Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effectiveness of community-directed diabetes prevention and control in a rural aboriginal population Daniel, Mark 1997

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1998-271285.pdf [ 20.06MB ]
JSON: 831-1.0088711.json
JSON-LD: 831-1.0088711-ld.json
RDF/XML (Pretty): 831-1.0088711-rdf.xml
RDF/JSON: 831-1.0088711-rdf.json
Turtle: 831-1.0088711-turtle.txt
N-Triples: 831-1.0088711-rdf-ntriples.txt
Original Record: 831-1.0088711-source.json
Full Text

Full Text

EFFECTIVENESS OF COMMUNITY-DIRECTED DIABETES PREVENTION AND CONTROL IN A RURAL ABORIGINAL POPULATION by MARK DANIEL B.Sc, Simon Fraser University, 1989 M.Sc, The University of Manitoba, 1991 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES Department of Health Care and Epidemiology We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA November 1997 © Mark Daniel 1997 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of V\g^vV4L <C^nxre - r.rJk ^E^\ Jvet^'v o \ o <s\^-| The University of British Columbia Vancouver, Canada Date DE-6 (2/88) ABSTRACT In response to the increasing prevalence and impact of non-insulin-dependent diabetes mellitus (NIDDM) in Canadian Aboriginal populations, a community-based diabetes prevention and control project of 24 months duration was implemented in the interior of British Columbia. A participatory approach was used to plan strategies by which diabetes could be addressed in ways acceptable and meaningful to the intervention community. The strategies emphasised a combination of changing behaviours and changing environments. Project workers implemented programme initiatives. Researchers served as facilitators and advocates for community change processes. The project was quasi-experimental. The intervention community was matched to two comparison communities. Workers in the intervention community conducted interviews of individuals with or at risk for diabetes during a seven-month pre-intervention phase (n = 59). Qualitative analyses were conducted to elucidate strategies for intervention. Baseline measures were obtained in each community, and implementation began in the eighth month of the project. A population approach was taken to diabetes prevention and control. Trend measurements of diabetes risk factors were obtained in each community for "high-risk" cohorts (persons with or at familial risk for NIDDM) (n = 105). Cohorts were tracked over the 16-month intervention phase, with measurements at baseline, the midpoint and completion of the study. Cross-sectional surveys of diabetes risk factors were conducted in each community at baseline and the end of the intervention phase (n = 295). Surveys of community systems were conducted during the pre-intervention and early and late intervention phases. The project yielded few changes in quantifiable outcomes. Activation of the intervention community was insufficient to enable individual and collective change through dissemination of quality interventions for diabetes prevention and control. Theory and previous research were not sufficiently integrated with information from pre-intervention interviews, nor were qualitative results brought to bear on activation and intervention planning. Interacting with these limitations were the short planning and intervention phases, just eight and sixteen months, respectively. The level of penetration of the interventions mounted was too limited to be effective. Attention to process is warranted, as is the feasibility of achieving effects within 24 months. iii TABLE OF CONTENTS Page Abstract ii Table of Contents iii List of Tables ; x List of Figures xii Acknowledgements xiii Dedication xiv Preface xv 1. Introduction 1 Policy Background 1 Purpose and Orientation 2 Statement of the Problem 6 Outline of the Dissertation 7 Part One — Review of Literature 9 2. Diabetes in the Canadian Aboriginal Population 1 0 Nature of the Disease 11 Relative Severity 15 Prevalence and Incidence 15 Mortality 18 Relation to Other Chronic Diseases in Aboriginal Populations 18 Weighing the Evidence on Importance 20 Aetiological Factors 21 Human Biology — Genetic Risk 21 Obesity 23 Physical Activity and Sedentary Lifestyle 26 Diet 28 iv Stress 30 Summary 32 Support for Community Interventions 35 Advantages of the Population-Based Approach 37 Limitations of the Population-Based Approach 38 Applied Prevention and Control Programmes in Canada 39 Socio-Cultural Considerations 42 Cultural Characteristics 43 Illness Interpretation and Health Belief Systems 43 Value Systems 45 Socio-Political Considerations 46 Quality of Life 48 Summary ....49 3. Theoretical Foundations 51 Accounting for a Relationship Between Acculturation and Diabetes 51 Social Determinants of Lifestyle and Behaviour 52 Acculturative Stress and Diabetes — a Theoretical Basis 58 Grand Theories 59 Process Theories 63 Theories to Explain and Predict Behavioural Change 66 The Health Belief Model 66 The Theory of Reasoned Action 67 Social Learning Theory 68 Strategies for Environmental Change 70 Community Organisation 71 Models of Community Change 72 Locality Development 72 Social Planning 73 Social Action 73 Implications for Prevention and Control 75 Linking Western Theory and Aboriginal Logic 76 Medical Rationality 78 Equilibrium Models 81 4. Methodological Issues in Community-Based Research 84 Community-Based Intervention Studies 85 Intervention Programmes As Social Experiments 87 Validity in Social Experiments 90 The Theoretical Model 93 Evaluative Design and Process 94 Treatment Theory and Data Analysis 96 Advancement of Treatment Effectiveness 97 Design Considerations 99 Unit of Analysis — Group versus Individual 99 Strategies for Analysis of Clustered Data 102 Analytic Efficiency and Studies Involving Few Social Units 105 Community Allocation and Sampling 110 Randomisation (Random Assignment) 110 Matching of Communities 111 Statistical Power 112 Sampling and Measurement in Communities 113 Data Analysis Issues 115 Fixed and Random Effects 115 Balanced and Unbalanced Data 117 Effects Models and Cell Means Models 120 Weighted and Unweighted Marginal Means 123 Analytic Approaches for Unbalanced Data 126 All-Cells-Filled Data 127 Some-Cells-Empty Data 127 Nested Classifications and Hypothesis Testing 129 A Prototypical Model 130 Omnibus Versus Planned Contrasts 131 Repeated Measures 133 Summary Measures 134 Univariate Analysis of Variance 138 vi Multivariate Analysis of Variance 140 Complex Models 141 Missing Data 143 Computer Software 146 Summary 147 Part Two — The Okanagan Diabetes Project 149 5. Context and Aims 150 Target Population 150 Provincial Context 150 Regional Context 151 Local Context 152 Planning Process 153 Background 153 Organisational Issues 156 Goals 158 Research Questions to be Addressed 158 6. Methodology • 160 Study Design 160 Sampling and Participant Selection 161 Cohort Surveys 161 Cross-Sectional Surveys 163 Community Surveys of Systems 165 Theoretical Model 166 Methods and Procedures 168 Development of Intervention Strategies 168 Worker Training 170 Diagnostic Pre-intervention Interviews 171 Qualitative Analysis 172 Knowledge, Attitudes and Cultural Values 174 Needs and Strategies for Diabetes Prevention and Control 176 Intervention Implementation 177 Interventions for Behavioural Change 177 vii Environmental Support for Behavioural Interventions 179 Monitoring and Programme Evaluation 180 Cohort Measurements 181 Physiological Variables 181 Analytic Methodology for Blood Measures 183 Laboratory Procedures 185 Anthropometric Variables 187 Waist and Hip Circumferences 188 Body Mass and Height 188 Behavioural Variables 189 Physical Activity 189 Dietary Behaviour 191 Psychosocial Variables 192 Self-Esteem 193 Mastery 194 Depression 194 Affect Balance 194 Social Support 195 Diabetes Quality of Life 195 Diabetes Health Beliefs 196 Participation in Intervention Initiatives 196 Cross-Sectional Measurements 198 Surveys of Community Systems 199 Data Preparation and Analysis 200 Statistical Analysis 202 Cohort Measurements 205 Cross-sectional Measurements 208 Software 209 7. Results 2 1 0 Cohort Surveys 210 Participation in Data Collection 210 Baseline Differences, Drop-outs, and Missing Data 211 viii Descriptive Characteristics of Cohort Samples 215 Participation in Interventions — Correlates and Impact 217 Physiological Outcomes 218 Diabetes Cases, Glycaemic Variables and Lipids 218 Blood Pressure 222 Anthropometric Outcomes 222 Behavioural Outcomes 224 Dietary Intake Over Three Days 224 Physical Activity 224 Alcohol Consumption 226 Smoking Status 226 Psychosocial Outcomes 226 Diabetes Knowledge, Quality of Life and Health Beliefs 226 Cross-Sectional (Community Level) Surveys 230 Sampling Fractions and Response Rates 230 Descriptive Characteristics of Samples 230 Diabetes Knowledge and Awareness 233 Behavioural and Anthropometric Outcomes 233 Co-Variation Between Changes in Knowledge and Risk Factors 233 Statistically Significant Results and Adjustments for Multiple Comparisons 235 Cohort Analyses 235 Cross-Sectional Analyses 237 Intervention Community Surveys of Systems 238 Sub-System Actions Involving Community Groups 238 Sub-System and Supra-System Relationships 238 Whole System Norms and Values 239 8. Discussion 244 Effects in Relation to Research Questions 244 Implications 247 Inadequate Implementation 247 Sufficiency of Community Activation 248 Adequate Time for Implementation : 252 ix Participation 254 Appropriateness of Programme Theory 256 Methodological Challenges and Limitations 261 Validity of Statistical Conclusions 262 Statistical Power 262 Appropriateness of Statistical Procedures 265 Internal Validity of Conclusions 267 Selection Bias and Allocation Bias 267 Sampling Bias 270 Validity of Self-Reported Data 270 Observation Bias and Misclassification Bias 271 Missing Data and Attrition from Cohorts 272 Cross-Contamination and Secular Trends 273 External Validity (Generalisability) 275 Synthesis 277 Recommendations for Future Initiatives 281 Summary and Conclusion 287 REFERENCES 292 APPENDICES A. Calculations of Sample Size and Statistical Power 351 B. Demographic Questionnaire 354 C. Anthropometric Proforma ....358 D. Lifestyle Questionnaire 360 E. Social Environment Questionnaire 365 F. Diabetes Quality-of-Life Questionnaire 370 G. Diabetes Health Beliefs and Knowledge Questionnaire 376 H. Cross-Sectional Community Survey for Interviews 380 I. Community Systems Surveys by Analytic Level 383 LIST OF TABLES Table Page 1. Diabetes Prevalence Rates Among Registered Indians in British Columbia at Year-End 1987, by Medical Services Branch Geographic Area, Age and Gender 17 2. Diabetes Therapy Among Registered Indians in British Columbia at Year-End 1987 18 3. Most Frequent Chronic Conditions Responsible for Follow-Up Among Registered Indians in British Columbia as of February, 1990 19 4. Most Frequently Reported Chronic Conditions for On-Reserve Indian Population in Saskatchewan, 1986 19 5. Risk Factors Associated with the Development and Control of Non-lnsulin-Dependent Diabetes Mellitus Among Aboriginal Peoples and Ratings of Each on Importance and Changeability 33 6. Project Timeline and Timing of Data Collection, Overall and by Month of Intervention 180 7. Programme Evaluation Model: Sampling and Measurement Strategies 182 8. Precision of Tests of Blood Measures 185 9. Intra-Class Correlations from the Minnesota Heart Health Program Used to Calculate Variance Inflation Factors to Adjust for Clustered Observations 204 10. Non-Response as a Proportion of All Persons Enrolled at Baseline in Cohorts by Condition, Community and Time of Data Collection 211 11. Characteristics of Persons Enrolled in Cohorts and Measured at Baseline by Condition and Participation Status 212 12. Characteristics of Cohort Samples by Condition and Community 216 xi 13. Glycaemic Variables and Lipids at Baseline, Mid-Project and End of Project, and Tests of Differential Change Over Time Between Conditions: Cohort Surveys 219 14. Prevalence of Clinically Elevated Metabolic Values at Baseline, Mid-Project and End of Project, and Tests of Change Over Time Within Conditions: Cohort Surveys 221 15. Body Mass Index, Waist-to-Hip Girth Ratio, and Blood Pressure at Baseline, Mid-Project and End of Project, and Tests of Differential Change Over Time Between Conditions: Cohort Surveys 223 16. Physical Activity and Dietary Behaviour at Baseline, Mid-Project and End of Project, and Tests of Differential Change Over Time Between Conditions: Cohort Surveys 225 17. Psychosocial Measures at Baseline, Mid-Project and End of Project, and Tests of Differential Change Over Time Between Conditions: Cohort Surveys 227 18. Mean Diabetes Knowledge, Quality of Life and Health Beliefs, and Net Intervention — Comparison Differences Over 16 Months from Baseline to End of Project: Cohort Surveys of Persons with Diabetes and Impaired Glucose Tolerance 228 19. Sampling Frames, Sample Sizes, and Response Rates by Condition and Community for Cross-Sectional Surveys at Baseline and End of Project 231 20. Characteristics of Simple Random Samples Surveyed at Baseline and 16 Months Later at End of Project, by Condition and Community 232 21. Diabetes Knowledge and Risk Factors, and Net Intervention - Comparison Differences Over 16 Months from Base-line to End of Project: Cross-Sectional Surveys of Adults Ages 18-87 234 22. Cohort Survey Means and 95% Confidence Intervals, Adjusted for Clustering: Variables for Which Tests of Change Over Time Between Conditions Were Statistically Significant 236 23. Cross-Sectional Survey Means and 95% Confidence Interval: Variables for Which Tests of Change Over Time Between Conditions Were Statistically Significant 237 24. Community Participation By Analytic Level: Systems Evaluation 241 LIST OF FIGURES Figure Page 1. Direct and Indirect Effects of Environmental and Behavioural and Lifestyle Factors on Health and Quality of Life 53 2. Theoretical Model by Month of Project 167 3. Distribution of Age for Longitudinal Cohorts by Community 215 4. Percentage Glycosylated Haemoglobin by Glycaemic Status, Condition and Time, for Longitudinal Cohort Samples 220 5. Body Mass Index by Condition and Time, for Longitudinal Cohort Samples 222 ACKNOWLEDGEMENTS I wish to express appreciation to all who assisted and contributed in the course of this work. Members of the project team included: Beatrice Bonneau, Okanagan Indian Band, Vernon; Sandy Burgess, R.N., Medical Services Branch, Health Canada, Salmon Arm; Laura Miller, Okanagan Indian Band, Vernon; Linda Mitchell, Spallumcheen Indian Band, Enderby; Sadie Muik, Okanagan Indian Band, Vernon; Jackie Pierre, Penticton Indian Band, Penticton; and Bernice Strachan, R.N., Diabetes Day Programme, Vernon Jubilee Hospital, Vernon. For the opportunity to undertake this work, I am grateful to Diane Gamble, M.N., R.N., and Joyce Henderson, M.P.H., R.N., both of the Division of Health and Social Programmes, Okanagan University College, Kelowna. Thanks are also due to Claire M. Budgen, Ph.D., R.N., Professor, Division of Health and Social Programmes, Okanagan University College, for helpful assistance throughout the project. For allowing and accepting me into their communities, I am indebted to the people of the Okanagan, Spallumcheen and Penticton Indian Bands. I am also thankful for the assistance and co-operation of Advisory Committee members and the Band Councils of all three communities. Special thanks for technical assistance are extended to Betty Carlson, R.T., and Lowell Laidlaw, R.T., Director, Laboratory, Vernon Jubilee Hospital, and to Suzanne Johnson, B.Sc, R.D.N., and Susi Wilkinson, B.Sc, R.D.N., South Okanagan Health Unit. I appreciate the service of student volunteers from the degree nursing programme at Okanagan University College, and the contributions of Research Assistants Deborah Gray, B.S.N., R.N., and Carol McLeod, B.S.N., R.N. I am thankful for the financial support of Health Canada, specifically, for the award by the National Health Research and Development Program (NHRDP) of a project grant (#6610-2022-ND) under the 1992 Special Initiative on Research on Diabetes in the Canadian Aboriginal Population, and a National Health Ph.D. Training Fellowship (1994-1997). I also acknowledge the Killam Trusts, as an Honourary Killam Scholar (1994-1996) and for the award of a one year Killam Pre-Doctoral Fellowship (1997-1998). Last, but far from least, I wish to thank the members of my supervisory committee: Lawrence W. Green, Dr.P.H., Professor, Department of Health Care & Epidemiology, Director, Institute of Health Promotion Research; Carol P. Herbert, M.D., Professor and Head, Department of Family Practice; Clyde Hertzman, M.D., M.Sc, Professor, Department of Health Care & Epidemiology; Stephen A. Marion, M.D., M.S., M.H.Sc., Associate Professor, Department of Health Care & Epidemiology, and Sam B. Sheps, M.D., M.Sc, Professor and Head, Department of Health Care & Epidemiology. During my tenure as a doctoral student I have faced a complex array of academic, intellectual, political, ethical and moral challenges. Each member of my supervisory committee has played a key role in enabling me to address and resolve such issues, and in so doing, stimulating my personal and professional development. For their commitment and the quality of their mentoring, I am most appreciative. xiv DEDICATION To my family. Read not to contradict and confute, nor to believe and take for granted, nor to find talk and discourse, but to weigh and consider. - Francis Bacon P R E F A C E This dissertation reports on and evaluates the effectiveness of the Okanagan Diabetes Project, an initiative funded by the National Health Research and Development Program (NHRDP), Health Canada (NHRDP #6610-2022-ND). A preliminary evaluation of the project was undertaken and submitted to the N H R D P in September 1996 (Daniel & Gamble, 1996). Aspects of the preliminary report are reviewed in this dissertation. This dissertation, however, with the prior knowledge and approval of the N H R D P , constitutes the main evaluation and statement on the effectiveness of the Okanagan Diabetes Project. Although the statistical techniques differed and the scope of the preliminary report was limited, its main conclusions are not altered by the additional data and broader review set forth in this dissertation. This evaluation builds on the preliminary report through a detailed appraisal of the circumstances and conditions associated with the project. CHAPTER 1 INTRODUCTION Policy Background In 1965 the World Health Organisation (WHO) released its First Report on diabetes mellitus* (World Health Organisation, 1965). Perhaps because the emphasis in public health at that time continued, at least in the WHO and developing countries, to be on communicable diseases, the First Report was of little consequence. Virtually all bodies ignored its recommendations (Alberti, 1980). The WHO'S Second Report on diabetes, released in 1980 (World Health Organisation, 1980), had considerably more impact. Influenced by the International Conference on Primary Health Care, held in Alma Ata in 1978 (World Health Organisation, 1979), the Second Report emphasised community level preventive, promotive, diagnostic and rehabilitative aspects of diabetes. Recommendations detailed in the WHO'S Second Report have been widely implemented. Nevertheless, in Canada, the specific recommendation of the Second Report that "the concept of primary prevention should be vigorously explored with particular attention to high-risk people and to environmental factors" (World Health Organisation, 1980) received little emphasis until recently. Only limited diabetes research had been undertaken in Canada in 1989, when the World Health Assembly adopted the landmark "Resolution on the Prevention and Control of Diabetes" (World Health Organisation, 1990). The Assembly resolution served to provide a mandate for joint action on diabetes involving governments, health care providers and diabetes organisations. As a basis for action the resolution made the points that diabetes is a chronic, debilitating and costly disease attended by severe complications, and that diabetes imposes a growing burden on public health in all countries. Accordingly, the Assembly invited member states to assess the national importance of diabetes and to implement population-based measures appropriate to the local context to prevent *The term diabetes is used hereafter, rather than diabetes mellitus, in reference to the disease. and control the disease. An aim of these actions was to establish models for an integrated approach to diabetes prevention and control at the community level. The growing problem of diabetes in the Canadian Aboriginal population requires a strong public health response, in keeping with the challenges laid out by the World Health Assembly. Since 1986, this need has been recognised by the National Health Research and Development Program (NHRDP) of Health Canada, through conferences and working groups. A national survey was undertaken in the late 1980s on the extent and impact of diabetes in the Canadian Aboriginal population (Young etal., 1990). Population-based actions were few, however, until the 1992 NHRDP Special Initiative on Research on Diabetes in the Canadian Aboriginal Population., This was a special research competition intended to support community-based diabetes interventions with strong evaluation components (Young & Ross, 1991). Purpose and Orientation This dissertation undertakes an in-depth appraisal of the issues associated with diabetes and its prevention and control in the Canadian Aboriginal population,* and reports on and evaluates the effectiveness in a rural Aboriginal population of a community-directed project that arose from the 1992 NHRDP Special Initiative to prevent and control diabetes. Methodological issues specific to undertaking and evaluating community-based research are also reviewed in detail. A prevalent theme throughout this dissertation is that diabetes in the Canadian Aboriginal population is both a consequence of environmental changes associated with acculturation, as "Different terms are used to describe broad ethnic groups (Young, 1994). In keeping with common usage, the term Native American refers to the indigenous populations of Canada and the United States, which are distinguished by country through the descriptors Aboriginal Canadian and either American Indian or Alaska Native. In differentiating Aboriginal Canadians, reference is made as necessary to Indians and Inuit for those whose health care services are administered by the federal government (i.e., "registered" Indians and Inuit), versus Metis and others reporting Aboriginal identity who are neither registered Indians nor Inuit. The above terms are used to simplify and ease communication, but it is acknowledged that the appropriateness of any particular term falls to the groups being discussed, and therein lie many group and individual differences. Any one term may be offensive to some people. For example, tribal names with a history of use (e.g., Cree, Ojibwa, Dogrib, Haida, and so forth) are used instead of less familiar tribal names given by individual Aboriginal languages, though there is a growing tendency, especially in British Columbia, towards the use among Aboriginal people of language-specific tribal names. The descriptor First Nation for Indian people has been avoided because it is awkward and does not include Inuit and Metis, though it is often used in a manner suggesting simultaneous reference to all Aboriginal peoples of Canada. well as an inherent genetic predisposition. In this context, acculturation is the rapid process by which the cultures of Aboriginal peoples are exposed to, diluted and made dependent on modern, "western" ways of living and external resources inconsistent with traditional patterns. It reflects the tremendous political, economic and social challenges faced by Aboriginal people primarily during the second half of the 20th century (Young, 1994). The impact of acculturation and the stress of rapid environmental change on the risk of developing diabetes is supported by ecological data from several migrating populations traditionally free of the disease (Dowse et al., 1990; Fujimoto etal., 1987; Kawate etal., 1979; Medalie etal., 1975; Shanghai Diabetes Research Co-operative Group, 1980; Tai etal., 1987). Thus, despite the important role of genetic factors, diabetes (at least the non-insulin-dependent type) can — and should — be considered a largely preventable disease (Manson & Spelsberg, 1994). An emphasis on acculturation is not to suggest that Aboriginal people are in some way deficient or incapable of responding effectively to environmental change. A litany of "culture bound syndrome" stereotypes (Webster & Nabigon, 1993) suggests that the health problems and social pathologies affecting indigenous populations represent a failure of the Aboriginal culture to adapt to stressful socio-economic conditions. This position assumes an unrealistically high level of self-responsibility for health, and ignores the role of environmental forces over which Aboriginal peoples have had little control. Rather, the stress of acculturation, and political, economic and social subjugation and relegation to low status identity, reflect the general failure of Canadian society to validate Aboriginal culture and, in doing so, promote Aboriginal health. There is an intimate relationship between culture and health. Devaluation of Aboriginal culture and the imposition of non-Aboriginal "solutions" for ill health associated with loss of culture and cultural identity does not recognise Aboriginal people as legitimate citizens with a history, traditions and values that may offer other solutions that are relevant, meaningful and effective (Colorado, 1988). For example, a deterioration in health status followed the loss of tribal status of the Klamath Indians of Oregon in 1954. Thirty years later, their level of unmet need and their absolute health status, relative to age, was poorer than that of the general population and other American Indian groups that retained their cultural identity (Joos & Ewart, 1988). The effects of acculturation on health are evident through Aboriginal versus non-Aboriginal population contrasts, as well as in variability within Aboriginal populations. Evidence from other populations of marginalised people (e.g., Hispanics in the United States) indicates that acculturation is a multi-dimensional phenomenon (Berry, 1980; Padilla, 1980). Among Hispanic samples, acculturation is associated with poor mental health status (Griffith, 1983), low levels of social support (Griffith & Villavicencio, 1985), alcohol and drug abuse (Graves, 1967) and premature mortality (Wei etal., 1996). Scribner and Dwyer (1989) found that acculturation among Hispanic women significantly predicted low birthweight status. In this study, U.S.-oriented mothers had a risk of ever having delivered a low birthweight infant that was 1.64 times greater than that of Mexican-oriented mothers. A secondary analysis of the same study using structural equation models indicated that smoking and dietary intake mediated indirectly the effect of acculturation on low birthweight status (Cobas etal., 1996). Psychosocial influences related to acculturation mediate determinants of smoking (Deosaransingh etal., 1995). Whether all dimensions of acculturation are equally relevant to particular health outcomes is unclear, as is the nature of the process by which acculturation influences health. As used throughout this dissertation, the term "health" refers to a state of physical, mental and social well-being. This conception derives from the Ottawa Charter for Health Promotion (1986), which defines health in relation to health promotion: Health promotion is the process of enabling people to increase control over, and to improve, their health. To reach a state of complete physical, mental and social well-being, an individual or group must be able to identify and realise aspirations, to satisfy needs, and to change or cope with the environment. Health is, therefore, seen as a resource for everyday life, not the objective of living. Health is a positive concept emphasising social and personal resources, as well as physical capacities. Therefore, health promotion is not just the responsibility of the health sector, but goes beyond healthy lifestyles to well-being. (Ottawa Charter on Health Promotion, 1986) This and other contemporary definitions of health differ from earlier conceptualisations in that (a) health is not simply the absence of disease and (b) health is linked to the social and economic environment, and to the way society operates. Health refers to a dynamic quality both individually and socially valued, with simultaneously objective and subjective dimensions. Overall health is a global quality, but health consists of particular qualities as well (e.g., physical, psychological, and social well-being and functioning) (Noack, 1991). "Sickness" is the sum of disease and illness, where "disease" is the objective, measurable aspect of sickness, and "illness" is the subjective, difficult-to-measure experience of disease (Eisenberg, 1977). Health status may be improved by attention to culture where its denial or oppression is associated with alienation (Axelos, 1976), anomie (Durkheim, 1933), or normlessness (Merton, 1949). That these conditions are relevant to the situation of Canadian Aboriginal people is supported by testimonials (Harris, 1995a; Mussell, 1992). It is also supported by demographic data describing outcomes and circumstances in keeping with those predicted by sociological theory (e.g., suicide, drug and alcohol abuse, mental illness, high rates of chronic disease and early mortality) (Cooper, 1995; Foster era/., 1995; MacMillan etal., 1996), as well as Aboriginal position papers and policy documents (Mardiros, 1987). For example, in a discussion paper written for the Assembly of First Nations Health Secretariat, Lemchuk-Favel (1995) distinguished the impact of socio-economic circumstances on health. It characterised these as interactive with ... cultural alienation or cultural stress ... caused by oppressive experiences such as loss of land, loss of control over living conditions, restricted economic opportunity, suppression of beliefs and spirituality, weakening of social institutions, displacement of political institutions, pervasive breakdown of cultural rules and values and diminished self-esteem, discrimination and institutional racism and their internalised effects, and voluntary or involuntary adoption of elements of an external culture and loss of identity. (Lemchuk-Favel, 1995, p. 7) Changes in behaviour and lifestyle interact with specific reactions associated with social, cultural and economic stress in predisposing the development of diabetes in vulnerable populations (World Health Organisation, 1994). Among indigenous peoples around the world, a growing body of evidence suggests that diabetes and other chronic diseases are a consequence of lifestyle changes associated with acculturation. In this context, the notion of "lifestyle" implies more complex, repetitive, if not habitual, patterns of behaviour conditioned by living standards and the set of social conditions that surround a social group, including cultural history and socio-economic circumstances (Green et al., 1996). Lifestyle is a composite expression of enduring patterns of behaviour that reflect the social and cultural circumstances that condition and constrain behaviour, as well as the consciously chosen, personal behaviour of individuals (Green & Kreuter, 1991). The public health application of this construction of lifestyle has been to seek policies and environmental regulations to support healthful living. The mechanisms for achieving change in lifestyle such as the development of personal skills and the strengthening of community action supported by social and economic policies, however, continue to be difficult to attain at both local and national levels (Gunning-Schepers & Gepkens, 1996). More than any other chronic disease, diabetes exemplifies the interaction among the four determinants of health — human biology, the environment, lifestyle and behavioural factors, and the health care system — considered in the milestone federal report "A New Perspective on the Health of Canadians" (Lalonde, 1974). The emergence of diabetes as one of the most pressing health issues facing Aboriginal Canadians today represents the outcome of an interaction among genetic factors, environmental, lifestyle and behavioural factors associated with acculturation. These, in turn, interact with the limited capacity of western biomedicine and the traditional health care system to meet the needs of Aboriginal peoples in culturally relevant and meaningful ways. The development of prevention and control strategies for diabetes among Aboriginal Canadians has become increasingly important because of its significant morbidity and mortality and the human and economic costs associated with its complications. A health promotion framework offers the potential to research culturally sensitive control and prevention initiatives to counter diabetes in Aboriginal communities. Uncertainties regarding the precise aetiological role of risk factors such as obesity, physical activity and diet in the development of diabetes lend support to a multiple-risk-factor intervention approach, through a healthful lifestyle strategy, for prevention and control. This approach has the potential to impact on other chronic diseases also influenced by lifestyle. Statement of the Problem This study responds to three problems: (a) the high prevalence of diabetes and the impact of the disease among the Canadian Aboriginal population; (b) the limited effectiveness of traditional health care approaches as applied to meet the needs of Aboriginal populations; and (c) the need to address through prevention the role of behavioural, lifestyle and environmental factors. It addresses a need to develop and implement, in collaboration with Aboriginal people, community-based, community-directed, culturally sensitive diabetes prevention and control programmes. It also seeks to evaluate the impact of such programmes on health outcomes, broadly defined to include psychosocial constructs and indicators of well-being, and on behavioural, environmental, physiological and metabolic risk factors for diabetes. Outline of the Dissertation This dissertation is organised according to the traditional format. Following the policy background, stated purpose and orientation provided as an introduction, it contains a literature review, a presentation of the methodology and procedures used, a presentation of results and a concluding discussion. Some flexibility has been required, however, with respect to form. This need is met by grouping chapters into two parts. Part One provides a systematic review of the literature, and a rationale for the development, implementation and evaluation of the project reported here. This review has three components, corresponding to Chapters 2, 3 and 4, respectively: (a) diabetes in the Canadian Aboriginal population; (b) theories accounting for a relationship between diabetes and acculturation and by which to explain and predict behavioural and environmental change; and (c) methodological issues involved in the design and analysis of community-based trials. Chapter 4 (the methodological review) could be skimmed in a general reading. The chapter exists to provide an orientation to methodological challenges peculiar to evaluating small-scale social interventions as applied research, and an explanatory background justifying decisions made in the design and analysis of the trial reported here. Were such issues to be reviewed in the description of the methodology used, the number of footnotes, qualifying statements, historical asides and other explanations would surely have compromised readability. The methodological review is relevant to this dissertation because there is no established orthodoxy to fall back on to evaluate small-scale community trials. Options for reconciling methodological difficulties are not widely known or well-accepted. Relevant strategies arise from and reflect a variety of paradigms, assumptions, methods and kinds of data that share some broad resemblances. Together, these do not suggest the coherence and consensus among researchers that would constitute a single paradigm. As the science of epidemiology provides an organising framework for this dissertation, it was necessary to integrate, from this perspective, diverse methodological strategies with roots in other disciplines, primarily the social sciences. The challenges involved in the work reported here required use of methods and analytic techniques not widely applied in epidemiological research. Part Two describes the community-based initiative evaluated here, known as the Okanagan Diabetes Project. Chapter 5 summarises the demographic characteristics of the population and communities involved in the project. Planning and organisational issues are reviewed, and the chapter concludes with a statement of goals and research questions. Chapter 6 covers the methodology used. The design of the project and sampling strategies are described first. This is followed by a review of the theoretical model on which the project was founded. A section on methods and procedures covers issues pertaining to each phase from the development and implementation of interventions to monitoring and programme evaluation procedures, data preparation and statistical analysis. Measurement procedures are described in detail. Chapter 7 presents the results of the project. Chapter 8 begins with a summary of results, and discusses the implications of these in relation to implementation and methodological issues. Contextual factors and previous research are also considered. The dissertation concludes with recommendations for future intervention research on diabetes or other health issues in indigenous populations. PART ONE — REVIEW OF LITERATURE C H A P T E R 2 DIABETES IN THE CANADIAN ABORIGINAL POPULATION Native Americans are a heterogeneous collection of indigenous peoples at substantial risk for the development of diabetes. Diabetes is an "indicator" disease of the "epidemiologic transition" that all societies undergo (Crews & McKeen, 1982; Kuberski & Bennett, 1980; Schooneveldt etal., 1988). Along with a decrease in the morbidity and mortality of infectious diseases over the last several decades has been an increase among Aboriginal Canadians in chronic diseases including diabetes (Muir, 1991; Young, 1988). Diabetes is a debilitating condition associated with severe complications that occur at frequencies often greater in Aboriginal than in non-Aboriginal populations (Bennett & Knowler, 1984). Further, the medical case-load associated with diabetic Native Americans is more than twice that of non-diabetic Native Americans, as determined by comparisons of rates of outpatient and inpatient utilisation patterns (Reinhard & Greenwalt, 1975). Complications observed among Aboriginal populations include those that arise from damage to arteries (macrovascular complications) such as cardiovascular disease (i.e., ischaemic heart disease, cerebrovascular disease and peripheral vascular disease) (Hoy etal., 1995; Macaulay etal., 1988; West etal., 1983) and hypertension (Montour & Macaulay, 1985; Young et al., 1985). Other complications arise from impaired blood supply to specific organs (microvascular complications) such as retinopathy (Lee etal., 1992; Nelson etal., 1989), nephropathy (Dyck & Tan, 1994; Newman etal., 1990; Young etal., 1989) and neuropathy (Nelson etal., 1988; Ross & Fick, 1990). The relative prevalence of these conditions in the general population, comparing diabetes cases with non-diabetics, ranges from 2.5 for hypertension to 10.3 for blindness (Huse era/., 1989). The prevention of complications is the primary rationale for glucose control among persons with established diabetes (Benjamin & Sacks, 1994; Nathan, 1992; Rossetti etal., 1990). The microvascular complications of diabetes are severe, and the macrovascular complications ischaemic heart disease and cerebrovascular disease are the leading causes of death in developed regions of the world. These causes of death are not always related to diabetes, although reducing the incidence of diabetes has the potential to avert more of the disease burden (both deaths and the impact of premature death and disability on a population) in developed regions than reductions in any other chronic disease (Murray & Lopez, 1996). Notwithstanding the WHO'S Second Report on diabetes mellitus (World Health Organisation, 1980) and the World Health Assembly's "Resolution on the Prevention and Control of Diabetes" (World Health Organisation, 1990), the extent of the problem among Canadian Aboriginal peoples indicates a clear need for research on community level preventive, promotive and rehabilitative aspects of diabetes (Daniel & Gamble, 1995). This research should parallel enquiry into genetic and environmental determinants of the disease. The purpose of this chapter is as follows: • To define diabetes and to describe its diagnosis; • To ascertain the importance of diabetes in relation to the health of Aboriginal Canadians and to justify the use of resources to develop and implement prevention and control strategies to counter diabetes among Canadian Aboriginal populations; • To outline aetiological factors related to the prevention and control of diabetes among Aboriginal and other populations; • To review the extent of support for population-based approaches to diabetes prevention and control, including an appraisal of programmes known to have been undertaken in Canada; • To consider cultural characteristics of Aboriginal people, and illness interpretation, health belief and value systems relevant to meaningful and acceptable diabetes prevention and health promotion initiatives; and • To appraise the problem in relation to quality of life. Nature of the Disease Blood glucose concentration is normally tightly regulated by the action of insulin and counter-regulatory hormones (primarily epinephrine and glucagon), with a balance maintained between glucose production by the liver and glucose clearance into peripheral tissues. Insulin released from p-cells of the pancreatic islet is constantly adjusted so that normoglycaemia is maintained. Among persons with diabetes, the regulation of blood glucose concentration is impaired (Shamoon, 1992). The hallmark of diabetes is hyperglycaemia, abnormally high concentrations of blood glucose consistent with either an absolute or relative deficiency of insulin (Wu, 1993). A deficit in insulin action can be caused by either a failure in pancreatic insulin secretion or cellular resistance to the action of insulin. Classic symptoms of diabetes include frequent urination, weight loss, excessive thirst and hunger, blurred vision and recurring infections, but in its early stages the disease is often asymptomatic. The term diabetes mellitus refers to a heterogeneous group of disease states with the common feature of hyperglycaemia due either to insulin deficiency or insulin resistance. Excess urine production resultant from hyperglycaemia is the origin of the word diabetes, which is Greek for "syphon," while mellitus is Latin for "honeyed," reflecting the high sugar content of urine in diabetes (glycosuria). Harris (1995b) classified four major types of diabetes defined by the U.S. National Diabetes Data Group (NDDG) (1979) and the World Health Organisation (1980,1985): (a) insulin-dependent diabetes mellitus (IDDM), also known as Type I diabetes mellitus; (b) non-insulin-dependent diabetes mellitus (NIDDM), also known as Type II diabetes mellitus; (c) gestational diabetes mellitus (GDM); and (d) other types, including diabetes secondary to or associated with pancreatic disease, hormonal disease, drug or chemical exposure, insulin receptor abnormalities, and certain genetic syndromes. IDDM usually, but not always, occurs before the age of 30 years. It presents with abrupt onset of classic diabetes symptoms and extreme hyperglycaemia associated with ketoacidosis that requires prompt medical treatment. Detectable concentrations of circulating endogenous insulin will be either low or absent. Approximately 10% of all people (adults and children) diagnosed with diabetes have this type. IDDM is thought to be caused by the autoimmune destruction of the pancreatic islet p-cells as a result of the combined effect of both genetic and environmental factors (Shamoon, 1992). Persons with IDDM require insulin replacement therapy to prevent diabetic ketoacidosis, coma, and death from insulin deficiency. NIDDM is usually found in adults older than 30 years, and it accounts for close to 90% of all diagnosed cases of diabetes. The incidence of the disease increases with age, but in high-risk populations susceptible persons develop NIDDM at earlier ages. The prevalence of diabetes in Pima Indians aged 25-29 years (13%) is, for example, as high as that for non-Hispanic Caucasians aged 60-64 years in the U.S. (Rewers & Hamman, 1995). Individuals with NIDDM may be symptom free for many years, as the onset and progression of symptoms occur slowly, and for this reason most unidentified cases of diabetes are NIDDM (American Diabetes Association, 1989). Diet and exercise are the mainstays of treatment for people with NIDDM, but oral hypoglycaemic agents (drugs that increase insulin release from the pancreas and utilisation of insulin by peripheral tissues) and insulin replacement therapy are also used to manage the disease (Shamoon, 1992). Insulin is usually not needed to prevent ketoacidosis in NIDDM, however, since some insulin continues to be produced by the pancreas (Rewers & Hamman, 1995). GDM is a disorder with onset of symptoms of glucose intolerance during pregnancy. It occurs in 2-5% of pregnancies and usually disappears after delivery (Coustan, 1995). Women with GDM may have an elevated risk for perinatal mortality and morbidity, and they are at greater risk for developing diabetes and coronary heart disease later (Benjamin etal., 1993). Pregnancy induces insulin resistance, which may precipitate overt hyperglycaemia among women with subclinical IDDM or NIDDM. The proportion of cases of GDM in which the disease is acquired during pregnancy rather than uncovered by testing, however, is unclear (Rewers & Hamman, 1995). GDM could either arise from the stresses of pregnancy or reflect pre-existing abnormal glucose tolerance. An often studied "adverse" outcome is macrosomia, variously defined as birth weight more than 4,000 g or 4,500 g (Coustan, 1995). In Pima Indian women submitting to oral glucose tolerance tests, maternal glucose concentrations were directly related to the likelihood of macrosomia and the relative weight of the off-spring up to 14 years of age (Pettitt et al., 1991).* Other types of diabetes associated with diseases, medications, chemicals, or genetic syndromes, account for approximately 1% to 2% of all disorders comprising the syndrome of diabetes (Ganda, 1995). Diabetes associated with other conditions may be secondary to the pathogenesis of these conditions, classified as: (a) pancreatic disorders (e.g., pancreatis, pancreatic malignancy, haemochromatosis, etc.); (b) endocrinopathies (e.g., acromegaly, Cushing's syndrome, phenochromocytoma, hyperthyroidism, etc.); (c) drugs, chemicals and *A tendency among the Pima toward obesity with increasing age seemed to obscure any effect of maternal glycaemia beyond age 14 years. toxins (e.g., diuretics and antihypertensive agents, psychoactive agents, antoprotozoals, anti-convulsants, rodenticides, etc.); and (d) genetic syndromes (e.g., pancreatic deficiencies, mutant insulin syndromes, glucokinase gene mutations, chromosomal defects, etc.) (Coustan, 1995).* Standardised criteria for diabetes defined by the NDDG and WHO permit discrimination among types of diabetes by different clinical presentations and genetic and environmental aetiologic factors (Harris, 1995b). It can be difficult to distinguish between types of diabetes, however, and classifying the disease involves both a clinical workup and an adequate medical history. Diagnosis of the disease among persons with classic diabetic symptoms is based on either random plasma glucose > 11.1 mmol/L or fasting plasma glucose > 7.8 mmol/L (Harris, 1995b). For individuals with equivocal symptoms or lower values for fasting plasma glucose concentrations, a diagnosis requires measurement of fasting plasma glucose concentrations as well as plasma glucose concentrations two hours following a 75-g oral glucose challenge (i.e., a two-hour oral glucose tolerance test (OGTT) with 75-g carbohydrate load). Criteria are then applied to classify individuals according to their OGTT as either: (a) normoglycemic (fasting and two hour plasma glucose < 7.8 mmol/L); (b) having impaired glucose tolerance (IGT) (fasting plasma glucose < 7.8 mmol/L and two hour plasma glucose 7.8-11.1 mmol/L); or (c) diabetic (fasting plasma glucose > 7.8 mmol/L and/or two hour plasma glucose > 11.1 mmol/L). The NDDG and WHO criteria were selected because studies in populations with a high prevalence of diabetes showed that two-hour post-load plasma glucose concentrations had a bi-modal distribution, with the antimode at - 11.1 mmol/L (Knowler et al., 1990). A further factor in decisions about criteria was the frequency of microvascular complications of diabetes in the same populations. For example, individuals with fasting plasma glucose > 7.8 mmol/L and/or two hour plasma glucose > 11.1 mmol/L were at high risk for retinopathy and nephropathy, whereas microvascular complications were rare among individuals with fasting plasma glucose < 7.8 mmol/L and two hour plasma glucose < 11.1 mmol/L. The classification IGT is defined by plasma glucose criteria, not by microvascular complications. IGT is a deviation of the normal metabolic state, but not a disease perse; it describes people with hyperglycaemia at a lower level than that qualifying as diabetes. Many people with IGT have risk factors that increase their risk of developing coronary 'See Coustan (1995) for a comprehensive classification of secondary forms of diabetes or impaired glucose tolerance. heart disease, and some of these people will progress to diabetes (Rewers & Hamman, 1995). Approximately 11% of asymptomatic adults have IGT when tested by OGTT. Relative Severity Prevalence and Incidence Morbidity trends reveal parallels in the epidemiologic transition faced by Native Americans in Canada and the United States. Diabetes prevalence has until recently tended to be much lower in Canada, but prevalence rates among Aboriginal Canadians have risen over time and now approximate the lower range of typical prevalence rates observed among American Indians.* Incidence rates are not available for Aboriginal populations in either country. In the United States general population, the incidence of all forms of diabetes is approximately 2.42 per 1,000 people per year (Kenny et al., 1995). Given the high prevalence of diabetes among Native American populations, the incidence of the disease among this segment of the general population is undoubtedly much greater than the average incidence rate for the general population. Whereas approximately 2-3% of the non-Aboriginal population in Canada (Statistics Canada, 1987) and the United States (Kenny etal., 1995) has diabetes, the prevalence of diabetes in many Native American populations is substantially greater. IDDM is relatively rare among Native Americans. NIDDM is the prevailing form of the disease (West, 1974, 1978). Virtually an unknown condition in any form before 1940, diabetes has become so increasingly prevalent since 1950 in Native Americans, that it has assumed "epidemic" status (Gohdes, 1995; Sievers & Fisher, 1985). Prevalence rates for Aboriginal populations in Canada and the United States are usually estimated from case registries maintained at health facilities, glucose testing at a community level, and surveys of self-reported diabetes. These estimates are thought to under-estimate the true prevalence of diabetes as would obtained from systematic, population-based screening (Young & Krahn, 1988). In the United States, diabetes prevalence rates range from less than 3% in Alaska Natives (Middaugh etal., 1991; Mouratoff etal., 1969; Schraer etal., 1988) to more than 50% in Pima Indians older than 35 years in Arizona (Knowler et al., 1990). Rates of between 10% and 30% are "Unless otherwise noted, prevalence reports are population-based (for adults and children together), though the majority of diabetes cases occur in middle-aged and older individuals. most common (Carter etal., 1989; Stahn etal., 1993; Sugarman & Percy, 1989; Sugarman etal., 1992). Few published data exist on the extent of diabetes in Canadian Aboriginal populations (Young, 1987), but the available literature indicates widely variable prevalence rates (Delisle & Ekoe, 1993; Evers etal., 1987; Montour & Macaulay, 1985; Schaefer, 1968; Szathmary & Holt, 1983; Young etal., 1985; Young & Shaw, 1987; Young etal., 1990). Only the national survey of Medical Services Branch units by Young etal. (1990) permits the direct comparison of diabetes prevalence rates across Canada. Based on known diabetic cases at year end 1987, age-adjusted rates for Canadian Indians (using age and gender specific totals by province or region as the population denominator, standardised directly against the national population) ranged from 0.8% in the Northwest Territories to 8.7% in the Atlantic region. Prevalence rates for Canadian Indians in the Northwest Territories, the Yukon Territory and British Columbia were, respectively (%), 0.8, 1.2 and 1.6. In all regions, both the crude and age-adjusted rates were slightly greater for females than for males. For the remainder of the country, diabetes prevalence rates for Canadian Indians were between 5.1% and 8.7%, two to four times greater than for all other Canadians, for whom the self-reported rate of diabetes is approximately 2.5% (Statistics Canada, 1987). Data indicating time trends in diabetes prevalence for Aboriginal Canadians are not available. Prevalence studies indicate, however, that the majority of Aboriginal Canadians developed their disease relatively recently, mainly within the last 20 to 30 years, suggesting that the incidence of diabetes is increasing in Canada's Aboriginal communities. Glycosuria was unknown among Athapaskan Indians in the Northwest Territories in the 1930s, despite "thousands" of urine glucose tests (Urquhart, 1935), and diabetes was not detected among more than 1,500 Aboriginal Canadians examined in the 1930s in Saskatchewan as part of a tuberculosis survey (Chase, 1937). The national survey by Young etal. (1990) found that the most important predictor of variation in diabetes prevalence rates was latitude; prevalence rates decreased as northern latitude increased, and were greatest in southern latitudes ecologically corresponding to areas of greatest urban development. These data support the notion that diabetes among Aboriginal Canadians is related to environmental changes associated with acculturation. Recent reports from Quebec (Delisle & Ekoe, 1993) and Saskatchewan (Dyck etal., 1995) indicate that the prevalence of diabetes is greater among Aboriginal communities with easier access to urban areas than among those communities with limited access to urban areas. In British Columbia, further analysis of the provincial survey carried out as part of the national effort by Young etal. (1990) showed that the prevalence of diabetes was 4.5% for registered Indians aged 35 years and older (Martin & Bell, 1990). Of the four Medical Services Branch regional zones in the province, diabetes prevalence was greatest in coastal and southern communities (Table 1). Again, the pattern of prevalence suggests a link between diabetes and acculturation related to urbanisation. The prevalence rate for registered Indians age 35 years and older in the Vancouver Island Zone was 6.3%, followed by the South Mainland Zone (5.2%), North West Zone (4.6%) and North East Zone (1.9%) (Martin & Bell, 1990). Of the total number of registered Indians in British Columbia, 25.5% were treated by diet alone, 35% were treated by diet plus oral hypoglycaemics, and 39.5% were treated by diet plus insulin (Table 2). The British Columbia Medical Services Branch prevalence survey was repeated in 1995. A preliminary analysis found that the total number of cases had increased from 348 to 833, corresponding to an overall increase among registered Indians aged 35 years and older of 2.0% for males and 2.6% for females (M