellular phone. It is crucial to

ensure an appropriate study population to obtain representative results.

Capture of data over a full year or multiple years is ideal; however, this

can be prohibitive due to budget and resource limitations. Approximations can

be made by selecting Gl high- and low-season time periods for the study (e.g.,

using historical surveillance data to determine when the highest and lowest

incidence of Gl occur in the population and administering the study during these

two time periods). This method makes data collection more vulnerable to

disease outbreaks, timing of holidays or special events and other unexpected

changes during survey time periods. It is necessary to have reliable historical

surveillance data from which to select time periods.

The scarcity of information from less developed parts of the world,

including South America, has resulted in a gap of valuable information on acute

gastrointestinal illness. The World Health Organization has recognized the need

for studies and the development of suitable methodologies to address this

knowledge gap (3,52). Identification of region and country-specific risk factors

and behaviours associated with acute gastrointestinal illness, in particular food

and water consumption and food handling behaviours, is important for effective

control and prevention strategies. The relative importance of these factors may

38

vary by location, gender, age and socio-economic status, and, thus, must be

explored with these variables in mind.

Generating pathogen-specific community level estimates for specific

countries or regions is important for better understanding of the burden of acute

gastrointestinal illness in the population. This information can highlight target

areas for prevention and control measures, as well as methods that can enhance

pathogen surveillance and reporting mechanisms.

Proper case definitions for diarrhea and acute gastrointestinal illness are

important for study comparisons. This has been the topic of much discussion in

the literature and is still evolving (53-56). Although not specifically addressed as

part of this thesis, a proposed standard case definition (53) was used when

possible.

Objectives

The main goal of this thesis was to describe the distribution and

population-level burden of acute gastrointestinal illness in select South American

communities, identify risk factors for Gl and evaluate the effect of different recall

periods on Gl incidence rates in population-level burden of acute gastrointestinal

illness studies. To achieve this goal, community-level burden of acute

gastrointestinal illness studies were conducted in Galvez, Santa Fe, Argentina

and in the Metropolitan Region, Chile. The specific objectives were:

39

• to determine the burden, distribution and associated risk factors of acute

gastrointestinal illness in Galvez, Argentina and the Metropolitan Region,

Chile;

• to examine the impact of different recall periods on precision and accuracy

of burden estimates;

• to evaluate food consumption trends and food sources within a food safety

context and their relationship with acute gastrointestinal illness in the

study areas;

• to evaluate water consumption trends and their relationship with acute

gastrointestinal illness in the study areas;

• to estimate the number of cases at the study population-level for specific

pathogens related to acute gastrointestinal illness and of public health

importance.

40

References

1. Guerrant RL, Kosek M, Moore S, Lorntz B, Brantley R, Lima AA. Magnitude and impact of diarrheal diseases. Arch Med Res 2002;33(4):351-5.

2. Kosek M, Bern C, Guerrant RL. The global burden of diarrhoeal disease, as estimated from studies published between 1992 and 2000. Bull World Health Organ 2003;81 (3): 197-204.

3. Stein C, Kuchenmuller T, Hendrickx S, Pruss-Ustun A, Wolfson L, Engels D, et al. The Global Burden of Disease assessments--WHO is responsible? PLoS Negl TropDis 2007; 1(3):e161.

4. Bryce J, Boschi-Pinto C, Shibuya K, Black RE, WHO Child Health Epidemiology Reference Group. WHO estimates of the causes of death in children. Lancet 2005;365(9465):1147-52.

5. Podewils LJ, Mintz ED, Nataro JP, Parashar UD. Acute, infectious diarrhea among children in developing countries. Semin Pediatr Infect Dis 2004;15(3):155-68.

6. Farthing MJ. Diarrhoea: a significant worldwide problem. Int J Antimicrob 4gente2000;14(1):65-9.

7. Kaferstein F. Foodbome diseases in developing countries: aetiology, epidemiology and strategies for prevention. Int J Environ Health Res 2003; 13 SuppM:S161-8.

8. Schlundt J. New directions in foodbome disease prevention. Int J Food Microbiol 2002;78(1 -2):3-17.

9. Herikstad H, Motarjemi Y, Tauxe RV. Salmonella surveillance: a global survey of public health serotyping. Epidemiol /nfecf 2002;129(1):1-8.

10. Guerrant RL, Kosek M, Lima AA, Lorntz B, Guyatt HL. Updating the DALYs for diarrhoeal disease. Trends Parasitol 2002; 18(5): 191 -3.

11. Flint JA, Van Duynhoven YT, Angulo FJ, DeLong SM, Braun P, Kirk M, et al. Estimating the burden of acute gastroenteritis, foodbome disease, and pathogens commonly transmitted by food: an international review. Clin Infect Dis 2005;41(5):698-704.

12. Jones TF, McMillian MB, Scallan E, Frenzen PD, Cronquist AB, Thomas S, et al. A population-based estimate of the substantial burden of diarrhoeal disease in the United States; FoodNet, 1996-2003. Epidemiol Infect 2007; 135(2):293-301.

41

13. Scallan E, Majowicz SE, Hall G, Banerjee A, Bowman CL, Daly L, et al. Prevalence of diarrhoea in the community in Australia, Canada, Ireland, and the United States. Int J Epidemiol 2005;34(2)A54-60.

14. Prado V, Solari V, Alvarez IM, Arellano C, Vidal R, Carreno M, et al. Epidemiological situation of foodborne diseases in Santiago, Chile in 1999-2000. Rev Med CA7//2002;130(5):495-501.

15. O'Brien SJ, Gillespie IA, Sivanesan MA, Elson R, Hughes C, Adak GK. Publication bias in foodborne outbreaks of infectious intestinal disease and its implications for evidence-based food policy. England and Wales 1992-2003. Epidemiol Infect 2006; 134(4):667-74.

16. Scallan E, Fitzgerald M, Collins C, Crowley D, Daly L, Devine M, et al. Acute gastroenteritis in northern Ireland and the Republic of Ireland: a telephone survey. Commun Dis Public Health 2004;7(1):61-7.

17. Majowicz SE, Dore K, Flint JA, Edge VL, Read S, Buffett MC, et al. Magnitude and distribution of acute, self-reported gastrointestinal illness in a Canadian community. Epidemiol Infect 2004; 132(4):607-17.

18. McGuire T, Wells KB, Bruce ML, Miranda J, Scheffler R, Durham M, et al. Burden of illness. Ment Health Serv Res 2002;4(4): 179-85.

19. Hall GV, Kirk MD, Ashbolt R, Stafford R, Lalor K. Frequency of infectious gastrointestinal illness in Australia, 2002: regional, seasonal and demographic variation. Epidemiol Infect 2006;134(1 ):111 -8.

20. Herikstad H, Yang S, Van Gilder TJ, Vugia D, Hadler J, Blake P, et al. A population-based estimate of the burden of diarrhoeal illness in the United States: FoodNet, 1996-7. Epidemiol Infect 2002;129(1):9-17.

21. Hoogenboom-Verdegaal AM, de Jong JC, During M, Hoogenveen R, Hoekstra JA. Community-based study of the incidence of gastrointestinal diseases in The Netherlands. Epidemiol Infect 1994;112(3):481-7.

22. de Wit MA, Koopmans MP, Kortbeek LM, Wannet WJ, Vinje J, van Leusden F, et al. Sensor, a population-based cohort study on gastroenteritis in the Netherlands: incidence and etiology. Am J Epidemiol 2001 ;154(7):666-74.

23. Imhoff B, Morse D, Shiferaw B, Hawkins M, Vugia D, Lance-Parker S, et al. Burden of self-reported acute diarrheal illness in FoodNet surveillance areas, 1998-1999. Clin Infect Dis 2004;38 Suppl 3:S219-26.

42

24. Kuusi M, Aavitsland P, Gondrosen B, Kapperud G. Incidence of gastroenteritis in Norway-a population-based survey. Epidemiol Infect 2003;131(1):591-7.

25. Sargeant JM, Majowicz SE, Snelgrove J. The burden of acute gastrointestinal illness in Ontario, Canada, 2005-2006. Epidemiol Infect 2007:1-10.

26. Thomas MK, Majowicz SE, MacDougall L, Sockett PN, Kovacs SJ, Fyfe M, et al. Population distribution and burden of acute gastrointestinal illness in British Columbia, Canada. BMC Public Health 2006;6:307.

27. Wheeler JG, Sethi D, Cowden JM, Wall PG, Rodrigues LC, Tompkins DS, et al. Study of infectious intestinal disease in England: rates in the community, presenting to general practice, and reported to national surveillance. The Infectious Intestinal Disease Study Executive. BMJ 1999;318(7190):1046-50.

28. Aguiar Prieto P, Finley RL, Muchaal PK, Guerin MT, Isaacs S, Castro Dominguez A, et al. Burden of self-reported acute gastrointestinal illness in Cuba. J Health Popul Nutr 2009;27:345-57.

29. Gargouri Darwaza N. Jordan burden of foodborne disease sentinel study, 2003 Technical Report.

30. Gargouri N, Walke H, Belbeisi A, Hadadin A, Salah S, Ellis A, et al. Estimated burden of human Salmonella, Shigella, and Brucella infections in Jordan, 2003-2004. Foodborne Pathog Dis 2009;6(4):481-6.

31. Voetsch AC, Van Gilder TJ, Angulo FJ, Farley MM, Shallow S, Marcus R, et al. FoodNet estimate of the burden of illness caused by nontyphoidal Salmonella infections in the United States. Clin Infect Dis 2004;38 Suppl 3:S127-34.

32. Majowicz SE, Edge VL, Fazil A, McNab WB, Dore KA, Sockett PN, et al. Estimating the under-reporting rate for infectious gastrointestinal illness in Ontario. Can J Public Health 2005;96(3):178-81.

33. Thomas MK, Majowicz SE, Sockett PN, Fazil A, Pollari F, Dore KA, et al. Estimated numbers of community cases of illness due to Salmonella, Campylobacter and verotoxigenic Escherichia coli: Pathogen-specific community rates. Can J Infect Dis Med Microbiol 2006; 17(4):229-34.

34. Batz MB, Doyle MP, Morris G,Jr, Painter J, Singh R, Tauxe RV, et al. Attributing illness to food. Emerg Infect Dis 2005;11(7):993-9.

35. Mead PS, Slutsker L, Dietz V, McCaig LF, Bresee JS, Shapiro C, et al. Food-related illness and death in the United States. Emerg Infect Dis 1999;5(5):607-25.

43

36. Nesbitt A, Majowicz S, Finley R, Pollari F, Pintar K, Marshall B, et al. Food consumption patterns in the Waterloo Region, Ontario, Canada: a cross-sectional telephone survey. BMC Public Health 2008;8:370.

37. Roy SL, Scallan E, Beach MJ. The rate of acute gastrointestinal illness in developed countries. J Water Health 2006;4 Suppl 2:31-69.

38. Klontz KC, Timbo BB, Fein SB, Levy AS. Prevalence of selected food consumption and preparation behaviors associated with increased risks of food-borne disease. J Food ProM995;58(8):927-930.

39. Yang S, Leff MG, McTague D, Horvath KA, Jackson-Thompson J, Murayi T, et al. Multistate surveillance for food-handling, preparation, and consumption behaviors associated with foodborne diseases: 1995 and 1996 BRFSS food-safety questions. MMWR CDC Surveill Summ 1998;47(4):33-57.

40. Samuel MC, Vugia D, Koehler KM, Marcus R, Deneen V, Damaske B, et al. Consumption of risky foods among adults at high risk for severe foodborne diseases: Room for improved targeted prevention messages. J Food Safety 2007;27:219-232.

41. Lee S, Levy D, Hightower A, Imhoff B, and the EIP FoodNet Working Group. Drinking water exposures and perceptions among 1998-1999 FoodNet survey respondents. International Conference on Emerging Infectious Diseases. Atlanta, GA, March 2002. Available at: http://www.cdc.gov/iceid/asm-iceid proqram.pdf. Accessed February 20, 2010.

42. Jones AQ, Dewey CE, Dore K, Majowicz SE, McEwen SA, Waltner-Toews D. Drinking water consumption patterns of residents in a Canadian community. J Water Health 2006;4(1): 125-38.

43. Jones AQ, Majowicz SE, Edge VL, Thomas MK, MacDougall L, Fyfe M, et al. Drinking water consumption patterns in British Columbia: an investigation of associations with demographic factors and acute gastrointestinal illness. Sci Total Environ 2007;388(1-3):54-65.

44. Westrell T, Andersson Y, Stenstrom TA. Drinking water consumption patterns in Sweden. J Water Health 2006;4(4):511-22.

45. Pintar KD, Waltner-Toews D, Charron D, Pollari F, Fazil A, McEwen SA, et al. Water consumption habits of a south-western Ontario community. J Water Health 2009;7(2):276-92.

46. Cowden JM. Foodborne infectious risks: do we need a wide system of data collection and survey? The lessons learned from the study of infectious intestinal disease in England. Rev Epidemiol Sante Publique 2002;50(1):89-92.

44

http://www.cdc.gov/iceid/asm-iceid

47. Thomas MK, Majowicz SE, Sockett PN, Fazil A, Pollari F, Dore K, et al. Estimated Numbers of Community Cases of Illness Due to Salmonella, Campylobacter and Verotoxigenic Escherichia Coli: Pathogen-specific Community Rates. Can J Infect Dis Med Microbiol 2006; 17(4):229-34.

48. Hall G, Yohannes K, Raupach J, Becker N, Kirk M. Estimating community incidence of Salmonella, Campylobacter, and Shiga toxin-producing Escherichia coli infections, Australia. Emerg Infect Dis 2008; 14(10): 1601 -9.

49. Adak GK, Long SM, O'Brien SJ. Trends in indigenous foodbome disease and deaths, England and Wales: 1992 to 2000. Gut 2002;51(6):832-41.

50. Adak GK, Meakins SM, Yip H, Lopman BA, O'Brien SJ. Disease risks from foods, England and Wales, 1996-2000. Emerg Infect Dis 2005;11(3):365-72.

51. Dohoo I, Martin W, Stryhn H. Veterinary Epidemiologic Research. Charlottetown, PEI, Canada: AVC Inc.; 2003.

52. World Health Organization. WHO consultation to develop a strategy to estimate the global burden of foodbome diseases: Taking stock and charting the way forward. 2007.

53. Majowicz SE, Hall G, Scallan E, Adak GK, Gauci C, Jones TF, et al. A common, symptom-based case definition for gastroenteritis. Epidemiol Infect 2008;136(7):889-94.

54. Baqui AH, Black RE, Yunus M, Hoque AR, Chowdhury HR, Sack RB. Methodological issues in diarrhoeal diseases epidemiology: definition of diarrhoeal episodes. Int J Epidemiol 1991 ;20(4): 1057-63.

55. Wright JA, Gundry SW, Conroy R, Wood D, Du Preez M, Ferro-Luzzi A, et al. Defining episodes of diarrhoea: results from a three-country study in Sub-Saharan Africa. J Health Popul Nutr 2006;24(1 ):8-16.

56. Wright J, Gundry SW, Conroy RM. A review of changing episode definitions and their effects on estimates of diarrhoeal morbidity. J Health Popul Nutr 2007;25(4):448-55.

45

Tabl

e 1.

1: S

umm

ary

of p

ublis

hed

popu

latio

n-ba

sed

burd

en o

f ac

ute

gast

roin

test

inal

illn

ess

(Gl)

stud

ies

1991

-2

00

6,

by

regi

on,

in c

hron

olog

ical

ord

er.

Stu

dy /

Loca

tion

D

ate

Stu

dy

des

ign

C

ase

def

init

ion

* S

amp

le

size

R

espo

nse

rate

In

cid

ence

(p

er

pers

on-

year

)

Gl R

isk

grou

ps a

nd

Key

res

ults

Eur

ope

The

19

91

Pro

spec

tive

(1)

D o

r V

plu

s tw

o or

mor

e N

ethe

rland

s co

hort

of

: N

, A,

C,

B o

r M

with

in o

ne

(21)

w

eek

(2)

sam

e as

1 b

ut o

n th

e sa

me

day

and

last

ing

at le

ast

two

days

with

in o

ne w

eek

2,25

7 36

%

0.63

(1)

0.18

(2)

• W

omen

•

Chi

ldre

n 0-

18 y

ears

I ID

- 19

93 to

P

rosp

ectiv

e A

ny D

or

V m

ore

than

onc

e in

E

ngla

nd

1996

co

hort

24

hou

rs,

inca

paci

tatin

g or

(2

7)

acco

mpa

nied

by

C o

r F

la

stin

g le

ss th

an tw

o w

eeks

9,77

6 40

%

0.19

4 •

1 in

5

peop

le in

ge

nera

l po

pula

tion

deve

lop

IID e

ach

year

S

en

sor-

19

98 to

th

e 19

99

Net

herla

nds

(22)

Pro

spec

tive

coho

rt

Thr

ee o

r m

ore

loos

e st

ools

or

vom

its in

24

hour

s or

D w

ith

two

or m

ore

of:

V, A

, C

, F,

N,

B,

M, w

ithin

24

hour

s

4,86

0 4

2%

0.

283

• C

hild

ren

0-11

yea

rs

• In

crea

sed

educ

atio

n •

Janu

ary

and

June

N

orw

ay

1999

to

Cro

ss-

Thr

ee o

r m

ore

loos

e st

ools

in

(24)

20

00

sect

iona

l 24

hou

rs o

r at

leas

t thr

ee o

f: su

rvey

(by

V

, N

, C,

F(>

38

°C)

with

in f

our

1,84

3 6

1%

1.

2 •

Wom

en

• A

dults

25-

39 y

ears

46

mai

l) w

eeks

Irela

nd (

16)

2000

to

2001

C

ross

-se

ctio

nal

surv

ey (

by

tele

phon

e)

D w

ith th

ree

or m

ore

loos

e st

ools

in 2

4 ho

urs

or B

or

V

with

at

leas

t one

of:

D, A

, C

, F

with

in fo

ur w

eeks

prio

r to

in

terv

iew

9,90

3 65

%

0.6

21-3

9 ye

ar

olds

tha

t tr

avel

ed

outs

ide

of

Nor

way

0-

14 y

ear

olds

who

co

nsum

e w

ater

fro

m

priv

ate

supp

ly

Wom

en

Chi

ldre

n 0-

4 ye

ars

Pro

fess

ion

al/n

on-

man

ual

wor

ker

Nor

th A

mer

ica

• C

hild

ren

0-4

year

s •

Incr

ease

d ed

ucat

ion

• U

rban

re

side

nts

• C

hild

ren

0-4

year

s •

Adu

lts 2

5-44

yea

rs

Foo

dNet

-

1996

to

Cro

ss-

D la

stin

g lo

nger

than

one

day

U

SA

(20

) 19

97

sect

iona

l or

D c

ausi

ng s

igni

fican

t su

rvey

(by

im

pairm

ent

of d

aily

act

iviti

es

tele

phon

e)

with

in fo

ur w

eeks

prio

r to

in

terv

iew

9,00

3 71

%

0.7

Fo

od

Ne

t-

1998

to

Cro

ss-

US

A (

23)

1999

se

ctio

nal

surv

ey (

by

tele

phon

e)

D w

ith th

ree

or m

ore

loos

e st

ools

in 2

4 ho

urs

with

in fo

ur

wee

ks p

rior

to i

nter

view

, w

here

D la

sted

mor

e th

an

one

day

12,7

55

41

%

0.7

47

Foo

d N

et -

US

A(1

2)

Fo

od

Ne

t-U

SA

(12

)

NS

AG

I-C

anad

a (1

7)

NS

AG

I -

Can

ada

(26)

NS

AG

I -

Can

ada

(25)

2000

to

2001

2002

to

2003

2001

to

2002

2002

to

2003

2005

to

2006

Cro

ss-

sect

iona

l su

rvey

(by

te

leph

one)

Cro

ss-

sect

iona

l su

rvey

(by

te

leph

one)

Cro

ss-

sect

iona

l su

rvey

(by

te

leph

one)

Cro

ss-

sect

iona

l su

rvey

(by

te

leph

one)

Cro

ss-

sect

iona

l su

rvey

(by

te

leph

one)

D w

ith th

ree

or m

ore

loos

e st

ools

in 2

4 ho

urs

with

in f

our

wee

ks p

rior

to i

nter

view

, w

here

D la

sted

mor

e th

an

one

day

D w

ith th

ree

or m

ore

loos

e st

ools

in 2

4 ho

urs

with

in fo

ur

wee

ks p

rior

to i

nter

view

, w

here

D la

sted

mor

e th

an

one

day

Any

D o

r V

with

in fo

ur w

eeks

pr

ior

to i

nter

view

Any

D o

r V

with

in fo

ur w

eeks

pr

ior

to i

nter

view

Any

D o

r V

with

in f

our

wee

ks

prio

r to

int

ervi

ew

14,1

39

15,5

78

3,49

6

4,61

2

2,09

0

NA

NA

37%

44%

37%

0.7

0.7

1.3

1.3

1.2

Au

stra

lia

OzF

oodN

et

- A

ustr

alia

20

01 to

20

02

Cro

ss-

sect

iona

l T

hree

loos

e st

ools

or

two

vom

its in

24

hour

s or

at

leas

t 6,

987

67%

0.

92

• C

hild

ren

0-4

year

s •

Adu

lts 1

8-35

yea

rs

» W

omen

•

Chi

ldre

n 0-

4 ye

ars

• A

dults

18-

35 y

ears

»

Wom

en

• C

hild

ren

0-9

year

s »

Adu

lts 2

0-24

yea

rs

» W

omen

»

Chi

ldre

n 0-

14 y

ears

•

Incr

ease

d ed

ucat

ion

• W

omen

•

Rur

al

resi

dent

s »

Feb

ruar

y an

d A

pril

» W

omen

»

Chi

ldre

n (1

9)

surv

ey (

by

four

loos

e st

ools

or

thre

e te

leph

one)

vo

mits

in 2

4 ho

urs

with

R,

0-4

year

s S

umm

er

48

with

in fo

ur w

eeks

prio

r to

in

terv

iew

Jo

rdan

Jo

rdan

(2

9,30

) 20

03 to

20

04

Cro

ss-

sect

iona

l su

rvey

(fa

ce-

to-f

ace)

Thr

ee o

r m

ore

loos

e st

ools

th

at to

ok th

e sh

ape

of a

co

ntai

ner

in 2

4 ho

urs

with

in

30 d

ays

prio

r to

int

ervi

ew

1,57

8 95

%

0.8

• 20

-40%

of

case

s so

ught

m

edic

al

care

La

tin A

mer

ica

Cub

a (2

8)

2005

to

2006

C

ross

-se

ctio

nal

surv

ey (

face

-to

-fac

e)

Thr

ee o

r m

ore

loos

e st

ools

in

24 h

ours

with

in 3

0 da

ys p

rior

to i

nter

view

6,39

9 97

%

1.4

Rai

ny

seas

on

Chi

ldre

n 0-

12 y

ears

T

eens

13-

17 y

ears

* D

=dia

rrhe

a, V

=vo

miti

ng,

N=n

ause

a, F

=fev

er,

A=a

bdom

inal

pai

n, C

=cra

mps

, B

=blo

od in

sto

ol, M

=muc

us in

sto

ol,

R=r

espi

rato

ry s

ympt

oms,

NA

=not

ava

ilabl

e

49

Tab

le 1

.2:

Sum

mar

y of

food

and

wat

er c

onsu

mpt

ion

data

from

pop

ulat

ion-

base

d st

udie

s, in

var

ious

cou

ntrie

s 19

92

2006

, in

chr

onol

ogic

al o

rder

.

Stu

dy

/ Lo

catio

n

Dat

e M

eth

od

olo

gy

Sam

ple

si

ze

Rec

all

peri

od

Key

res

ults

Food

Co

nsu

mp

tion

F

DA

-US

A

1992

to

Cro

ss-s

ectio

nal

(38)

19

93

tele

phon

e su

rvey

(a

dults

> 1

7 ye

ars)

1,26

0 N

A

Mor

e th

an 5

0% c

onsu

med

raw

egg

s, 2

3%

unde

rcoo

ked

ham

burg

ers,

17%

raw

she

llfis

h M

ales

and

18-

39 y

ear

olds

mor

e lik

ely

to e

at r

aw

anim

al p

rote

in a

nd r

euse

dirt

y cu

tting

boa

rds

Tho

se w

ith m

ore

than

hig

h sc

hool

edu

catio

n m

ore

likel

y to

eat

raw

ani

mal

pro

tein

B

RF

SS

- 19

95 to

C

ross

-sec

tiona

l U

SA

(39

) 19

96

tele

phon

e su

rvey

(a

dults

> 1

7 ye

ars)

19,3

56

12

mon

ths

50%

con

sum

ed u

nder

cook

ed e

ggs

Con

sum

ptio

n of

und

erco

oked

ham

burg

ers

com

mon

am

ong

mal

es, d

ecre

ased

with

in

crea

sing

age

, inc

reas

ed w

ith in

crea

sing

sal

ary

and

educ

atio

n N

early

20%

did

not

was

h ha

nds

or c

uttin

g bo

ard

with

soa

p af

ter

hand

ling

raw

mea

t or

poul

try

Foo

dNet

19

98 to

C

ross

-sec

tiona

l (4

0)

1999

te

leph

one

surv

ey

(adu

lts >

17

year

s)

12.7

55

7 da

ys

18%

con

sum

ed r

unny

egg

s in

wee

k pr

ior

to

inte

rvie

w

Mal

es m

ore

likel

y to

con

sum

e hi

gh r

isk

food

s th

an f

emal

es

Adu

lts >

65 w

ere

less

like

ly to

con

sum

e hi

gh r

isk

food

s C

-Ent

erN

et

- C

anad

a (3

6)

2005

to

Cro

ss-s

ectio

nal

2006

te

leph

one

surv

ey

2,33

2 7

days

•

Adu

lts >

65 y

ears

mor

e lik

ely

to c

onsu

med

egg

s in

clud

ing

unde

rcoo

ked

eggs

•

Con

sum

ptio

n of

chi

cken

nug

gets

, ha

mbu

rger

s an

d de

li m

eats

com

mon

am

ong

child

ren,

teen

s an

d yo

ung

adul

ts

50

• H

ome

is m

ost

com

mon

loca

tion

to e

at m

eals

W

ater

Co

nsu

mp

tion

F

oo

dN

et-

19

98 to

C

ross

-sec

tiona

l 12

,755

U

SA

(41

) 19

99

tele

phon

e su

rvey

NS

AG

I-

2001

to

Cro

ss-s

ectio

nal

1757

C

anad

a 20

02

tele

phon

e su

rvey

(4

2)

NS

AG

I-

20

02

- C

ross

-sec

tiona

l 46

12

Can

ada

2003

te

leph

one

surv

ey

(43)

Sw

eden

1

99

9-

(1).

Cro

ss-

(1).

10,9

57

(44)

20

03

sect

iona

l na

tiona

l (2

). 1

57

mai

led

surv

ey

(3).

75

(2).

Out

brea

k in

vest

igat

ion

(3).

Cro

ss-

sect

iona

l m

aile

d su

rvey

C

-Ent

erN

et

2005

to

Cro

ss-s

ectio

nal

2189

-

Can

ada

2006

te

leph

one

surv

ey

(45)

NA

•

Prim

ary

wat

er s

ourc

e id

entif

ied

as m

unic

ipal

w

ater

(63

.8%

), b

ottle

d w

ater

(17

.8%

) an

d pr

ivat

e w

ell w

ater

(15

.0%

) •

Dia

rrhe

al il

lnes

s no

t as

soci

ated

with

wat

er

expo

sure

var

iabl

es

24

• W

ater

con

sum

ptio

n de

crea

sed

with

inc

reas

ing

hour

s ag

e •

Wat

er c

onsu

mpt

ion

incr

ease

d w

ith i

ncre

asin

g ed

ucat

ion,

and

use

of

in-h

ome

wat

er t

reat

men

t de

vice

24

•

Med

ian

wat

er c

onsu

mpt

ion

of 4

ser

ving

s (1

.0L)

ho

urs

• W

ater

con

sum

ptio

n hi

gher

on

wee

kday

s, a

mon

g bo

ttled

wat

er u

sers

, am

ong

thos

e w

ith h

ighe

r ed

ucat

ion,

and

fem

ales

>25

yea

rs

• O

dds

of G

l inc

reas

ed w

ith in

crea

sed

wat

er

cons

umpt

ion

(1).

•

Ave

rage

col

d ta

p w

ater

con

sum

ptio

n 0.

86L

Dai

ly

• M

ore

wat

er c

onsu

med

by

wom

en t

han

men

and

(2

).

adul

ts £

70 y

ears

of

age

Dai

ly

(3).

W

eekl

y

24

• A

vera

ge ta

p w

ater

con

sum

ptio

n of

1.3

9L

hour

s •

Men

con

sum

ed l

ess

wat

er th

an w

omen

•

Incr

ease

wat

er c

onsu

mpt

ion

with

hig

her

educ

atio

n

51

• F

requ

ent

bottl

ed w

ater

use

rs c

onsu

med

less

w

ater

NA

=not

ava

ilabl

e

52

CHAPTER TWO

Burden of acute gastrointestinal illness in Galvez, Argentina, 2007

As Published: Journal of Health, Population and Nutrition, 2010; 28(2): 149-58.

Abstract

This study evaluated the magnitude and distribution of acute gastrointestinal

illness (Gl) in Galvez, Argentina, and assessed the outcome of a 7-day versus

30-day recall period in survey methodology. A cross-sectional population survey,

with either a 7-day or a 30-day retrospective recall period, was administered

door-to-door to randomly selected residents during the 'high' and 'low' season of

Gl in the community. Comparisons were made between the annual incidence

rate obtained using the 7-day and the 30-day recall periods. Using the 30-day

recall period, the mean annual incidence rate was 0.43 (low Gl season) and 0.49

(high Gl season) episodes per person-year. Using the 7-day recall period the

mean annual incidence rate was 0.76 (low Gl season) and 2.66 (high Gl season)

episodes per person-year. This study highlights the significant burden of Gl in a

South American community, confirms the importance of seasonality when

investigating Gl in the population and presents evidence suggesting that a longer

recall period may underestimate the burden of Gl in retrospective population

surveys of Gl.

53

Introduction

Acute gastrointestinal illness (Gl) causes significant morbidity, mortality,

and socio-economic burden worldwide (1,2). Clean water, sanitation, and food

safety are key components to preventing and controlling Gl in the population (3).

These public health areas are at the forefront of international public health

organizations' objectives and priorities as well as local public health workers

concerns (4-7). Understanding the magnitude, distribution and demographic

factors associated with Gl is key for its mitigation (8). However, Gl cases tend to

be under-reported by traditional surveillance techniques which require cases to

seek medical attention in order to be captured. To address this, numerous

countries have conducted population based studies to better estimate the

disease burden (8-19). With population-level baseline information, interventions,

targeted surveillance and research activities can be accurately evaluated.

Likewise, the impacts of broader worldwide trends such as globalization, climate

change and international travel and trade, on the magnitude and distribution of

disease can be gauged. Additionally, within population based study

methodology, discussions on prospective and retrospective methods, recall

period selection and recall bias are ongoing (18,20,21). Further research to

evaluate these issues within the burden of Gl context is needed.

In September 2006, Argentina's Ministry of Health completed their first

pilot burden of Gl study in Diamante, (Entre Rios province), which estimated a

monthly Gl prevalence of 8.2% (Dr. Oswaldo Rico, Argentina Ministry of Health,

personal communication 2006). Building from the pilot, we conducted a study in

54

Galvez (Santa Fe province) in 2007. The objectives of the Galvez study were to

determine the magnitude and distribution of Gl in the population, describe the

burden and clinical presentation of Gl, evaluate under-reporting and identify risk

factors associated with Gl. An additional objective was to assess the differences

between a 7-day and a 30-day recall period.

Materials and Methods

Population baseline study

A cross-sectional, door-to-door survey of randomly selected residents of

Galvez, Santa Fe, Argentina was administered 30 April 2007 - 21 May 2007

(phase 1: high Gl season) and 1 -12 October 2007 (phase 2: low Gl season).

Galvez and the pilot location Diamante, were conveniently selected by the

Argentine Ministry of Health based on their suitability, willingness of local and

regional authorities, feasibility of completing the studies, and availability of data

based on local and regional surveillance activities. Galvez has a population of

approximately 18 500 people, is primarily an urban area surrounded by farmland

and rural areas, and is divided into 15 neighborhoods3,4. 'High' and 'low' Gl

season designation was based on the data contained in the municipal

surveillance system housed at the Centro de Desarrollo de Agroalimentario

(CeDA) Galvez, Argentina. This surveillance system collects the monthly

number of cases of Gl in the community presenting at the local hospital and

clinics.

3 2001 Census data, Institute) Nacional de Estadistica y Censos, www.indec.mecon.qov.ar 42000 Ciudad de Galvez, www.unimedio.com/galvez

55

http://www.indec.mecon.qov.ar

http://www.unimedio.com/galvez

Face-to-face interviews were conducted by trained interviewers from the

community. Households were randomly selected proportionally by neighborhood

population from a community census using Epidat 3.1 (Pan American Health

Organization, 2006). The individual in the household with the next birthday was

selected to participate in the survey as is commonly done in population surveys

to achieve a random sample (10,14-17). If the selected individual declined or no

one lived at the residence, the neighboring house, that being the next closest

house, was selected conveniently by the surveyor, as the replacement. If the

selected individual was under the age of 12, the parent or guardian answered the

survey on their behalf. If the selected individual was between the ages of 12 and

18, the parent, guardian or child answered the survey at the discretion of the

parent or guardian. All surveys were administered in Spanish.

Sample size

Sample sizes were calculated using Epilnfo 3.0 (Centers for Disease

Control and Prevention, Atlanta, Georgia, 2000) with a 2% allowable error and a

95% confidence level in a population of 18 500. In phase 1 (high Gl season), the

target sample sizes of 681 respondents (30-day recall period) and 725

respondents (7-day recall period) were based on expected monthly (8%) and

weekly (2%) prevalences estimated from a prior Diamante, Argentina pilot study.

The prevalences estimated from phase 1 were used as expected prevalences in

phase 2 (low Gl season), yielding target sample sizes of 753 respondents for

both the 30- and 7-day recall periods. The total target sample size for the study

was 2 912.

56

Data gathering

The survey instrument (Appendix I and II) was developed by modifying the

survey tools used previously in Diamante, Argentina. Modifications to the

Diamante pilot survey included revisions to some questions to improve their

clarity and utility, while additional questions pertaining to potential risk factors and

recent antibiotic use were incorporated. Respondents were asked if they had

experienced any symptoms of diarrhea in the previous 7 or 30 days, depending

on the survey recall period, where diarrhea was defined as three or more loose

stools in 24 hours. Individuals who suffered chronic diarrhea or diarrhea caused

by use of medications, laxatives, alcohol or medical conditions, were considered

non-cases. Additional questions asked about socio-demographic factors,

secondary symptoms, number of days of missed school or work and whether

hospitalization was required.

Under-reporting estimation

From the population survey, the percentage of cases that visited the local

clinics and hospital were used to estimate the magnitude of under-reporting from

the community level to the CeDA managed municipal surveillance system, using

the model shown in the burden of illness pyramid (Figure 2.1).

Ethics

The study was approved by the Human Subjects Committee of the

University of Guelph Research Ethics Board (Guelph, Ontario, Canada) in

partnership with Argentina's Ministry of Health. Signed, informed consent was

57

obtained from all participants or the parent/guardian in the event the participant

was a minor.

Statistics

Data were manually entered into Epilnfo 3.0 and managed using Microsoft

Access. Analysis was performed using SAS 9.0 (SAS Institute Inc., Cary North

Carolina, 2004). Individuals responding 'don't know' or 'unsure' were excluded

from the analysis of that question. Whether cases had used antibiotics in the 4

weeks prior to illness was compared with whether non-cases had used antibiotics

in the 4 weeks prior to interview to assess the effect of recent antibiotic use.

Univariable analysis was performed on the overall dataset (both recall

periods and study phases). The null hypothesis of no association between

presence of Gl and individual potential risk factors was tested using the Fisher's

Exact test or the Monte Carlo estimation of the Fisher's Exact test in SAS. A

weighted multivariate logistic regression model was built manually beginning

with those variables that had a Fisher's Exact test p-value <0.25 on univariable

analysis (22). Weighting was used to correct for differences in neighbourhood

sampling fractions. All remaining variables were offered to the model, however

only variables with a p-value <0.05 (Wald's test) were kept in the final model.

Differences between medians were tested using the Median test in SAS.

The primary outcome measures of monthly and weekly prevalence were

defined as the number of respondents reporting Gl in the previous 30 or 7 days,

respectively, divided by the total number of respondents for the 30 or 7 day

58

surveys. Prevalence, incidence rate and incidence proportion calculations were

also performed (Modern Epidemiology; (23)); the formulas are shown in

Appendix III.

Using the burden of illness model shown in Figure 2.1, the under-reporting

estimate was generated via stochastic modeling in @RISK student version

(Palisade Corporation, Newfield New York) as an add-on to Microsoft Excel. The

Beta form (a, b) where a= number of cases that seek medical care +1 and b=

number of cases - number of cases that seek medical care + 1, was used to

estimate underreporting between the bottom and middle step of the pyramid (per

cent of cases who seek care) (24). The per cent of cases reported to the

municipal surveillance system was assumed to be 100%, therefore the inverse of

the per cent of cases who seek care was considered to be the estimated under

reporting fraction.

To facilitate international comparisons Majowicz et al. (25) proposed a

minimum set of reported results and a standard symptom-based case definition

for Gl of three or more loose stools or any vomiting in 24 hours excluding those

(a) with cancer of the bowel, irritable bowel syndrome, Crohn's disease,

ulcerative colitis, cystic fibrosis, celiac disease, or any other chronic illness with

symptoms of diarrhea or vomiting, or (b) who report their symptoms were due to

drugs, alcohol or pregnancy. Although our study definition did not capture

'vomiting only' cases, we still report the suggested minimum set of results, using

the study definition to facilitate international study comparisons.

59

Results

Magnitude, distribution and burden

The demographic distribution of Galvez residents versus survey

respondents are shown in Table 2.1 along with the prevalence, annual incidence

rate, annual incidence proportion, and prevalence by demographic

characteristics. The overall annual incidence rate varied between 0.46 and 1.68

episodes per person year, for the 30-day and 7-day recall periods, respectively.

Statistically significant higher annual incidence proportions were seen in phase 1

(high Gl season) compared to phase 2 (low Gl season) for the 7-day recall

period.

The proportion of the study population that was female or over 19 years of

age was larger than the target population of Galvez. The median age of cases

(46.5 years) and non-cases (46.6 years) for the full dataset was not statistically

different (p-value=0.92). The response rate of 61.1% for phase 2 was calculated

by dividing the number of completed surveys by the number of households

visited. Denominator data were not available for phase 1 and thus the response

rate was not calculated.

The type and frequency of secondary symptoms are shown in Table 2.2.

Headache and muscle pain were most often reported, followed by vomiting and

fever. Bloody diarrhea was only reported by cases in phase 1 (high Gl season),

30-day recall period.

The overall number of days of missed work and school of cases and of

caretakers is shown in Table 2.3. In phase 1 (high Gl season) a greater

60

proportion of cases missed work or school, and with a higher maximum number

of days missed, compared to phase 2 (low Gl season). However, in phase 2, a

larger proportion of cases had family members miss work or school to take care

of them.

Univariable and multivariable analysis

In the overall dataset (n=2915) study phase (p<0.05), age (p<0.05),

neighbourhood of residence (p<0.05), level of education (p=0.08), occupation

(p=0.25), number of people in the household (p=0.19), ownership of a rabbit

(p=0.08), ownership of a dog (p=0.22), and ownership of a chicken (p=0.14) were

associated at the preliminary univariable level (p<0.25) with being a case of Gl.

Variables not associated were sex (p=0.46), number of bedrooms in the

household (p=0.82), use of antibiotics in the 4 weeks prior to illness (p=1.00), a

cat (p=0.26), a bird (p=0.87), a cow (p=1.00), a horse (p=1.00), a sheep (p=1.00),

a goat (p=1.00), a turtle (p=1.00), a fish (p=1.00) and ownership of any pet

(p=0.32). A final multivariable model included significant predictor variables (p-

value <0.05) of study phase, age and neighbourhood of residence (Table 2.4).

Medical system use

Medications used by cases to treat symptoms, medical facilities visited by

cases and reasons for not seeking medical care are reported in Table 2.5.

Antidiarrheals and analgesics were used most frequently, followed by antibiotics

with and without prescription. Among those cases that sought medical care,

61

private clinics and the public hospital were most frequently visited. In total, two

cases required hospitalization for their illness, for a duration of 2 days and 8 days

respectively, both during phase 1. 'Self-medication' and 'not thinking the illness

was important enough to seek medical care' were the most common reasons for

not seeking medical attention.

Under-reporting estimation

The mean, minimum and maximum percentage of cases that sought

medical care are reported in Table 2.6. Assuming that all cases that sought

medical care are reported to the surveillance system, the average number of

cases of Gl in the community for each case in the surveillance system ranged

from 2.6 (minimum=1.5, maximum=7.4) to 4.3 (minimum=1.7, maximum=90.1)

depending on the study phase and recall period.

Standard case definition comparison

Table 2.7 reports the proposed minimum set of results of this study, thus

allowing for international comparisons. Using a subset of the proposed standard

case definition, no statistically significant differences were observed between the

incidence of Gl in males and females within a given recall period nor in the

percentage of cases with symptoms on the day of interview between study

phases and recall periods.

Discussion

62

This study provides the first population-based estimates of the magnitude,

distribution, and burden of Gl in an Argentinean community. Additionally, this

study provided an opportunity to evaluate the effect of retrospective recall period

(7-day versus 30-day recall) on estimates generated from a Gl survey.

In both phases of the study, the 7-day recall period yielded higher annual

estimates of Gl than the 30-day recall period. Assuming recall bias is minimized

the shorter the retrospective observation period, this is contrary to the suggestion

that 'telescoping' past illnesses into the observation period causes overestimates

of disease in the population when using retrospective methods, as suggested by

Wheeler et al. (18). These results may be evidence of a recall bias effect in the

opposite direction, such that the true burden of disease is actually under

estimated when a longer recall period is used. This may be due to forgetting

episodes of 'familiar illnesses' such as Gl or more easily remembering illnesses

that are perceived as severe (26). Further research into the mechanisms of this

potential bias is warranted.

We found that age, study phase and neighbourhood of residence were all

significantly associated with Gl. The odds of Gl were 2.14 times higher in the

'high' season (phase 1) compared to the 'low' season (phase 2). The odds of Gl

were greatest among the young (those under the age of 20 years) and the elderly

(those over the age of 59 years) when compared to the referent group (20 - 59

years of age) which is similar to other reported studies (9,12,14,16,17,19). Three

neighbourhoods were found to have significantly lower odds of Gl compared to

the referent neighbourhood. These three neighbourhoods are located on the

63

North-west, East, and South-east borders of the referent neighbourhood. Socio-

demographic information is not available at the neighbourhood level.

Municipal surveillance data for Galvez support the seasonal trend

observed in this study; during the same timeframe, surveillance data showed a

peak of Gl prevalence in the high season (phase 1) that was approximately three

times the prevalence seen in the low season (phase 2). A seasonal effect was

also observed in a Cuban study conducted in 2005-2006, where the prevalence

of Gl was approximately 2-5 fold times greater in the rainy season compared to

the dry season (10). Likewise, in Galvez, high Gl season coincided with more

rainfall and low Gl season coincided with less rainfall5. Interestingly, the

significantly higher odds associated with the 'high' season in this study was more

pronounced for the 7-day versus the 30-day recall period. This phenomenon

warrants more investigation

Gender was not significantly associated at the univariable level with Gl in

any recall periods or study phases. However, it was striking that in phase 1 (but

not phase 2), all cases <15 years of age were male (n=8, data not shown).

Similarly, results from a Cuban study (10) indicate that when controlling for

season, sentinel site and age group, there was a higher risk for males than for

females, supporting this potential relationship. Additionally, research in England

and Wales on demographic determinants of Campylobacter infections found an

increased risk among males between birth and 17 years of age (27). The

5 2005-2008 Meteorological data, Oliveros weather station, Santa Fe, Argentina, Instituto Nacional de Tecnologia Agropecuaria, www.inta.qov.ar

64

http://www.inta.qov.ar

potential higher risk of Gl of young males in the high season should be pursued

in further research into behavioural and other risk factors.

Our results indicate that there are more cases in the community than are

captured by local Gl surveillance systems, demonstrating that the true burden of

Gl is larger than typically detected by surveillance. Similar under-reporting has

been found by several other studies in developed countries (9,12,14,15,17-

19,28). We assumed that all cases that sought medical care were captured by

the municipal surveillance system but could not verify this. Any human error in

reporting of cases or misclassification of cases at the hospital or clinic level

would contribute to further under-estimation of the true burden.

The strict case definition used here was selected to be consistent with the

previous pilot study conducted in Argentina, and was specifically chosen to

reduce potential misclassifications of cases of non-infectious causes of Gl

symptoms (e.g., alcohol consumption). However, some infectious Gl cases with

vomiting as the sole symptom or less than 3 episodes of diarrhea in 24 hours

may have been excluded using this definition and if so, this would cause some

under-estimation of the true burden in the community.

Our findings are similar to those of others that have applied the proposed

symptom-based case definition (25), with the exception of the incidence

calculations for phase 1, 7-day recall period. However, our results are based on

two time periods selected to represent the 'high Gl season' and the 'low Gl

season' in the community and thus cannot be applied directly as full annual

estimates.

65

In phase 1 of the study, we observed more cases in the 7-day recall

period than in the 30-day recall period. This is surprising given that these two

survey recall periods occurred during the same calendar time period. Further

investigation of this is necessary, potentially examining multiple different recall

periods, study locations and times.

A potential limitation of this study was the retrospective methodology

used. Retrospective methods may be more subject to recall bias and thus under

ideal conditions, prospective methodology is preferred (18). This is somewhat

compensated by the advantage that we used similar methods to numerous other

retrospective studies, thereby enabling comparison with these studies.

Another limitation of this study may be selection bias, as the age and

gender distributions of study participants differed from those of the reference

community. Additionally, lack of denominator data for phase 1 prevented

calculation of the response rate. However, since the structure and management

of both study phases were identical, it is likely that there is not a large difference

between response rates of the two phases. Moreover, a response rate of 61 %

was achieved for phase 2 of the study, which is on the high-end of the range of

response rates from other published retrospective surveys (25). The door-to-

door methodology likely contributed to the relatively high response rate.

Provided that there are no differences between responders and non-responders

in terms of confounding characteristics and the risk of Gl, then non-response

should not impact our results. Additionally, to improve sample size in the

multivariable analysis, the outcome of being a case of Gl combined cases from

66

both the 7-day and the 30-day recall periods, thus interpretation of the model

odds ratios is not as straight forward.

Institutions and hospitals were not included as part of the study

population. Thus it is possible that cases of Gl that resided in these locations

were missed and may cause an under-estimation of the true burden.

This study builds on the pilot burden of Gl research conducted by the

Argentina Ministry of Health and is the first publication of this kind from

Argentina. It contributes to the growing understanding of Gl in the population

and highlights the significant burden of Gl in this Argentine community. It

presents evidence suggesting that a shorter recall period may be more valid for

retrospective population surveys of Gl. It demonstrates associations between Gl

and age, neighbourhood of residence and season. It provides the proposed

required results for international comparison using a subset of the proposed

standard case of Gl definition.

67

References

1. Guerrant RL, Kosek M, Moore S, Lorntz B, Brantley R, Lima AA. Magnitude and impact of diarrheal diseases. Arch Med Res 2002;33(4):351-5.

2. Kosek M, Bern C, Guerrant RL. The global burden of diarrhoeal disease, as estimated from studies published between 1992 and 2000. Bull World Health Organ 2003;81 (3): 197-204.

3. Kaferstein F. Foodborne diseases in developing countries: aetiology, epidemiology and strategies for prevention. Int J Environ Health Res 2003; 13 Suppl 1:S161-8.

4. Schlundt J. New directions in foodborne disease prevention. Int J Food Microbiol 2002;78(1 -2):3-17.

5. Herikstad H, Motarjemi Y, Tauxe RV. Salmonella surveillance: a global survey of public health serotyping. Epidemiol Infect 2002;129(1):1-8.

6. Guerrant RL, Kosek M, Lima AA, Lorntz B, Guyatt HL. Updating the DALYs for diarrhoeal disease. Trends Parasitol 2002; 18(5): 191 -3.

7. Flint JA, Van Duynhoven YT, Angulo FJ, DeLong SM, Braun P, Kirk M, et al. Estimating the burden of acute gastroenteritis, foodborne disease, and pathogens commonly transmitted by food: an international review. Clin Infect Dis 2005;41(5):698-704.

8. Jones TF, McMillian MB, Scallan E, Frenzen PD, Cronquist AB, Thomas S, et al. A population-based estimate of the substantial burden of diarrhoeal disease in the United States; FoodNet, 1996-2003. Epidemiol Infect 2007; 135(2):293-301.

9. de Wit MA, Koopmans MP, Kortbeek LM, Wannet WJ, Vinje J, van Leusden F, et al. Sensor, a population-based cohort study on gastroenteritis in the Netherlands: incidence and etiology. Am J Epidemiol 2001 ;154(7):666-74.

10. Aguiar Prieto P, Finley RL, Muchaal PK, Guerin MT, Isaacs S, Castro Dominguez A, et al. Burden of self-reported acute gastrointestinal illness in Cuba. J Health Popul Nutr 2009.

11. Hall GV, Kirk MD, Ashbolt R, Stafford R, Lalor K. Frequency of infectious gastrointestinal illness in Australia, 2002: regional, seasonal and demographic variation. Epidemiol Infect 2006;134(1 ):111 -8.

68

12. Imhoff B, Morse D, Shiferaw B, Hawkins M, Vugia D, Lance-Parker S, et al. Burden of self-reported acute diarrheal illness in FoodNet surveillance areas, 1998-1999. Clin Infect Dis 2004;38 Suppl 3:S219-26.

13. Kuusi M, Aavitsland P, Gondrosen B, Kapperud G. Incidence of gastroenteritis in Norway--a population-based survey. Epidemiol Infect 2003;131(1):591-7.

14. Majowicz SE, Dore K, Flint JA, Edge VL, Read S, Buffett MC, et al. Magnitude and distribution of acute, self-reported gastrointestinal illness in a Canadian community. Epidemiol Infect 2004; 132(4):607-17.

15. Scallan E, Fitzgerald M, Collins C, Crowley D, Daly L, Devine M, et al. Acute gastroenteritis in northern Ireland and the Republic of Ireland: a telephone survey. Commun Dis Public Health 2004;7(1):61-7.

16. Sargeant JM, Majowicz SE, Snelgrove J. The burden of acute gastrointestinal illness in Ontario, Canada, 2005-2006. Epidemiol Infect 2008, 136(4): 451-460.

17. Thomas MK, Majowicz SE, MacDougall L, Sockett PN, Kovacs SJ, Fyfe M, et al. Population distribution and burden of acute gastrointestinal illness in British Columbia, Canada. BMC Public Health 2006;6(307).

18. Wheeler JG, Sethi D, Cowden JM, Wall PG, Rodrigues LC, Tompkins DS, et al. Study of infectious intestinal disease in England: rates in the community, presenting to general practice, and reported to national surveillance. The Infectious Intestinal Disease Study Executive. BMJ 1999;318(7190):1046-50.

19. Herikstad H, Yang S, Van Gilder TJ, Vugia D, Hadler J, Blake P, et al. A population-based estimate of the burden of diarrhoeal illness in the United States: FoodNet, 1996-7. Epidemiol Infect 2002; 129(1 ):9-17.

20. Boland M, Sweeney MR, Scallan E, Harrington M, Staines A. Emerging advantages and drawbacks of telephone surveying in public health research in Ireland and the U.K. BMC Public Health 2006;6:208.

21. Rodrigues LC. Let us not forget telescoping as a major risk of telephone surveys. Comment on Boland et al. BMC Public Health 2006;6(208).

22. Mickey RM, Greenland S. The impact of confounder selection criteria on effect estimation. Am J Epidemiol 1989;129(1 ):125-37.

23. Rothman KJ, Greenland S. Modern Epidemiology 2nd Edition. 2nd ed. Philadelphia, Pennsylvania: Lippincott-Raven Publishers; 1998.

69

24. Vose D. Risk Analysis: A quantitative guide. 2nd ed. Chichester: John Wiley & Sons ltd.; 2000.

25. Majowicz SE, Hall G, Scallan E, Adak GK, Gauci C, Jones TF, et al. A common, symptom-based case definition for gastroenteritis. Epidemiol Infect 2008;136(7):889-94.

26. Roy SL, Scallan E, Beach MJ. The rate of acute gastrointestinal illness in developed countries. J Water Health 2006;4 Suppl 2:31-69.

27. Gillespie IA, O'Brien SJ, Penman C, Tompkins D, Cowden J, Humphrey TJ. Demographic determinants for Campylobacter infection in England and Wales: implications for future epidemiological studies. Epidemiol Infect 2008;136(12):1717-25.

28. Hall G, Kirk MD, Becker N, Gregory JE, Unicomb L, Millard G, et al. Estimating foodborne gastroenteritis, Australia. Emerg Infect Dis 2005;11(8):1257-64.

70

Tab

le 2

.1:

Res

pond

ent

repr

esen

tativ

enes

s, d

emog

raph

ic d

istr

ibut

ion

and

the

prev

alen

ce o

f ac

ute

gast

roin

test

inal

illn

ess

per

stud

y ph

ase

and

reca

ll pe

riod

in G

alve

z A

rgen

tina,

200

7.

Ph

asel

: 30

-Day

P

hase

1:

7-D

ay

Pha

se 2

: 30

-Day

P

hase

2:

7-D

ay

Gal

vez

Su

rvey

P

reva

len

c S

urv

ey

Pre

vale

nc

Su

rvey

P

reva

len

c S

urv

ey

Pre

vale

nce

R

esid

ents

* re

spon

dent

s e

(n=2

7)

resp

on

den

ts

e (n

=36)

re

spon

dent

s e

(n=2

6)

resp

on

den

ts

(n=1

1)

(n=1

8542

) (n

=680

) (n

=724

) (n

=755

) (n

=756

) D

emo

gra

ph

ic v

aria

ble

s (%

) G

ende

r Mal

e F

emal

e A

ge

0-4

5-9

10-1

9 20

-59

60+

E

duca

tion

<6

yea

rs

old

Illite

rate

P

rimar

y S

econ

dary

T

ertia

ry

Uni

vers

ity

48.3

51

.7

7.2

7.7

16.9

49

.1

19.1

NA

f

NA

N

A

NA

N

A

NA

38.2

61

.8

1.7

1.1

5.1

69.0

23

.2

3.9

1.1

43.4

37

.2

9.1

3.2

Num

ber

of p

eopl

e in

hou

seho

ld

1-4 5+

M

agn

itud

e P

reva

lenc

e *(

95

%C

I)

Inci

denc

e ra

te1

(95%

CI)

In

cide

nce

prop

ortio

n§

NA

N

A

NA

NA

NA

79.1

20

.9

3.97

% (

2.5

-

0.49

(0.

31 -

38.9

% (

26.5

5.4

4.7

14.3

6.

7 0.

0 3.

7 9.

0

14.3

12.5

6.

1 3.

5 1.

5 4.

4

5.0

4.7

-5.4

)

0.68

)

-49

.1)

40.7

59

.3

2.0

2.1

7.3

65.2

23

.5

2.0

1.5

43.2

46

.1

7.8

3.2

76.2

23

.8

4.97

% (

3.4

-

2.66

(1.

83 -

93.0

% (

83.5

3.9

4.1

9.1

28.6

14

.7

3.3

2.6

15.4

0.0

3.9

4.7

0.0

0.0

3.5

5.6

-6.6

)

3.58

)

-97

.2)

34.0

66

.0

2.1

1.7

10.5

52

.7

32.9

1.5

1.5

49.5

37

.2

7.3

2.0

76.0

24

.0

3.44

% (

2.3

-

0.43

(0.

28 -

34.7

% (

24.2

1.4

1.5

0.0

0.0

1.5

1.6

1.3

0.0

0.0

1.4

1.4

3.7

0.0

1.2

2.2

-5.0

)

0.63

)

-46

.7)

36.8

63

.2

1.1

1.7

9.0

58.6

29

.6

3.6

1.1

48.4

36

.6

9.3

2.5

80.5

19

.5

1.46

% (

0.7

-

0.76

(0.

35 -

53.4

% (

29.2

4.3

3.0

6.3

7.7

10.1

2.

3 2.

8

11.5

25.0

2.

8 3.

4 2.

9 0.

0

3.0

5.4

•2.6

)

1.40

)

- 75

.3)

71

(95%

CI)

* 20

01 C

ensu

s da

ta,

Inst

ituto

Nac

iona

l de

Est

adis

tica

y C

enso

s, w

ww

.inde

c.m

econ

.gov

.ar.

f N

A =

Not

ava

ilabl

e/ap

plic

able

. *

Wee

kly

or m

onth

ly p

reva

lenc

e ac

cord

ing

to 7

-day

or

30-d

ay r

ecal

l per

iod.

N

A =

Not

ava

ilabl

e/ap

plic

able

U

Ann

ual

inci

denc

e ra

te p

er p

erso

n ye

ar.

§ A

nnua

l in

cide

nce

prop

ortio

n.

72

http://www.indec.mecon.gov.ar

Table 2.2: Symptoms and their duration for both study phases and recall periods combined, Galvez, Argentina, 2007.

Secondary Number of cases reporting secondary symptoms symptom (n=100) Headache 23

Fever 10 Muscle Pain 23

Nausea 4 Vomiting 6 Cramps 4

Stomach Pain 1 Bloody diarrhea 4

Duration (days) Range 0.5 - 28

Median 3 Mean 3.4

73

Table 2.3: Days of missed work and school by cases and care-givers for both study phases and recall periods combined, Galvez, Argentina, 2007.

Variable Number of cases (n=100)

Number of cases that missed work due to illness 19 Median number and range of missed days 2 ( 1 - 8 ) Number of cases that missed school due to illness 10 Median number and range of missed days 2.5 (1 - 7) Number of cases with care-givers who missed work or school 7 Median number and range of care-giver missed days 1(1—3)

74

Table 2.4: Final multivariable model of risk factors associated with acute gastrointestinal illness in Galvez, Argentina, 2007 (n=2871).

Variable Study Phase

Phase 1 ('High') Phase 2 ('Low')

Age 0-4 5-9

10-19 20-59 60+

Neighbourhood* A B C D E F G H I J K L M N O

Frequency

1376 1504

49 48 233 1759 791

38 164 235 435 261 160 172 87 212 196 267 175 113 295 65

Odds ratio (95% CI)

2.14(1.41 -3.24) Referent

3.25(1.15-9.17) 2.87(0.89-9.30) 3.24(1.75-6.02) Referent 1.65(1.04-2.16)

0.53 (0.00 - 74.44) 0.94(0.32-2.80) 0.25(0.08-0.82) Referent 0.14(0.03-0.54) 0.59(0.19-1.81) 0.87(0.33-2.29) 0.63(0.09-4.58) 0.53(0.21 -1.37) 0.70(0.28-1.73) 0.35(0.14-0.85) 0.56(0.19-1.65) 0.17(0.01 -2.96) 1.24(0.72-2.14) 1.44(0.22-9.26)

P-value 0.0003

0.0009

0.0445

* Neighbourhoods have been given an identifying letter to maintain confidentiality.

75

Table 2.5: Medications and access to medical care, for both study phases and recall periods combined, Galvez, Argentina, 2007.

Variable Medications to treat symptoms

Analgesics

Antibiotics (with and without prescription) Antidiarrheals

Antiinflammatories Diet

Sought medical care: Yes No

Location of medical care sought* Private clinics Public clinics

Public hospital Unsure/did not respond

Reasons for not seeking medical care Self-medication

Natural remedies Didn't have time

Didn't think it was important Unsure/did not respond

Number of cases (n=100)

11

7 16 2 1

26 74

16 1 7 3

(n=74) 9 5 1

20 39

Some cases visited more than one location, so totals may exceed 100%.

76

Tab

le 2

.6:

Nu

mb

er

an

d m

ea

n,

min

imu

m,

an

d m

axi

mu

m p

erc

en

tag

e of

ca

ses

tha

t so

ug

ht

me

dic

al

att

en

tion

an

d th

e e

stim

ate

d u

nd

er-

rep

ort

ing

, fo

r bo

th s

tud

y p

ha

ses

an

d re

call

pe

rio

ds,

Ga

lve

z, A

rge

nti

na

, 2

00

7.

Pyr

amid

P

hase

1

Pha

se 2

st

ep

30-D

ay

7-D

ay

30-D

ay

7-D

ay

Num

ber

Mea

n %

N

umbe

r M

ean

%

Num

ber

Mea

n %

N

umbe

r M

ean

%

(min

.%,

(min

.%,

(min

.%,

(min

.%,

max

.%)

max

.%)

max

.%)

max

.%)

To

tal

680

724

755

756

surv

eyed

C

ases

27

36

26

11

V

isit

MD

10

38

.4

8 23

.6

6 24

.9

2 23

.1

(13.

6,65

.2)

(6.0

,51.

9)

(4.4

,52.

0)

(1.1

,60.

2)

Und

er-

2.6

4.2

4.0

4.3

rep

ort

ing

(1

.5-7

.4)

(1.9

-16

.7)

(1.9

-22

.7)

(1.7

-90

.1)

fact

or*

* E

stim

ate

d n

um

be

r of

ca

ses

in t

he

po

pu

latio

n pe

r ca

se r

ep

ort

ed

to m

un

icip

al

surv

eill

an

ce s

yste

m.

77

Table 2.7: Minimum set of results proposed for studies of acute gastrointestinal illness (25) for both study phases and recall periods, Galvez, Argentina, 2007.*

Categories of minimum set of results Annual incidence per person-year (95% CI)

Annual incidence per person-year in males Annual incidence per person-year in females Mean age of cases (years) Mean duration of illness (days) Cases with bloody diarrhea (%) Cases who saw a physician (%) Cases submitting a stool sample for testing (%) Cases with respiratory symptomsf (%) Cases with symptoms still ongoing at time of interview (%)

Phase 1 30-Day

0.49 (0.31 -0.68) 0.48

0.50

37

4.4

15

37

11

(...)

15

7-Day 2.66

(1.83-3.58) 2.92

2.50

52

2.4

0

22

6

(...)

14

Phase 2 30-Day

0.43 (0.28-0.63) 0.53

0.37

39

3.0

0

23

0

(...)

15

7-Day 0.76

(0.35-1.40) 0.76

0.77

46

5.9

0

18

0

(...)

18

* Study definition for case of Gl was anyone who had experienced 3 or more loose stools in 24 hours. t (...) = Data not collected. Survey respondents were not asked about respiratory symptoms.

78

Figure 2.1: Theoretical burden of illness pyramid for Galvez, Argentina, 2007

Number of cases reported to the

municipality

Number of cases that seek medical care

Number of cases in the population

79

CHAPTER THREE

Burden of acute gastrointestinal illness in the Metropolitan region, Chile,

2008

Accepted: Epidemiology and Infection

Summary

The purpose of this study was to determine the magnitude and distribution

of acute gastrointestinal illness (Gl) in the population, describe its burden and

presentation, identify risk factors associated with Gl and assess the differences

between a 7-day, 15-day and a 30-day recall period in the population-based

burden of illness study design. Face-to-face surveys were conducted on 6047

randomly selected residents in the Metropolitan region, Chile, with an average

response rate of 75.8%. The age-adjusted monthly prevalence of Gl was 9.2%.

The 7-day recall period provided annual incidence rate estimates approximately

2.2 times those of the 30-day recall period. Age, occupation, health care system,

sewer system, antibiotic use and cat ownership were all found to be significant

predictors for being a case of Gl. This study expands on the discussion of recall

bias in retrospective population studies and reports the first Chilean population-

based estimates of the burden and distribution of Gl.

Introduction

80

The World Health Organization (WHO) ranks diarrhoeal diseases fifth

among the world's top causes of mortality, responsible for 2.2 million deaths

worldwide [1]. In low-income countries, diarrhoeal diseases rank third, illustrating

the large burden of acute gastrointestinal illnesses (Gl) on the global population,

particularly in developing countries. Clean water, sanitation and food safety are

key components to preventing and controlling Gl in the population [2]. These

public health areas remain priorities for international public health organizations

and public health workers [3-6]. Accurately determining the burden of Gl is

important for its mitigation [7]. However, Gl cases tend to be under-reported by

traditional surveillance techniques, which require cases to seek medical attention

in order to be captured. To address this, numerous countries have conducted

population-based studies to better estimate the disease burden [7-18]. With

population-level baseline information, interventions, targeted surveillance and

research activities can be implemented and evaluated. However, there are still

unresolved issues in population-based burden of illness study methodology,

including recall period selection and recall bias [17, 19, 20].

In 2008, a partnership of the Pan-American Health Organization, the

Public Health Agency of Canada, the University of Guelph and the Ministry of

Health in Chile completed the first population burden of Gl study in the

Metropolitan region of Chile. The objectives of this study were to determine the

magnitude and distribution of Gl in the population, describe the burden and

clinical presentation of Gl and identify risk factors associated with Gl. An

81

additional objective was to assess the differences between a 7-day, 15-day and a

30-day recall period in the population-based burden of illness study design.

Methods


A cross-sectional, door-to-door survey of randomly selected residents of

the Metropolitan region of Chile was administered July 21 - August 25, 2008

(phase 1: low Gl season) and November 14 - December 21, 2008 (phase 2: high

Gl season). The Metropolitan region of Chile was selected as it is a diverse

region consisting of 6061185 residents which account for 40.1 % of the total

population of Chile. 'High' and 'low' Gl season designation was based on data

from the Ministry of Health surveillance system on reported cases of Gl and

outbreaks related to food and water.

The Metropolitan region is divided into 52 neighbourhoods which are

further divided into districts, zones and blocks. Neighbourhoods are classified by

the Instituto Nacional de Estadisticas6 into 5 categories according to socio

economic level. The number of surveys administered per socio-economic

category, was determined proportional to population size per category. Three

neighbourhoods were excluded from the sample due to concerns for surveyor

safety. Blocks were randomly selected proportional to the number of households

in each block. SAS 9.1 (SAS Institute Inc., Cary, North Carolina, 2004) was used

to conduct the proportional random selection. A convenience sample of

6 Instituto Nacional de Estadisticas, www.ine.cl 2009.

82

http://www.ine.cl

households was generated from within the randomly selected block at the

discretion of the surveyor in the field.

Face-to-face interviews were conducted by trained surveyors from the

region. The individual in the household with the next birthday was selected to

participate in the survey. If the selected individual declined or no one lived at the

residence, the neighboring house was selected as the replacement. If the





Sample size

Sample sizes were calculated using Epilnfo 3.4.1 (Centers for Disease

Control and Prevention, Atlanta, Georgia, 2007). Using an expected monthly

prevalence of 8%, a 1% allowable error and a 95% confidence interval, a target

sample size of 2826 was calculated, which was rounded to 3000 surveys per

phase for an overall total sample size of 6000 surveys.

Data gathering

The survey tool (Appendix IV) was developed by modifying the survey

tools used previously in Argentina (Appendix I and II, Chapter 2, [21]) and using

other similar cross-sectional population burden of Gl studies as models [7, 9,11,

13-16]. Respondents were asked if they had experienced any symptoms of

diarrhoea or vomiting in the previous 7, 15 and 30 days, where diarrhoea was

defined as three or more loose stools in 24 hours. Individuals who suffered

83

chronic diarrhoea or diarrhoea caused by use of medications, laxatives, alcohol

or medical conditions, were considered non-cases. Additional questions asked

about socio-demographic factors, secondary symptoms, number of days of

missed school or work, whether hospitalization was required and potential risk

factors.

Ethics


University of Guelph Research Ethics Board (Guelph, Ontario, Canada) and by

the Servicio de Salud Metropolitano Oriente scientific ethics committee of the

Government of Chile. Signed, informed consent was obtained from all

participants or the parent/guardian in the event the participant was a minor.

Statistics

Data were manually entered into Epilnfo 3.4.1 and managed using

Microsoft Access. Analysis was performed using SAS 9.1. Individuals

responding 'don't know' or 'unsure' were excluded from the analysis of that

question. Whether cases had used antibiotics in the four weeks prior to illness

was compared with whether non-cases had used antibiotics in the four weeks

prior to interview to assess impact of recent antibiotic use.

The primary outcome measures of weekly, 15-day and monthly

prevalence were defined as the number of respondents reporting Gl in the

previous 7, 15 or 30 days, respectively, divided by the total number of

respondents. Prevalence, incidence rate and incidence proportion calculations

84

(Modern Epidemiology; [22]) were performed for all three recall periods; the

formulas are shown in Appendix III.

Analysis was performed on the total dataset using monthly cases as the

outcome. The null hypothesis of no association between presence of Gl and

individual potential risk factors was tested using the Fisher's Exact test or the

Monte Carlo estimation of the Fisher's Exact test in SAS. A multivariate logistic

regression model was built manually using backwards elimination. Only

variables with a p-value <0.05 (Wald's test) were kept in the final model. All

variables that were initially screened out of the final model were re-introduced to

test for significance and visually assess confounding. Confounding was

determined by looking for a change of 30% or more in model coefficients. All

possible interactions between variables in the final model were assessed at a p-

value <0.05 (Wald's test). The Hosmer-Lemeshow test in SAS was used to

assess goodness of fit of the model, where a significant p-value (p-value <0.05)

indicates poor fit of the model. Differences between medians were tested using

the Median test in SAS.

Recently Majowicz et al. [23] proposed a standard symptom-based case

definition for Gl of three or more loose stools or any vomiting in 24 hours

excluding those (a) with cancer of the bowel, irritable bowel syndrome, Crohn's

disease, ulcerative colitis, cystic fibrosis, celiac disease, or any other chronic

illness with symptoms of diarrhoea or vomiting, or (b) who report their symptoms

were due to drugs, alcohol or pregnancy. We calculated the minimum set of

results, and report them here to facilitate future international comparisons.

85

Recall period comparison

There is no proper, accepted statistical test available to compare annual

estimates generated from different recall periods as we did here. Thus, we used

two approaches: A. compare 95% confidence intervals of the annual incidence

rate and proportion estimates generated from the three different recall periods, 7-

day, 15-day, and 30-day; and B., a simple binomial test as described below. To

compare thel 5-day and 30-day recall periods to the 7-day recall period, the

observed number of Gl cases of a recall period was tested by conducting a

simple binomial test using as the expected probability the prevalence of the 7-

day recall (fixed). P-values for the tests were computed by using the PROBBNML

function of SAS 9.1 (p-value < 0.05 indicates a significant result). The expected

probability of a case for the 15-day and 30-day recall periods was calculated

using the formula shown in Appendix V, assuming that the probability of being a

Gl case in the 15-day and 30-day recall was 2.14 and 4.29 times greater than the

probability in the 7-day recall respectively. By fixing the 7-day results as the

referent, we are ignoring the fact that the observed number of cases is a random

variable (even holding the number of surveys fixed) and that there is dependence

between the 7-day result and each of the 15-day and 30-day results.

Results

Burden and case description

In total, 6047 surveys were completed, 3033 in phase 1 and 3014 in

phase 2, with an overall average response rate of 75.8%. The demographic

86

distribution of residents of the Metropolitan region, along with survey respondents

and monthly cases is illustrated in Table 3.1. In general, survey respondents

were older, more educated and more likely to be female than residents. Overall,

cases were significantly younger than non-cases, with median ages of 33 and 39

respectively (p<0.001).

Of the 6047 respondents, in total 467 (7.7%; 7.1 - 8.4), 384 (6.4%; 5.8 -

7.0) and 262 (4.3%; 3.8 - 4.9) had symptoms of vomiting or diarrhoea in the 30,

15 and 7 days prior to interview, respectively (Table 3.2). The overall age-

adjusted monthly prevalence was 9.2%.

Symptoms and severity

The majority of monthly cases suffered symptoms of 'only vomiting'

followed by those with 'only diarrhoea' in phase 1 and overall; the reverse was

seen in phase 2, with the number of cases experiencing 'only diarrhoea' being

greater than the number of cases with 'only vomiting' (Table 3.2). Symptoms of

'both diarrhoea and vomiting' was highest in cases 0-4 years of age, while

symptoms of 'only diarrhoea' and 'only vomiting' peaked in cases 10-19 years of

age (Figure 3.1).

Of the 467 monthly cases, 110 (23.6%) had more than one episode of

diarrhoea and 78 (14.3%) had more than one episode of vomiting in the 30 days

prior to interview. Of the 292 cases that experienced diarrhoea, 11 (3.8%) had

blood in their diarrhoea. On the day of interview, 43 cases had diarrhoea and 20

cases had vomiting.

87

The most commonly experienced secondary symptoms were headache,

nausea and muscle pain (Table 3.3). As a result of their illness, 74 and 46 cases

missed work or school, respectively, and 25 cases required someone else to

miss work or school in order to provide care. Overall the median duration of

diarrhoea and vomiting were 2 days and 1 day, respectively. Cases reported the

maximum number of diarrhoeal events to be an average of 4.4 loose stools

(range 3 - 20) and a maximum number of vomiting events to be an average of

3.0 (range 1-20) in a 24 hour period.

Medical system use

Medications used by cases to treat symptoms, medical facilities visited by

cases and reasons for not seeking medical care are reported in Table 3.4.

Liquids, antidarrheals and analgesics were the top choices for medications used

by cases. Among those cases that sought medical care (n=99), public clinics,

private clinics and public hospitals were the most frequently visited. Only one

case required hospitalization for their illness, for a total of 24 hours. In total, 11

(11%) cases were asked to submit a sample of which 9 (82%) complied. Of the

nine samples submitted only two cases knew that their test results were negative

and the rest did not know the result. 'Self-medicating' and 'not thinking the

illness important' were the most common reasons for not seeking medical care.

Univariable and multivariable analysis

From univariable analysis of the full dataset (n=6047) age, socio-economic

level, occupation, education, ownership of a cat, ownership of a cow, ownership

of any pet, health system plan, sewer system and use of antibiotics in the four

88

weeks prior to illness/interview were all significantly associated with being a

monthly case of Gl p<0.05 (Table 3.5). A final multivariable model included

significant (p<0.05) predictor variables of age, health system, occupation, sewer

system, antibiotic use, and ownership of a cat (Table 3.6). Interaction between

'antibiotic use' and 'ownership of a cat', though significant in the final model, was

excluded due to low frequency of respondents included by this interaction. The

Hosmer-Lemeshow goodness of fit test p-value was 0.86 indicating the model fit

the data well.

Standard case definition comparison

A summary of this study's results using the suggested symptom-based

case definition are outlined in Table 3.7.

Recall period comparison

Significant differences in annual incidence rate and annual incidence

proportion estimates occurred between the 7-day, 15-day and 30-day recall

periods, within each study phase and within the combined overall estimates

(Appendix V). The annual incidence rate estimate from the 7-day recall was

approximately 2.2 times greater than the annual incidence rate estimate from the

30-day recall, for both study phases and overall (Table 3.2).

Discussion

This study provides population based estimates of Gl for the Metropolitan

region of Chile, illustrating that the incidence of Gl is comparable to international

estimates. An advantage of this study was using face-to-face methodology to

89

administer surveys, and a high average response rate of 75.8% was achieved

compared to other telephone administered surveys [12-16]. Another unique

contribution of this study, we evaluated the impact of a 7-day, 15-day and 30-day

recall period on estimates of the magnitude of Gl from a survey.

Overall the age-adjusted monthly prevalence of 9.2% is similar to

prevalences from studies in Canada and Cuba [7, 9, 13, 15, 16], and somewhat

higher than prevalences from studies in the United States, Australia and

Argentina (Chapter 2, [7, 10, 21]). The overall range of annual incidence rates

reported here of 0.98 - 2.3 episodes per person-year are higher than those

reported in Ireland, the United Kingdom and the Netherlands [14, 17, 24]. In part,

variations in case definition may explain differences in the magnitude of Gl

reported here compared with other studies.

A standard case definition has been proposed [23] to assist with

international comparisons and reports estimates from five countries: Australia,

Canada, Ireland, Malta and the United States. The incidence per person-year

overall, incidence per person-year in females and the percentage of cases that

sought medical care reported here are similar to those reported in Australia and

Canada. The mean age of cases, 36 years, is equivalent to that of Canada,

which is the highest of all the countries. Mean duration of illness, and

percentage of cases with respiratory symptoms reported here is lower than other

countries, additionally, percentage of cases submitting a stool sample reported

here is second lowest compared to the other countries. Percentage of cases

with symptoms at time of interview (12.85%) is similar to Canada and in the

90

middle of the range reported by other countries (8.22% -18.20%). Similarly, the

percentage of cases with bloody diarrhoea (2.36%) is in the middle of the range

reported by other countries (0.87% - 5.10%). The annual incidence per person-

year in males of 0.95 reported here was higher than all other rates reported

(range 0.31 -0.87).

The symptoms, severity and duration of illness were similar to other

studies, [7, 9, 11, 13-16, 18] though median duration and percentage of cases

with bloody diarrhoea were at the lower end of the spectrum of reported results.

Similarly, percentages of cases that sought medical care, had a stool sample

requested and that submitted a stool sample are at the low end of the range of

percentages reported from other studies [7, 11, 14-16, 18]. However, we had a

higher response rate, therefore it may not be a true difference or may be a bias in

other surveys where response rates are low. The lower percentage of individuals

seeking care means fewer cases would be captured in a surveillance system,

resulting in a larger under-estimation of the true burden of illness.

From multivariable modeling, age, health system, occupation, sewer

system, antibiotic use in the four weeks prior to illness/interview, and ownership

of a cat were significant risk factors to being a case of Gl. Children 0-4 years of

age and 10-19 years of age were 2.98 and 1.55 times more likely, respectively,

to be cases compared to the referent group of 20-59 year olds, which is similar to

other reported studies where children and youths had higher odds of being a

case of Gl [7-9, 13-16].

91

Residents who belonged to private-system health care had significantly

lower odds (OR=0.68) of being a case when compared to those belonging to the

public system, whereas those who did not belong to any health system had 1.54

times the odds of being a case compared to those belonging to the public health

system. This result is similar to studies in the United States, where those without

medical insurance reported higher rates of Gl than those with insurance and

where rural residents had an increased rate of Gl compared to urban residents

[7]. Likewise, those using a septic tank or outdoor latrine in place of the

municipal sewer system had 4.18 times the odds of being a case of Gl. Though

socio-economic level was not significant in the final model, these results may in

part reflect differences in socio-economic status; however, this needs to be

explored further.

Those who did not take antibiotics in the four weeks prior to illness or

interview had 1.64 times the odds of being a case as those that took antibiotics.

This result does not support the theory that antibiotic-associated diarrhoea is a

common side-effect of antibiotic use [25]. However, the occurrence of antibiotic-

associated diarrhoea is dependent on host factors as well as the type of antibiotic

and method in which it is taken which may in part explain the contrasting result

found here [26-28]. Information on type of antibiotic and method of

administration was not collected in this survey. Additionally, this information was

self-reported and not verified with medical records or prescriptions.

Of the other non-socio-demographic risk factors, only ownership of a cat

was significant, resulting in a 1.36 times higher odds of being a case of Gl. Cats

92

are known to carry a number of pathogens including Toxocara, Ancylostoma sp.,

Uncinaria, Dipylidium, Spirometra, Giardia, Toxoplasma, Cryptosporidium spp.

Campylobacter spp., Salmonella spp., and rabies [29, 30] which can be

transmitted to humans. In particular, contact with cats has been documented to

be associated with being a case of campylobacteriosis [31], and recent work has

looked at pet cats as a potential risk factor for enteric infection in the home [32,

33].

Though 'phase' was not significantly associated with being a case of Gl, it

is of interest that 'vomiting only' was more frequent among cases in phase 1, July

and August ('winter'), whereas 'diarrhoea only' was more frequent among cases

in phase 2, November and December ('summer'). This may indicate a seasonal

difference in Gl pathogens, where viral infections are associated with winter [34]

and bacterial and parasitic infections are associated with summer [35].

The shorter 7-day recall period yielded significantly greater annual

estimates compared to the 15-day and 30-day recall periods. This result is

similar to that reported for our population survey of an Argentine community

where the 7-day recall period yielded 1.7- 5.4 times the annual incidence rate

compared to the longer 30-day recall period (Chapter 2, [21]). However, this is

contrary to the suggestion that 'telescoping' past illnesses into the observation

period causes overestimation of disease in the population when using

retrospective methods, as suggested by Wheeler et al. [17]. These results

suggest an opposite effect of recall bias, such that the true burden of disease is

under-estimated when a longer recall period is used. Further investigation and

93

international comparisons are needed to explore the impact of different recall

periods.

The retrospective methodology of our study was a potential limitation.

Retrospective studies are more subject to recall bias and prospective

methodology is preferred [17]. The exploration of shorter recall periods was an

attempt to evaluate the impact of recall bias in our study and population-based

burden of illness studies in general. Methods similar to those used in previous

retrospective burden of illness studies were used, enabling comparison among

studies.

Selection bias due to differences in age and gender of respondents and

the referent community is another possible limitation of the study. Additionally,

institutions and hospitals were not included as part of the study population; it is

therefore possible that cases of Gl that resided in these locations were missed

and may cause an under-estimation of the true burden.

This study expands on the discussion of recall bias in retrospective

population burden of illness studies. It reports the first Chilean population-based

burden of Gl, the distribution of Gl estimates relative to various demographic

characteristics and is one of only a handful of these types of studies to have

been conducted in a developing country.

94

References

1. World Health Organization. The global burden of disease: 2004 update. 2008.

2. Kaferstein F. Foodborne diseases in developing countries: aetiology, epidemiology and strategies for prevention. Int J Environ Health Res 2003; 13 Suppl 1:S161-8.

3. Schlundt J. New directions in foodborne disease prevention. Int J Food Microbiol 2002; 78:3-17.

4. Herikstad H, Motarjemi Y, Tauxe RV. Salmonella surveillance: a global survey of public health serotyping. Epidemiol Infect 2002; 129: 1-8.

5. Guerrant RL, et al. Updating the DALYs for diarrhoeal disease. Trends Parasitol 2002; 18: 191-193.

6. Flint JA, et al. Estimating the burden of acute gastroenteritis, foodborne disease, and pathogens commonly transmitted by food: an international review. Clin Infect Dis 2005; 41: 698-704.

7. Jones TF, etal. A population-based estimate of the substantial burden of diarrhoeal disease in the United States; FoodNet, 1996-2003. Epidemiol Infect 2007; 135:293-301.

8. de Wit MA, et al. Sensor, a population-based cohort study on gastroenteritis in the Netherlands: incidence and etiology. Am J Epidemiol 2001; 154: 666-674.

9. Aguiar Prieto P, et al. Burden of self-reported acute gastrointestinal illness in Cuba. J Health Popul Nutr 2009.

10. Hall GV, et al. Frequency of infectious gastrointestinal illness in Australia, 2002: regional, seasonal and demographic variation. Epidemiol Infect 2006; 134: 111-118.

11. Imhoff B, et al. Burden of self-reported acute diarrheal illness in FoodNet surveillance areas, 1998-1999. Clin Infect Dis 2004; 38 Suppl 3: S219-26.

12. Kuusi M, etal. Incidence of gastroenteritis in Norway--a population-based survey. Epidemiol Infect 2003; 131: 591-597.

13. Majowicz SE, etal. Magnitude and distribution of acute, self-reported gastrointestinal illness in a Canadian community. Epidemiol Infect 2004; 132: 607-617.

95

14. Scallan E, etal. Acute gastroenteritis in northern Ireland and the Republic of Ireland: a telephone survey. Commun Dis Public Health 2004; 7: 61-67.

15. Sargeant JM, Majowicz SE, Snelgrove J. The burden of acute gastrointestinal illness in Ontario, Canada, 2005-2006. Epidemiol Infect 2008, 136(4): 451-460.

16. Thomas MK, etal. Population distribution and burden of acute gastrointestinal illness in British Columbia, Canada. BMC Public Health 2006; 6.

17. Wheeler JG, etal. Study of infectious intestinal disease in England: rates in the community, presenting to general practice, and reported to national surveillance. The Infectious Intestinal Disease Study Executive. BMJ 1999; 318: 1046-1050.

18. Herikstad H, etal. A population-based estimate of the burden of diarrhoeal illness in the United States: FoodNet, 1996-7. Epidemiol Infect 2002; 129: 9-17.

19. Boland M, etal. Emerging advantages and drawbacks of telephone surveying in public health research in Ireland and the U.K. BMC Public Health 2006; 6: 208.

20. Rodrigues LC. Let us not forget telescoping as a major risk of telephone surveys. Comment on Boland et al. BMC Public Health 2006; 6.

21. Thomas MK, etal. The burden and impact of acute gastrointestinal illness in Galvez, Argentina, 2007. J Health Popul Nutr{\n press, 2010).

22. Rothman KJ, Greenland S. Modern Epidemiology 2nd Edition, 2nd edn. Philadelphia, Pennsylvania: Lippincott-Raven Publishers, 1998.

23. Majowicz SE, etal. A common, symptom-based case definition for gastroenteritis. Epidemiol Infect 2008; 136: 889-94.

24. de Wit MA, et al. Gastroenteritis in sentinel general practices.The Netherlands. Emerg Infect Dis 2001; 7: 82-91.

25. Hogenauer C, etal. Mechanisms and management of antibiotic-associated diarrhea. Clin Infect Dis 1998; 27: 702-710.

26. Bartlett JG. Antibiotic-associated diarrhea. Clin Infect Dis 1992; 15: 573-581.

27. Bergogne-Berezin E. Treatment and prevention of antibiotic associated diarrhea. I'nt J Antimicrob Agents 2000; 16: 521-526.

28. Barbut F, Meynard JL. Managing antibiotic associated diarrhoea. BMJ 2002; 324: 1345-1346.

96

29. Robertson ID, Thompson RC. Enteric parasitic zoonoses of domesticated dogs and cats. Microbes Infect 2002; 4: 867-873.

30. Lappin MR. General concepts in zoonotic disease control. Vet Clin North Am Small Anim Pract 2005; 35: 1-20.

31. Deming MS, et al. Campylobacter enteritis at a university: transmission from eating chicken and from cats. Am J Epidemiol 1987; 126: 526-534.

32. Hill SL, et al. Prevalence of enteric zoonotic organisms in cats. J Am Vet Med Assoc 2000; 216: 687-692.

33. Spain CV, et al. Prevalence of enteric zoonotic agents in cats less than 1 year old in central New York State. J Vet Intern Med 2001; 15: 33-38.

34. Mounts AW, et al. Cold weather seasonality of gastroenteritis associated with Norwalk-like viruses. J Infect Dis 2000; 181 Suppl 2: S284-7.

35. Naumova EN, et al. Seasonality in six enterically transmitted diseases and ambient temperature. Epidemiol Infect 2007; 135: 281-292.

97

Table 3.1: Socio-demographic distribution of Metropolitan region residents, survey respondents and monthly prevalence of acute gastrointestinal illness by category, Chile 2008.

Variable

Age (years) 0-4 5-9

10-19 20-59

60+ C o y O C A

Male Female

Education Illiterate Primary

Secondary Technical University

Residents (n=6,061,185)

451,995(7.5%) 511,864(8.4%) 1,046,091 (17.3%) 3,381,732(55.8%) 669,543(11.0%)

2,937,193(48.5%) 3,123,992(51.5%)

NA 1,887,649(31.1%) 2,167,683(35.8%) 503,105(8.3%) 707,563(11.7%)

Respondents (n=6047)

96(1.6%) 107(1.8%) 545 (9.0%) 4361 (72.1%) 879(14.5%)

2594 (42.9%) 3451 (57.1%)

35 (0.6%) 978(16.2%) 2398 (39.7%) 1176(19.4%) 1247(20.6%)

Monthly prevalence

*20.8 (13.2-30.3) 8.4(3.9-15.4) *13.6(10.8-16.7) *6.9(6.1 -7.7) 6.6(5.1 -8.5)

7.5(6.5-8.6) 7.9(7.0-8.9)

5.7(0.7-19.2) 9.2(7.5-11.2) *7.8(6.8-9.0) * 7.2 (5.8-8.9) *6.7(5.3-8.2)

* Proportion per category significantly different to all other categories combined (p<0.05).

98

Tab

le 3

.2:

Num

ber

of c

ases

by

gast

roin

test

inal

illn

ess

sym

ptom

, pre

vale

nce,

ann

ual i

ncid

ence

rat

e an

d an

nual

inc

iden

ce

prop

ortio

n, b

y re

call

perio

d an

d ph

ase,

Met

ropo

litan

reg

ion,

Chi

le 2

008.

Pha

se 1

(n=3

033)

P

hase

2 (

n=3

014)

C

om

bin

ed (

n=6

047)

7

days

15

day

s 30

day

s 7

days

15

day

s 30

day

s 7

days

15

day

s 30

day

s D

iarr

hoea

on

ly

Vom

iting

on

ly

Bot

h D

iarr

hoea

an

d vo

miti

ng

Tot

al C

ases

O

vera

ll pr

eval

ence

(9

5% C

I)

Ann

ual

Inci

denc

e R

ate

(95%

CI)

A

nnua

l In

cide

nce

Pro

port

ion

(95%

CI)

46

60

11

117

3.9%

(3

.2-4

.6)

2.1

(1.8

-2.5

)

87.1

%

(82.

6 -

91.4

)

72

93

19

184

6.1%

(5

.2-

7.0)

1.5

(1.3

- 1.

8)

78.2

%

(72

.7-

82.9

)

93

118

26

237

7.8%

(6

.9-

8.8)

0.99

(0

.87-

1.1)

62.8

%

(58.

1 -

67.4

)

62

52

31

145

4.8%

(4

.1-5

.6)

2.6

(2.2

- 3.

0)

92.4

%

(88.

7 -

95.0

)

84

72

44

200

6.6%

(5

.8-

7.6)

1.7

(1.5

-1.9

)

81.2

%

(76

.6-

85.4

)

100

82

48

230

7.6%

(6

.7-

8.6)

0.97

(0

.84-

1.1)

61.9

%

(57

.0-

66.5

)

108

112

42

262

4.3%

(3

.8-

4.9)

2.3

(2.0

- 2.

6)

90.1

%

(86

.7-

92.7

)

156

165

63

384

6.4%

(5

.8-

7.0)

1.6

(1.5

-1.8

)

79.7

%

(76

.6-

82.9

)

193

200

74

467

7.7%

(7

.1-8

.4)

0.98

(0

.90-

1.1)

62.4

%

(59.

2 -

65.6

)

99

Table 3.3: Number and percent of cases (n=467) by secondary symptoms, duration of gastrointestinal symptoms and duration of missed activities due to gastrointestinal illness, Metropolitan region, Chile, 2008.

Symptom Number of cases (%) Nausea Headache Muscle pain Fever Cramps/sore stomach Sore throat Runny nose Cough Dizzy/vertigo Heartburn Lack of energy Chills Bloated With a cold Sweaty Chest pain Thirsty

185(40) 165(35) 86(18) 62(13) 48(10) 29(6) 28(6) 28(6) 5(1) 3(1) 2(<1) 2(<1) 1(<1) 1(<1) 1(<1) 1(<1)

K<D Duration of illness (days) Mean duration of diarrhoea Median duration of diarrhoea (range) Mean duration of vomiting Median duration of vomiting (range)

2.6 2(1-22) 1.6 1 (1-8)

Missed activities of cases (days) Mean duration of missed work Median duration of missed work (range) Mean duration of missed school Median duration of missed school (range)

1.7 1 (1-7) 2.6 2(1-21)

Missed activities of caregivers (days) Mean duration of missed work/school Median duration of missed work/school (range)

2.4 1 (1-14)

100

Table 3.4: Number and percent of cases (n=467) by treatments, use of medical care and reasons for not seeking medical care by Gl cases, Metropolitan region, Chile, 2008.

Variable Number of cases (%) Medications to treat symptoms:

Liquids 88(19) Antidiarrhoeals 76(16)

Analgesic 53(11) Antibiotic (with or without prescription) 25 (5)

Antispasmodics 33 (7) Antiemetic 16(3)

Antacid 6(1) Herbal 6(1)

Cold medication 4(1) Carbon 2(<1)

Sought medical care: Yes 99(21) No 368(79)

Medical facility visited (n=98): Private clinic 29(33) Public clinic 45 (46)

Hospital (institutional) 1(1) Private hospital 2 (2) Public hospital 21(21)

Reasons for not seeking medical are: Self medicated 160(43)

Illness not important enough to seek medical 95 (26) care

Natural remedies 47 (13)

101

Table 3.5: Univarable analysis results of association with acute gastrointestinal illness, Metropolitan region, Chile. 2008.

Variable Phase Age* Socio-economic level* Sex Occupation* (unemployed, housewife, student, retired, self-employed, private sector, public sector, general employer, not applicable i.e. child) Education* Ownership of dog Ownership of cat* Ownership of bird Ownership of cow* Ownership of sheep Ownership of horse Ownership of goat Ownership of chicken Ownership of rabbit Ownership of turtle Ownership of fish Ownership of hamster Ownership of reptiles Ownership of any pet* Health system* (military, private-system individual, no insurance, public) Number of people (categorized 1-4, 5-9 Number of bedrooms (categorized 0-2, House type (house, apartment, room in primitive cabin, shelter)

i, private-

,10+) 3-5, 6+) home,

Sewer system* (municipal, septic tank/latrine) Water source (municipal, well) Antibiotic use*

P-value 0.8097 <0.0001 0.0083 0.5593 <0.0001

0.0029 0.2271 0.0018 0.0507 0.0169 1.000 0.4304 0.1485 0.0797 0.7187 0.1210 1.000 0.7660 1.000 0.0065 0.0053

0.1149 0.2847 0.5604

0.0098 1.000 <0.0001

Indicates p-value <0.05.

102

Table 3.6: Final multivariable model of risk factors associated with acute gastrointestinal illness, Metropolitan region, Chile, 2008.

Variable Frequency Odds Ratio P-value Age

0-4 5-9

10-19 20-59

60+ Health System

Military Private - System

Private - Individual No insurance

Public Occupation

Not applicable (i.e. child) Unemployed

Self-employed Private sector Public sector

General employer Student Retired

Housewife Sewer system

Septic tank/Latrine Municipal

Antibiotic use Yes No

Ownership of cat Yes No

88 99 491 4174 845

126 1254 70 409 3834

71 245 771 1105 656 181 1036 441 1186

18 5679

790 4907

1214 4483

2.98(1.32-0.99(0.46-1.55(1.08-Referent 0.83(0.56-

1.02(0.52-0.68(0.53-0.94(0.37-1.54(1.10-Referent

1.11 (0.41 -1.30(0.79-0.74 (0.49 -1.06(0.75-1.39(0.95-1.53(0.88-1.42(0.99-1.44(0.87-Referent

4.18(1.42-Referent

0.61 (0.43-Referent

1.36(1.08-Referent

-6.69) -2.13) -2.22)

-1.22)

-1.97) -0.90) -2.37) -2.16)

-3.03) -2.15) -1.10) -1.50) -2.03) -2.68) -2.06) -2.36)

-12.25)

-0.86)

-1.71)

0.0117

0.0040

0.0486

0.0092

0.0045

0.0080

103

Table 3.7: Descriptive statistics of acute gastrointestinal illness based on 30-day recall period following the proposed standard case definition of gastrointestinal illness, Metropolitan region, Chile, 2008.

Annual incidence per person-year (95% CI) 0.98 (0.89 - 1.07) Annual incidence per person-year in males 0.95 Annual incidence per person-year in females 1.00 Mean age of cases (years) 36 Mean duration of illness (days) 2.09 Cases with bloody diarrhoea (%) 2.36 Cases who sought medical care (%) 21.20 Cases submitting a stool sample for testing 1.93 (%) Cases with respiratory symptoms (%) 14.13 Cases with symptoms still ongoing at time of 12.85 interview (%)

104

Figure 3.1: Monthly prevalence of Gl by symptoms and age group, Metropolitan region, Chile, 2008.

25% -

a, 2 0 % -u c 5 15% -2> a 1" 10% -•#-»

c o

S 5% -

0% -

• 0-4

• 5-9

• 10-19

Age

B Both Vomiting and Diarrhea

• Vomiting only

• Diarrhea Only

*

• 20-59

• 60+

105

CHAPTER FOUR

Risk factors for acute gastrointestinal illness related to food consumption

trends, food purchasing and hygiene habits among residents of the

Metropolitan Region, Chile, 2008

Prepared for submission: Foodborne Pathogens and Disease

Abstract

Background: This paper describes food consumption patterns, common food

purchasing habits, hygiene behaviours related to food safety and their

associations with acute gastrointestinal illness (Gl) in the Metropolitan region of

Chile.

Methods: A cross-sectional survey was administered door-to-door 21 July to 25

August and 14 November to 21 December 2008 in the Metropolitan region, Chile.

Respondents were randomly selected after socio-economic level stratification of

the region. Food consumption and food purchasing habits were assessed using

7-day recall, as well as general hand washing and food preparation habits.

Logistic regression was used to assess the association of Gl with food

consumption and hygiene behaviours.

Results: In total, 6047 residents participated with an average response rate of

75.8%. Supermarket and home were the most commonly identified locations of

food purchasing and meal eating, respectively. Respondents reported

consumption of an average of 15.2 meals per week. Most respondents

106

answered that they 'always' washed hands before eating and after using the

bathroom; used soap when washing hands; used a specific cutting board for

cutting meat and washed fruits and vegetables prior to eating. After controlling

for the effect of age, being a case of GI in the month prior to interview was

associated with consumption of undercooked poultry (OR=2.99), undercooked

beef (OR=2.18) and cheese made from unpasteurized milk (OR=2.09) in the 7

days prior to interview.

Conclusions: This study assessed food consumption and purchasing and

hygiene habits in the Metropolitan region, Chile from a food safety and Gl

perspective. These results can assist in developing Gl prevention and control

strategies.

Introduction

Access to 'safe food' is identified as a basic human right, which is not

always achieved by everyone, particularly in less developed countries, and this is

not fully appreciated by all public health authorities (Kaferstein, 2003, Kaferstein,

et al., 1999). Food consumption studies that assess nutritional intake with a

focus on malnutrition, obesity and chronic diseases have been conducted in

many countries. The safety of food, however, is a function of the presence or

absence of harmful concentrations of contaminants, including pathogens, which

may be especially important when food quantity is low (Kaferstein, 2003).

Recent research in Canada and the United States has focused on population

food consumption patterns in terms of food safety and risk of infectious diseases,

107

specifically acute gastrointestinal illnesses (Gl) (Nesbitt, etal., 2008, Altekruse,

etal., 1999, Yang, etal., 1998).

In Chile, surveys are done to monitor factors that affect quality of life,

including food intake (Ministerio de Salud Gobierno de Chile, 2006, Castillo, et

a/., 2002, Olivares, etal., 2004). However, associations between food

consumption and Gl are mainly derived from foodborne disease outbreak

reports, which represent only a fraction of all foodborne disease cases (Ministerio

de Salud (Chile), 2007, Prado, etal., 2002).

In 2008, the first population burden of Gl study in the Metropolitan region

of Chile was conducted (Chapter 3, (Thomas, etal., (Accepted 2010))). This

survey included specific questions related to food consumption and food

purchasing habits, as well as hygiene behaviours. The survey formed the basis

of this study, the objectives of which were to describe food consumption patterns,

common food purchasing habits and hygiene behaviours related to food safety,

and their associations with Gl.

Methods


Detailed methodology of the survey with results of basic burden of Gl is

reported elsewhere (Chapter 3, (Thomas, et ai, (Accepted 2010))). In brief, a

cross-sectional, door-to-door survey of randomly selected residents of the

Metropolitan region of Chile was administered July 21 - August 25, 2008 (Phase

1, low Gl season) and November 14 - December 21, 2008 (Phase 2, high Gl

season).

108

The 52 neighbourhoods of the region were classified by Instituto Nacional

de Estadisticas (INE)7 according to socio-economic level into five categories.

The number of surveys administered per socio-economic category was

proportional to population size. Blocks within neighbourhoods were randomly

selected proportional to the number of households within each block.

Households were selected from within the randomly selected block at the

convenience and discretion of the surveyor in the field, as within each block the

area was considered to be homogeneous.









Sample size



prevalence of Gl (defined below) of 8%, a 1% allowable error, a 95% confidence

interval and a population of 6,000,000, a target sample size of 2826 was

calculated, which was rounded to 3000 surveys per phase for an overall total

sample size of 6000 surveys.


109

http://www.ine.cl

Data gathering

The survey tool (Appendix IV) was developed by modifying that used

previously in Argentina (Appendix I and II, Chapter 2, (Thomas, etai, 2010)) and

using other similar cross-sectional population burden of Gl surveys (Aguiar

Prieto, etai, 2009, Majowicz, etai, 2004, Sargeant, etai, 2007, Thomas, etai,

2008) and food consumption surveys (Nesbitt, etai, 2008, Yang, etai, 1998,

Samuel, etai, 2007) as models. In addition to questions about Gl symptoms

and general socio-demographic factors, respondents were asked, with regard to

the seven days prior to interview, (a) if they had consumed specific foods which

had been identified by the Ministry of Health as 'high Gl risk', (b) where they had

purchased food (by product category), (c) the number of meals they had eaten

and where, as well as (d) general hand washing and food preparation habits.

Ethics





participants or their parent/guardian.

Statistics


Microsoft Access (Microsoft Corporation, 2003). Analysis was performed using

SAS 9.1 (SAS Institute Inc., Cary North Carolina, 2004). Individuals responding

'don't know' or 'unsure' were excluded from the analysis of that question.

110

The case definition of Gl in a respondent was defined as experiencing

symptoms of vomiting or diarrhea in the 30 days prior to interview, where

diarrhea was defined as three or more loose stools in 24 hours. Monthly

prevalence of Gl was defined as the number of cases of Gl in the 30 days prior to

interview, divided by the total number of survey respondents. The estimate using

the 30-day recall period (i.e., rather than using the 7-day or 15-day recall periods)

was selected to maximize sample size. Exact confidence intervals were

calculated at the 95% level.

Differences in distribution of socio-demographic factors between

respondents and residents of the Metropolitan region (as determined from 2002

Census, INE) were tested using PROBBNML test in SAS, where a p-value <0.05

was considered to be statistically significant.

Logistic regression was used to test the association of percentage of

meals consumed by location with Gl. Each location of meal consumption was

assessed individually; associations with a p-value <0.05 (Wald's test) were

considered to be significant.

Odds ratios (ORs) were used to estimate the strength of unconditional

association of Gl with food purchasing habits, and consumption of high risk food

items. Unconditional associations between Gl and hand washing and hygiene

behaviours were assessed using ORs where responses of 'sometimes' were

combined with responses of (a) 'always' and compared to responses of 'never'

and (b) 'never' and compared to responses of 'always'. Additionally, ORs were

calculated to assess the strengths of association of age and gender (adjusted by

111

each other), with consumption of each high risk food item. Age was categorized

for analysis and referent groups were females and respondents between the

ages of 20 and 59 years (this age category contained the largest number of

respondents, ensuring more stable estimates). The 95% confidence intervals

(CI) and p-values of the ORs were adjusted for multiple comparisons when

appropriate using an in-house fortran program.8 Evaluation of 95% CI of ORs

was used to identify statistically significant differences.

The Cochran-Armitage trend test in SAS was used to evaluate the linear

trend relationship between education level and (a) consumption of each food and

(b) hand washing and hygiene habits, and the relationship between age and

hand washing and hygiene habits. The Fisher's Exact test was used to evaluate

the relationship between gender and hand washing and hygiene habits.

Responses of 'always' and 'sometimes' were combined for questions on hand

washing and hygiene habits. A p-value <0.05 was considered to be significant

for the above comparisons.

Multivariate analysis was conducted using manual backwards elimination

logistic regression to assess the association of Gl with food consumption and

hygiene behaviours. All individual hygiene behaviours, food consumption and

socio-demographic (age, gender and education level) variables were evaluated

as potential predictor variables. Only variables with a p-value <0.05 (Wald's test)

or considered to be confounders a priori were kept in the final model. All

variables that were initially screened out of the final model (i.e., not considered to

8 DISTRIB, William C Sears, University of Guelph, 2009, based on the information in Hochberg, Y. and Tamhane, A. C, 1987: Multiple Comparison Procedures. Wiley, New York.

112

be confounders a priori or not significant) were re-introduced to test for

significance and visually assess confounding. Confounding was considered

present when a change of 30% or more, a change in sign or a change in

significance in model coefficients was observed. Multicollinearity was assessed

and considered present if two variables were individually significant, but became

non-significant when both were present in the model. Interaction among all

variables in the final model was assessed and considered present if the

interaction term had a p-value <0.05 (Wald's test). The Hosmer-Lemeshow test

in SAS was used to assess goodness of fit of the model, where a significant p-

value (p-value <0.05) indicated poor fit of the model.

Results


phase 2, with an overall average response rate of 75.8%. The demographic

distribution of survey respondents and residents of the Metropolitan region is

illustrated in Table 4.1. In general, survey respondents were older, more

educated and more likely to be female than residents.

Univariable (unconditional) significant associations

Food purchasing

'Supermarket' was the most common source reported for meat, fish and

dairy and egg purchases, while 'farmers market' was the most common source

for fruits and vegetable purchases in the week prior to interview (Table 4.2).

Purchasing dairy and eggs from a 'stationary or mobile market/kiosk' was

associated with increased odds of Gl in the month prior to interview, whereas

113

purchasing dairy and eggs from a 'supermarket' was associated with decreased

odds of Gl. Ninety-nine percent of respondents said that their responses

reflected where they habitually purchased their groceries.

Eating habits

Respondents reported consuming an average of 15.2 meals in the week

prior to interview (range 1-66). The largest average proportion of weekly meals

was consumed in the home (84.8%, range 0-100%) followed by a

cafeteria/casino (6.0%, range 0-100%), from the street (cart/kiosk; 3.2%, range

0-100%), in a sit-down restaurant (2.4%, range 0-100%), in a fast food restaurant

(2.4%, range 0-100%), and delivered (1.3%, range 0-100%).

Eating at home (p<0.01, Wald's test), in a cafeteria/casino (p=0.01) or

from the street (p=0.01) in the 7 days prior to interview were significantly

associated with Gl in the month prior to interview (Figure 4.1). As the proportion

of meals consumed in a cafeteria or in the street increased, so did the odds of

Gl, while as the proportion of meals consumed at home increased, the odds of Gl

decreased. Ninety-seven percent of respondents reported that their responses

reflected eating habits of a typical week.

Hand washing and hygiene habits

The majority of respondents responded 'always' to questions of washing

hands before eating, after using the bathroom, using soap when washing hands,

using a specific cutting board for cutting meat, and washing fruits and vegetables

prior to eating (Table 4.3). Increasing level of education was significantly

associated with 'always' or 'sometimes' washing hands prior to eating (p=0.012,

114

Cochran-Armitage trend test), using soap when washing hands prior to eating

(p<0.001), washing hands after using the bathroom (p<0.001) and using soap

when washing hands after using the bathroom (p<0.001). Males were more

likely than females to 'never' wash their hands after using the bathroom

(p=0.027, Fisher's Exact test), 'never' use a specific cutting board for meat

(p=0.005), and 'never' wash their fruits and vegetables prior to eating them

(p=0.022). Increasing age was significantly associated with 'always' or

'sometimes' performing all of the six hand washing and hygiene behaviours that

were questioned in the survey.

Compared to those who responded 'sometimes' and 'never' combined for

each question, the odds of Gl in the month prior to interview were significantly

lower in respondents who reported they 'always' wash their hands prior to eating,

use soap when washing hands prior to eating, use a specific cutting board for

cutting meat and wash fruits and vegetables prior to eating. Those who reported

they 'never' use soap when washing their hands after using the bathroom and

those who reported they 'never' wash fruits and vegetables prior to eating, had a

significantly higher odds of Gl compared to those who responded 'always' and

'sometimes' combined for each question.

High risk food consumption

In total, 71.5% (n=4323) of respondents reported eating at least one 'high

Gl risk' food (as defined by Ministry of Health) in the week prior to interview

(Table 4.4). The most commonly consumed 'high Gl risk' foods were commercial

mayonnaise, undercooked eggs and homemade mayonnaise. Gl was most

115

strongly associated with consumption of undercooked poultry (OR=2.99),

undercooked beef (OR=2.18) and cheese made from unpasteurized milk

(OR=2.09). Males were at significantly higher odds of consuming several 'high

Gl risk' food items in the week prior to interview compared to females (Table 4.5).

Children less than 5 years of age and adults 60 years and older were at

significantly lower odds of consuming any 'high Gl risk' food compared to

respondents 20 - 59 years of age, whereas respondents between the ages of 10

and 19 years had significantly greater odds of consuming any 'high Gl risk' food

compared to the referent age group (Table 4.6). Increasing level of education

was significantly associated with consumption of commercial mayonnaise

(p<0.001, Cochran-Armitage trend test), undercooked beef (p<0.001), raw fish

(p<0.001), oysters, mussels or molluscs (p<0.001), undercooked eggs (p<0.001),

undercooked chicken (p<0.001) and any 'high Gl risk' food (p<0.001). Eighty-

four percent of respondents said that their responses reflected what they ate in a

normal week.

High risk food consumption was significantly greater in phase 1 than in

phase 2 (p<0.001, Fisher's Exact test). Specifically, it was significantly more

common for respondents to report that they consumed raw eggs (p<0.001), raw

seafood (p<0.001), undercooked beef (p<0.001), undercooked hamburgers

(p<0.001), cheese made from unpasteurized milk (p<0.001) and commercial

mayonnaise (p<0.001) in July-August compared to November-December.

Multivariable results

116

In the final multivariable model, age and consumption of undercooked

beef, undercooked chicken and cheese made from unpasteurized milk in the 7

days prior to interview were all significantly (p<0.05, Wald's test) associated with

being a case of Gl in the 30 days prior to interview (Table 4.7). Those 0-4 and

10-19 years of age were at increased odds of Gl. None of the excluded

variables were confounders, multicollinearty was not present and no interactions

were significant in the final model. The Hosmer-Lemeshow goodness of fit test

p-value was 0.78, indicating the model fit the data.

Discussion

Previously published food consumption and dietary intake studies in Chile

typically had a nutritional focus rather than one of food safety concerns

(Ministerio de Salud Gobiemo de Chile, 2006, Castillo, etal., 2002, Olivares, et

al., 2004). This study assessed food consumption and purchasing as well as

general hygiene habits in the population of the Metropolitan region of Chile, from

a food safety and Gl perspective. The results deepen our understanding of

behavioural and food risk factors for Gl in Chile which could assist in developing

prevention and control measures. As well, these results can be used in food

safety-related risk assessments by contributing estimates of the prevalence of

specific food consumption and hand-washing and hygiene behaviours.

In the final multivariable logistic model, age and consumption of

undercooked chicken, beef and cheese made from unpasteurized milk in the 7

days prior to interview were all found to be significantly associated with being a

case of Gl in the 30 days prior to interview. Consumption of these food products

117

has been associated with numerous foodborne disease outbreaks in Chile

(Ministerio de Salud (Chile), 2007, Prado, et al., 2002). Evidence from foodborne

disease outbreaks and surveillance provided the basis for classification of 'high

Gl risk' foods in this study (Ministerio de Salud (Chile), 2007, Prado, et al., 2002).

The highest statistically significant odds of being a case of Gl in the 30

days prior to interview in this study were associated with consuming

undercooked poultry, undercooked beef, and cheese made from unpasteurized

milk in the seven days prior to interview. Due to the cross-sectional design of

this study, it is not possible to infer causal associations, as we are unsure

whether consumption of these foods occurred prior to becoming a case of Gl.

We note, however, that these are certainly biologically plausible associations that

are consistent with other studies (Oliver, et al., 2009, Smith, et al., 2008, Currie,

etal., 2005, Centers for Disease Control and Prevention (CDC), 2009, Rivas, et

al., 2008), and that 84% of respondents said that their responses reflected what

they ate in a normal week, indicating a potentially large, habitual exposure to

'high Gl risk' foods. In an effort to educate the public, the Ministry of Health

currently provides information on how to prevent Gl with proper hand-washing,

food preparation and general hygiene.9 Expansion of prevention and education

campaigns on the risks associated with 'high Gl risk' foods could be beneficial in

reducing Gl in the population.

Significant differences in 'high Gl risk' food consumption were observed by

age group; young children and the elderly were less likely to consume a 'high Gl

risk' food item. Respondents 10 to 19 years of age had the greatest number (2

9 Chile Ministry of Health, http://epi.minsal.cl/epi/html/enfer/veranosalud.html Accessed March 2010.

118

http://epi.minsal.cl/epi/html/enfer/veranosalud.html

of possible 16 food items were significant) of 'high Gl risk' food items with

significantly increased odds of consumption compared to the referent group.

Adolescents in this age group typically experience an increase in independence

(Christie, et al., 2005), which may be reflected in their food choices. A peak in

monthly Gl prevalence was observed in this age group in the burden of Gl study

in Chile (Chapter 3, (Thomas, et al., (Accepted 2010))), which might, in part, be

attributed to their food choices and perhaps also to unsafe food handling and

preparation associated with the phenomenon of 'second weaning', where

individuals in this age group are beginning to cook food for themselves and move

out of the family home.

The location of food purchases varied depending on the food item.

'Supermarkets' were identified most frequently for purchases of meat, fish, and

dairy and eggs, while 'farmers markets' were most frequently reported for

purchases of fruits and vegetables. For all food product categories, purchase at

'supermarkets' was associated with lower odds of Gl in the 30 days prior to

interview, though this was only statistically significant for dairy and egg

purchases. Food from supermarkets may be derived from different sources or

more consistently refrigerated than food from farmers markets, but further

research is needed to characterize the basis for these associations.

Furthermore, purchasing dairy and eggs from a 'stationary or mobile

market/kiosk' was associated with increased odds of Gl in the 30 days prior to

interview. This may also be due to cross-contamination, failure to refrigerate

perishable foods, acquisition from less safe sources, or alternatively, may be a

119

reflection of socio-economic level, however further investigation is needed. This

information could be of use for food safety risk assessments and for targeted

food safety education campaigns.

Respondents most commonly reported consuming meals in the home.

The higher percentage of meals consumed at home was unconditionally

associated with a decrease in the odds of Gl while the reverse relationship was

seen with increasing percentage of meals consumed in the street or from a

cafeteria/casino. This too could affect exposure to potential hazards in foods and

should be considered in food safety risk assessments and education campaigns.

Of interest, the most frequently cited location of outbreaks in Chile is the home

(Ministerio de Salud (Chile), 2007, Prado, etal., 2002), though this is probably a

reflection of the large proportion of meals consumed in the home.

Increasing education, age and being female were generally associated

with reported 'good' hand washing and hygiene behaviours. A similar

association of reported 'good' hand washing and hygiene habits with being

female and with increasing age was previously reported in the United States

(Altekruse, etai, 1999, Yang, etal., 1998); however, the opposite relationship

between education and hand washing and hygiene habits (i.e., increasing

education was associated with poor hand washing and hygiene habits) was

reported in these same studies. These results may be useful for more targeted

awareness and prevention campaigns aimed at Chilean residents.

Results from this study illustrate that those who report 'good' hand

washing and hygiene practices have reduced levels of Gl. This association has

120

been seen in other studies (Kaferstein, 2003, Gorter, etal., 1998, Huttly, etal.,

1997, Larsen, 2003). Promotion and education on proper hand hygiene as a

method to prevent infectious diseases should be continued.

One or more food items identified to be of 'high Gl risk' by the Ministry of

Health were consumed in the seven days prior to interview by over 70% of the

survey respondents, with commercial mayonnaise, undercooked eggs and

homemade mayonnaise representing the most frequently consumed. However,

these three food products were not found to be significantly associated with Gl in

the 30 days prior to interview in this study in univariable or multivariable model

tests. Nevertheless, eggs and homemade mayonnaise are often identified as a

source of Salmonella (Mitchell, etal., 1989, Alexandre, etal., 2000, Callaway, et

al, 2008). In Chile in 2006, mayonnaise and eggs ranked 4th and 7th respectively

as sources of foodbome disease outbreaks with an identified food source

(Ministerio de Salud (Chile), 2007). Other foods designated as 'high Gl risk'

(e.g., raw fish, undercooked hamburgers, unpasteurized milk, etc.) were not

found to be significantly associated with Gl in this study; however, this may be

due to small sample sizes of consumers and the adjustment for multiple

comparisons creating conservative 95% confidence intervals and p-values.

Additional studies and more detailed information on food consumption, sporadic

cases and outbreaks of Gl would be useful in refining hypotheses of attribution of

Gl to specific food products.

Significant differences by gender were observed for 4 of the 16 'high Gl

risk' foods, with males being more likely to consume these items. These results

121

are similar to those from studies in Canada and the United States where males

were more likely to consume foods considered 'high Gl risk' (Nesbitt, etal., 2008,

Altekruse, etal., 1999, Samuel, etal., 2007). This could be important for food

safety interventions, which may have more impact if targeted to males. Of

interest, from burden of Gl studies, the monthly prevalence of Gl is reported to be

higher among females than males in Canada and the United States (Majowicz, et

al., 2004, Thomas, etal., 2006, Jones, etal., 2007); however, in this Chilean

study there was no difference in prevalence of Gl between males and females

(Chapter 3, (Thomas, et al., (Accepted 2010))). This may be a reflection of the

high response rate in this study or may indicate that the source of illness may be

different for genders in different regions; requires further investigation.

The prevalence of consumption of several specific 'high Gl risk' foods as

well as consumption of any identified 'high Gl risk' food increased with increasing

education. This trend has also been seen in the United States (Altekruse, et al.,

1999, Yang, etal., 1998). Yang et al (Yang, etal., 1998) hypothesized that

highly educated individuals are either unaware of or choose to ignore risks

associated with the consumption of 'high Gl risk' foods, while Levy et al (Levy, et

al., 2008) proposed that self-confidence and complacency associated with higher

education leads to a lower risk perception. Another possible explanation is that

certain 'high risk foods' (e.g., seafood, rare steak etc.) could be related to a more

affluent lifestyle and that higher education is a surrogate for wealth in this

population.

122

'High Gl risk' food consumption was greater in phase 1 of this study than

in phase 2, which is opposite to the pattern of seasonality associated with Gl in

Chile (Ministerio de Salud (Chile), 2007, Thomas, etal., (Accepted 2010)). This

is somewhat surprising; however, the pattern of seasonality associated with Gl

may be due to the temporal or behavioural factors influencing the presence or

growth of microbial pathogens in food, or consumption of foods in settings less

conducive to good hygiene and food preparation practices (i.e., lack of

refrigeration, barbecue season, inadequate cooking, etc.).

Recall bias is a potential limitation of this study as we asked about food

consumption in the seven days prior to interview, and symptoms of Gl in the 30

days prior to interview. Use of a 1-day recall of food consumption may have

reduced recall bias, but this can be subject to day-to-day fluctuations that can be

limited by multiple-day recall (Block, 1982). Choice of recall period length for

burden of Gl studies is an on-going topic of discussion and no clear consensus

has yet been reached (Wheeler, etal., 1999, Boland, etal., 2006, Rodrigues,

2006). Additionally, proxy respondents who answer on behalf of children may

not know exactly what the child ate and children may not be able to accurately

recall what they ate, which may also result in recall bias.

Selection bias due to differences in the distribution of age, gender and

education level among respondents and the referent community may be another

limitation of this study. As well, misclassification bias could have occurred, given

that responses were self-reported and they are only as accurate as respondents

123

are honest. Certain responses may have been over or under-reported due to

perceived pressure to give the 'right' answer (Nederhof, 1985).

Conclusions

This study assessed food consumption, purchasing and hygiene habits

from a food safety perspective in Chile. A key advantage of this study was the

face-to-face methodology which contributed to its high response rate, which

minimized the risk of non-response bias. The results generated here will be

useful for public health practitioners, risk assessors and food safety professionals

in Chile and around the world.

124

References

Aguiar Prieto P, Finley RL, Muchaal PK, Guerin MT, Isaacs S, Castro Domfnguez A, Gisele Coutin M, Perez E. Burden of self-reported acute gastrointestinal illness in Cuba. J Health Popul Nutr 2009; 27:345-57.

Alexandre M, Pozo C, Gonzalez V, Martinez MC, Prat S, Fernandez A, Fica A, Fernandez J, Heitmann I. Detection of Salmonella enteritidis in samples of poultry products for human consumption in the Chilean metropolitan area. Rev.Med.Chil. 2000; 128:1075-1083.

Altekruse SF, Yang S, Timbo BB, Angulo FJ. A multi-state survey of consumer food-handling and food-consumption practices. Am.J.Prev.Med. 1999; 16:216-221.

Block G. A review of validations of dietary assessment methods. Am.J.Epidemiol. 1982; 115:492-505.

Boland M, Sweeney MR, Scallan E, Harrington M, Staines A. Emerging advantages and drawbacks of telephone surveying in public health research in Ireland and the U.K. BMC Public Health 2006; 6:208.

Callaway TR, Edrington TS, Anderson RC, Byrd JA, Nisbet DJ. Gastrointestinal microbial ecology and the safety of our food supply as related to Salmonella. J.Anim.Sci. 2008; 86:E163-72.

Castillo O, Rozowski J, Cuevas A, Maiz A, Soto M, Mardones F, Leighton F. Nutrients intake in elderly people living in Providence, Santiago de Chile. Rev.Med.Chil. 2002; 130:1335-1342.

Centers for Disease Control and Prevention (CDC). Campylobacter jejuni infection associated with unpasteurized milk and cheese-Kansas, 2007. MMWR Morb.Mortal.Wkly.Rep. 2009; 57:1377-1379.

Christie D and Viner R. Adolescent development. BMJ 2005; 330:301-304.

Currie A, MacDougall L, Aramini J, Gaulin C, Ahmed R, Isaacs S. Frozen chicken nuggets and strips and eggs are leading risk factors for Salmonella Heidelberg infections in Canada. Epidemiol.Infect. 2005; 133:809-816.

Gorter AC, Sandiford P, Pauw J, Morales P, Perez RM, Alberts H. Hygiene behaviour in rural Nicaragua in relation to diarrhoea. Int.J.Epidemiol. 1998; 27:1090-1100.

125

Huttly SR, Morris SS, Pisani V. Prevention of diarrhoea in young children in developing countries. Bull.World Health Organ. 1997; 75:163-174.

Jones TF, McMillian MB, Scallan E, Frenzen PD, Cronquist AB, Thomas S, Angulo FJ. A population-based estimate of the substantial burden of diarrhoeal disease in the United States; FoodNet, 1996-2003. Epidemiol.Infect. 2007; 135:293-301.

Kaferstein F. Foodborne diseases in developing countries: aetiology, epidemiology and strategies for prevention. Int.J.Environ.Health Res. 2003; 13 SuppM:S161-8.

Kaferstein F and Abdussalam M. Food safety in the 21st century. Bull.World Health Organ. 1999; 77:347-351.

Larsen B. Hygiene and health in developing countries: defining priorities through cost- benefit assessments. Int.J.Environ.Health Res. 2003; 13 Suppl 1:S37-46.

Levy AS, Choiniere CJ, Fein SB. Practice-specific risk perceptions and self-reported food safety practices. Risk Anal. 2008; 28:749-761.

Majowicz SE, Dore K, Flint JA, Edge VL, Read S, Buffett MC, McEwen S, McNab WB, Stacey D, Sockett P, Wilson JB. Magnitude and distribution of acute, self-reported gastrointestinal illness in a Canadian community. Epidemiol.Infect. 2004; 132:607-617.

Ministerio de Salud (Chile). Boletin de vigilancia de salud publica en Chile. 2007; 10(25):.

Ministerio de Salud Gobiemo de Chile. II Encuesta de Calidad de Vida y Salud Chile 2006 Informe de Resultados Total Nacional. 2006; .

Mitchell E, O'Mahony M, Lynch D, Ward LR, Rowe B, Uttley A, Rogers T, Cunningham DG, Watson R. Large outbreak of food poisoning caused by Salmonella typhimurium definitive type 49 in mayonnaise. BMJ 1989; 298:99-101.

Nederhof AJ. Methods of coping with social desirability bias: a review. Eur J Soc Psychol 1985; 15:263-283.

Nesbitt A, Majowicz S, Finley R, Pollari F, Pintar K, Marshall B, Cook A, Sargeant J, Wilson J, Ribble C, Knowles L. Food consumption patterns in the Waterloo Region, Ontario, Canada: a cross-sectional telephone survey. BMC Public Health 2008; 8:370.

126

Olivares S, Kain J, Lera L, Pizarro F, Vio F, Moron C. Nutritional status, food consumption and physical activity among Chilean school children: a descriptive study. Eur.J.CIin.Nutr. 2004; 58:1278-1285.

Oliver SP, Boor KJ, Murphy SC, Murinda SE. Food safety hazards associated with consumption of raw milk. Foodborne Pathog.Dis. 2009; 6:793-806.

Prado V, Solari V, Alvarez IM, Arellano C, Vidal R, Carreno M, Mamani N, Fuentes D, O'Ryan M, Munoz V. Epidemiological situation of foodborne diseases in Santiago, Chile in 1999-2000. Rev.Med.Chil. 2002; 130:495-501.

Rivas M, Sosa-Estani S, Rangel J, Caletti MG, Valles P, Roldan CD, Balbi L, Marsano de Mollar MC, Amoedo D, Miliwebsky E, Chinen I, Hoekstra RM, Mead P, Griffin PM. Risk factors for sporadic Shiga toxin-producing Escherichia coli infections in children, Argentina. Emerg.lnfect.Dis. 2008; 14:763-771.

Rodrigues LC. Let us not forget telescoping as a major risk of telephone surveys. Comment on Boland et al. BMC Public Health 2006; 6:.

Samuel MC, Vugia D, Koehler KM, Marcus R, Deneen V, Damaske B, Shiferaw B, Hadler J, Henao O, Angulo F. Consumption of risky foods among adults at high risk for severe foodborne diseases: Room for improved targeted prevention messages. J Food Safety 2007; 27:219-232.

Sargeant JM, Majowicz SE, Snelgrove J. The burden of acute gastrointestinal illness in Ontario, Canada, 2005-2006. Epidemiol.Infect. 2008;136: 451-460.

Smith KE, Medus C, Meyer SD, Boxrud DJ, Leano F, Hedberg CW, Elfering K, Braymen C, Bender JB, Danila RN. Outbreaks of salmonellosis in Minnesota (1998 through 2006) associated with frozen, microwaveable, breaded, stuffed chicken products. J.Food Prot. 2008; 71:2153-2160.

Thomas MK, Perez E, Majowicz SE, Reid-Smith R, Olea A, Diaz J, Solari V, McEwen SA. Burden of acute gastrointestinal illness in the Metropolitan region, Chile, 2008. Epidemiol.Infect. (Accepted 2010); .

Thomas MK, Majowicz SE, MacDougall L, Sockett PN, Kovacs SJ, Fyfe M, Edge VL, Flint JA, Henson S, Jones AQ. Population distribution and burden of acute gastrointestinal illness in British Columbia, Canada. BMC Public Health 2006; 6:307.

Thomas MK, Perez E, Majowicz SE, Reid-Smith R, Albil S, Monteverde M, McEwen SA. Burden of acute gastrointestinal illness in Galvez, Argentina, 2007. J Health Popul Nutr. 2010;28: 149-158. .

127

Thomas MK, Majowicz SE, Pollari F, Sockett PN. Burden of acute gastrointestinal illness in Canada, 1999-2007: interim summary of NSAGI activities. Can.Commun.Dis.Rep. 2008; 34:8-15.

Wheeler JG, Sethi D, Cowden JM, Wall PG, Rodrigues LC, Tompkins DS, Hudson MJ, Roderick PJ. Study of infectious intestinal disease in England: rates in the community, presenting to general practice, and reported to national surveillance. The Infectious Intestinal Disease Study Executive. BMJ 1999; 318:1046-1050.

Yang S, Leff MG, McTague D, Horvath KA, Jackson-Thompson J, Murayi T, Boeselager GK, Melnik TA, Gildemaster MC, Ridings DL, Altekruse SF, Angulo FJ. Multistate surveillance for food-handling, preparation, and consumption behaviors associated with foodborne diseases: 1995 and 1996 BRFSS food-safety questions. MMWR CDC Surveill.Summ. 1998; 47:33-57.

128

Table 4.1: Age, gender and education distribution of Metropolitan region residents and survey respondents, Chile 2008.

Variable

Age (years) 0-4 5-9

10-19 20-59

60+ Sex

Male Female



Community residents (from census) (N=6,061,185)

451,995 (7.5%) 511,864 (8.4%) 1,046,091 (17.3%) 3,381,732(55.8%) 669,543(11.0%)

2,937,193(48.5%) 3,123,992(51.5%)

NA 1,887,649(31.1%) 2,167,683(35.8%) 503,105(8.3%) 707,563(11.7%)

Study respondents (n=6047)

96(1.6%)* 107(1.8%)* 545 (9.0%)* 4361 (72.1%)* 879(14.5%)*

2594 (42.9%)* 3451 (57.1%)*

35 (0.6%) 978(16.2%)* 2398 (39.7%)* 1176(19.4%)* 1247(20.6%)*

Indicates significant (p<0.05) difference between survey respondents and residents.

129

Tabl

e 4.

2: D

istr

ibut

ion

of fo

od p

urch

asin

g ha

bits

in th

e se

ven

days

prio

r to

inte

rvie

w,

mon

thly

pre

vale

nce

of a

cute

ga

stro

inte

stin

al i

llnes

s (G

l), a

nd u

ncon

ditio

nal a

ssoc

iatio

ns (

odds

rat

io w

here

ref

eren

t is

all

othe

r so

urce

cat

egor

ies

com

bine

d) b

etw

een

loca

tion

of fo

od it

em p

urch

ase

Gl i

n th

e 30

day

s pr

ior

to in

terv

iew

(ad

just

ed f

or m

ultip

le c

ompa

rison

s),

Met

ropo

litan

reg

ion,

Chi

le, 2

008.

Food

ite

m

Mea

t (b

eef,

poul

try,

por

k,

lam

b)

Fish

Dai

ry a

nd

eggs

So

urc

e

Sta

tiona

ry o

r m

obile

mar

ket/k

iosk

F

arm

F

arm

ers

mar

ket

Spe

cial

ized

sto

re

Sup

erm

arke

t S

tatio

nary

or

mob

ile m

arke

t/kio

sk

Far

mer

s m

arke

t S

peci

aliz

ed s

tore

S

uper

mar

ket

Sta

tiona

ry o

r m

obile

mar

ket/k

iosk

F

arm

F

arm

ers

mar

ket

Spe

cial

ized

sto

re

Sup

erm

arke

t

Per

cen

t of

re

spo

nd

ent

s(n

)

1.8

(10

6)

0.1

(5)

2.0(

117)

26

.7(1

584)

69

.5(4

128)

1

.8(9

0)

36.5

(181

8)

11.0

(547

) 50

.8 (

2532

) 22

.7(1

342)

0.3

(16

) 7.

5 (4

40)

4.6

(274

) 64

.9 (

3832

)

Mon

thly

Gl

pre

vale

nce

* am

ong

re

spo

nd

ents

wit

h

pu

rch

asin

g h

abit

13.2

1 (7

.41

-21

.17

)

40

.00

(5.2

7-8

5.3

4)

5.1

3(1

.90

-10

.83

) 8

.59

(7.2

5-1

0.0

8)

7.1

9(6

.42

-8.0

3)

13

.33

(7.0

8-2

2.1

3)

9.4

7(7

.23

-9.8

5)

6.9

5(4

.96

-9.4

1)

6.7

9(5

.84

-7.8

4)

10

.88

(9.2

6-1

2.6

7)

18

.75

(4.0

5-4

5.6

5)

6.1

4(4

.08

-8.8

0)

8.3

9(5

.40

-12

.33

) 6

.63

(5.8

6-7

.46

)

Un

con

diti

on

al a

sso

ciat

ion

be

twee

n f

oo

d p

urch

as

and

Gl i

n r

esp

on

den

ts

Od

ds

ratio

(O

R)

(95%

CI)

1.8

6(0

.85

-3.8

2)

8.0

7(0

.42

-10

0.2

3)

0.6

5(0

.19

-1.7

4)

1.1

9(0

.90

-1.5

6)

0.81

(0

.62

-1.0

6)

1.91

(0

.83

-4.0

8)

1.2

3(0

.94

-1.6

1)

0.91

(0

.57

-1.3

9)

0.80

(0.6

1 -1

.05

) 1.

71 (

1.2

9-2

.24

) f

2.81

(0

.37

-12

.94

) 0

.78

(0.4

5-1

.29

) 1

.12

(0.6

2-1

.95

) 0.

68 (

0.5

3-0

.88

) f

ing

hab

it

Ad

just

ed

P-v

alu

e of

O

R

0.17

0.21

0.

87

0.41

0.

18

0.15

0.21

0.

97

0.16

<0

.01

0.43

0.

73

0.99

<0

.01

130

Frui

ts a

nd

veg

etab

les

Sta

tiona

ry o

r m

obile

mar

ket/k

iosk

F

arm

F

arm

ers

mar

ket

Spe

cial

ized

sto

re

Sup

erm

arke

t

1.6

(98

)

<0.1

(2)

66

.9 (

3995

) 3.

5 (2

05)

28.0

(167

1)

11

.22

(5.7

4-1

9.2

0)

50

.00

(1.2

6-9

8.7

4)

8.01

(7

.19

-8.9

0)

9.7

6(6

.06

-14

.67

) 6

.34

(5.2

2-7

.62

)

1.5

7(0

.65

-3.5

1)

12

.35

(0.0

7-2

30

7.1

1)

1.2

6(0

.95

-1.6

7)

1.3

5(0

.69

-2.4

4)

0.7

8(0

.58

-1.0

5)

0.56

0.50

0.

16

0.67

0.

16

* D

efin

ed a

s th

e nu

mbe

r of

cas

es o

f G

I in

the

30 d

ays

prio

r to

inte

rvie

w,

divi

ded

by th

e to

tal n

umbe

r of

sur

vey

resp

onde

nts.

t

Indi

cate

s si

gnifi

cant

odd

s ra

tio.

131

Tab

le 4

.3:

Fre

quen

cy o

f ha

nd w

ashi

ng a

nd h

ygie

ne b

ehav

iour

s, m

onth

ly a

cute

gas

troi

ntes

tinal

illn

ess

(Gl)

prev

alen

ce a

nd

unco

nditi

onal

ass

ocia

tions

(od

ds r

atio

) be

twee

n ha

nd w

ashi

ng a

nd h

ygie

ne b

ehav

iour

s an

d G

l in

the

30 d

ays

prio

r to

in

terv

iew

, M

etro

polit

an r

egio

n, C

hile

, 200

8.

Beh

avio

ur

Was

h ha

nds

befo

re e

atin

g?

Alw

ays

Som

etim

es

Nev

er

Use

soa

p w

hen

was

hing

ha

nds

befo

re e

atin

g?

Alw

ays

Som

etim

es

Nev

er

Was

h ha

nds

afte

r go

ing

to t

he

bath

room

? A

lway

s S

omet

imes

N

ever

U

se s

oap

whe

n w

ashi

ng

hand

s af

ter

goin

g to

the

bath

room

? A

lway

s S

omet

imes

N

ever

U

se s

peci

fic c

uttin

g bo

ard

for

cutti

ng m

eat?

A

lway

s S

omet

imes

N

ever

Per

cen

t of

re

spo

nd

ents

(n)

90.2

(53

82)

9.4

(563

) 0.

4 (2

3)

85.2

(50

78)

13.8

(821

) 1

.0(6

1)

94.9

(56

50)

4.8

(286

) 0

.3(1

7)

89.0

(52

92)

10.3

(610

) 0

.7(4

1)

73.2

(414

4)

14.8

(836

) 12

.1 (

683)

Mon

thly

Gl

pre

vale

nce

7.1

9(6

.51

-7.9

1)

13.4

1 (1

0.4

6-1

6.2

2)

13

.04

(2.7

8-3

3.5

9)

7.3

7(6

.66

-8.1

2)

10

.48

(8.4

6-1

2.7

7)

8.2

(2.7

2-1

8.1

0)

7.7

2(7

.03

-8.4

4)

9.0

9(6

.02

-13

.04

) 1

1.7

6(1

.46

-36

.44

)

7.69

(6.

99 -

8.

44)

8.0

3(6

.00

-10

.48

) 1

7.0

7(7

.15

-32

.06

)

7.0

9(6

.33

-7.9

2)

9.5

7(7

.66

-11

.77

) 8

.20

(6.2

5-1

0.5

2)

Odd

s R

atio

(O

R)

(95%

CI)

0.51

(0

.39

-N

A

1.7

8(0

.45

-

0.6

9(0

.54

-N

A

1.0

6(0

.40

-

0.8

2(0

.55

-N

A

1.5

8(0

.26

-

0.89

(0.

66 -

NA

2

.46

(1.0

5-

0.7

8(0

.63

-N

A

1.1

0(0

.82

-- 0.

67)*

- 5.

74)

- 0.

88)*

-2.6

0)

-1.2

5)

-6.4

3)

-1.1

9)

-5.5

8)*

-0.9

6)*

-1.4

8)

P-v

alu

e of

OR

<0.0

1

0.42

<0.0

1

0.81

0.32

0.39

0.39

0.04

0.02

0.54

132

Was

h fr

uits

and

veg

etab

les

befo

re e

atin

g?

Alw

ays

88.9

(527

2)

Som

etim

es

10.4

(616

) N

ever

0

.7(3

9)

* In

dica

tes

sign

ifica

nt o

dds

ratio

. N

A =

Not

app

licab

le.

6.9

8(6

.31

-7.7

0)

0.4

6(0

.36

-0.5

9)*

<

0.01

1

3.0

8(1

1.1

7-1

6.7

8)

NA

1

7.9

5(7

.54

-33

.53

) 2

.62

(1.1

2-6

.03

)*

0-03

133

Tabl

e 4.

4:

Pro

port

ion

of r

espo

nden

ts t

hat

cons

umed

hig

h ris

k fo

od*

at le

ast

once

in th

e se

ven

days

prio

r to

int

ervi

ew,

mon

thly

pre

vale

nce

of a

cute

gas

troi

ntes

tinal

illn

ess

(Gl)

and

unco

nditi

onal

ass

ocia

tions

(od

ds r

atio

) be

twee

n co

nsum

ptio

n of

indi

vidu

al h

igh

risk

food

and

Gl (

adju

sted

for

mul

tiple

com

paris

ons)

, M

etro

polit

an r

egio

n, C

hile

, 200

8 (n

=604

7).

Hig

h r

isk

foo

d

% (n

) M

onth

ly G

l p

reva

len

ce

OR

(95

% C

I)

Ad

just

ed P

-va

lue

Egg

s -

raw

E

ggs

- un

derc

ooke

d R

aw s

eafo

od

Oys

ters

, m

usse

ls,

mol

lusc

s F

ish

- ra

w

Fis

h -

unde

rcoo

ked

Bee

f - r

aw

Bee

f -

unde

rcoo

ked

Ham

burg

ers

-un

derc

ooke

d P

ork

- un

derc

ooke

d P

oultr

y -

unde

rcoo

ked

Lam

b -

unde

rcoo

ked

Milk

-

unpa

steu

rized

C

hees

e -

from

un

past

euriz

ed m

ilk

May

onna

ise

-co

mm

erci

al/b

ottle

d M

ayon

nais

e -

hom

emad

e A

ny h

igh

risk

food

}:

6.8

(414

) 19

.9(1

206)

5.

7 (3

46)

0.9

(54)

2.

1 (1

32)

0.3

(20)

1.

2 (7

4)

6.6

(398

) 0.

9 (5

6)

0.8

(49)

1

.2(7

1)

0.2(

13)

0.9

(57)

5.

9 (3

59)

56.6

(34

20)

8.7

(528

) 71

.5(4

323)

10.1

4(7.

41 -

13

.47

) 9

.45

(7.8

6-1

1.2

5)

7.9

2(5

.40

-11

.11

) 7

.41

(2.0

6-1

7.8

9)

11

.36

(6.5

0-1

8.0

5)

10

.00

(1.2

3-3

1.7

0)

8.11

(3

.03

-16

.82

) 1

4.5

7(1

1.2

6-1

8.4

3)

14

.29

(6.3

8-2

6.2

2)

16.3

3 (7

.32

-29

.66

) 1

9.7

2(1

1.2

2-3

0.8

6)

15

.38

(1.9

2-4

5.4

5)

15

.79

(7.4

8-2

7.8

7)

14.2

1 (1

0.7

6-1

8.2

5)

7.6

0(6

.74

-8.5

4)

9.47

(7.1

1 -1

2.2

9)

8.21

(7.

41 -

9.0

7)

1.38

(0.8

1 -

1.3

3(0

.94

-1

.03

(0.5

5-

0.9

5(0

.13

-1.

54(0

.61

-1

.32

(0.0

5-

1.0

5(0

.23

-2

.18

(1.3

8-

2.0

0(0

.55

-

2.3

5(0

.64

-2

.99

(1.1

2-

2.1

7(0

.08

-2

.26

(0.6

4-

2.0

9(1

.28

-

0.96

(0.

72 -

1.2

8(0

.79

-1

.29

(1.0

3--2

.24

) -1

.85

) -1

.80

) -3

.65

) -3

.35

) -8

.62

) -

3.35

) -

3.3

8)f

-5

.89

)

-6.9

9)

-7.1

1)t

-1

6.0

6)

-6.5

0)

- 3

.32

)f

-1.2

8)

-2.0

0)

-1.6

1)t

0.69

0.

20

1.00

1.

00

0.90

1.

00

1.00

<0

.01

0.72

0.57

0.

01

0.99

0.

48

<0.0

1

1.00

0.88

0.

02

* H

igh

risk

of G

l foo

d as

det

erm

ined

by

Min

istr

y of

Hea

lth, C

hile

. f

Indi

cate

s si

gnifi

cant

odd

s ra

tio.

t U

nadj

uste

d 95

% c

onfid

ence

int

erva

l and

p-v

alue

.

134

Table 4.5: Associations between reported consumption of high risk* food items and gender (males compared to females) in the seven days prior to interview (controlling for age and adjusted for multiple comparisons), Metropolitan region Chile, 2008.

Food item

Eggs - raw Eggs - undercooked Raw seafood Oysters, mussels, molluscs Fish - raw Fish - undercooked Meat - raw Meat - undercooked Hamburgers - undercooked Pork - undercooked Poultry - undercooked Lamb - undercooked Milk - unpasteurized Cheese - made from unpasteurized milk Mayonnaise -commercial/bottled Mayonnaise - homemade Any high risk foodt

Consumption Females (n=3451) 209 639 184 22 65 14 23 169 29 25 32 4 31 182

1936

267 2441

frequency

Males (n=2594) 205 566 195 32 67 6 51 229 27 24 39 9 26 177

1484

261 1881

OR

1.35 1.21 1.47t 2.06 1.40 0.56 3.02f 1.90f 1.21 1.33 1.63 3.00 1.09 1.31

1.05

1.35f 1.09

Adjusted 95% CI

1.00-1.82 1.00-1.47 1.07-2.01 0.90 - 4.74 0.83-2.37 0.13-2.38 1.43-6.36 1.40-2.60 0.55-2.67 0.57-3.13 0.80-3.30 0.51 -17.67 0.49-2.39 0.95-1.81

0.90-1.23

1.03-1.77 0.98-1.23

* High Gl risk food identified by Ministry of Health. f Indicates significant difference in odds ratio compared to referent group (females). j : Unadjusted 95% confidence interval.

135

Tabl

e 4.

6: A

ssoc

iatio

ns o

f re

port

ed c

onsu

mpt

ion

of h

igh

risk*

food

item

s an

d ag

e gr

oup

in th

e se

ven

days

prio

r to

in

terv

iew

(co

ntro

lling

for

gen

der

and

adju

sted

for

mul

tiple

com

paris

ons)

com

pare

d w

ith r

efer

ent

grou

p of

indi

vidu

als

20-5

9 ye

ars

of a

ge,

Met

ropo

litan

reg

ion,

Chi

le, 2

008.

Food

Ite

m

Egg

s -

raw

E

ggs

- un

derc

ooke

d R

aw s

eafo

od

Oys

ters

, m

usse

ls,

mol

lusc

s F

ish

- ra

w

Fis

h -

unde

rcoo

ked

Mea

t - r

aw

Mea

t -

unde

rcoo

ked

Ham

burg

ers

- un

derc

ooke

d P

ork

- un

derc

ooke

d P

oultr

y -

unde

rcoo

ked

Lam

b -

unde

rcoo

ked

Milk

-

unpa

steu

rized

C

hees

e -

from

unp

aste

uriz

ed m

ilk

May

onna

ise

- co

mm

erci

al/b

ottle

d M

ayon

nais

e -

hom

emad

e A

ny h

igh

risk

foo

dt

OR

0.

28

1.55

0.

28

NA

1.

30

NA

N

A

NA

N

A

NA

N

A

NA

N

A

0.63

0.

18f

0.21

0

.24

t

0-4

95%

CI

0.0

3-

0.7

2-

0.0

3-

0.1

9-

0.1

2-

0.0

8-

0.0

2-

0.1

6--2

.89

-3.3

4 -2

.95

-9.1

6

-3.3

9 -0

.41

-2.1

9 -0

.37

OR

0.

62

1.93

0.

12

NA

0.

75

NA

N

A

0.33

2.

17

NA

N

A

NA

3.

00

0.41

0.

95

0.89

1.

21

Age

of

resp

on

den

t 5-

9 95

% C

I 0

.14

-0

.96

-0.

01-

0.0

7-

0.0

5-

0.2

0-

0.41

-0.

06-

0.50

-0.

28-

0.77

--2.8

4 -3

.85

-3.2

9

-8.0

0

-2.2

7 -

23.9

9

-21

.95

-2.8

3 -1

.81

-2.8

1 -1

.90

OR

1.

07

1.74

f 0.

62

0.98

1.

40

2.16

0.

47

0.81

2.

83

1.72

1.

77

1.56

1.

76

0.70

1.

44f

1.03

1

.41

|

10-1

9 95

% C

I 0.

61 -

1.24

-0

.30

-0.

21 -

0.6

0-

0.3

4-

0.0

9-

0.4

4-

0.9

7-

0.4

8-

0.6

2-

0.3

4-

0.5

2-

0.3

5-

1.05

-0.

61 -

1.13

--1.8

8 -2

.42

-1.2

6 -4

.68

-3.2

7 -1

3.7

7 -2

.58

-1.4

9 -8

.24

-6.2

4 -5

.10

-19

.84

-5.9

4 -1

.42

-1.9

8 -1

.71

-1.7

5

OR

0.

91

0.85

0.

82

1.00

0.

19

0.33

0.

39

0.5

2f

0.54

0.

40

0.46

0.

51

0.48

0.

84

0.6

6|

0.85

0

.64

f

>60

95%

CI

0.5

5-

0.6

1-

0.4

9-

0.2

8-

0.0

4-

0.0

1-

0.0

8-

0.2

9-

0.1

0-

0.0

6-

0.1

0-

0.0

2-

0.0

9-

0.4

9-

0.5

2-

0.5

4-

0.5

5--1

.48

-1.1

7 -1

.38

-3.6

0 -1

.01

-9.6

3 -1

.77

-0.9

4 -3

.02

-2.8

9 -2

.14

-15

.78

-2.6

9 -1

.42

-0.8

4 -1

.32

-0.7

5 *

Hig

h G

l ris

k fo

od id

entif

ied

by M

inis

try

of H

ealth

. t

Indi

cate

s si

gnifi

cant

diff

eren

ce i

n od

ds r

atio

com

pare

d to

ref

eren

t gro

up (

20-5

9 ye

ar o

lds)

. t

Una

djus

ted

95%

con

fiden

ce in

terv

al.

136

Table 4.7: Final multivariable model of socio-demographic, food consumption and hygiene behaviour factors associated with Gl, Metropolitan region, Chile, 2008.

Variable Age

0-4 5-9

10-19 20-59

60+ Consumption of undercooked beef

Yes No

Consumption of undercooked chicken

Yes No

Consumption of cheese made from unpasteurized milk

Yes No

Frequency

96 107 545 4361 879

397 5576

71 5896

356 5617

OR (95% CI)

4.00(2.40-6.66) 1.37(0.68-2.74) 2.19(1.67-2.89) Referent 1.01 (0.76-1.36)

1.98(1.46-2.71) Referent

2.44(1.33-4.50) Referent

1.95(1.41 -2.69) Referent

P value <0.001

<0.001

0.004

<0.001

137

Figure 4.1: Unconditional odds of acute gastrointestinal illness (Gl) and proportion of meals consumed in different locations, Metropolitan region, Chile, 2008.

0.3 -

0.25 -

0.2 -

O > • -

° 0.15 -TJ T3 O

0.1 -

0.05

n U i

- '

— - - Cafeteria

Street

Home

^r.cr:

0 10 20 30 40 50 60 70 80 90 100

Proportion of meals

138

CHAPTER FIVE

Water consumption trends and associations with acute gastrointestinal

illness in the Metropolitan Region, Chile, 2008

Prepared for submission: BMC Public Health

Abstract

Background: An adequate, safe water supply is essential, and contaminated

water has been a global cause of morbidity and mortality. In 2008, the first

population survey on the burden of acute gastrointestinal illness (Gl) in the

Metropolitan region of Chile was conducted. The objectives of the study

presented here were to describe water consumption patterns, use of bottled

water and water treatment, and to identify associations of drinking water-related

risk factors with Gl.

Methods: A cross-sectional survey was administered door-to-door in the

Metropolitan region, Chile from 21 July to 25 August and 14 November to 21

December, 2008. Respondents were randomly selected proportional to socio

economic level stratification. Survey questions related to water consumption,

source of water, use of bottled water, water treatment and recent symptoms of

Gl.

Results: Overall, 6047 respondents participated (76% response rate). The

majority of respondents did nothing to treat their drinking water. Approximately

139

18% of respondents consumed 75% or more of their daily water as bottled water.

Males, respondents < 10 years of age, and non-bottled water users tended to

consume more water. Boiling water and increased water consumption were

associated with increased odds of Gl.

Conclusions: This study describes the water consumption patterns of the

Metropolitan region, Chile and their associations with Gl. This information

provides region-specific data that can inform quantitative microbial risk

assessments on drinking water for this population.

Background

An adequate, safe supply of water is a necessity [1, 2], and contaminated

water has been a cause of morbidity and mortality around the world, often due to

acute gastrointestinal illness (Gl)from bacteria (e.g., Campylobacter spp., Vibrio

spp.), parasites (e.g., Giardia spp., Cryptosporidium spp.) and viruses (e.g.,

hepatitis A virus) [3-13]. To minimize the adverse health impacts of

microbiological contaminants in drinking water, several control measures can be

implemented. These include boiling of potentially contaminated water in homes

prior to consumption, delivery of piped water to households and other buildings,

treating water by chlorination, sedimentation, flocculation, filtration or other

treatment procedures or devices, and commercial delivery of potable water in

containers (e.g., bottled water). At the international level the World Health

Organization (WHO) has established recommended guidelines for drinking water

quality to ensure drinking water safety and support the development and

140

implementation of waterborne risk management strategies [1]. A decision

support tool that is increasingly being promoted for use by waterborne disease

risk managers is quantitative microbial risk assessment (QMRA [1]). QMRA can

provide a structured approach for assembling scientific information that is

important for making sound drinking water risk management decisions [14].

However, to effectively conduct a QMRA on drinking water, key data are

required, including water consumption and water source information.

The WHO estimated that in 2006, 98% and 72% of the urban and rural

residents of Chile, respectively, had access to 'improved' drinking water sources

(i.e., water piped into residence or yard or public taps, wells or protected water

sources [15]). However, there is little published information related to the amount

of water consumed in Chile, how much of that water is from bottled sources, and

which water treatment methods are used at the household level.

In 2008, the first population survey on the burden of acute gastrointestinal

illness (Gl) in the Metropolitan region of Chile was conducted. This survey

included specific questions related to water consumption, source of water, and

the use of bottled water and water treatment. The objectives of the study

presented here were to describe the distribution of water consumption patterns,

use of bottled water and water treatment and to identify associations of drinking

water-related risk factors with Gl.


Population study

141

Detailed methodology of the survey is reported elsewhere (Chapter 3,

[16]). In summary, a cross-sectional, door-to-door survey of randomly selected

residents of the Metropolitan region of Chile was administered in two phases,

July 21 - August 25, 2008 (Phase 1, low Gl season) and November 14 -

December 21, 2008 (Phase 2, high Gl season). These phases were selected to

correspond with low and high Gl seasons (based on surveillance and outbreak

data for the Metropolitan region from the Ministry of Health).

The 52 neighbourhoods of the region were classified by Instituto Nacional

de Estadisticas10 according to socio-economic level into five categories. The

number of surveys administered per socio-economic category was selected

proportional to population size. Blocks within neighbourhoods were randomly

selected proportional to the number of households within each block. A

convenience sample of households was selected from within the randomly

selected block at the discretion of the surveyor in the field.









Sample size


142

http://www.ine.cl



prevalence of Gl of 8%, a 1% allowable error and a 95% confidence interval, a

target sample size of 2826 was calculated, which was rounded to 3000 surveys

per study phase for an overall total sample size of 6000 surveys.

Data collection

The survey (Appendix IV) was developed by modifying the survey tools

used previously in Argentina (Appendix I and II, Chapter 2, [17]) with reference to

other similar cross-sectional population burden of Gl studies [18-21] and water

consumption studies [22-24] as models. In addition to questions about Gl

symptoms and general socio-demographic factors, respondents were asked to

estimate the number of 250ml servings of cold water (i.e. not including water

used to make tea or coffee) they had consumed in the 24 hours prior to interview,

and to identify their source of drinking water (i.e., tap, bottled water, other), any

processes or devices used to treat their water at home (i.e., boiling, filtering,

other) and the origin of their drinking water (i.e., municipal water supply, well,

other).

Ethics





participants or the parent/guardian in the event the participant was a minor.

143

Statistics


Microsoft Access (Microsoft Corporation, 2003). Analysis was performed using

SAS 9.1 (SAS Institute Inc., Cary North Carolina, 2004). Individuals responding

'don't know' or 'unsure' were excluded from the analysis of that question.

Respondents who consumed 75% or more of their total water intake as 'bottled

water' were classified as 'bottled water users' for analytic purposes.

Gl in a respondent was defined as experiencing symptoms of vomiting or

diarrhea in the 30 days prior to interview, where diarrhea was defined as three or

more loose stools in 24 hours. Monthly prevalence of Gl was defined as the

number of cases of Gl in the 30 days prior to interview, divided by the total

number of survey respondents.

Multivariable analyses

A manual backwards elimination procedure was used for all multivariable

models and interactions were considered. Negative binomial regression was

used for the multivariable analysis of explanatory variables associated with the

number of 250 ml servings of cold drinking water consumed, using Proc

GENMOD in SAS. All socio-demographic (age, gender, education, socio

economic level, occupation) and water-related (source of water, water treatment

use, being a bottled water user) variables and the study phase variable were

offered to the model. Two-way interactions were created using all main effect

variables which had a likelihood ratio p-value <0.05. Only variables and two-way

144

interaction terms with likelihood ratio p-values <0.01 were kept in the final model.

All variables that were initially screened out of the final model (i.e., not

considered to be confounders a priori or not statistically significant) were re

introduced to test for significance and visually assess confounding. Confounding

was considered present when a change of 30% or more, a change in sign or a

change in significance in model coefficients was observed.

Logistic regression was used for the multivariable analysis of risk factors

for Gl using Proc GENMOD in SAS. All socio-demographic (age, gender,

education, socio-economic level, occupation) and water-related (source of water,

water treatment use, being a bottled water user, number of servings of water

consumed) variables along with the study phase and health insurance system

variables were offered to the model. Two-way interactions were created among

all main effect variables, with the exception of health insurance system and water

source due to sparseness of data. Main effect variables and interactions were

kept in the model if significant (p-value <0.05, Wald's test). All variables that were

initially screened out of the final model (i.e., not considered to be confounders a

priori or not significant) were re-introduced to test for significance and visually

assess confounding, as described above. The Hosmer-Lemeshow test in SAS

was used to assess goodness of fit of the model, where a significant p-value (p-

value <0.05) indicated poor fit of the model.

Results

145


phase 2 with an overall average response rate of 75.8%. The demographic

distribution of residents of the Metropolitan region, along with survey respondents

and Gl cases is presented in Table 5.1. In general, survey respondents were

older, more educated and more likely to be female than residents. The majority

of respondents received their water from the municipal water supply (99.8%).

Water treatment

In total, 6024 respondents answered questions about processes and

devices they used to treat their water. The majority indicated that they did

nothing to treat their water (82.6%), followed by boiling (12.1%), use of tap filters

(4.8%) or use of jug filters (0.4%). One respondent reported use of 'osmosis'.

Bottled water consumption

The total amount of bottled water consumed was reported by 5927

respondents. Bottled water was not consumed the day prior to interview by

53.7% (n=3184) of respondents. However 1.9% (n=110) of respondents

consumed a 'small' amount (<25% of total daily consumption) of bottled water. A

'moderate' amount (25% - 50% of total daily consumption) of bottled water was

consumed by 10.8% (n=643) of respondents. A 'large' amount (50% - 74% of

total daily consumption) of bottled water was consumed by 15.1% (n=895) of

respondents. 'Nearly all' (75% - 99% of total daily water consumption) water

consumed as bottled water was reported by 2.8% (n=166) of respondents and

146

15.7% (n=929) of respondents consumed 'all' (100% of total daily water

consumption) water as bottled water. Respondents that consumed 75% or

greater of their total water consumption as bottled water (n=1095) were classified

as 'bottled water users'. Of the 1095 'bottled water users', total water

consumption (including both bottled and other water) ranged from 1 to 24

servings (250 ml) with a median of 4 servings consumed in the day prior to

interview.

Factors associated with total water consumption

Responses regarding the amount of water consumed were received from

6025 respondents. Total daily water intake ranged from 0 to 32 servings (250

ml) per day with a median of 4 servings per day. A few respondents (n=77,

1.3%) reportedly consumed no water, 31.4% (n=1898) consumed a 'small'

amount of water (less than 4 servings), 48.8% (n=2952) consumed a 'moderate'

amount of water (between 4 and 8 servings) and 18.2% (n=1098) consumed a

'large' amount of water (8 or more servings) in the day prior to interview. In total,

93.1% of respondents said that the amount of water consumed in the 24 hours

prior to interview was the amount of water that they consume in a day on a

regular basis.

The final multivariable model of the amount of water consumed contained

the variables study phase, age, gender, socio-economic level, education,

occupation, being a bottled water user and three 2-way interaction terms (study

phase by education and by gender, and age by gender) and a 3-way interaction

147

term (study phase by socio-economic level by being a bottled water user). Table

5.2 shows the estimated mean number of servings by occupation, the only main

effect with no significant interactions, where 'general employers' consumed the

most (5.51 servings) and those who worked in the public sector consumed the

least (4.37 servings).

Figure 5.1 shows the estimated mean number of servings by gender and

study phase. Males tended to consume more servings than females; this

difference was greater in phase 2, where more servings were consumed than in

phase 1. Figure 5.2 shows the estimated mean number of servings by education

and phase. Those with higher education tended to consume more water in

phase 1 while the reverse was seen in phase 2, though the variation was smaller

across the different education levels in phase 2. Figure 5.3 shows the estimated

mean number of servings by gender and age. Respondents <10 years of age

drank more than older responders, and males consumed more than females in

all age groups except those 60 years and older. Figure 5.4 shows the estimated

mean number of servings by phase, socio-economic level and bottled water use.

Non-bottled water users typically consumed more water than bottled water users

in phase 2 and most of phase 1, though this varied by socio-economic level.

Differences in socio-economic level were greatest in phase 2. Those in the

lowest socio-economic level consumed the highest amount of water in phase 2,

while in phase 1 bottled water users in the mid-low and mid-high socio-economic

levels consumed the greatest amount of water.

148

Factors associated with acute gastrointestinal illness

In total 7.7% (n=467) of respondents experienced symptoms of Gl in the

30 days prior to interview. In the final logistic regression model with Gl as the

outcome, study phase, occupation, health system, age, water treatment, number

of portions of water consumed and the interaction of socio-economic level by

gender were significant (Table 5.3, Figure 5.5). The Hosmer-Lemeshow

goodness of fit test p-value was 0.11, indicating the model fit the data.

Those who treated their water by boiling had higher odds of Gl compared

to those did nothing to their water. For every 250 ml serving of water consumed

in the day prior to interview, the odds of Gl in the 30 days prior to interview

increased by 1.10 (95% CI 1.06, 1.13) times. Water source (p=0.37) and being a

bottled water user (p=0.76) were not significantly associated with Gl.

Of the non-water-related variables, participants surveyed in study phase 1

(low Gl season) had higher odds of Gl compared to those in study phase 2 (high

Gl season). Students had higher odds of Gl compared to housewives. Children

0-4 years and youths 10-19 years of age had the highest odds of Gl. Having no

health insurance was associated with higher odds of Gl, while those subscribing

to a private health insurance system had lower odds of Gl, compared to those

that subscribed to a public health insurance system. Females in the lowest

socio-economic level had higher odds of Gl compared to females in the highest

socio-economic level.

Discussion

149

This study reports the drinking water habits of residents of the

Metropolitan region, Chile during two time periods of 2008 and evaluates

associations among water-related factors (i.e., water source, amount of water

consumed, being a bottled water user and in-home water treatment) and Gl.

Nearly 8% of the respondents reported experiencing Gl in the 30-days

prior to interview. The amount of water consumed (number of 250 ml servings)

on the day before interview and water treatment in the home were both

significantly associated with being a case of Gl in the 30 days prior to interview.

For each serving of water consumed, the odds of Gl increased by 1.10 times, a

finding similar to a Canadian water consumption study [23]. This suggests the

possibility that there may be waterborne Gl in the community, especially in light

of an apparent dose-response relationship between the amount of water

consumed and odds of Gl. However, due to the cross-sectional study design,

and since data on water consumption pertained to the 24 hours prior to interview

while data on Gl pertained to the 30 days prior to interview, it is not possible to

be sure that increasing water consumption was causally associated with being a

case of Gl. We cannot rule out the possibility that cases of Gl consumed more

water to alleviate symptoms of their illness (e.g., dehydration).

Those who used boiling as a water treatment had increased odds of Gl.

We do not, however, believe that this means that water boiling caused Gl, rather,

we think water boiling may be an indicator of poor water quality in affected

homes, associated with a perceived potential for water contamination among

participants experiencing Gl or associated with the possible occurrence of boil

150

water advisories to control waterborne disease in regions of the community with

water treatment problems. In any case, as discussed above, due to the study

design we cannot conclude that boiling water was causally associated with being

a case of Gl. Interactions by water source could not be explored due to

sparseness of data, though water source was not significant as an individual

variable in this model and socio-economic level was controlled analytically in the

model. A Canadian study found an increase in Gl among those consuming

untreated tap water compared to filtered tap water [25], while a recent study in

the USA found no association of endemic cryptosporidiosis with drinking filtered,

bottled or plain water compared to boiled water [26], underscoring the need for

further investigation into the association of type of water treatment and water

consumption with Gl in the general population.

Respondents in phase 1 had modestly higher odds of Gl compared to

respondents in phase 2, controlling for other variables in the model. The time

period for study phase 1 took place within the 'low Gl season' in the Metropolitan

region, Chile, thus, this finding is somewhat surprising. The water consumption

patterns reported here may in part explain this result. We found that the amount

of water consumed was greater in phase 2 than phase 1 and that increasing

water consumption was associated with Gl. It is possible that the increased

water consumption in phase 2 may in part explain the typical seasonal pattern of

Gl seen in this population and that by including water consumption in the

multivariable model we have, in part, accounted for the Gl seasonality.

Additional explanations could be that the typical Gl seasonality is due to certain

151

pathogens (e.g., bacteria) that are more often reported than others (e.g., viruses)

and their inherent seasonality. Thus what is captured by surveillance and used to

determine 'high' and 'low' Gl season reflects seasonality of certain Gl-related

pathogens rather than seasonality of Gl caused by all pathogens. Moreover, it is

possible that severe cases of Gl that may be more likely to occur in the high Gl

season (phase 2) were hospitalized and thus not captured in the survey. Further

investigation of this association is needed.

More detailed discussion about the socio-demographic and health

insurance system associations with Gl is discussed elsewhere (Chapter 3, [16]).

In brief, children 0-4 years of age and youths 10-19 years of age had increased

odds of Gl, and this is similar to findings in other studies [18-21, 27-29].

Children's behaviour may increase their exposure to pathogens through

environment and person-to-person contact [30] and we found that youths 10-19

years of age were more likely to consume high Gl risk food than adults 20-59

years of age in our analysis of risk factors for Gl related to food consumption

trends, food purchasing and hygiene habits (Chapter 4).

Students had higher odds of Gl compared to housewives, adjusting for

age. This may in part be attributed to food choices as well as the phenomenon

of 'second weaning', where individuals in this age group are beginning to cook

food for themselves and move out of the family home. In Ireland, increased odds

of Gl was observed among professional/non-manual occupations compared to

skilled/unskilled manual occupations [29], and in Cuba [18] health service and

administrative occupations were at increased odds of Gl compared to being a

152

housewife, although there was no difference in odds of Gl between housewives

and students, as was found in this study. Direct comparison of these results is

difficult due to differences in occupation categorization; however, these results

may be evidence of country-specific differences in risk of Gl related to

occupation.

Respondents without health insurance had higher odds of Gl and those

subscribing to private health insurance had lower odds of Gl compared to those

subscribing to the public health insurance system. This is similar to studies in the

United States where a lack of medical insurance was associated with Gl [28].

This may in part reflect differences in socio-economic level. Furthermore,

significant interaction of gender and socio-economic level indicated that females

in the lowest socio-economic level had the highest odds of Gl. Increase odds of

Gl were found among low income females in Canada [31] and in low income

households in developing countries [32-34].

The median number of servings (250 ml) of water consumed was 4 with a

range of 0 to 32 servings. This result is similar to those from water consumption

studies from Canada where the median number of servings was twice reported

as 4 (250 ml servings) [22, 23] and the mean volume of water consumed was

1.39L [24]. The multiple significant interactions and variables in the final model

predicting the amount of water consumed underline the complexity in the factors

associated with water consumption. In general, the number of servings

consumed was higher among males, respondents less than 10 years old, non-

bottled water users, respondents in phase 2 of the study, those with a higher

153

level of education in phase 1, and those in the lower socio-economic levels in

phase 2. Water consumption studies from Canada, Sweden and Italy found that

increased water consumption was associated with being female [23, 24, 35] and

being a bottled water user [23, 24] all of which are opposite to the findings

reported here. These study differences may indicate country-specific or cultural

variations in water consumption patterns, including preference for bottled water

that may reflect necessity (i.e., lack of potable water) in some locations versus

convenience, and factors that affect amount of water consumed. This

underscores the need for population-specific estimates to ensure accuracy in risk

assessments.

Furthermore, decreasing age was associated with increased water

consumption as was seen in Canadian studies [22, 23] but adults 70 years and

older and 65 years and older consumed the largest amount of water in Sweden

[35] and Italy [36], respectively. Additionally, the use of water treatment devices

was significantly associated with increased consumption of water in Canada [22,

23] in contrast to the results of this study. Once again, these international

differences highlight the need for population and region-specific estimates.

Increased level of education was associated with increased water

consumption in phase 1, as was also observed in Canadian studies [22-24], and

in the United States where more educated, older adults were more likely than

younger and less educated adults to have a healthy diet pattern including high

water consumption [37].

154

Within-home water treatment was not practiced by over 80% of

respondents in this study. Furthermore, no bottled water was consumed in the

day prior to interview by more than half (54%) of the respondents, while 18% of

respondents were classified as bottled water users, consuming 75% or more of

their total daily water intake as bottled water. This may be a reflection of the high

proportion of respondents with a municipal water supply assumed to be of

adequate quality, or the potential lack of in-home water filtration devices available

and affordable for the general population in Chile. In Canadian water

consumption studies, roughly 50% of respondents did not employ in-home water

treatment, while the bottled water results are similar to those presented here [22-

24]. This information could be used in waterborne disease and drinking water-

related risk assessments to better describe the potential exposures of specific

populations.

Potential limitations of this study include the inability to evaluate the role of

unmeasured variables such as participant weight, physical activity, and health

condition as well as weather conditions which may have contributed to a more

complete model of the key determinants of water consumption. Water-related

survey questions pertained to 'cold' water consumption and did not include tea or

coffee or other hot beverages because we assumed that boiling would affect the

probabilities of exposure to waterborne microbiological hazards. Future studies

may wish to include consumption of these other types of water to assess

exposure to other types of hazards (e.g., chemicals, heavy metals etc.). Though

93% of respondents claimed that the reported amount of water consumed in the

155

day prior to interview reflected their normal daily consumption, it is likely that a 24

hour recall does not fully capture water consumption patterns, which is another

limitation of this study.

Conclusions

In conclusion, this study provides insight into the water consumption

patterns of the general population in the Metropolitan region, Chile, and its

associations with Gl. Average daily water consumption was 1.2 I. Age, gender,

socio-economic level, education, occupation, being a bottled water user and

study phase all were significantly associated with number of servings of water

consumed. Approximately 8% of the survey population experienced symptoms

of Gl in the 30 days prior to interview. In addition to socio-demographic, study

phase and health insurance system variables, the use of water treatment (i.e.,

boiling) and increasing amount of water consumed were associated with Gl. The

majority of respondents received water from a municipal water supply and did not

treat their water. Bottled water use was common in less than one quarter of the

respondents. This information can be used by waterbome disease risk

managers and public health officials to better describe and understand potential

drinking water health hazards.

156

References

1. World Health Organization: Guidelines for Drinking-water quality, incorporating first and second addenda. Volume 1 Recommendations. 2008:515.

2. World Health Organization, UNICEF: Meeting the MDG drinking water and sanitation target: the urban and rural challenge of the decade. 2006:42.

3. Yoder JS, Hlavsa MC, Craun GF, Hill V, Roberts V, Yu PA, Hicks LA, Alexander NT, Calderon RL, Roy SL, Beach MJ, Centers for Disease Control and Prevention (CDC): Surveillance for waterborne disease and outbreaks associated with recreational water use and other aquatic facility-associated health events-United States, 2005-2006. MMWR Surveill Summ 2008, 57(9):1-29.

4. Yoder J, Roberts V, Craun GF, Hill V, Hicks LA, Alexander NT, Radke V, Calderon RL, Hlavsa MC, Beach MJ, Roy SL, Centers for Disease Control and Prevention (CDC): Surveillance for waterborne disease and outbreaks associated with drinking water and water not intended for drinking-United States, 2005-2006. MMWR Surveill Summ 2008, 57(9):39-62.

5. Sailaja B, Murhekar MV, Hutin YJ, Kuruva S, Murthy SP, Reddy KS, Rao GM, Gupte MD: Outbreak of waterborne hepatitis E in Hyderabad, India, 2005. Epidemiol Infect 2009, 137(2):234-240.

6. Franklin LJ, Fielding JE, Gregory J, Gullan L, Lightfoot D, Poznanski SY, Vally H: An outbreak of Salmonella Typhimurium 9 at a school camp linked to contamination of rainwater tanks. Epidemiol Infect 2009, 137(3):434-440.

7. Bhunia R, Hutin Y, Ramakrishnan R, Pal N, Sen T, Murhekar M: A typhoid fever outbreak in a slum of South Dumdum municipality, West Bengal, India, 2007: evidence for foodborne and waterborne transmission. BMC Public Health 2009, 9:115.

8. Jones JL, Dubey JP: Waterborne toxoplasmosis - Recent developments. Exp Parasitol 2009, .

9. Centers for Disease Control and Prevention (CDC): Outbreak of cryptosporidiosis associated with a splash park - Idaho, 2007. MMWR Morb Mortal Wkly Rep 2009, 58(22):615-618.

10. Hrudey SE, Payment P, Huck PM, Gillham RW, Hrudey EJ: A fatal waterborne disease epidemic in Walkerton, Ontario: comparison with other

157

waterborne outbreaks in the developed world. Water Sci Technol 2003, 47(3):7-14.

11. Hewitt J, Bell D, Simmons GC, Rivera-Aban M, Wolf S, Greening GE: Gastroenteritis outbreak caused by waterborne norovirus at a New Zealand ski resort. Appl Environ Microbiol 2007, 73(24):7853-7857.

12. Kopilovic B, Ucakar V, Koren N, Krek M, Kraigher A: Waterborne outbreak of acute gastroenteritis in a costal area in Slovenia in June and July 2008. Euro Surveill 2008, 13(34):18957.

13. Jakopanec I, Borgen K, Void L, Lund H, Forseth T, Hannula R, Nygard K: A large waterborne outbreak of campylobacteriosis in Norway: the need to focus on distribution system safety. BMC Infect Dis 2008, 8:128.

14. Walls I, International Life Sciences Institute Research Foundation: Framework for identification and collection of data useful for risk assessments of microbial foodborne or waterborne hazards: a report from the International Life Sciences Institute Research Foundation Advisory Committee on data collection for microbial risk assessment. J Food Prot 2007, 70(7):1744-1751.

15. UNICEF, World Health Organization: Progress on drinking water and sanitation. Special focus on sanitation. 2008, :.

16. Thomas MK, Perez E, Majowicz SE, Reid-Smith R, Olea A, Diaz J, Solari V, McEwen SA: Burden of acute gastrointestinal illness in the Metropolitan region, Chile, 2008. Epidemiol Infect (Accepted 2010).

17. Thomas MK, Perez E, Majowicz SE, Reid-Smith R, Albil S, Monteverde M, McEwen SA: Burden of acute gastrointestinal illness in Galvez, Argentina, 2007. J Health Popul Nutr. 2010; 28:149-158.

18. Aguiar Prieto P, Finley RL, Muchaal PK, Guerin MT, Isaacs S, Castro Dominguez A, Gisele Coutin M, Perez E: Burden of self-reported acute gastrointestinal illness in Cuba. J Health Popul Nutr 2009, 27:345-57.

19. Majowicz SE, Dore K, Flint JA, Edge VL, Read S, Buffett MC, McEwen S, McNab WB, Stacey D, Sockett P, Wilson JB: Magnitude and distribution of acute, self-reported gastrointestinal illness in a Canadian community. Epidemiol Infect 2004,132(4):607-617.

20. Thomas MK, Majowicz SE, MacDougall L, Sockett PN, Kovacs SJ, Fyfe M, Edge VL, Flint JA, Henson S, Jones AQ: Population distribution and burden

158

of acute gastrointestinal illness in British Columbia, Canada. BMC Public Health 2006, 6:307.

21. Sargeant JM, Majowicz SE, Snelgrove J: The burden of acute gastrointestinal illness in Ontario, Canada, 2005-2006. Epidemiol Infect 2008;136(4):451-460.

22. Jones AQ, Dewey CE, Dore K, Majowicz SE, McEwen SA, Waltner-Toews D: Drinking water consumption patterns of residents in a Canadian community. J Water Health 2006, 4(1 ):125-138.

23. Jones AQ, Majowicz SE, Edge VL, Thomas MK, MacDougall L, Fyfe M, Atashband S, Kovacs SJ: Drinking water consumption patterns in British Columbia: an investigation of associations with demographic factors and acute gastrointestinal illness. Sci Total Environ 2007, 388(1 -3):54-65.

24. Pintar KD, Waltner-Toews D, Charron D, Pollari F, Fazil A, McEwen SA, Nesbitt A, Majowicz S: Water consumption habits of a south-western Ontario community. J Water Health 2009, 7(2):276-292.

25. Payment P, Richardson L, Siemiatycki J, Dewar R, Edwardes M, Franco E: A randomized trial to evaluate the risk of gastrointestinal disease due to consumption of drinking water meeting current microbiological standards. Am J Public Health 1991, 81(6):703-708.

26. Khalakdina A, Vugia DJ, Nadle J, Rothrock GA, Colford JM.Jr: Is drinking water a risk factor for endemic cryptosporidiosis? A case-control study in the immunocompetent general population of the San Francisco Bay Area. BMC Public Health 2003, 3:11.

27. de Wit MA, Koopmans MP, Kortbeek LM, Wannet WJ, Vinje J, van Leusden F, Bartelds Al, van Duynhoven YT: Sensor, a population-based cohort study on gastroenteritis in the Netherlands: incidence and etiology. Am J Epidemiol 2001, 154(7):666-674.

28. Jones TF, McMillian MB, Scallan E, Frenzen PD, Cronquist AB, Thomas S, Angulo FJ: A population-based estimate of the substantial burden of diarrhoeal disease in the United States; FoodNet, 1996-2003. Epidemiol Infect 2007, 135(2):293-301.

29. Scallan E, Fitzgerald M, Collins C, Crowley D, Daly L, Devine M, Igoe D, Quigley T, Robinson T, Smyth B: Acute gastroenteritis in northern Ireland and the Republic of Ireland: a telephone survey. Commun Dis Public Health 2004, 7(1 ):61-67.

159

30. Hall GV, Kirk MD, Ashbolt R, Stafford R, Lalor K: Frequency of infectious gastrointestinal illness in Australia, 2002: regional, seasonal and demographic variation. Epidemiol Infect 2006, 134(1 ):111-118.

31. Majowicz SE, Horrocks J, Booking K: Demographic determinants of acute gastrointestinal illness in Canada: a population study. BMC Public Health 2007,7:162.

32. Woldemicael G: Diarrhoeal morbidity among young children in Eritrea: environmental and socioeconomic determinants. J Health Popul Nutr 2001, 19(2):83-90.

33. Hatt LE, Waters HR: Determinants of child morbidity in Latin America: a pooled analysis of interactions between parental education and economic status. Soc Sci Med 2006, 62(2):375-386.

34. Genser B, Strina A, Teles CA, Prado MS, Barreto ML: Risk factors for childhood diarrhea incidence: dynamic analysis of a longitudinal study. Epidemiology 2006, 17(6):658-667.

35. Westrell T, Andersson Y, Stenstrom TA: Drinking water consumption patterns in Sweden. J Water Health 2006, 4(4):511-522.

36. Turrini A, Saba A, Perrone D, Cialfa E, D'Amicis A: Food consumption patterns in Italy: the INN-CA Study 1994-1996. Eur J Clin Nutr 2001, 55(7):571-588.

37. Popkin BM, Barclay DV, Nielsen SJ: Water and food consumption patterns of U.S. adults from 1999 to 2001. Obes Res 2005,13(12):2146-2152.

160

Table 5.1: Age, gender and education distribution of Metropolitan region residents and survey respondents, Chile 2008.

Variable

Age (years) 0-4 5-9

10-19 20-59

60+ Sex

Male Female



Metropolitan region residents (N=6,061,185)

451,995 (7.5%) 511,864 (8.4%) 1,046,091 (17.3%) 3,381,732(55.8%) 669,543(11.0%)

2,937,193(48.5%) 3,123,992(51.5%)

NA 1,887,649(31.1%) 2,167,683(35.8%) 503,105(8.3%) 707,563(11.7%)

Survey respondents (n=6047)

96(1.6%)* 107(1.8%)* 545 (9.0%)* 4361 (72.1%)* 879(14.5%)*

2594 (42.9%)* 3451 (57.1%)*

35 (0.6%) 978(16.2%)* 2398 (39.7%)* 1176(19.4%)* 1247(20.6%)*

* Indicates significant (p<0.05) difference between survey respondents and residents.

161

Table 5.2: Mean number of drinking water servings consumed, by occupation, adjusted for other variables in the multivariable model, Metropolitan region, Chile, 2008 (n=5765).

Occupation General employer Self-employed Housewife Student Private sector Child not yet in school Unemployed Retired Public sector

Number of drinking water servings

5.51 4.98 4.81 4.80 4.60 4.57 4.55 4.46 4.37

95% CI (4.98, 6.08) (4.61,5.38) (4.45, 5.19) (4.48, 5.14) (4.27, 4.95) (4.08, 5.12) (4.15,5.00) (4.08, 4.88) (4.05, 4.72)

162

Tab

le 5

.3:

Fin

al m

ultiv

aria

ble

logi

stic

reg

ress

ion

mod

el o

f ris

k fa

ctor

s fo

r G

l and

pro

port

ion

of r

espo

nden

ts w

ith G

l by

risk

fact

or,

Met

ropo

litan

reg

ion,

Chi

le, 2

008

(n=5

709)

.

Var

iab

le

Leve

l O

R (

95%

CI)

P

rop

ort

ion

wit

h G

l (95

% C

I)

P-v

alu

e S

tudy

pha

se

Occ

upat

ion

1 2 Une

mpl

oyed

G

ener

al e

mpl

oyer

S

tude

nt

Ret

ired

Chi

ld n

ot ye

t in

sch

ool

Priv

ate

sect

or

Sel

f-em

ploy

ed

Pub

lic s

ecto

r H

ouse

wife

W

ater

tre

atm

ent

Age

Hea

lth s

yste

m

Filt

er

Boi

ling

Not

hing

0-4

5-9

10-1

9 20

-59

60+

1.25

(1.0

1, 1

.54)

R

efer

ent

1.51

(0.

90,2

.53)

1.

71 (

0.96

,3.0

6)

1.6

3(1

.11

,2.3

9)

1.3

5(0

.81

,2.2

6)

1.5

7(0

.73

,3.3

8)

1.20

(0.8

2, 1

.75)

0.

80(0

.52,

1.2

4)

1.4

5(0

.99

,2.1

2)

Ref

eren

t

0.93

(0.5

7, 1

.54)

1

.98

(1.5

2,2

.58

) R

efer

ent

2.8

3(1

.36

,5.8

8)

1.01

(0.

47,2

.15)

1

.50

(1.0

5,2

.14

) R

efer

ent

0.91

(0.

61, 1.

34)

0.1

16

(0.0

83

,0.1

59

) 0

.09

5(0

.06

7,0

.13

3)

0.1

18

(0.0

70

,0.1

93

) 0

.13

2(0

.07

4,0

.22

4)

0.1

26

(0.0

90

,0.1

74

) 0

.10

7(0

.06

3,0

.17

4)

0.1

22

(0.0

66

,0.2

14

) 0

.09

6(0

.06

2,0

.14

4)

0.0

66

(0.0

41

,0.1

06

) 0

.11

3(0

.07

5,0

.16

9)

0.0

81

(0.0

53

,0.1

23

)

0.0

82

(0.0

47

,0.1

37

) 0

.15

9(0

.11

3,0

.22

0)

0.0

87

(0.0

65

,0.1

17

)

0.2

02

(0.1

09

,0.3

44

) 0

.08

3(0

.03

9,0

.16

6)

0.1

18

(0.0

79

,0.1

72

) 0

.08

2(0

.06

1,0

.10

9)

0.0

75

(0.0

50

,0.1

12

)

0.04

0.04

<0.

01

0.03

0.02

163

Var

iab

le

Lev

el

OR

(95

% C

I)

Pro

po

rtio

n w

ith

Gl (

95%

CI)

P-v

alu

e M

ilita

ry

Priv

ate

- S

yste

m

No

insu

ranc

e P

rivat

e -

Indi

vidu

al

Pub

lic

Soc

io-e

cono

mic

leve

l*f

A

B

C

D

E

Gen

der*

F

emal

e M

ale

Soc

io-e

cono

mic

leve

l*G

ende

r A

*Fem

ale

A*M

ale

B*F

emal

e B

*Mal

e C

*Fem

ale

C*M

ale

D*F

emal

e D

*Mal

e E

*Fem

ale

E*M

ale

Per

ser

ving

of w

ater

con

sum

ed

1.0

5(0

.54

,2.0

7)

0.7

3(0

.55

,0.9

7)

1.5

3(1

.09

,2.1

7)

0.90

(0.

35, 2

.30)

R

efer

ent

Ref

eren

t 1.

24(0

.90,

1.7

1)

0.9

7(0

.68

, 1.

38)

1.35

(0.9

5, 1

.91)

1.

20(0

.80,

1.8

0)

1.17

(0.9

3, 1

.47)

R

efer

ent

Ref

eren

t R

efer

ent

1.3

5(0

.87

,2.0

7)

1.1

3(0

.71

,1.8

1)

1.51

(0

.97

,2.3

4)

0.6

3(0

.37

, 1.

06)

1.3

3(0

.82

,2.1

4)

1.3

6(0

.84

,2.2

2)

1.8

3(1

.11

,3.0

3)

0.7

9(0

.43

, 1.

45)

1.10

(1.0

6, 1

.13)

0.1

09

(0.0

57

,0.2

00

) 0

.07

8(0

.05

5,0

.10

9)

0.15

1 (0

.107

,0.2

09)

0.0

94

(0.0

38

,0.2

15

) 0

.10

4(0

.07

9,0

.13

4)

0.9

30

(0.0

62

,0.1

36

) 0

.11

2(0

.07

9,0

.15

8)

0.09

1 (0

.062

,0.1

31)

0.12

1 (0

.083

,0.1

74)

0.1

10

(0.0

73

,0.1

61

)

0.1

12

(0.0

80

,0.1

56

) 0

.09

8(0

.06

9,0

.13

6)

0.0

84

(0.0

53

,0.1

31

) 0

.10

3(0

.06

5,0

.15

9)

0.1

10

(0.0

74

,0.1

60

) 0

.11

5(0

.07

7,0

.16

8)

0.1

22

(0.0

82

,0.1

76

) 0

.06

7(0

.04

1,0

.10

6)

0.1

09

(0.0

70

,0.1

65

) 0

.13

5(0

.08

8,0

.20

1)

0.1

44

(0.0

93

,0.2

16

) 0

.08

3(0

.04

9,0

.13

7)

0.22

0.19

<0.0

1

<0.0

1

164

* V

aria

bles

par

t of

inte

ract

ion

term

. t

Soc

io-e

cono

mic

lev

el, w

ith A

= H

igh

and

E=L

ow.

165

Figure 5.1: Estimated mean number of servings of water consumed in the 24 hours prior to interview by gender and study phase from multivariable model with 95% confidence interval error bars, Metropolitan region, Chile, 2008 (n=5765).

D) C

E w

n E 3 Z

6 -

5 -

4 -

3

2

1 -

0 -

«!!• Phase 1

-•—Phase 2 i i — -

• — i

— -i

Female Male

Gender

166

Figure 5.2: Estimated mean number of servings of water consumed in the 24 hours prior to interview by study phase and education level with from multivariable model with 95% confidence interval error bars, Metropolitan region, Chile, 2008 (n=5765).

8

7 in

D) 6

I 5 in

O ^ i .

f 2 z

1

0

Phase 1

• Phase 2

• * - + -&- ~ -̂

III iterate Not applicable (<6 years old)

Primary Secondary

Education Level

Technical University

167

Figure 5.3: Estimated mean number of servings of water consumed in the 24 hours prior to interview by gender and age category from multivariable model with 95% confidence interval error bars, Metropolitan region, Chile, 2008 (n=5765).

8 -

7 -

TO 6 -

0) Q Si O -E 3 2

1 -

o -

_ ^ _ Female

—•—Male

•f • Y —

0-4 5-9

i r

10-19

Age

20-59

___£

I

60+

168

Figure 5.4: Estimated mean number of servings of water consumed in the 24 hours prior to interview by socio-economic level and study phase and being a bottled water user from multivariable model, Metropolitan region, Chile, 2008 (n=5765).

8 -

7 -

3 "K 6 -

**-( f l o ra c 'E 4 -V (A **-

V

n

I 2-z

1

- - A -

A'

A

Phase 1 - Bottled water user * —• Phase 1 - Non-bottled water user

—=~^£i— ~~~* ~ * " ^ ^ '

• • • • ' ' ' • * . • • • " ' • • - . B

. , . - ' • • ' ' • - - . . . . • • ' & . = • — — • —

• • ~ . . ~ „ ^ , . . . — " • #

B C D E

Socio-economic level

169

Figure 5.5: Probability of being a case of Gl by socio-economic level and gender from multivariable model with 95% confidence interval error bars, Metropolitan region, Chile, 2008 (5709).

0.25 -

0.2 -O

o 0.15 -2?

Pro

bab

il

p

0.05

0 -

•>- Female

—m— Male

I ^ ^

1

A B

„.„....._..—•-- ̂ *••- ^ ^

r-~~~ ..

C D

Socio-economic level

„.-'*• *

E

170

CHAPTER SIX

Estimated numbers of human infections due to Salmonella, Campylobacter

and Shigella, Metropolitan region, Chile, 2008

Prepared for submission: BMC Public Health

Abstract

Background: Acute gastrointestinal illness (Gl) causes global morbidity and

mortality; however, the true pathogen-specific burden at the community level is

not always known. As part of a larger burden of Gl study, we estimated the

number of human infections due to Salmonella, Campylobacter and Shigella in

the Metropolitan region of Chile, 2008, using under-reporting multipliers. The

Instituto de Salud Publica (ISP) reference laboratory collated the number of

pathogen-specific infections in Chile and provided the total number of reported

laboratory-confirmed infections due to Salmonella, Campylobacter and Shigella

in the Metropolitan region of Chile, 2008.

Methods: The proportions of cases expected to progress at each step of the

reporting system were used as input values in a stochastic model. For some

steps, Gl data were used when pathogen-specific data were not available. To

better approximate Gl symptoms by pathogen, data from international outbreaks

were used to estimate the proportion of bloody diarrhea by pathogen. The

171

inverse of the proportions at each step were multiplied together to derive the

overall estimated under-reporting multiplier. The number of pathogen-specific

laboratory confirmed cases from ISP was multiplied by the estimated under

reporting multiplier. The model outcome was expressed as the mean under

reporting multiplier with 5th and 95th percentiles to illustrate the uncertainty in the

estimate.

Results: The overall estimated mean multiplier (5th - 95th percentile) for

Salmonella infections was 66 (30 - 129), for Campylobacter 470 (179 - 1006),

and for Shigella 77 (35 - 148). The estimated incidence rates in 2008 per

100,000 person-years were 1140 (513 - 2227), 233 (89 - 498), and 278 (126 -

538) for Salmonella, Campylobacter and Shigella, respectively.

Conclusions: A significant number of infections occur in the Metropolitan

region, Chile due to Salmonella, Campylobacter and Shigella, and under

reporting of their true burden exists. Better understanding of the true burden can

help to mitigate the burden and inform prevention and control activities.

Background

Acute gastrointestinal illness (Gl) causes significant morbidity, mortality,

and socio-economic burden around the world [1-3]. Outbreak reports and

surveillance data provide estimates of this burden; however, these are typically

subject to under-reporting [4]. To address this, numerous countries have

conducted population-based studies to better estimate the burden of Gl [5-16].

Reportable disease data can be adjusted by under-reporting multipliers

172

generated, in part, from population-based survey information, to better estimate

the pathogen-associated burden in the population [17-20].

In Chile, the Instituto de Salud Publica national reference laboratory (ISP)

is legislated to record the number of samples tested and confirmed positive for

select pathogens, including Salmonella, Campylobacter and Shigella. In order to

be captured in this database, an ill individual must seek medical care, have a

stool sample requested and submit a stool sample for testing. Local laboratories

conduct preliminary testing and send samples as appropriate to ISP for further

testing or confirmation of results. Typically, compliance with each of these steps

is less than perfect, thus, the database is subject to under-estimation of the true

number of cases of illness caused by specific pathogens.

In 2008, a partnership of the Pan-American Health Organization, the

Public Health Agency of Canada, the University of Guelph and the Ministry of

Health in Chile completed the first population burden of Gl study in the

Metropolitan region of Chile (Chapter 3, [21]) and its main objective was to

determine the burden of Gl in the population. Data arising from that burden of

illness study were used to support the present study, whose purpose was to

estimate the number of human infections due to Salmonella, Campylobacter and

Shigella in the Metropolitan region of Chile, 2008.


Analytic approach

173

For each pathogen, data from various sources were used to estimate the

proportion of cases reported at each step of the reporting system. Where

pathogen-specific data were not available, data on Gl were used, accounting for

differences in severity (i.e., bloody versus non-bloody diarrhea). A stochastic,

probability distribution approach taking into account inherent uncertainty of the

input values at each step, was used (Table 6.1). To generate population-level,

pathogen-specific estimates, the numbers of cases confirmed by the ISP for the

Metropolitan region for each pathogen were multiplied by estimated multiplication

factors, generated by using Equation 1 adapted from Thomas et al [18].

Equation 1:

1 (pBD x ai x b\x ci x d\x e\x f\) + (pNBD x ai x bi x ci x di x ei x fi)

Where:

y = Estimated under-reporting multiplier

pBD = Proportion of cases with bloody diarrhea

pNBD = Proportion of cases with non bloody diarrhea

a = Proportion of cases that visit a physician*

b = Proportion of cases that have a stool sample requested of them*

c = Proportion of cases that submit a stool sample*

d = Proportion of laboratories that test for pathogen

e = Sensitivity of laboratory pathogen test

f = Proportion of samples that are forwarded to ISP for confirmation or further

testing

174

* Where 1 and 2 represent the proportions for bloody and non bloody diarrhea,

respectively

Stochastic simulations were performed using @RISK version 5.0

(Palisade Corporation, Ithaca, New York) as an add-on to Microsoft Excel

(Microsoft Corporation, 2003) with 20 000 iterations and Latin Hypercube

sampling. The mean, median, 5th and 95th percentile values for the estimated

multipliers, number of cases and incidence per 100,000 person-years were

reported. Calculation of input values was performed using SAS 9.1 (SAS

Institute Inc., Cary North Carolina, 2004). Input distributions (type and input

parameters) are described in Table 6.1.

A sensitivity analysis was done to determine which input distributions had

the greatest influence on the overall under-reporting multiplier, by pathogen. The

correlation between each input distribution with the under-reporting multiplier was

calculated. The five most influential input distributions by pathogen are

presented.

Data

Data from a recently conducted population survey (Chapter 3, [21]) were

used to determine the proportion of cases that sought medical care, had a stool

sample requested and submitted a stool sample in the Metropolitan region, Chile

(Table 6.1, steps A-C) [21]. Individuals responding 'don't know' or 'unsure' were

excluded from the analysis of that question. Proportions were segregated by

cases experiencing bloody versus non-bloody diarrhea. For this analysis, cases

of Gl were defined as people experiencing three or more loose stools within a 24

175

hour period, in the 30 days prior to interview. Individuals who reported that their

symptoms were due to a chronic or medical condition were excluded from the

case group. The expected range in proportion of cases that experience bloody

diarrhea for each pathogen were derived from published reports of outbreaks

(Table 6.1, step G) [22-39].

A laboratory survey was conducted as part of the burden of Gl study

conducted in the Metropolitan region, Chile (Appendix VI). In total, 106

laboratories from the Metropolitan region were contacted to participate in the

survey and responses were obtained from 49 (46%). Questions in the survey

pertained to the number of samples received, tested and found positive during

July 1-31 and November 15-December 15, 2008 (these time periods were

selected to coincide with the timing of the above population survey). Information

in the survey was used to determine the estimated proportion of laboratories that

routinely (80%-100% of the time) test for each pathogen, and the proportion of

positive samples that are referred to ISP for confirmation or further testing, by

pathogen (Table 6.1, steps D, F). The ranges in sensitivity of laboratory tests, by

pathogen were obtained from international literature (Table 6.1, step E) [40-42].

Results

Table 6.2 shows the reported and estimated annual number of human

cases of Salmonella, Campylobacter and Shigella infection in the Metropolitan

region, Chile in 2008. For every case of Salmonella infection reported by ISP

there was an estimated annual average of 66 (5th and 95th percentile: 30 - 129)

176

cases of Salmonella in 2008 in the Metropolitan region, Chile, corresponding to

an incidence rate of 1140 (5th and 95th percentile: 513 - 2227) cases per 100,000

person-years. Likewise, for every case of Campylobacter and Shigella infection

reported by ISP there were estimated annual averages of 470 (5th and 95th

percentile: 179 - 1006) and 76 (5th and 95th percentile: 35 - 148) cases in 2008

in the Metropolitan region, Chile, respectively. These values correspond to

incidence rates of 233 (5th and 95th percentile: 89 - 498) and 278 (5th and 95th

percentile: 126 - 538) cases per 100,000 person-years.

From the sensitivity analysis, the proportion of stool samples requested

(Table 6.1, step B) had the highest correlation with the estimated under-reporting

multiplier for all three of the pathogens (Table 6.3). Thus this input distribution

contributed the most to the uncertainty of the overall under-reporting estimate for

each of the pathogens.

Discussion

This study provides population estimates of three important enteric

pathogens using stochastic methods for the Metropolitan region, Chile in 2008.

To date, similar results have been published from the United States, England

Canada, Australia, Jordan, and Japan as well as an international estimate of the

global burden of Salmonella (Table 6.4) [15, 17-20, 43-45]. Assessing

differences in annual incidence estimates is useful for comparing relative

pathogen

177

The mean incidence rate of Shigella infection reported here is similar to

the rate reported from Jordan [20]; however, it is higher than the rates reported

from the United States [44] and England [15]. This may be due to a true

difference in the incidence of Shigella infection in these different populations. It

is also possible that differences in methodology may, in part, explain the variation

between study results. The study in England used prospective cohort

methodology, which can lead to lower estimates of burden compared to cross-

sectional study designs, possibly due to the requirement that ill study participants

to submit stool samples [15]. The study from the United States [44] combined

data from outbreak reports with those from active and passive surveillance

reports to estimate the average annual number of reported cases, to which a

multiplier of 20 was applied. This multiplier was used for any pathogen typically

associated with bloody diarrhea and was based on the under-reporting of

Escherichia coli 0157:H7 found in a previous internal report [46]. Consequently,

it may not have fully accounted for the under-reporting of Shigella in the

American study. Furthermore, we utilised a stochastic model with distributions to

account for uncertainty in the estimates, which generated a mean value higher

than the median value.

The mean incidence rate of Campylobacter infection reported here is

similar to that reported by Japan [43], but lower than those reported by the United

States [44], England [15], Canada [18] and Australia [19]. It is possible that the

incidence of Campylobacter infection is lower in Chile compared to these

locations; however, this low incidence is likely, to some extent, a reflection of the

178

low percentage (i.e., 16.3%) of laboratories that routinely test for Campylobacter

in the Metropolitan region, Chile. From the sensitivity analysis, the second most

influential input distribution was the percentage of laboratories that test for

Campylobacter, contributing to the uncertainty of the overall under-reporting

multiplier for Campylobacter.

The mean population-based incidence rate of Salmonella infection

estimated here (i.e., 1140 per 100,000 person-years) is equal to the estimate of

the global burden of Salmonella [45]. However, this international estimate was

based on regional estimates, including one of 80 cases per 100,000 person

years for the Latin American - South region. Note that our estimate for Chile is

considerably higher than this regional estimate. Due to a lack of studies such as

our recent work (Chapter 3, [21]), the global burden of Salmonella study relied on

data from a returning traveler survey, with adjustment for under-ascertainment

and differences in susceptibility of residents versus travelers, to generate this

value. These differences in methodology may in part explain the differences in

results and highlight the importance of region-specific data. Moreover, the mean

estimated rate for Chile is higher than all other reported studies [15, 17-20, 43-

45]. This may reflect a true difference in the burden of Salmonella by country or

region, or perhaps a difference in healthcare seeking or laboratory reporting

behaviours such that the under-reporting of Salmonella is much greater in Chile

than other regions. This may be due to differences in access to healthcare or

public health and laboratory surveillance system designs.

179

In the case of Salmonella, a large proportion (i.e., 83.7%) of laboratories in

the Metropolitan region routinely tested for Salmonella and the ISP confirmed

incidence rate of Salmonella infections was 17.2 per 100,000 person-years in

2008, in the Metropolitan region, Chile. The increased rate of testing likely

contributed to an increased rate of recovery of positive samples and the

confirmed number of Salmonella infections.

The Chilean laboratory confirmed rate of Salmonella (17.2 per 100,000

person-years) is similar to the laboratory reported incidence rate of Salmonella

from Canada (19 per 100,000 person-years [18] and the United States (13.6 per

100,000 person-years [47], thus the differences in estimated rates may be

attributed to differences in healthcare seeking behaviours in the different

populations or different estimate methodologies. Our estimates incorporated

uncertainty with distributions at each step, while the Canada study provided a low

and high estimated value, and the United States study generated a point

estimate. Furthermore, the estimates from the United States used a point

estimate of 50.3% of cases to have bloody diarrhea, while a range of values were

used here and in the Canadian study.

From the sensitivity analysis, healthcare seeking behaviours were most

influential on the uncertainty of the overall under-reporting multiplier for both

Salmonella and Shigella, with four of the top five measures of correlation related

to the under-reporting multiplier (Table 6.3). For Campylobacter, the second and

third most influential input distributions were related to laboratory testing and

180

reporting to ISP. Further research would be useful to provide better data on

these steps.

Pathogen-specific rates can be useful for public health policy makers. It is

possible that the prioritization or ranking of pathogens, for the purposes of

control, prevention or surveillance activities may differ depending upon which

estimates are used: estimated population rates versus reported laboratory

confirmed rates. Laboratory confirmed rates are only a subset of the true

pathogen-specific burden in the population. For example, pathogens previously

not given high priority may become more important to public health officials when

the population level estimates are known. This can result in increased laboratory

testing or specific studies to better understand their burden in the community. In

Chile, this may prove useful in the case of Campylobacter as it appears to be

considerably under-reported in this population and had the lowest rate of

laboratory testing of the three pathogens examined. Additionally, assessing

country-specific under-reporting multipliers by pathogen enables comparison of

reporting and surveillance systems which may help to explain where differences

exist between pathogen-specific international incidence estimates.

We attempted to estimate the population-level burden of Escherichia coli

0157:H7; however, due to scarcity of data at the laboratory level the range in

estimated values exceeded what was considered sensible and biologically

plausible (data not shown). It would be useful to explore this in future research

activities.

181

There are some limitations of this study. The low response rate (46%) on

the laboratory survey. This could have created non-response bias if the activities

of the responding laboratories differed from those of the non-responding

laboratories. The location of the responding laboratories was geographically

representative of the Metropolitan region (data not shown). The ranges of

laboratory test sensitivities and proportions of cases that suffer bloody diarrhea,

by pathogen were based on those observed in published studies in other

countries and are similar to those used in studies published from the United

States, Canada and Australia [17-19]. The use of outbreak data may not reflect

the true distribution of severity of sporadic cases in the population. Less severe

cases may be captured due to the recognized outbreak (i.e., whereas under non-

outbreak conditions they may not have been identified) or the opposite could

occur where an outbreak is identified because of the extreme nature of the

symptoms of individual cases, thus shifting the severity distribution. No

adjustment was made for test specificity and the possibility of false positives due

to insufficient information (i.e., we do not know the true prevalence nor the

number of test negatives). This is a common limitation to these types of studies

and can present an issue when the specificity of a test is poor, otherwise the

impact of false positives would be small. Based on previous publications, test

specificity for Salmonella and Shigella is 100% but test specificity for

Campylobacter ranges from 32.5%-99%, and could therefore have a large impact

on the Campylobacter incidence rate [40-42].

182

To account for differences in case behaviour given severity of illness, we

segregated our analysis by bloody and non-bloody diarrhea in an attempt to

better describe pathogen-specific illness behaviours as this information is not

currently available for Chile. However, this segregation created small numbers

for some of the calculations and thus contributed to the wider range in values

obtained.

Conclusions

This study provides the first population-based estimates of enteric

bacterial infections in Chile. The uncertainty of these estimates was captured

using the 5th and 95th percentile values as upper and lower bounds to the

estimates. These results will be of use not only to local public health workers

and policy makers and should enable international comparisons that better reflect

the true population burden by pathogen. This sort of information has been

sought after by the World Health Organization and will be useful in generating

burden of disease estimates [3, 48].

183

References

1. Guerrant RL, Kosek M, Moore S, Lorntz B, Brantley R, Lima AA: Magnitude and impact of diarrheal diseases. Arch Med Res 2002, 33(4):351-355.

2. Kosek M, Bern C, Guerrant RL: The global burden of diarrhoeal disease, as estimated from studies published between 1992 and 2000. Bull World Health Organ 2003, 81(3):197-204.

3. Stein C, KuchenmullerT, Hendrickx S, Pruss-Ustun A, Wolfson L, Engels D, Schlundt J: The Global Burden of Disease assessments-WHO is responsible? PLoS Negl Trap Dis 2007, 1(3):e161.

4. Scallan E, Majowicz SE, Hall G, Banerjee A, Bowman CL, Daly L, Jones T, Kirk MD, Fitzgerald M, Angulo FJ: Prevalence of diarrhoea in the community in Australia, Canada, Ireland, and the United States. Int J Epidemiol 2005, 34(2):454-460.

5. de Wit MA, Koopmans MP, Kortbeek LM, Wannet WJ, Vinje J, van Leusden F, Bartelds Al, van Duynhoven YT: Sensor, a population-based cohort study on gastroenteritis in the Netherlands: incidence and etiology. Am J Epidemiol 2001, 154(7):666-674.

6. Aguiar Prieto P, Finley RL, Muchaal PK, Guerin MT, Isaacs S, Castro Dominguez A, Gisele Coutin M, Perez E: Burden of self-reported acute gastrointestinal illness in Cuba. J Health Popul Nutr 2009, 27:345-57.

7. Hall GV, Kirk MD, Ashbolt R, Stafford R, Lalor K: Frequency of infectious gastrointestinal illness in Australia, 2002: regional, seasonal and demographic variation. Epidemiol Infect 2006, 134(1 ):111-118.

8. Imhoff B, Morse D, Shiferaw B, Hawkins M, Vugia D, Lance-Parker S, Hadler J, Medus C, Kennedy M, Moore MR, Van Gilder T, Emerging Infections Program Food Net Working Group: Burden of self-reported acute diarrheal illness in FoodNet surveillance areas, 1998-1999. Clin Infect Dis 2004, 38 Suppl 3:S219-26.

9. Jones TF, McMillian MB, Scallan E, Frenzen PD, Cronquist AB, Thomas S, Angulo FJ: A population-based estimate of the substantial burden of diarrhoeal disease in the United States; FoodNet, 1996-2003. Epidemiol Infect 2007, 135(2):293-301.

10. Kuusi M, Aavitsland P, Gondrosen B, Kapperud G: Incidence of gastroenteritis in Norway-a population-based survey. Epidemiol Infect 2003, 131(1):591-597.

184

11. Majowicz SE, Dore K, Flint JA, Edge VL, Read S, Buffett MC, McEwen S, McNab WB, Stacey D, Sockett P, Wilson JB: Magnitude and distribution of acute, self-reported gastrointestinal illness in a Canadian community. Epidemiol Infect 2004, 132(4):607-617.

12. Scallan E, Fitzgerald M, Collins C, Crowley D, Daly L, Devine M, Igoe D, Quigley T, Robinson T, Smyth B: Acute gastroenteritis in northern Ireland and the Republic of Ireland: a telephone survey. Commun Dis Public Health 2004, 7(1):61-67.

13. Sargeant JM, Majowicz SE, Snelgrove J: The burden of acute gastrointestinal illness in Ontario, Canada, 2005-2006. Epidemiol Infect 2008, 136(4): 451-460.

14. Thomas MK, Majowicz SE, MacDougall L, Sockett PN, Kovacs SJ, Fyfe M, Edge VL, Flint JA, Henson S, Jones AQ: Population distribution and burden of acute gastrointestinal illness in British Columbia, Canada. BMC Public Health 2006, 6(307).

15. Wheeler JG, Sethi D, Cowden JM, Wall PG, Rodrigues LC, Tompkins DS, Hudson MJ, Roderick PJ: Study of infectious intestinal disease in England: rates in the community, presenting to general practice, and reported to national surveillance. The Infectious Intestinal Disease Study Executive. BMJ 1999, 318(7190):1046-1050.

16. Herikstad H, Yang S, Van Gilder TJ, Vugia D, Hadler J, Blake P, Deneen V, Shiferaw B, Angulo FJ: A population-based estimate of the burden of diarrhoeal illness in the United States: FoodNet, 1996-7. Epidemiol Infect 2002, 129(1):9-17.

17. Voetsch AC, Van Gilder TJ, Angulo FJ, Farley MM, Shallow S, Marcus R, Cieslak PR, Deneen VC, Tauxe RV, Emerging Infections Program FoodNet Working Group: FoodNet estimate of the burden of illness caused by nontyphoidal Salmonella infections in the United States. Clin Infect Dis 2004, 38 Suppl3:S127-34.

18. Thomas MK, Majowicz SE, Sockett PN, Fazil A, Pollari F, Dore KA, Flint JA, Edge VL: Estimated numbers of community cases of illness due to Salmonella, Campylobacter and verotoxigenic Escherichia coli: Pathogen-specific community rates. Can J Infect Dis Med Microbiol 2006, 17(4):229-234.

19. Hall G, Yohannes K, Raupach J, Becker N, Kirk M: Estimating community incidence of Salmonella, Campylobacter, and Shiga toxin-producing Escherichia coli infections, Australia. Emerg Infect Dis 2008, 14(10):1601-1609.

185

20. Gargouri N, Walke H, Belbeisi A, Hadadin A, Salah S, Ellis A, Braam HP, Angulo FJ: Estimated burden of human Salmonella, Shigella, and Brucella infections in Jordan, 2003-2004. Foodborne Pathog Dis 2009, 6(4):481-486.

21. Thomas MK, Perez E, Majowicz SE, Reid-Smith R, Olea A, Diaz J, Solari V, McEwen SA: Burden of acute gastrointestinal illness in the Metropolitan region, Chile, 2008. Epidemiol Infect (Submitted 2010).

22. Gupta SK, Nalluswami K, Snider C, Perch M, Balasegaram M, Burmeister D, Lockett J, Sandt C, Hoekstra RM, Montgomery S: Outbreak of Salmonella Braenderup infections associated with Roma tomatoes, northeastern United States, 2004: a useful method for subtyping exposures in field investigations. Epidemiol Infect 2007, 135(7): 1165-1173.

23. Ellis A, Preston M, Borczyk A, Miller B, Stone P, Hatton B, Chagla A, Hockin J: A community outbreak of Salmonella berta associated with a soft cheese product. Epidemiol Infect 1998, 120(1 ):29-35.

24. Shane AL, Roels TH, Goldoft M, Herikstad H, Angulo FJ: Foodborne disease in our global village: a multinational investigation of an outbreak of Salmonella serotype Enteritidis phage type 4 infection in Puerto Vallarta, Mexico. Int J Infect Dis 2002, 6(2):98-102.

25. Anonymous Restaurant foodhandler-associated outbreak of Salmonella Heidelberg gastroenteritis identified by calls to a local telehealth service, Edmonton, Alberta, 2004. Can Commun Dis Rep 2005, 31(10):105-110.

26. Centers for Disease Control and Prevention (CDC): Outbreak of Salmonella serotype Muenchen infections associated with unpasteurized orange juice--United States and Canada, June 1999. MMWR Morb Mortal Wkly Rep 1999, 48(27):582-585.

27. Richardson G, Thomas DR, Smith RM, Nehaul L, Ribeiro CD, Brown AG, Salmon RL: A community outbreak of Campylobacter jejuni infection from a chlorinated public water supply. Epidemiol Infect 2007,135(7):1151-1158.

28. Olsen SJ, Hansen GR, Bartlett L, Fitzgerald C, Sonder A, Manjrekar R, Riggs T, Kim J, Flahart R, Pezzino G, Swerdlow DL: An outbreak of Campylobacter jejuni infections associated with food handler contamination: the use of pulsed-field gel electrophoresis. J Infect Dis 2001, 183(1):164-167.

29. Kuusi M, Klemets P, Miettinen I, Laaksonen I, Sarkkinen H, Hanninen ML, Rautelin H, Kela E, Nuorti JP: An outbreak of gastroenteritis from a non-chlorinated community water supply. J Epidemiol Community Health 2004, 58(4):273-277.

186

30. Gillespie IA, O'Brien SJ, Adak GK, Tarn CC, Frost JA, Bolton FJ, Tompkins DS, Campylobacter Sentinel Surveillance Scheme Collaborators: Point source outbreaks of Campylobacter jejuni infection-are they more common than we think and what might cause them? Epidemiol Infect 2003, 130(3):367-375.

31. Centers for Disease Control and Prevention (CDC): Outbreak of Campylobacter enteritis associated with cross-contamination of food-Oklahoma, 1996. MMWR Morb Mortal Wkly Rep 1998, 47(7):129-131.

32. Centers for Disease Control and Prevention (CDC): Outbreak of Campylobacter jejuni infections associated with drinking unpasteurized milk procured through a cow-leasing program-Wisconsin, 2001. MMWR Morb Mortal Wkly Rep 2002, 51(25):548-549.

33. Centers for Disease Control and Prevention (CDC): Shigellosis outbreak associated with an unchlorinated fill-and-drain wading pool-Iowa, 2001. MMWR Morb Mortal Wkly Rep 2001, 50(37):797-800.

34. Davis H, Taylor JP, Perdue JN, Stelma GN,Jr, Humphreys JM.Jr, Rowntree R,3rd, Greene KD: A shigellosis outbreak traced to commercially distributed shredded lettuce. Am J Epidemiol 1988, 128(6):1312-1321.

35. Frost JA, McEvoy MB, Bentley CA, Andersson Y: An outbreak of Shigella sonnei infection associated with consumption of iceberg lettuce. Emerg Infect Dis 1995, 1(1):26-29.

36. Huq I, Alam AK, Morris GK, Wathen G, Merson M: Foodborne outbreak of shigellosis caused by an unusual Shigella strain. J Clin Microbiol 1980, 11(4):337-339.

37. Makintubee S, Mallonee J, Istre GR: Shigellosis outbreak associated with swimming. Am J Public Health 1987, 77(2):166-168.

38. Reller ME, Nelson JM, Molbak K, Ackman DM, Schoonmaker-Bopp DJ, Root TP, Mintz ED: A large, multiple-restaurant outbreak of infection with Shigella flexneri serotype 2a traced to tomatoes. Clin Infect Dis 2006, 42(2):163-169.

39. Werber D, Dreesman J, Feil F, van Treeck U, Fell G, Ethelberg S, Hauri AM, Roggentin P, Prager R, Fisher IS, Behnke SC, Bartelt E, Weise E, Ellis A, Siitonen A, Andersson Y, Tschape H, Kramer MH, Ammon A: International outbreak of Salmonella Oranienburg due to German chocolate. BMC Infect Dis 2005, 5(1 ):7.

40. Odumeru JA, Steele M, Fruhner L, Larkin C, Jiang J, Mann E, McNab WB: Evaluation of accuracy and repeatability of identification of food-borne

187

pathogens by automated bacterial identification systems. J Clin Microbiol 1999, 37(4):944-949.

41. Hindiyeh M, Jense S, Hohmann S, Benett H, Edwards C, Aldeen W, Croft A, Daly J, Mottice S, Carroll KC: Rapid detection of Campylobacter jejuni in stool specimens by an enzyme immunoassay and surveillance for Campylobacter upsaliensis in the greater Salt Lake City area. J Clin Microbiol 2000, 38(8):3076-3079.

42. Dutta S, Chatterjee A, Dutta P, Rajendran K, Roy S, Pramanik KC, Bhattacharya SK: Sensitivity and performance characteristics of a direct PCR with stool samples in comparison to conventional techniques for diagnosis of Shigella and enteroinvasive Escherichia coli infection in children with acute diarrhoea in Calcutta, India. J Med Microbiol 2001, 50(8):667-674.

43. Kubota K, Iwasaki E, Inagaki S, Nokubo T, Sakurai Y, Komatsu M, Toyofuku H, Kasuga F, Angulo FJ, Morikawa K: The human health burden of foodborne infections caused by Campylobacter, Salmonella, and Vibrio parahaemolyticus in Miyagi Prefecture, Japan. Foodborne Pathog Dis 2008, 5(5):641-648.

44. Mead PS, Slutsker L, Dietz V, McCaig LF, Bresee JS, Shapiro C, Griffin PM, Tauxe RV: Food-related illness and death in the United States. Emerg Infect Dis 1999, 5(5):607-625.

45. Majowicz SE, Musto J, Scallan E, Angulo FJ, Kirk M, O'Brien SJ, Jones TF, Fazil A, Hoekstra RM, International Collaboration on Enteric Disease 'Burden of Illness' Studies: The global burden of nontyphoidal Salmonella gastroenteritis. Clin Infect Dis 2010, 50(6):882-889.

46. Hedberg C, Angulo F, Townes J, Vugia D, Farley M, FoodNet.: Differences in Escherichia co/# 0157:H7 annual incidence among FoodNet active surveillance sites. Baltimore, MD. 1997.

47. Centers for Disease Control and Prevention: Salmonella surveillance: annual summary, 2006. 2008.

48. World Health Organization: WHO consultation to develop a strategy to estimate the global burden of foodborne diseases: Taking stock and charting the way forward. 2007.

188

Tab

le 6

.1:

Dat

a so

urce

s an

d in

put d

istr

ibut

ions

use

d to

est

imat

e un

der-

repo

rtin

g of

gas

tro-

inte

stin

al i

nfec

tions

(S

alm

onel

la,

Cam

pylo

bact

er,

and

Shi

gella

) in

the

Met

ropo

litan

reg

ion,

Chi

le, 2

008.

Ste

p

G

F

E

D

C

B

A

Mod

el s

tep

Pro

port

ion

with

blo

ody

diar

rhea

Labo

rato

ry s

ends

po

sitiv

e sa

mpl

e to

IS

P

for

conf

irmat

ion

Sen

sitiv

ity o

f tes

t

Pro

port

ion

of

labo

rato

ries

that

test

fo

r pa

thog

en

Sam

ple

is s

ubm

itted

B

lood

y di

arrh

ea

Non

-blo

ody

diar

rhea

S

ampl

e is

req

uest

ed

Blo

ody

diar

rhea

N

on-b

lood

y di

arrh

ea

Cas

e vi

sits

phy

sici

an

Blo

ody

diar

rhea

N

on-b

lood

y di

arrh

ea

Dis

trib

utio

n

Cum

ulat

iveT

Be

ta1

Uni

form

**

Bet

a

Bet

a

Bet

a

Bet

a

Inpu

t va

lues

by

pat

ho

gen

S

alm

on

ella

6.

5%,

13.3

%,

18.8

%, 2

4.4%

, 43

.8%

, 59.

0%

88.9

% (

64+1

, 72

-64+

1)

Min

=72.

5%,

Max

=89

%

83.7

% (

41+1

, 49

-41+

1)

Cam

pyl

ob

acte

r 3.

6%,

7.8%

, 10

.9%

, 21.

4%,

24.3

%, 2

9.2%

87

.5%

(7+

1, 8

-7+

1)

Min

=85%

, M

ax=1

00%

16

.3%

(8+

1, 4

9-8+

1)

50%

(1+

1, 2

-1+

1)

100%

(6+

1, 6

-6+

1)

100%

(2+

1, 2

-2+

1)

10.3

% (

6+1,

58-

6+1)

18.2

% (

2+1,

11-

2+1)

22

.7%

(58

+1,

256-

58+

Sh

igel

la

14.8

%, 3

3.3%

, 36

.3%

, 39

.1%

, 39

.7%

, 40.

3%

98.1

% (

51+1

, 52

-51+

1)

Min

=54%

, M

ax=9

6%

81.6

% (

40+1

, 49

-40+

1)

D

So

urc

e /

Ref

eren

ce

[22-

39].

Labo

rato

ry

surv

ey

Inte

rnat

iona

l lit

erat

ure

[40-

42]

Labo

rato

ry

surv

ey

Pop

ulat

ion

surv

ey

Pop

ulat

ion

surv

ey

Pop

ulat

ion

surv

ey

If th

e in

put v

alue

is th

e sa

me

for

all t

hree

pat

hoge

ns it

is p

rese

nted

onl

y on

ce.

t Cum

ulat

ive

form

(m

inim

um, m

axim

um,{X

j}, {

pi})

, i =

1 to

n,

whe

re x

i is

the

prop

ortio

n of

cas

es w

ith b

lood

y di

arrh

ea p

er

outb

reak

rep

ort

and

pi is

the

cum

ulat

ive

prob

abili

ty a

nd n

=num

ber

of o

utbr

eak

repo

rts.

jB

eta

for

m (

a,b)

whe

re a

=(n

umbe

r su

cces

ses)

+ 1

and

b=(

tota

l num

ber

of a

ttem

pts)

- (

num

ber

of s

ucce

sses

) +

1.

** U

nifo

rm fo

rm (

min

imum

, m

axim

um).

189

Tab

le 6

.2:

The

ann

ual r

epor

ted

num

ber

of in

fect

ions

and

inci

denc

e by

Ins

titut

o de

Sal

ud P

ublic

a (I

SP

) an

d es

timat

ed

annu

al n

umbe

r of

cas

es,

inci

denc

e an

d un

der-

repo

rtin

g m

ultip

liers

for

Sal

mon

ella

, C

ampy

loba

cter

and

Shi

gella

infe

ctio

ns,

Met

ropo

litan

reg

ion,

Chi

le, 2

008.

Pat

hoge

n

Rep

orte

d

Est

imat

ed m

ean

, med

ian

(5

- 95

p

erce

ntil

e)

Nu

mb

er

Inci

den

ce p

er

Num

ber

of

case

s In

cid

ence

per

U

nd

er-r

epo

rtin

g

of c

ases

10

0,00

0 p

erso

n-

100,

000

per

son

- m

ulti

plie

r ye

ars

year

s S

alm

on

ella

10

45

17.2

69

,075

,60,

088

1140

,995

66

.1,5

7.5

(31

,03

7-1

35

,01

4)

(51

3-2

22

7)

(29

.7-1

29

.2)

Cam

pyl

ob

acte

r 30

0.

5 14

,094

,11,

754

23

3,1

94

469.

8,39

1.8

(53

73

-30

,18

9)

(89

-49

8)

(17

9.1

-10

06

.3)

Sh

igel

la

220

3.6

16,8

30,1

4,49

8 2

78

,23

9 76

.5,6

5.9

(76

56

-32

,58

2)

(12

6-5

38

) (3

4.8

-14

8.1

)

190

Tabl

e 6.

3: S

ensi

tivity

ana

lysi

s of

cor

rela

tion

of in

put d

istr

ibut

ions

and

ove

rall

unde

r-re

port

ing

mul

tiplie

r by

pat

hoge

n fo

r th

e M

etro

polit

an r

egio

n, C

hile

, 200

8.

Ran

k In

put

step

and

co

rrel

atio

n w

ith

ove

rall

un

der

-rep

ort

ing

mu

ltip

lier

Sal

mo

nel

la

Cam

pyl

ob

acte

r S

hig

ella

1 2 3 4 5

Inpu

t S

tep

S

ampl

e re

ques

ted:

N

on-b

lood

y S

ampl

e su

bmitt

ed:

Non

-blo

ody

Sam

ple

subm

itted

: B

lood

y P

ropo

rtio

n of

cas

es

with

blo

ody

diar

rhea

C

ase

visi

ts

phys

icia

n: N

on-

bloo

dy

Co

rrel

atio

n

(-0.

53)

(-0.

24)

(-0.

21)

(-0.

20)

(-0.

17)

Inpu

t S

tep

S

ampl

e re

ques

ted:

Non

-bl

oody

La

bora

tory

tes

ts fo

r pa

thog

en

Pos

itive

s se

nt to

IS

P

Sam

ple

subm

itted

: Non

-bl

oody

C

ase

visi

ts p

hysi

cian

: N

on-b

lood

y

Co

rrel

atio

n

(-0.

39)

(-0.

34)

(-0.

21)

(-0.

17)

(-0.

13)

Inpu

t S

tep

S

ampl

e re

ques

ted:

N

on-b

lood

y S

ampl

e su

bmitt

ed:

Blo

ody

Sen

sitiv

ity o

f la

bora

tory

tes

ts

Sam

ple

subm

itted

: N

on-b

lood

y C

ase

visi

ts

phys

icia

n: B

lood

y

Co

rrel

atio

n

(-0.

42)

(-0.

22)

(-0.

19)

(-0.

18)

(-0.

17)

191

Table 6.4: Summary of international pathogen-specific incidence rates for comparison with results from the Metropolitan region, Chile, 2008.

Study/Location Estimated incidence per 100,000 person-years Source Salmonella Campylobacter Shigella

Global Burden of 1140 [45] Salmonella (GBS) GBS-LAM South 80 USA* 530 920 England (95% CI) * 220 (110-430) 870 (610-1230) USA 520 Australia (5th - 95th 262 (150-624) 1184 (756-2670) percentile) Canada 250-690 910-1930 (min. - max.)* Jordan 124 Japan 32 237 * Incidence rates were calculated based on the 1997 United States Census population of 266,490,000. t CI = Confidence interval. $ min.= minimum value, max. = maximum value.

168 27(4-190)

306

[45] [44] [15] [17] [19]

[18]

[20] [43]

192

CHAPTER SEVEN

Summary Discussion and Conclusions

Although acute gastrointestinal illness (Gl) remains an important public

health burden in both developed and developing countries, few population-level

studies of Gl and associated risk factors have been conducted in developing

regions; regions that, according to the World Health Organization (WHO), carry

the greater part of the global burden. The shortage of population-based

information from less developed parts of the world, including South America, has

resulted in important gaps in available information on Gl, including: (a) the

population-level burden of Gl in South America; (b) the population-level food and

water-related risk factors for Gl in South America; (c) the community-level enteric

pathogen-specific burden estimates in South America. In addition, there is a

general limit in our understanding of the impact of using different recall periods in

a population-level burden of Gl study. The objectives of the research described

in this dissertation were: (1) to describe the distribution and population-level

burden of Gl; (2) the risk factors associated with Gl in Galvez, Argentina and the

Metropolitan region, Chile; and (3) to evaluate the effect of different recall periods

in population-level burden of Gl studies.

We examined the burden of Gl in two locations: first, a small Argentinean

community, Galvez, Argentina, population of approximately 18,500 in 2007, and

second, the large Metropolitan region, Chile, population of approximately

193

6,000,000 in 2008. To do this, proportional random samples from the

communities were obtained using residential listings and census data. Surveys

were administered door-to-door within each of these populations and the door-to-

door methodology required geographical information of the region to be

surveyed. A certain amount of creativity and resourcefulness was needed to

develop a sampling frame and provide the interviewers with appropriate lists of

households to survey. Often the ideal data are not available, especially in a

developing country. In Galvez, there was a residential listing that was used to

randomly select households and generate a sample. In the Metropolitan region,

this did not exist; however, there were maps and a listing to the level of

neighbourhood block from a recent census. These were used to randomly select

blocks. Maps were photocopied and selected blocks were manually highlighted

and provided to interviewers along with the list of selected blocks in order to go

into the field and conduct the surveys.

As it was not feasible to conduct surveys year-round, two time periods

(high and low Gl season) were selected during which to conduct the surveys.

This can present some research challenges as it requires sufficient and reliable

historical data to select the time periods appropriately. Once selected, the

validity of the study results is vulnerable to any Gl outbreak that might take place

during the study periods, or a shift in the seasonality of Gl creating a lack of

synchronization between study periods and peaks of illness, as this may inflate

or decrease the resulting estimates. Furthermore, holidays and special

occasions need to be taken into account as they may result in overall over- or

194

under-estimates of the true burden and complicate data collection. In addition,

extreme weather (e.g., heat, rain, natural disasters) can also present tremendous

challenges, particularly when surveys are being done door-to-door. Interviewers

typically do not enjoy working in extreme weather conditions and extra time must

be included in the schedule in order to prepare for such unpredictable

circumstances. Other events such as political strikes and protests can

complicate data collection. Having a strong team on the ground prepared for any

sort of contingency is essential to achieving success.

For the work in Galvez, we partnered with a municipal organization

(Centra de Desarrollo Agroalimentario, CeDA) whose mandate included food

safety activities. Their knowledge and local contacts proved valuable in obtaining

interviewers, selecting appropriate time periods and managing the survey.

Extreme heat and heavy rainfall during the data collection period proved

challenging for some interviewers; however, the desired sample sizes were

achieved in a timely fashion. Likewise, in Chile the Ministry of Health, the

Metropolitan Region Health Agency (SEREMI), the local Pan American Health

Organization (PAHO) officials and contacts at a local university provided

expertise to the study coordination, management, interviewer and time period

selection. Issues related to holidays, political strikes, large distances to travel for

interviewers and weather conditions were overcome.

Keeping the survey short, with only the essential questions, and thinking

ahead about how the questions will be analysed will help in the structure of each

question and of the overall survey. Asking others to review, edit and pilot-test the

195

survey are good ways to ensure the survey is acceptable, makes sense and

addresses the main questions that need to be answered. Language should be

clear and in line with local terms and colloquialisms to ensure that the survey is

understandable. Additionally, it is necessary to be cognisant of issues that might

arise due to translation differences if working in an area that speaks a language

other than the first-language of the researchers. For the work in Galvez, the

survey was kept to one page (double-sided) and questions were organized in an

appropriate sequence to facilitate the interview process. In the Metropolitan

region, surveys were more complex, but questions were kept to a minimum and

thought out in advance in terms of analysis, with many closed and semi-

structured questions, rather than entirely open-ended questions. This allowed for

data entry templates to include drop down menus as well as open areas for

answers outside the available list. Furthermore, much discussion took place

around ensuring that the difference between acute and chronic diarrhea was well

understood by interviewers. In Argentina for example, "colitis" in Spanish was

not understood as a chronic illness, as is normally the case for the general

population in English.

High response rates were achieved in both study locations; 61% in Galvez

and 76% in the Metropolitan region. These are among the highest response

rates for cross-sectional population-based burden of Gl studies. Campaigns to

raise awareness within the community can help in ensuring high response rates

and facilitate the work of the interviewers. Television, radio and newspapers or

individual letters sent to the homes in the area could be useful. In Galvez, the

196

local television and radio station conducted interviews with CeDA representatives

and me to inform the community about the study.

It is essential that all interviewers for Gl surveys have clearly visible

identification (e.g., badge or nametag) identifying them so respondents are not

suspicious or afraid to participate. Furthermore, hiring interviewers from the

community will help ensure they know the area, how to get from one place to

another and are familiar with local customs and language. Students and

workers in the areas of public health, medicine, nursing and veterinary medicine

are suitable interviewers as they may have a better understanding of research

principles and the importance of Gl than the general population. Nevertheless,

proper training is essential. We conducted role-playing and trouble-shooting

exercises which were a very useful way to discuss potential scenarios that might

occur out in the field. Using examples from previous studies can also help

illustrate the importance of certain practices and methods for data collection.

In total, 2915 and 6047 surveys were completed in Galvez and the

Metropolitan region, respectively. The information that was collected pertained to

symptoms of diarrhea and vomiting, secondary symptoms, socio-demographic

factors as well as medical system use, and in the Metropolitan region, Chile,

survey information on food and water consumption and general hygiene risk

factors was collected.

Key results from the analysis of the survey data include the observations

that the monthly prevalence of diarrhea in Galvez was 3.4% and the monthly

prevalence of Gl in the Metropolitan region was 7.7%. The annual incidence rate

197

of Gl in the Metropolitan region, Chile ranged from 0.98 to 2.3 episodes per

person-year, for a 30-day and a 7-day recall period, respectively. This illustrates

the significant burden of GI as well as the effect of shorter and longer recall

periods in burden of Gl surveys, in these communities.

In Galvez, comparisons were made of a 7-day and a 30-day recall period,

while in the Metropolitan region, Chile, 7-day, 15-day and 30-day recall periods

were included. Discussion and debate in the literature on recall period selection

and study methodology is ongoing. Prospective cohort studies are subject to

less recall bias and thus likely to generate more accurate burden estimates;

however, they are more expensive and time consuming to conduct. Cross-

sectional studies on the other hand, offer the advantages of being relatively

inexpensive and less time and labour intensive, and require the selection of an

appropriate recall period that minimizes potential recall bias but is feasible to

administer. Stool sample requests that are often incorporated in prospective

cohort studies have the potential to cause under-estimation of the burden due to

participants' aversion to providing a sample, while longer recall periods in cross-

sectional studies can lead to 'telescoping' and resultant over-estimation of the

burden. This research illustrates the potential under-estimation of the burden of

Gl when using longer recall periods, assuming shorter recall periods are less

subject to recall bias. This observation contradicts the thinking that 'telescoping'

of respondents in cross-sectional surveys occurs and results in over-estimation

of the burden.

198

As indicated by the WHO, syndromic surveillance of Gl requires, at

minimum, that someone seek medical care to be captured and counted in the

surveillance system. This source of information can provide basic case counts

and potentially can be used for outbreak identification. Of greater use is

laboratory-based surveillance that requires stool sample collection, submission

and testing. However, as more detailed information is sought, fewer cases are

captured. For example, my research in Galvez, Argentina, 26% of cases sought

medical care, and in the Metropolitan region, Chile 21% sought medical care,

highlighting the under-estimation of the burden of Gl in passively collected

surveillance data which rely solely on records of patient presentation at a medical

facility. This underscores the need for burden of Gl studies that are conducted

within the general population.

Of note, in most other population-level burden of Gl studies, women

typically had higher rates of Gl than men; however, in both study locations of this

research, men had nearly the same rate of Gl as women. This could indicate a

potential cultural-gender difference in the epidemiology of Gl in these

populations, perhaps due to differences in eating habits, occupations, or general

hygiene. Furthermore, men were more likely than women to have poor hygiene

and food safety habits and to consume foods identified to be of 'high Gl risk' in

the Metropolitan region, Chile. This information could be useful for more

targeted, population-specific interventions and education activities in the future.

Information specific to behaviours associated with Gl could be useful for

prevention and control activities. My research identified that risk factors

199

associated with Gl in the Chilean study included consumption of undercooked

poultry, beef and cheese made from unpasteurized milk, as well as increasing

consumption of water and using boiling as a method to treat drinking water. It

was not possible to infer causal associations as we are not sure if any of these

behaviours occurred prior to or perhaps as a result of, developing Gl.

Nevertheless, it is useful to highlight potential areas for improvement in terms of

foodborne and waterborne disease prevention and several of these associations

are in line with other measures of causality including biological plausibility and

consistency with other research.

Application of some of the results from the population survey, combined

with information from the literature and a laboratory survey enabled us to

estimate the pathogen-specific burden in the Metropolitan region, Chile, 2008.

We estimated community-level incidence of infections due to Salmonella,

Campylobacter and Shigella to be 1140, 233 and 278 per 100,000 person-years,

respectively. These results highlight the considerable pathogen-specific under

reporting that exists and could be useful for public health interventions.

Key strengths of the research described in this dissertation are the large

sample sizes in the surveys with all ages and both genders represented and the

high response rates. These factors contribute to the representativeness of the

sample and the internal validity of the results. Door-to-door surveys, although

time consuming, are likely to have contributed to achieving such high response

rates (61% and 76%). Exploring the impact of different recall periods in cross-

sectional population burden of Gl studies provides an innovative element to this

200

dissertation. Much discussion has arisen around study methodology and how to

best determine accurate burden of Gl estimates at the population level. The

issue of recall bias has been at the center of these discussions. By exploring

different recall periods we have shed light on this topic and furthered population-

based burden of Gl study methodology by demonstrating the utility of a shorter

recall period in population-based burden of Gl studies.

Population-level information on the burden of Gl from South America was

lacking prior to this work. Providing information from an under-represented part

of the world is a major contribution to the international literature. Current efforts

from the WHO to determine the global burden of foodborne disease can utilise

research like mine to better inform estimates and continue to elevate the level of

research being done in less developed parts of the world. Additionally, English

language publication of research from Spanish-speaking countries broadens the

level of international awareness.

One drawback of this study is the cross-sectional study design that was

used for the surveys. This design is not ideal for drawing causal associations

and formal hypothesis testing, as timing of events is not always clear. However,

it provides considerable breadth of information in a convenient and efficient

manner that would be difficult and costly to obtain using other study designs

(e.g., prospective cohort).

Another potential limitation was the lack of survey questions aimed at

determining details on costs associated with lost productivity and medical

expenses; however, information on number of medical visits, hospitalizations and

201

number of days missed from school or work due to illness was collected. These

values can be compared to other studies to help establish the economic burden,

and in the future could be incorporated with local average medical costs and

salary information to estimate the economic burden associated with Gl in these

communities.

Additionally, despite making considerable efforts to obtain a truly

representative, random sample, the study populations in both locations were

more likely to be female and older than the general population in the respective

communities. This weakness was somewhat expected as it is seen in many

population-based burden of Gl studies. This may be due to women being more

willing to participate and that parents are protective of their children so do not

wish them to participate in such surveys. The practice of selecting the person

with the next birthday was easy to implement in the field and overall did allow for

both genders and all ages to participate and be represented in the data.

As is often the case, we are left with as many questions when we finish as

when we started. Going forward from this research there are several projects

and recommendations for further study that I would suggest. An international

comparison of different recall periods used in burden of Gl cross-sectional

studies is needed in order to strengthen the evidence on the effect of shorter and

longer recall periods and determine if the results of this dissertation are

comparable to studies from other countries. This could have broad-reaching

effects on study design methodology in the future.

202

Additional studies in Latin American countries and developing countries

are needed. Of particular interest would be investigation of the gender

differences in rates of Gl in other locations and determination if, at least in some

populations, rates of Gl are similar for males and females, or potentially higher

among males, a finding thus far not seen in developed countries. Furthermore,

additional cross-sectional population surveys in other developing countries would

be useful for estimating the population-level burden of Gl, and can be

recommended in the WHO Foodbome Disease Burden Epidemiology Reference

Group (FERG) country protocols.

The burden of Gl in the study communities might be reduced by

developing or enhancing existing food safety campaigns, perhaps specifically

focusing on the demographic groups found to have higher rates of poor food

safety and hygiene behaviours. Additionally, evaluating the costs associated

with Gl (i.e., medical costs and lost productivity) to assess the economic impact

could help justify and support such campaigns to minimize Gl in the population

and to promote cost-effective ways to improve food safety.

If funds and resources permitted, it would be useful to conduct a year-

round survey to assess any bias that might result from selecting two time periods

to administer the survey. Furthermore, obtaining more information at the

laboratory level and assessing the under-reporting and under-estimation of other

important enteric pathogens would be of benefit.

On a more personal note, the work described in this dissertation, as is

likely the case in most doctoral research, was at once a 'labour of love' and a set

203

of considerable challenges. Patience and persistence are essential to success in

these sorts of situations. For me, it reaffirmed the importance of personal

relationships and networks in order to get a job done. Without these networks

and connections this work would not have been possible. We brought together

local, regional, national and international organizations, partners from North and

South America to make this project a success. It taught me the value of clear

communication and to never assume that things are they way you expect them to

be. This especially applies when working in different cultures, languages and

countries. What may seem obvious is often not, and how you phrase a question

is key to obtaining the answer you need.

In closing, the research described in this dissertation has demonstrated

the significant burden of Gl that is not fully captured in surveillance and outbreak

registries. It has taken steps to fill the void of population-level burden of Gl

information from developing countries, in particular from South America and it

has contributed to burden of Gl research methodology, in particular for use in

developing countries.

204

APPENDICES

205

APPENDIX I

Survey tool - Argentina (7 day recall period)

No. de encuesta Fecha Encuestador:

INFORMACldN GENERAL

Nombre:

Fecha de nacimiento _

Aoellido:

• " • ' " • " ' • ' • ' ' —

Sexo: Masculino Femenino

Domicilio: Calle:_

Entre:

No. Dpto._

Barrio: Telefono:

Nivel de escolaridad: D No corresponde

1 Analfabeto

G Primaria Incompleta G Primaria Completa G Secundaria Incompleta G Secundaria Completa l i Terciario Incompleto G Terciario Completo G Universitario G No sabe/No contesta

Numero de personas que viven en la casa:

Cantidad habitaciones que dispone:

Cuales animates tiene: No sabe I No contesta

ninguno perro gato pajaro vaca oveja caballo cabra gallina otro

Ocupacion: G No corresponde

G Obrero o empleado G Patron l; Trabajador por cuenta propia i Personal domestico

G Estudiante G Jubilado o pensionado G Ama de casa G Trabajador en restaurante G Cuida ninos

G Trabajador en residencia de ancianos 0 Enfermero o medico G Desocupado • No sabe/ No contesta

iPor que decidio hacer esta encuesta? Esta interesado en salud en general Tiene tiempo para hacerla Otra

Esta interesado en diarrea particularmente No sabe /No contesta

DATOS ESPECIFICOS DE LA ENFERMEDAD

1. ^Tiene diarrea cronica?

Si l: No 11 No sabe/ No contesta

2. &Uso algun antibiotico en los ultimos 30 dias?

Si No No sabe/ No contesta

206

3. i,Presento cuadros de diarreas en los ultimos 7 dias?

Si j No Q No sabe/ No contests

4. &Cuantos episodios de diarreas en los ultimos 30 dias ?

1 2 no contesta

no sabe/

5. ^Tiene diarrea hoy dia?

n Si l l No n No sabe/ No contesta

6. ^Relaciono su diarrea al consumo de n No sabe / No contesta • Alimentos

• Agua 0 Otros Cual?

• No sabe / No contesta

7. i,Algun miembro de su familia tuvo diarreas durante la semana previa a que Ud. tuvo diarrea?

n Si n No

n No recuerda D No sabe /No contesta

8. ^Cuantos dias duraron las diarreas?

dias n No recuerda 0 No sabe /No contesta

9. i,Habia sangre en su diarrea?

n Si D No D No sabe / No contesta

10. £Por alguna causa uso antibioticos en las 4 semanas antes de que presento cuadros de diarreas?

D Si D No • No sabe/ No contesta

11. ^Durante s u ultimo episodio de diarrea, tuvo/tiene alguno de los siguientes sintomas?

I Vomito I Fiebre l i Nauseas n Calambre in Dolor de cabeza G Dolores imusculares

I"': Otros ,i,Cuales?_ otros sintomas

[J No sabe / No contesta n No tuvo

12. i,Uso algun remedio (antibiotico, analgesico, antidiarreico, fluidos etc.) para tratar su sintomas 0 diarrea?

IJ No (pasara preg. 14 ) D Antibiotico D Analgesico D Antidiarreicos 0 Fluidos III Otro

1 No sabe/ No contesta

13. <j,Como obtuvo su remedio? C No sabe/ No contesta

Li Con receta medica • Farmacia sin receta • Otro

14. i,Acudio a un servicio de salud / medico por su cuadro de diarrea?

207

I Si n No sabe /No contesta D No Porque? (pasar a preg 22)

n Se automedico D No le dio importancia a las diarreas

1.1 Uso remedios naturales o caseros por su cuenta n Otra causa <j,Cual?

15. ^Cuantas veces fue a un servicio de salud / medico por las diarreas?

D No corresponde

16. £,A que servicio de salud consulto? 0 No corresponde

n Consultorio particular D Consultorio de Centra de Salud

li Consultorio de guardia privada D Consultorio de guardia de hospital publico

Otro <j,Cual? n No sabe /No contesta

17. £,Necesito hospitalizacion durante el episodio de diarrea? U Si

No (pasar a pregunta 19)

l ] No sabe /No contesta

18. d,Cuanto tiempo estuvo hospitalizado? hs(<1dia) dias

D No sabe /No contesta

19. ^Le solicitaron analisis de materia fecal (coprocultivo)? D No corresponde

IJ Si • No D No recuerda D No sabe /No contesta

20. i,Se realizo el analisis de materia fecal (coprocultivo)? • No corresponde

n Si (pasar a preg 22) D No D No recuerda D No sabe /No contesta

21. £Por que causa no se realizo el coprocultivo? • No corresponde

n Desaparecieron las diarreas • El laboratorio esta lejos

No le dio importancia l.l No se hacen coprocultivos en el laboratorio

Otras ^Cual? • No sabe /No contesta (pasar a preg 23)

22. i,Cual fue el resultado del analisis de materia fecal? D No corresponde

Positivo iQue patogeno?

n Negativo II No recuerda • No sabe /No contesta

208

23 t&iantos dias de trabajo perdio por las diarreas?

dias U No recuerda n No sabe /No contesta

24. ^Cuantos dias de escuela perdio por las diarreas?

dias G No recuerda in No sabe /No contesta

25. t&iantos dias dejaron de trabajar sus familiares para cuidarlo?

dias • No recuerda D No sabe /No contesta

Comentarios

209

APPENDIX II

Survey tool - Argentina (30 day recall period)

No. de encuesta Fecha. Encuestador:

INFORhMOONGENERAL

Nombre: _Apellido:.

Fecha de nacimiento Sexo: Masculino Femenino

Domicilio: Calle:

Entre:

No. Dpto..

Barrio: Telefono: Region:

Nivel de escolaridad: D No corresponde

Analfabeto Primaria Incompleta Secundaria Incompleta Secundaria Completa

Terciario Incomplete) Universitario

D No sabe/No contesta D Primaria Completa

D 0 Terciario Completo

Numero de personas que viven en la casa:

Cantidad habitaciones que dispone: _

Cuales animates tiene: No sabe I No contesta

ninguno perro gato pajaro vaca caballo oveja cabra gallina otro

Ocupacion: n No corresponde

Obrero o empleado Patron Trabajador por cuenta propia Personal domestico

D Estudiante • Jubilado o pensionado n Ama de casa L'J Trabajador en restaurante D Cuida nifios

• Trabajador en residencia de ancianos D Enfermero o medico n Desocupado G No sabe/ No contesta

iPor que decidio hacer esta encuesta? Esta interesado en salud en general Tiene tiempo para hacerla Otra

Esta interesado en diarrea particularmente No sabe /No contesta

DATOS ESPECIFICOS DE LA ENFERMEDAD

1. ^Tiene diarrea cronica?

n Si D No D No sabe/ No contesta

2. i,Uso algiin antibiotico en los ultimos 30 dias?


3. ^Presento cuadros de diarreas en los ultimos 30 dias? 0 Si • No • No sabe/ No

4. £,Cuantos episodios de diarreas en los ultimos 30 dias ? (episodios separados por 7 dias)

210

contests


n Si • No • No sabe/ No contesta

7. iAIgiin miembro de su familia tuvo diarreas 7 dias antes que Ud? 0 SI D No

n No recuerda G No sabe /No contesta

9. i,Habia sangre en su diarrea?

I Si I: No l.'i No sabe / No contesta

1 2 3 4 5 no sabe/ no contesta

6. i,Relaciono su diarrea al consumo de

D Agua • Alimentos • Otros Cual? n No sabe / No contesta

8. ^Cuantos dias duraron las diarreas?

dias • No recuerda • No sabe /No contesta

10. <,Por alguna causa uso antibioticos en las 4 semanas antes de que presento cuadros de diarreas?

Li Si D No Li No sabe/ No contesta

11. ^Durante su ultimo episodio de diarrea, tuvo/tiene alguno de los siguientes sintomas?

l i: Vomito u Fiebre U Nauseas D Calambre D Dolor de cabeza • Dolores musculares

D Otros iCuales? • No sabe / No contesta D No tuvo otros sintomas

12. <,Uso algun remedio (antibiotico, analgesico, antidiarreico, fluidos etc.) para tratar su sintomas 0 diarrea?

• No (pasar a preg. 12 ) • Antibiotico • Analgesico D Antidiarreicos D Fluidos LI Otro

1 J No sabe/ No contesta

13. i,Como obtuvo su remedio?

I ! Con receta medica 0 Farmacia sin receta Li Otro

jl: No sabe/ No contesta

14. i,Acudio a un servicio de salud / medico por su cuadro de diarrea?

in Si Q No sabe/No contesta LI No Por que? (pasar a preg 21) LI Se automedico

211

• No le dio importancia a las diarreas • Uso remedios naturales o caseros por su cuenta

n Otra causa iCua\?

15. i.Cuantas veces fue a un servicio de salud / medico por las diarreas?

LI No corresponde

16. <j,A que servicio de salud consulto?

LI Consultorio particular

L! Consultorio de guardia privada

l; Otro <j,Cual?

Ll No sabe /No contesta

17. tNecesito hospitalizacion durante el episodio 16. ^Cuanto tiempo estuvo hospitalizado? de diarrea?

hs(<1dia) n Si Li No (pasar a pregunta dias 18)

Ll No sabe /No contesta I! No sabe /No contesta

18. <,Le solicitaron analisis de materia fecal (coprocultivo)? Ll No corresponde

! Si 11 No M No recuerda IJ No sabe/No contesta

19. ^Se realizo el analisis de materia fecal (coprocultivo)? D No corresponde

n Si (pasar a preg 21) Li No D No recuerda • No sabe /No contesta

20. ^Por que causa no se realizo el coprocultivo? • No corresponde

Ll Desaparecieron las diarreas D El laboratorio esta lejos

Ll No le dio importancia Li No se hacen coprocultivos en el laboratorio

G Otras iCual? Q No sabe /No contesta

21. i,Cual fue el resultado del analisis de materia fecal? D No corresponde

l Positivo iQue patogeno?

l; Negativo Li No recuerda Ll No sabe /No contesta

D No corresponde

0 Consultorio de Centra de Salud

n Consultorio de guardia de hospital publico

212

22 i,Cuantos dias de trabajo perdio por las diarreas? 23. iCuantos dias de escuela perdio por las diarreas?

dias • No recuerda dias D No recuerda

n No sabe /No contesta D No sabe /No contesta

24. ^Cuantos dias dejaron de trabajar sus familiares para cuidarlo?

dias • No recuerda 0 No sabe /No contesta

iComentarios

213

APPENDIX III

Formulas for calculating prevalence, incidence rate and incidence proportion

Prevalence:

# of cases

Total # at risk

Annual incidence rate

# of cases 365

\/2[(Total#at risk) + (Total# at risk-#of cases)] #of days of recall period

Annual incidence proportion

A (] __ y\ (3651'Hofdays ofrecall period)

# ofcases where x =

Total # at risk - 1/2 # of withdrawals

214

APPENDIX IV

Survey tool - Chile (7, 15 and 30 day recall periods)

No. de encuesta Fecha.... Encuestador:

A. INFORMAClON GENERAL

Nombre:

Fecha de nacimiento _

Domicilio: Calle:

Apellido:

o Edad Sexo: Masculino

No.

Femenino

Dpto.

iComuna: Ciudad: Telefono:

Niimero de personas que viven en la casa: Tipo de vivienda:

Cantidad piezas para dormir de que dispone:

La vivienda que usted ocupa es:

Propia ( pagada completo) Propia (pagada a plazo) Arrendada Cedida por trabajo o servicio Gratuita Otra

Que animales tiene: ninguno perro

oveja gallina otro

No sabe I No contesta gato pajaro vaca

caballo cabra

Casa Departamento en edificio Pieza en una casa o conventillo Mediagua, mejora Rancho o choza Otro

Su servicio de higienico es de o esta: Conectado a alcantarillado Conectado a fosa septica Cajon sobre pozo negro Cajon sobre acequia o canal Quimico No tiene servicio higienico Otro

<,De donde obtiene su agua? Red publica Pozo Rio Otro

Nivel de educacion:

No corresponde Analfabeto Kinder o Jardin

infantil Basico o

Primaria Media comun Media tecnico-

Profesional Humanidades

Normal Centra formacion tecnica Instituto profesional Universitario Post-Grado No sabe/No contesta Otro

Ocupacion: No corresponde Empleador o Patron Trabajador por cuenta propia

[independiente) Asalariado sector privado Asalariado sector publico Personal domestico Estudiante Jubilado o pensionado

Ama de casa Trabajador en

restaurante Cuida nihos Trabajador en

residencia de ancianos Enfermero o medico Desocupado No sabe/No contesta Otro

215

1A que sistema de salud pertenece usted? FONASA CAPREDENA ISAPRE Particular No tiene No Sabe/No Contesta Otro

,/Por que decidio hacer esta encuesta?

Esta interesado en salud en general Esta interesado en diarrea particularmente Tiene tiempo para hacerla No sabe /No contesta Otra

B. DATOS ESPECIFICOS DE LA ENFERMEDAD

1. ^Tiene diarrea cronica o sintomas de diarrea o vomito por alguna condicion diagnosticada por un medico (enfermedad de Crohn, colitis ulcerosa, colon irritable, cirugia biliodigestiva, abuso de laxantes, alcoholismo, radio o quirnioterapia durante los dos meses anteriores, inmunodeficiencia (SIDA, otras) o si esta embarazada y ha presentado diarrea o vomito por esta razon)?

I Si (pasa a pregunta 2 y parte D solamente) [ No I No sabe/ No contesta

2. i,Uso algun antibiotico en los ultimos 30 dias (por cualquier causa)?


3. {.Presento cuadros de diarreas en los ultimos:

3a. ij dias?

i; Si 11 No No contesta

Ll No sabe/

3b. ^15 dias?

C Si r.l No n No sabe/ No contesta

3c. ^30 dias?

Li Si (J No No contesta

G No sabe/

B.1 SOLO PARA LOS ENCUESTADOS QUE RESPONDEN 'SI' EN ALGUNA PARTE DE PREGUNTA 3, EL RESTO PASE A PREGUNTA 11.

{4. <,Cuantos episodios (separado por 7 dias) tuvo de diarreas en los ultimos 30 dias?

1 2 contesta

no sabe/ no


D Si D No Li No sabe/ No contesta

6. Ên el dia peor, cuantos cuadros de diarreas tuvo?

veces/dia

7. ^Relaciono su diarrea al consumo de

[j Alimentos Li Alcohol

Remedios/laxantes Otro

Ll Agua

l: No sabe / No contesta

8. iLa semana antes de que usted enfermada, hubo algun familiar en su casa con diarreas o vomito?

11 SI I No i No sabe /No contesta

9. ^Cuantos dias duraron las diarreas? j j dias I".] No recuerda D No sabe /No ;contesta

10. £Habia sangre en su diarrea?

L' Si n No Li No sabe / No contesta

1̂1 ̂ Presento vomito en los ultimos:

216

11a. <j,7 dias?

11 Si n No No contests

n No sabe/

11b. <s,15 dias?

D Si D No n No sabe/ No contests

11c. i.30 dias?

• Si n No No contests

D No sabe/

B.2 SOLO PARA LOS ENCUESTADOS QUE RESPONDEN 'SI' EN ALGUNA PARTE DE PREGUNTA 11, EL RESTQ PASE A PARTES C v P.

12. &Cuantos episodios (separado por 7 dias) tuvo de vomito en los ultimos 30 dias?

1 2 contesta

no sabe/ no

13. ^Tiene vomito hoy dia?

D Si 0 No D No sabe/ No contesta

14.«;,En el dia peor, cuantos cuadros de vomito tuvo?

veces/dia

15. i,Relaciono su vomito al consumo de

D Alimentos D Alcohol

0 Remedios/laxantes 0 Otro

• Agua

D No sabe / No contesta

16. 6La semana antes de que usted enfermada, hubo algun familiar en su casa con diarreas o jvomito?

n si f] No D No sabe /No contesta

17. ^.Cuantos dias duraron los vomitos?

dias D No recuerda D No sabe /No contesta

C. SOLO PARA LOS ENCUESTADOS QUE RESPONDE 'SI'A ALGUNA PARTE.DE LAS PREGUNTAS 3 o 11

Si no, pase a la parte V de los habitos

18. iPor alguna causa uso antibioticos en las 4 semanas antes de que presento cuadros de diarreas o vomito?

G Si • No D No sabe/ No contesta

19. ^Durante su ultimo episodio de diarrea o vomito, tuvo/tiene alguno de los siguientes sintomas?

n Fiebre • Nauseas • Calambre D Dolor de cabeza • Dolores musculares D Dolor garganta

Romadizo (nariz con mucosa) • Tos/estornudo Otros ^Cuales?

No sabe / No contesta • No tuvo otros sintomas

20. Ûso algun remedio (antibiotico, analgesico, antidiarreico, liquidos, antiemetico (para controlar nausea), antiespasmodico, etc.) para tratar sus sintomas de diarrea o vomito?

I No (pasar a pregunta 22) IJ Antibiotico • Analgesico G Antidiarreicos r.'l Liquidos G Antiemetico r! Antiespasmodico

G Otro G No sabe/ No contesta

|21. ,f,C6mo obtuvo su remedio? G No sabe/ No contesta

l: Con receta medica L' l Farmacia sin receta G Otro

217

22. i,Acudio a un servicio de salud / medico por su

GSi • No,

D No sabe /No contesta (pasar a pregunta 31) • No vomito (pasar a pregunta 31)

(pasar a pregunta 31) • No

D No (pasar a orequnta 31)

diarrea o vomito?

Se automedico (pasar a pregunta

No le dio importancia a las diarreas o

31)

Uso remedios naturales o caseros por su cuenta

Otra causa <j,Cual?

23. £Cuantas veces fue a un servicio de salud / medico por su diarrea o vomito?

i l No corresponde

24. Â que servicio de salud consulto? n No corresponde

lj Consulta privado (clinica privada) • Consulta publico (posta medica) D Mutualidades

D Hospital publico • Hospital D Hospital privado institutional

No sabe /No contesta • Otro <^Cual?.

25. ^Necesito hospitalizacion durante el episodio de diarrea o vomito?

nsi

D, No (pasar a pregunta 27)

D No sabe /No contesta (pasar a pregunta 27)

26. i,Cuanto tiempo estuvo hospitalizado?

hs(<1dia) dias

• No sabe /No contesta

27. i,Le solicitaron anaiisis de materia fecal (coprocultivo)?

n No corresponde n Si

I: No (pasar a pregunta 31)

n No sabe /No contesta (pasar a pregunta 31)

28. (,Se realizo el anaiisis de materia fecal (coprocultivo)?

LI No corresponde pregunta 30)

D No

r.1 No sabe /No contesta

D Si (pasara

|29. iPor que causa no se realizo el coprocultivo (pasar a pregunta 31)?

G Desaparecieron las diarreas • El laboratorio esta lejos

• No le dio importancia • No se hacen coprocultivos en el laboratorio

D Otras <j,Cual? D No sabe /No contesta

No corresponde

30. i,Cual fue el resultado del anaiisis de materia fecal? n No corresponde

Li Positivo i,Que patogeno? (pide por el resultado)

l :> Negativo D No sabe /No contesta

31 iCuantos dias de trabajo perdio por su diarrea o vomito?

dias D No sabe /No contesta

32. tCuantos dias de escuela perdio por su diarrea o vomito?

dias • No sabe /No contesta

33. i,Cuantos dias dejaron de trabajar sus familiares para cuidarlo?

dias • No sabe /No contesta

218

D. DATOS ESPECIFICOS DE HABITOS

34. i,Lava sus manos antes de comer?

Ll Siempre D A veces n Nunca D No Sabe/No contesta

36. 6Lava sus manos despues de ir al bano?

I! Siempre G A veces 0 Nunca D No Sabe/No contesta

38. <j,Usa una tabla especifica para cortar su carne? (solo para carne).

I: Siempre i A veces I'] Nunca I1 No Sabe/No contesta

[] Si n No sabe /No contesta G No ^Por que?

42. i,En la ultima semana, donde com pro su carne, polio, cerdo, cordero?

I.I Un supermercado (Jumbo, Lideretc.) I: Un negocio especializado (Carniceria) i.i Un local de abarrotes (kiosco, almacen de Barrio), mercado G En la feria libre I: Otro

35. <s,Usa jabon para lavar sus manos antes de comer?

D Siempre D A veces • Nunca 0 No Sabe/No contesta

37. iUsa jabon para lavar sus manos despues de ir al bano?

D Siempre D A veces • Nunca D No Sabe/No contesta

39. iLava sus frutas y verduras antes de comer?

• Siempre G A veces D Nunca G No Sabe/No contesta

43. i,En la ultima semana, donde com pro su pescado o mariscos?

[J Un supermercado (Jumbo, Lider etc.) G Un negocio especializado (Pescaderia) n Un local de abarrotes (kiosco, almacen de Barrio), mercado G En la feria libre G Otro

40. <,En la ultima semana (7 dias), consumio:

I ] Huevos crudos (tiramisu, bebidas, masa de postre, etc.) G Huevos no cocidos completamente (huevos a la copa o revueltos)

I Mariscos crudos G Ostras, mejillones, molusco G Pescado crudo G Pescado mal cocido

G Carne cruda G Carne no cocida completamente (con centra rosado o con sangre)

II Hamburguesas no cocidas completamente (con centra rosado o con sangre)

G Cerdo no cocido completamente (con centra rosado o con sangre)

l j Pollo/ave no cocido completamente (con centra rosado o con sangre)

l; Cordero no cocido completamente (con centra rosado o con sangre)

Li Leche no pasteurizada G Queso artesanal/casero/ de campo (hecho con leche no pasteurizada)

G Mayonesa envasado G Mayonesa hecho en casa

41. i,Sus respuestas en la pregunta 40, es lo que habitualmente consume usted en una semana normal?

219

44. £,En la ultima semana, donde compro sus huevos y leche, queso otro productos de leche?

n Un supermercado (Jumbo, Lider etc.) I! Un negocio especializado (Codipra,) L: Un local de abarrotes (kiosco, almacen de Barrio), mercado l ] En la feria libre • Otro

45. £En la ultima semana, donde compro sus frutas y verduras?

D Un supermercado (Jumbo, Lider etc.) • Un negocio especializado (Fruteria, Verduleria) • Un local de abarrotes (kiosco, almacen de Barrio), mercado • En la feria libre • Otro

46. <j,Sus respuestas en las preguntas 42, 43, 44 y 45, corresponden a donde usted habitualmente compra alimentos normalmente?

Si D No sabe /No contesta que?

• No <s,Por

47. £En promedio, cuantas comidas/veces come en un dia?

veces/dia

48. <,En la ultima semana (7 dias), cuantas veces comio en una restaurante (sentado)?

veces/semana

49. £En la ultima semana (7 dias), cuantas veces comio en un casino/comedor o buffet?

veces/semana

50. <;,En la ultima semana (7 dias), cuantas veces comio comida rapida (McDonald's, etc.)?

veces/semana

51. £En la ultima semana (7 dias), cuantas veces comio algo preparado en la calle (hot dog, completa, etc.)?

veces/semana

52. i,En la ultima semana (7 dias), cuantas veces comio algo pedido a su casa/oficina (despacho al domicilio)?

veces/semana

53. <,En la ultima semana (7 dias), cuantas veces comio algo que prepare en su casa?

veces/semana

54. i,Sus respuestas en las preguntas 48, 49, 50, 51, 52 y 53, corresponden a donde usted habitualmente come?

LI Si D No sabe /No contesta

I ] No <,Por que?

55. <,Ayer, cuantas porciones (vasos, 250ml) de agua consumio (incluyendo jugos o bebidas que estan hechos con agua, pero sin considerar cafe o te?

porciones/ayer

56. i,Esta cantidad (pregunta 57) es lo que habitualmente Ud. consuma de agua diario?

• Si • No sabe /No contesta

n No <j,Por Ique?

57. ^Cuantas de estas porciones (vasos, 250ml) de agua son de:

jAgua de una botella Agua de la casa (de la Have)

ptro (numero de vasos)

58. iQue usa para tratar su agua de la casa?

Nada P Hervir n Filtro del lavaplatos D Filtro de la jarra de agua (ej. Brita)

I" Otro

220

Comentarios:

221

APPENDIX V

Expected probability formulas - Chile

observed # of cases (7 - day recall) {recall period) Expected probability = *

total # at risk 7

Recall

period

7-day

15-day

30-day

Phase

1

2

combined

1

2

combined

1

2

combined

Observed

number

of cases

117

145

262

184

200

384

237

230

467

Total

number

at risk

3033

3014

6047

3033

3014

6047

3033

3014

6047

Observed

probability

0.039

0.048

0.043

0.061

0.066

0.064

0.078

0.076

0.077

Expected

probability

NA

NA

NA

0.083

0.103

0.093

0.165

0.206

0.186

p-value

NA

NA

NA

<0.001

<0.001

O.001

O.001

<0.001

<0.001

222

APPENDIX VI

Survey tool - Chile (Laboratory survey)

Estimados Laboratories:

Junto con saludarles, les agradecenamos su colaboracion para contestar la siguiente encuesta, la cual, forma parte del Estudio sobre Carga de Gastroenteritis segun Agente Etiologico en la Region Metropolitana, este estudio se esta realizando en conjunto con el Ministerio de Salud de Chile, la Organizacion Panamericana de la Salud y la Universidad de Guelph de Canada.

La presente encuesta consta de tres partes A= Bacteriologia, B= Parasitology, C= Virologia.

Usted debera responder preguntas breves generales y preguntas especificas en relacion a dos periodos: 1 de Julio al 31 de Julio y 15 de noviembre al 15 de diciembre de 2008

El plazo de entrega de la encuesta es el 15 de enero de 2009, para esto usted dispone de dos vias para hacerla llegar:

1. correo electronico: [email protected] , [email protected] 2. Correo convencional: Instituto de Salud Publica, Av Maraton 1000

Departamento de Parasitologia (Dra. Carolina Marchant)

Si usted presenta alguna duda puede comunicarse con Cecilia Parada Carvallo al 8-1716459 o [email protected]

De Antemano, Agradecemos su Colaboracion.

No. de Encuesta Fecha: '."?./"•: 7/

INFORMACION GENERAL

Nombre del Laboratorio:

Nombre de Jefe/Jefa del Laboratorio:

223

mailto:[email protected]



Direccion del Laboratorio: Calle Numero Comuna Ciudad Telefono

Region

Comentarios:

A. Bacteriologia Nombre del Encuestado:

Durante los dos periodos (1) 1 - 31 de julio 2008 y (2) 15 de noviembre al 15 de diciembre 2008 cuantos coprocultivos: 1. £Se recibieron para hacer analisis de bacteriologia en ese laboratorio? 2. i,De los recibidos, cuantos efectivamente se analizaron para estudio bacteriologico? 3. &De los analizados, cuantos fueron positivos para un patogeno bacteriano?

1. Recibidos en el periodo

2. Analizados en el periodo

3. Positivos en el periodo

( 1 ) 1 - 3 1 de julio 2008 (2) 15 de noviembre al 15 de diciembre 2008

4. i,En los periodos senalados, cuantas muestras fueron positivas para las siguientes bacterias? ( 1 ) 1 - 3 1 de

(D julio 2008 y (2) 15 de noviembre al 15 de diciembre 2008

(2) Bacterias Aeromonas spp. Campylobacter spp. Clostridium spp. Escherichia coli

- Escherichia coli 0157 - VTEC/STEC (verotoxigenica) - EPEC (E coli enteropatogenica) - EHEC (£. coli enterohemorragica) - ETEC (E. coli enterotoxigenica) - EIEC (E colienteroinvasiva ) - Otro, cual

Listeria spp. Plesiomonas spp. Salmonella spp. i Shigella Vibrio spp

- parahemolyticus - cholerae - Otro, cual

Yersinia Otro, cual

224

5 . ; En los periodos senalados, de las cepas enteropatoqenas cuantas envio al ISP? (1) 1 - 31 de julio 2008 y (2) 15 de noviembre al 15 de diciembre 2008

(D (2) Bacterias Aeromonas spp. Campylobacter spp. Clostridium spp. Escherichia coli

- Escherichia coli 0157 - VTEC/STEC (verotoxigenica) - EPEC (E. coli enteropatogenica) - EHEC (E. coli enterohemorragica) - ETEC (E. coli enterotoxigenica) - EIEC (E. coli enteroinvasiva ) - Otro, cual

Listeria spp. Plesiomonas spp. Salmonella spp. Shigella Vibrio spp

- parahemolyticus - cholerae - Otro, cual

Yersinia Otro, cual

6. Marque las bacterias que investiga en los coprocultivos en su laboratorio en forma (a) RUTINA (b) solo cuando se solicita en forma ESPECIAL (c) NUNCA

Aeromonas spp. Campylobacter spp. Clostridium spp.

Escherichia coli

Listeria spp. Plesiomonas spp. Salmonella spp. Shigella

.0157

. VTEC/STEC (E. coli verotoxigenica) EPEC (E. coli enteropatogenica EHEC (E. coli enterohemorragica ETEC (E. coli enterotoxigenica) EIEC (E. coli enteroinvasiva )

" Otro Ninguno

Vibrio spp.

Yersinia Otro, cual

parahemolyticus cholerae

\ Otro Ninguno

225

7. Cuando no fue posible hacer un analisis de bacteriologia, ^cuaies fueron las razones para rechazar una muestra? Marque mas de una respuesta si es necesario.

Muestra insuficiente El frasco venia danado Demora entre toma de muestra y recepcion en el laboratorio Muestra en solucion inadecuada Falta de medios de cultivo apropiados Otra

8. £Con que frecuencia, hacen analisis de la resistencia de antibioticos/antimicrobianos a las siguientes bacterias? (Siempre =100%, casi siempre =80-99%, a veces =50-79%, casi nunca 10-49%, nunca = menos de 10%)

Aeromonas spp. i Campylobacter spp. 0157 Clostridium spp. VTEC/STEC (E. coli verotoxigenica)

EPEC (E coli enteropatogenica) Escherichia coli • EH EC (E. coli enterohemorragica)

ETEC (E coli enterotoxigenica) EIEC (E colienteroinvasiva ) Otro Ninguno

parahemolyticus cholerae Otro Ninguno

Yersinia Otra, identifique

9. i,Que metodos usan normalmente para identificar Ja resistencia de antibioticos/antimicrobianos? (Indicar todos los metodos que usan)

Difusion en disco (Kirby-Bauer) Dilucion en caldo (CIM) Dilucion en agar (CIM) Sistema Vitek Sistema MicroScan Sistema Sensititre E-test Sistema Fenix Otro, cual No sabe

Comentarios:

Listeria spp. Plesiomonas spp. Salmonella spp. Shigella

Vibrio spp.

226

Nombre del Encuestado: B. Parasitologia

Durante los dos periodos (1) 1 - 3 1 de julio 2008 y (2) 15 de noviembre al 15 de diciembre 2008 cuantos coproparasitologicos: 1. £Se recibieron para hacer analisis de parasitologia en ese laboratorio? 2. i,De los recibidos, fueron efectivamente analizadas para la busqueda de enteroparasitos? 3. <j,De los analizados, fueron positivos para algun enteroparasito patogeno?

( 1 ) 1 - 3 1 de julio 2008

(2) 15 de noviembre al 15 de diciembre 2008

1. Recibidos en el

periodo

2. Analizados en el

periodo


4. Ên los periodos senalados, cuantas muestras fueron positivas para los siguientes parasitos? (1) 1 - 31 de julio 2008 y (2) 15 de noviembre al 15 de diciembre 2008

(1) (2) Parasitos Entamoeba histolytica/dispar Cryptosporidium spp. Cyclospora spp Isospora belli Giardia lamblia Microsporidium spp Otro, cual

5. Ên los periodos senalados, de los analizados y positivos, por patogeno, cuantas de las muestras fueron enviadas al Instituto de Salud Publica? (1) 1 - 31 de julio 2008 y (2) 15 de noviembre al 15 de diciembre 2008

(1) (2) Parasitos Entamoeba histolytica/dispar Cryptosporidium spp. Cyclospora spp Isospora belli Giardia lamblia Microsporidium spp Otro, cual

227

6. ^Cuales de los siguientes parasitos son buscados en un analisis en forma: (a) RUTINA (b) solo cuando se solicita en forma ESPECIAL (c) NUNCA

Entamoeba histolytica/dispar Cryptosporidium spp. Cyclospora spp Isospora belli Giardia lamblia Microsporidium spp Otro, cual

7. i,Que metodo de rutina usa para la busqueda de enteroparasitos? Burrows modificado Telemann modificado Otro, cual

8. ^Cuantas muestras por paciente procesa normalmente? 1 2 3 Otro, cual

9. ^Cuando no fue posible hacer un analisis, cuales fueron las razones para rechazar una muestra? Marque mas de una respuesta si es necesario.

Muestra insuficiente El frasco venia danado Muestra en solucion inadecuada Falta de reactivos qtra

Comentarios:

228

Nombre del Encuestado: C. Virologia

Durante los dos periodos ( 1 ) 1 - 31 de julio 2008 y (2) 15 de noviembre al 15 de diciembre 2008 cuantas muestras fecates: 1. £Se recibieron para hacer analisis de virologia en este laboratorio? 2. iDe las recibidas, cuantas fueron efectivamente analizadas para detectar algun virus? 3. &De las analizadas, fueron positivos para algun virus?

1. Recibidos en el

periodo

2. Analizados en el

periodo


( 1 ) 1 - 3 1 de julio 2008

(2) 15 de noviembre al 15 de diciembre 2008

4. i,En los periodos senalados, cuantas muestras fueron positivas para los siguientes virus? 1) 1 - 3 1 de julio 2008 y (2) 15 de noviembre al 15 de diciembre 2008

(1) (2) Virus Astrovirus Adenovirus entericos Calicivirus, Norwalk-like, Norwalk, SRSV Rotavirus Otro, cual

5. iEn los periodos senalados, de las analizadas, por patogeno, cuantas de las muestras fueron derivadas a otro Laboratorio. (1) 1 - 31 de julio 2008 y (2) 15 de noviembre al 15 de diciembre 2008

(1) (2) Virus Astrovirus Adenovirus entericos Calicivirus, Norwalk-like, Norwalk, SRSV Rotavirus Otro, cual

6. i,Cuando no fue posible hacer un analisis de virologia, cuales fueron las razones para rechazar una muestra? Marque mas de una respuesta si es necesario

Muestra insuficiente El frasco venia danado Muestra mal conservada Sin disponibilidad de reactivos Otra

229

7. <,Cuales de los siguientes virus son buscados en su laboratorio por un analisis (a) RUTINA (b) solo cuando se solicita en forma ESPECIAL (c) NUNCA (d) DERIVA, identifique a donde?

Deriva: Astrovirus Donde

Adenovirus entericos Donde

Calicivirus, Norwalk-like, Norwalk, SRSV Donde

Rotavirus Donde

Otro, cual , Donde

8. £lndique el metodo utilizado y la sensibilidad y fa especificidad para detectar los siguientes virus (segun especificaciones del inserto provisto por el fabricante)? Astrovirus Metodo

Sens= Espec= Adenovirus entericos Metodo

Sens= Espec= Calicivirus, Norwalk-like, Norwalk, SRSV Metodo

Sens= Espec= Rotavirus Metodo

Sens= Espec= Otro, cual , Metodo

Sens=_ Especf Comentarios:

230

APPENDIX VII

Sentinel Clinic Study

In conjunction with the population survey, a sentinel clinic study of

diarrheal cases and non-cases was completed in the Metropolitan region, Chile.

However, in the end this was not part of the final research dissertation. In this

study, cases of diarrheal illness presenting at seven sentinel clinics located in the

Metropolitan region, Chile during July 15 - August 31 and November 3 -

December 16, were invited to participate in a survey and submit a stool sample

for testing. Concurrently, non-cases presenting at the sentinel clinics were asked

to complete a risk factor survey. In the first phase, 25 cases and 26 non-cases

that participated while in the second phase, 128 cases and 299 non-cases

participated. This portion of the research was inconvenienced by protests and

strikes that shut down clinics for extended periods of time as well as difficulties

with compliance for stool submission from adults. The data from these surveys

and the corresponding laboratory results has not yet been analysed, however the

surveys are presented here for study completeness.

231

Survey Tool - Sentinel clinic study (cases)

No. de encuesta Encuestador:

A. INFORMAClON GENERAL

Nombre: _Apellido:

Nombre de la Clinica:

Domicilio: Calle:_

Comuna:

No. Dpto..

Ciudad: Telefono:

Numero de personas que viven en la casa:. JTipo de vivienda:

Cantidad piezas para dormir que dispone:

La vivienda que usted ocupa es: Propia (pagado completo) Propia (pagado a plazo) Arrendada Cedida por trabajo o servicio Gratuita Otra

\Que animates tiene: ninguno perro oveja caballo otro

No sabe I No contesta gato pajaro vaca

cabra gallina



i,De donde obtiene su agua? Red publico Pozo Rio Otro

Nivel de educacion: No corresponde Analfabeto Kinder Basico o Primaria Media comiin Media tecnico-


Normal Centra formacion tecnica Institute profesional Universitario Post-Grado No sabe/No contesta Otro


(independiente) Asalariado sector privado Asalariado sector publico Personal domestico Estudiante Jubilado o pensionado


restaurante Cuida nihos Trabajador en

residencia de ancianos Enfermero o medico Desocupado No sabe/No contesta Otro

232

IA que sistema de salud pertenece usted? FONASA CAPREDENA ISAPRE Particular No tiene No Sabe/No Contesta Otro

iPor que decidio hacer esta encuesta?


B. DATOS ESPECIFICOS DE LA ENFERMEDAD

1. i,Tiene diarrea cronica o sintomas de diarrea o vomito por alguna condicion diagnosticada por un medico (enfermedad de Crohn, colitis ulcerosa, colon irritable, cirugia biliodigestiva, abuso de laxantes, alcoholismo, radio o quimioterapia durante los dos meses anteriores, inmunodeficiencia (SIDA, otras) o si esta embarazada y ha presentado diarrea o vomito por esta razon)?

n Si n No O No sabe/ No contesta

2. Cuando fue a la clinica: / /2008

3. Indique cual era la razon por la que fue a la clinica ese dia.

diarrea sin vomito (pasa a partes B.1 y C y D)

vomito sin diarrea (pasa a partes B.2 y C y D)

diarrea y vomito (pasa partes B.1 y B.2 y C y D)

B.1 DATOS ESPECIFICOS DE LAS DIARREAS

4. ^Relaciono su diarrea al consumo de

i Alimentos IT! Aqua • Alcohol IJ Remedios/laxantes

I j Otro n No sabe / No contesta

5. <,La semana antes de que usted enfermada, hubo algun familiar en su casa con diarreas o vomito?

I"] SI Li No D No sabe /No contesta

6. ^Habia sangre en su diarrea?

D Si • No D No sabe / No contesta

7. i,Cuantos episodios (separado por 7 dias) de diarreas tuvo en los ultimos 30 dias?

1 2 3 4 5 no sabe/ no contesta

B.2 DATOS ESPECIFICOS DEL VOMITO

8. i,Relaciono su vomito al consumo de

H Alimentos Remedios/laxantes

G Otro

9. <i,La semana antes de que usted enfermada, hubc

• SI IJ No G No sabe /No contesta

D Agua G Alcohol G

0 No sabe / No contesta

> algun familiar en su casa con diarreas o vomito?

233

10. ^Cuantos episodios (separado por 7 dias) de vomito en los ultimos 30 dias?

1 2 3 4 5 no sabe/ no contests

C. DATOS ESPECIFICOS DELAS DIARREAS Y EL VOMITO

11. iPor alguna causa uso antibioticos en las 4 semanas antes de que presento cuadros de diarreas o vomito?

GlSi • No D No sabe/ No contests

12. ^Durante este episodio de diarrea o vomito, tuvo/tiene alguno de los siguientes sintomas?

Fiebre D Nauseas G Calambre 0 Dolor de cabeza G Dolores musculares D Dolor garganta

I Romadizo (nariz con mucosa) G Tos/estornudo Otros ^Cuales?

i No sabe / No contests G No tuvo otros sintomas

13. <,Us6 algun remedio (antibiotico, analgesico, antidiarreico, liquidos, antiemetico (para controlar nausea), antiespasmodico, etc.) para tratar sus sintomas de diarrea o vomito?

I ] No G Antibiotico CJ Analgesico 0 Antidiarreicos D Liquidos D Antiemetico • Antiespasmodico

Otro D No sabe/ No contests

14 tCuantos dias de trabajo perdio por su diarrea o vomito?

dias D No sabe /No contests

15. ^Cuantos dias de escuela perdio por su diarrea o vomito?

dias 0 No sabe /No contesta

16. i,Cuantos dias dejaron de trabajar sus familiares para cuidarlo?

dias Q No sabe /No contesta

D. DATOS ESPECIFICOS DE HABITOS

17. c,Lava sus manos antes de comer?

Siempre I ; A veces G Nunca G No Sabe/No contesta

18. <j,Usa jabon para lavar sus manos antes de comer?

0 Siempre G A veces Sabe/No contesta

G Nunca G No


Siempre G A veces G Nunca D No Sabe/No contesta

20. Ûsa jabon para lavar sus manos despues de ir al bano?

G Siempre Q A veces Sabe/No contesta

0 Nuncs G No

21. i,Usa una tabla especifica para cortar su carne? (solo para carne).

22. i,Lava sus frutas y verduras antes de comer?

• Siempre G A veces icontesta

G Nunca D No Sabe/No G Siempre Q A veces Sabe/No contesta

Q Nunca D No

234

23. 6En la ultima semana (7 dias), consumio:

n Huevos crudos (tiramisu, bebidas, masa de postre, etc.) • Huevos no cocidos completamente (huevos a la copa o revueltos)

r: Mariscos crudos I i Ostras, mejillones, molusco 0 Pescado crudo n Pescado mal cocido

n Carne caida • Carne no cocida completamente (con centra rosado o con sangre)

I: Hamburguesas no cocidas completamente (con centra rosado o con sangre)

I ] Cerdo no cocido completamente (con centra rosado o con sangre)

n Pollo/ave no cocido completamente (con centra rosado o con sangre)

Ll Cordero no cocido completamente (con centra rosado o con sangre)

• Queso artesanal/casero/ de campo (hecho con leche no • Leche no pasteurizada pasteuhzada)

i Mayonesa envasado D Mayonesa hecho en casa

24. i,Sus respuestas en la pregunta 23, es lo que habitualmente consume usted en una semana normal?

I : Si r i No sabe /No contesta que?

No iPor

25. i,En la ultima semana, donde compro su carne, polio, cerdo, cordero?

I Un supermercado (Jumbo, Lider etc.) I Un negocio especializado (Carniceria) |J Un local de abarrotes (kiosco, almacen de Barrio), mercado [.; En la feria libre [j Otro

26. <j,En la ultima semana, donde compro su pescado o mariscos?

• Un supermercado (Jumbo, Lider etc.) G Un negocio especializado (Pescaderia) • Un local de abarrotes (kiosco, almacen de Barrio), mercado D En la feria libre • Otro

27. i,En la ultima semana, donde compro sus huevos y leche, queso otro productos de leche?

l l Un supermercado (Jumbo, Lider etc.) Un negocio especializado (codipra,) Un local de abarrotes (kiosco, almacen de Barrio),

mercado l.".1 En la feria libre • Otro

28. &En la ultima semana, donde compro sus frutas y verduras?

• Un supermercado (Jumbo, Lider etc.) D Un negocio especializado (fruteria, verduleria) D Un local de abarrotes (kiosco, almacen de Barrio), mercado • En la feria libre • Otro

29. i,Sus respuestas en las preguntas 25, 26, 27 y 28, corresponden a donde usted habitualmente compra alimentos normalmente?

l Si que?

• No sabe /No contesta D No ^Por

30. i,En promedio, cuantas comidas/veces come en un dia?

veces/dia

31. £,En la ultima semana (7 dias), cuantas veces comio en una restaurante (sentado)?

veces/semana

32. i,En la ultima semana (7 dias), cuantas veces comio en un casino/comedor o buffet?

veces/semana

235

33. i,En la ultima semana (7 dias), cuantas veces comio comida rapida (McDonald's, etc.)?

veces/semana

34. Ên la ultima semana (7 dias), cuantas veces comio algo preparado en la calle (hot dog, completa, etc.)?

veces/semana

35. £,En la ultima semana (7 dias), cuantas veces comio algo pedido a su casa/oficina (despacho al domicilio)?

veces/semana

36. i,En la ultima semana (7 dias), cuantas veces comio algo que preparo en su cocina?

veces/semana

37. <i,Sus respuestas en las preguntas 31, 32, 33, 34, 35 y 36, corresponden a donde usted habitualmente come?

n Si que?

n No sabe /No contesta D No iPor

38. iAyer, cuantas porciones (vasos, 250ml) de agua consumio (incluyendo jugos o bebidas que estan hechos con agua, pero sin considerar cafe o te?

porciones/ayer

39. iEsta cantidad (pregunta 38) es lo que habitualmente Ud. consuma de agua diario?

D Si • No sabe /No contesta

D No <,Por que?


Agua de una botella Agua de la casa (de la Nave).

Otro (numero de vasos)

141. i,Que usa para tratar su agua de la casa?

I : Nada G Hervir U Filtro del lavaplatos • Filtro de la jarra de agua (ej. Brita)

Otro

Comentarios:

236

Survey Tool - Sentinel clinic study (non-cases)

No. de encuesta Fecha... Encuestador:

A. INFORMACI6N GENERAL

Nombre:

Nombre de la Clinica:

Apellido:

Fecha de nacimiento Edad

0 Sexo: Masculino Femenino

Domicilio: Calle:

iComuna:

No. Dpto..

La vivienda que usted ocupa es: Propia ( pagado completo) Propia (pagado a plazo) Arrendada Cedida por trabajo o servicio Gratuita Otra

Ciudad: Telefono:

Numero de personas que viven en la casa: Tipo de vivienda:

Cantidad piezas para dormir que dispone:

Que animates tiene: ninguno perro oveja caballo otro

No sabe I No contesta gato pajaro vaca cabra gallina

Nivel de educacion:

No corresponde Analfabeto Kinder Basico o Primaria Media comun Media tecnico-


Normal Centra formacion tecnica Instituto profesional Universitario Post-Grado No sabe/No contesta Otro



i,De donde obtiene su agua? Red publico Pozo Rio Otro


(independiente) Asalariado sector privado Asalariado sector publico Personal domestico Estudiante Jubilado o pensionado


restaurante Cuida niflos Trabajador en

residencia de ancianos

Enfermero o medico Desocupado No sabe/No

contesta Otro

237

c A que sistema de salud pertenece usted? FONASA CAPREDENA ISAPRE Particular No tiene No Sabe/No Contesta Otro

iPor que decidio hacer esta encuesta?


6. DATOS ESPECIFICOS DE SU SALUD

1. i/Tiene diarrea cronica o sintomas de diarrea o vomito por alguna condicion diagnosticada por un medico (enfermedad de Crohn, colitis ulcerosa, colon irritable, cirugia biliodigestiva, abuso de laxantes, alcoholismo, radio o quimioterapia durante los dos meses anteriores, inmunodeficiencia (SIDA, otras) o si esta embarazada y ha presentado diarrea o vomito por esta razon)?

Ii Si n No D No sabe/ No contesta

2. <j,Us6 algiin antibiotico en los ultimos 30 dias (por cualquier causa)?


3. Indique cual es la razon por la que esta en la clinica hoy.

otra (Personas que no tuvo sintomas de diarrea o vomito en los ultimos 15 dias, pasa a parte C)

C. DATOS ESPECIFICOS DE HABITOS

A. i,Lava sus manos antes de comer?

11 Siempre iJ A veces I'J Nunca D No Sabe/No contesta

5. i,Usa jabon para lavar sus manos antes de comer?

D Siempre D A veces Sabe/No contesta

D Nunca 0 No


[J Siempre D A veces • Nunca • No Sabe/No contesta

7. iUsa jabon para lavar sus manos despues de ir al bano?

• Siempre D A veces Sabe/No contesta

D Nunca D No

8. £Usa una tabla especifica para cortar su carne? (solo para carne).

9. i,Lava sus frutas y verduras antes de comer?

Li Siempre contesta

D A veces D Nunca • No Sabe/No D Siempre D A veces Sabe/No contesta

• Nunca • No

238

10. £,En la ultima semana (7 dias), consumio:

• Huevos crudos (tiramisu, bebidas, masa de postre, etc.) n Huevos no cocidos completamente (huevos a la copa o revueltos)

I Mariscos crudos G Ostras, mejillones, molusco G Pescado crudo il Pescado mal cocido

l Came cruda LI Carne no cocida completamente (con centra rosado o con sangre)

r: Hamburguesas no cocidas completamente (con centra rosado o con sangre)

n Cerdo no cocido completamente (con centra rosado o con sangre)

I..) Pollo/ave no cocido completamente (con centra rosado o con sangre)

n Cordero no cocido completamente (con centra rosado o con sangre)

• Queso artesanal/casero/ de campo (hecho con leche no I] Leche no pasteurizada pasteurizada)

n Mayonesa envasado • Mayonesa hecho en casa

11. ^Sus respuestas en la pregunta 10, es lo que habitualmente consume usted en una semana normal?

u Si D No sabe /No contesta D No ,̂Por que?

12. £En la ultima semana, donde compro su carne, polio, cerdo, cordero?

I i Un supermercado (Jumbo, Lideretc.) J Un negocio especializado (Camiceria) 1 Un local de abarrotes (kiosco, almacen de Barrio), mercado I i En la feria libra l": Otro

14. i,En la ultima semana, donde compro sus huevos y leche, queso otro productos de leche?

i; Un supermercado (Jumbo, Lider etc.) I.: Un negocio especializado (Codipra,) l"i Un local de abarrotes (kiosco, almacen de Barrio), mercado Ll En la feria libra LJ Otro

13. <,En la ultima semana, donde compro su pescado o mariscos?

D Un supermercado (Jumbo, Lider etc.) D Un negocio especializado (Pescaderia) • Un local de abarrotes (kiosco, almacen de Barrio), mercado • En la feria libra a Otro

15. i,En la ultima semana, donde compro sus frutas y verduras?

Un supermercado (Jumbo, Lideretc.) D Un negocio especializado (Fruteria, Verduleria) D Un local de abarrotes (kiosco, almacen de Barrio), mercado • En la feria libra • Otro

16. <,Sus respuestas en las preguntas 12,13,14 y 15, corresponden a donde usted habitualmente compra alimentos normalmente?

Si que?

C No sabe /No contesta • No ,£,Por

17. Ên promedio, cuantas comidas/veces come en un dia?

veces/dia

18. <i,En la ultima semana (7 dias), cuantas veces comio en una restaurante (sentado)?

veces/semana

19. i,En la ultima semana (7 dias), cuantas veces comio en un casino/comedor o buffet?

veces/semana

239

20. iEn la ultima semana (7 dias), cuantas veces comio comida rapida (McDonald's, etc.)?

veces/semana

21. «|,En la ultima semana (7 dias), cuantas veces comio algo preparado en la calle (hot dog, completa, etc.)?

veces/semana

22. Ên la ultima semana (7 dias), cuantas veces comio algo pedido a su casa/oficina (despacho al domicilio)?

veces/semana

23. Ên la ultima semana (7 dias), cuantas veces comio algo que preparo en su cocina?

veces/semana

24. i,Sus respuestas en las preguntas 18,19, 20, 21, 22 y 23, corresponden a donde usted habitualmente come?

i: Si que?

11 No sabe /No contests l l No <j,Por

25. i,Ayer, cuantas porciones (vasos, 250ml) de agua consumio (incluyendo jugos o bebidas que estan hechos con agua, pero sin considerar cafe o te?

porciones/ayer

26. <j,Esta cantidad (pregunta 25) es lo que habitualmente Ud. consuma de agua diario?

• Si 111 No sabe /No contesta

n No i,Por que?


Agua de una botella Agua de la casa (de la Have).

Otro (numero de vasos)

28. i,Que usa para tratar su agua de la casa?

Nada D Hervir • Filtro del lavaplatos • Filtro de la jarra de agua (ej. Brita)

Otro

iComentarios:

240

1*1 Library and Archives - University of Guelph Atrium

Documents