Risk Factors for Diabetes Mellitus: a Comparative Analysis of Subpopulation Differences in a Large Canadian Sample

RISK FACTORS FOR DIABETES MELLITUS: A COMPARATIVE ANALYSIS OF SUBPOPULATION DIFFERENCES IN A LARGE CANADIAN SAMPLE

(Thesis format: Monograph)

Michael James Taylor

Graduate Program in Epidemiology and Biostatistics

A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science

The School of Graduate and Postdoctoral Studies The University of Western Ontario London, Ontario, Canada

Abstract

Objectives: Certain Canadian subpopulations observe numerous modifiable and non- modifiable risk factors for diabetes. This study compares immigrants and Aboriginals (First Nations, Inuit, and Métis) with Canada-born individuals at higher risks for diabetes, and deciphers the determinant differences between them.

Methods: Pooled Canadian Community Health Survey data (2001-2010) were used. Time trends for diabetes within each subsample were calculated using individual survey year prevalence rates; diabetes diagnoses were self-reported (N=33,565). Various risk factors were also examined using logistic regression.

Results: Diabetes prevalence rates significantly increased from 2001 to 2010 for each subpopulation, as well as the total sample: Canada-Born individuals (3.9% to 5.7%), Immigrants (5.0% to 8.5%), Aboriginals (5.4% to 7.4%), and Canadians overall (4.1% to 6.4%).

Conclusions: All Canadians, regardless of risk, experienced and will continue to experience a rise in diabetes. Future diabetes research involving the impact of race, culture, and ethnicity in Canadian immigrants should be holistically explored.

Keywords diabetes mellitus, type 2 diabetes, risk factor, ethnicity, race, culture, immigrant, healthy immigrant effect, acculturation, Aboriginal, First Nations, Inuit, Métis, Canadian Community Health Survey (CCHS), pooling

Acknowledgments

I would first like to thank Dr. Amardeep Thind for his unwavering encouragement and support throughout my time at the University of Western Ontario. From the very beginning, he has offered consistently helpful advice, reassurance, and enthusiasm. Dr. Thind has honed my research interests and changed my academic perspective for the better. He has taught me that skepticism is not the primary tenet of the academic; rather, that it is an effective tool to judge the methodological and experimental frameworks I will encounter throughout my future career.

I would also like to thank Dr. Stewart Harris for his optimistic and supportive contributions to my research process. Dr. Harris has dependably posed new and exciting ways of examining pertinent literature, certain subpopulations, and has helped a great deal by ensuring that my current work will lead to a fruitful career in research. His clinical experience, diabetes research mastery, and overall wealth of knowledge has provided me with an invaluable perspective throughout the thesis process.

In addition to my supervisory committee, I would like to thank Dr. Kathy Speechley for her support throughout my entire time in the Epidemiology and Biostatistics program. As Graduate Chair, and as a mentor, she has been an incredible source of encouragement, reason, and understanding.

Lastly, I would like to thank my family for their continuing support of my academic and professional interests; especially my mother, who has inspired my passion for diabetes research.

iii

Table of Contents

Abstract ...... ii

Acknowledgments ...... iii

List of Tables ...... viii

List of Figures ...... iix

List of Appendices ...... x

List of Abbreviations ...... xi

Chapter 1 ...... 1

1 Introduction ...... 1 1.1 Diabetes Mellitus ...... 1 1.1.1 Impact as a Chronic Disease ...... 3 1.2 Immigrant Health ...... 4 1.3 Aboriginal Health ...... 5 1.4 Aim of the Present Study ...... 7

Chapter 2 ...... 9

2 Literature Review ...... 9 2.1 Modifiable Risk Factors for T2DM and Diabetes Related Complications ...... 10 2.2 Non-Modifiable Risk Factors for T2DM and Diabetes Related Complications... 16 2.2.1 The Definition of Ethnicity in Epidemiological Research ...... 18 2.2.2 Ethnic Origin and Obesity ...... 19 2.2.3 Ethnic Origin and Insulin Resistance ...... 20 2.2.4 Findings from Canada: Ethnic Origin and T2DM ...... 20 2.2.5 Findings from the United Kingdom: Ethnic Origin and T2DM ...... 21 2.2.6 Findings from the United States: Ethnic Origin and T2DM ...... 22 2.2.7 The Thrifty Gene Effect ...... 23 2.3 Migration, Health, and Birth Origin ...... 25 2.3.1 The Healthy Immigrant Effect and Acculturation in North America ...... 26

2.3.2 Potential Barriers to Immigrant Health, Diabetes Management, and Healthcare Utilization ...... 29 2.4 T2DM in Immigrants and Migration in Canada ...... 31 2.5 Canadian Aboriginals: Unique Characteristics ...... 33 2.5.1 Modifiable Risk Factors for T2DM in Canadian Aboriginals ...... 34 2.5.2 Non-Modifiable Risk Factors for T2DM in Canadian Aboriginals ...... 36 2.5.3 Canadian Aboriginals: Diabetes Burden and Diabetes Related Complications ...... 37 2.6 Summary ...... 38

Chapter 3 ...... 41

3 Objectives ...... 41

Chapter 4 ...... 43

4 Methodology ...... 43 4.1 Data Source and Sampling Design ...... 43 4.2 Data Collection ...... 48 4.2.1 Nonresponse and Data Quality ...... 49 4.2.2 Self-Reporting Bias and Interview Mode Effects ...... 49 4.3 Weighting ...... 51 4.3.1 Area Frame Weight ...... 51 4.3.2 Telephone Frame Weight ...... 52 4.3.3 Final Weight Integration ...... 52 4.4 Combining CCHS Cycles ...... 53 4.5 CCHS Data Access ...... 55 4.6 Outcome Variable and Total Sample ...... 55 4.7 Conceptual Framework ...... 56 4.8 Statistical Analyses ...... 61 4.8.1 Objective One ...... 62 4.8.2 Objective Two ...... 63

Chapter 5 ...... 65

5 Results ...... 65

5.1 Diabetes Mellitus Sample Characteristics ...... 65 5.2 Period Prevalence of Diabetes Mellitus (2001 to 2010) ...... 76 5.3 Comparative Impact of Diabetes Risk Factors - Crude Results ...... 79 5.3.1 Smoking Status ...... 79 5.3.2 Alcohol Consumption Frequency ...... 79 5.3.3 Body Mass Index ...... 80 5.3.4 Household and Personal Income ...... 80 5.3.5 Household and Personal Education ...... 80 5.3.6 Diet ...... 80 5.3.7 National Language Competency ...... 80 5.3.8 Age ...... 81 5.3.9 Sex ...... 81 5.3.10 Racial/Cultural and Ethnic Origin ...... 81 5.3.11 Immigrant-Specific Risk Factors ...... 82 5.4 Comparative Impact of Diabetes Risk Factors - Adjusted Results ...... 93 5.4.1 Smoking Status ...... 93 5.4.2 Alcohol Consumption Frequency ...... 93 5.4.3 Body Mass Index ...... 93 5.4.4 Household and Personal Income ...... 93 5.4.5 Household and Personal Education ...... 94 5.4.6 Diet ...... 94 5.4.7 National Language Competency ...... 95 5.4.8 Age ...... 95 5.4.9 Sex ...... 95 5.4.10 Racial/Cultural and Ethnic Origin ...... 95 5.4.11 Immigrant-Specific Risk Factors ...... 96

Chapter 6 ...... 107

6 Discussion ...... 107 6.1 Overview of the Findings ...... 107 6.1.1 Main Subpopulation Differences ...... 107 6.1.2 Modifiable Risk Factor Differences ...... 110

6.1.3 Non-Modifiable Risk Factor Differences ...... 116 6.1.4 Immigrant-Specific Risk Factors ...... 118 6.2 Challenges in Subpopulation Research ...... 121 6.2.1 Challenges in Studying Immigrant Health ...... 121 6.2.2 Challenges in Studying Aboriginal Health ...... 122 6.3 Strengths ...... 122 6.4 Limitations ...... 123 6.4.1 Self-Report and Limitations with Measures ...... 123 6.4.2 Non-Response, Response Rate, and Interview Mode ...... 124 6.4.3 Pooling Methodology ...... 124 6.4.4 Cross-sectional Design and Temporality ...... 125 6.5 Implications and Conclusions ...... 125

References ...... 137

vii

List of Tables

Table 1: MeSH Term Search Strategies ...... 9

Table 2: Country Origin of Recent Immigrants by Canadian Census Year...... 32

Table 3: Summary of CCHS Annual Content Modifications from Cycle to Cycle ...... 45

Table 4: Number of Health Regions and Sample Sizes Required by CCHS Cycle...... 47

Table 5: Variable Coding………………………………………………………………...59

Table 6: Distribution of Variables for the Total Sample (2001 to 2010 Inclusive)…...…69

Table 7: Diabetes Mellitus Period Prevalence Rates………………………………….....78

Table 8: Results from Bivariate Logistic Regression Models (Crude)…………………..84

Table 9: Results from Multivariable Logistic Regression Models (Adjusted)…………..98

viii

List of Figures

Figure 1: Thrifty Hypothesis Schema…………………………………………………… 24

Figure 2: Example of Southern Ontario Health Region Divisions for Cycle 1.1 (2001)... 47

Figure 3: Conceptual Framework………………………………………………………... 58

Figure 4: Diabetes Trends from 2001 to 2010…………………………...... 79

List of Appendices

Appendix 1: Distribution of Variables by CCHS Survey Year (2001 to 2010) ...... 129

List of Abbreviations

APS Aboriginal Peoples Survey CANRISK Canadian Diabetes Risk Assessment Questionnaire CCDSS Canadian Chronic Disease Surveillance System CCHS Canadian Community Health Survey CKD Chronic Kidney Disease CVD Cardiovascular Disease DCCT Diabetes Control and Complications Trial EDIC Epidemiology of Diabetes Interventions and Complications ESRD End Stage Renal Disease FNRLHS First Nations Regional Longitudinal Health Survey HNF Hepatic Nuclear Factor ICES Institute for Clinical Evaluative Sciences IGF1 Insulin-like Growth Factor 1 IGFBP1 Insulin-like Growth Factor Binding Protein 1 MeSH Medical Subject Heading NDSS National Diabetes Surveillance System NPHS National Population Health Survey SES Socioeconomic status T2DM Type 2 Diabetes Mellitus

xi 1

Chapter 1 1 Introduction 1.1 Diabetes Mellitus

As a chronic disease, diabetes mellitus has emerged as a worldwide epidemic. In 2010, the global diabetes prevalence rate (age- and sex- adjusted) was approximately 6.4% of adults aged 20 to 79 (285 million). 1 It has been estimated that this rate will reach 7.7% (approximately 439 million) within the next 17 years; 1 a growth of 69% in developing nations and 20% growth in first-world nations, such as Canada, has also been predicted. 1 Approximately 6.8% (2.4 million) Canadians aged one year and older are currently diagnosed with diabetes; 200,000 being diagnosed for the first time in 2008/2009 (6.3 diabetes cases/1000 individuals).2 Over the next seven years, another 1.2 million people are expected to be diagnosed in Canada, increasing the total prevalence of the disease from 4.2% in the year 2000 to approximately 9.9% in the year 2020. 2 While diabetes can be managed effectively through dedicated treatment regimens, it still presents a large economic burden on the Canadian healthcare system. In 2010, the cost of diabetes on the healthcare system and the overall economy was approximately $11.7 billion; by 2020, this figure is expected to rise towards $16 billion per year. 3,4

Diabetes mellitus will manifest either as type 1 (less than 10% of all disease cases), type 2 (approximately 90% of all disease cases), or gestational (approximately 3-5% of all successful pregnancies). Type 1 diabetics are typically diagnosed prior to late- adolescence (hence the former name “juvenile” diabetes) and are insulin dependent (the pancreas is unable to produce insulin); causes for this type of diabetes are not exactly known, but are considered to be autoimmune in origin. Type 2 diabetics are often diagnosed during later adulthood (hence the former name “adult-onset” diabetes), and are not normally insulin dependent during the earlier stages of the disease (the body either has difficulties utilizing the insulin that is produced by the pancreas, or the pancreas is unable to produce sufficient amounts of insulin); causes for this type of diabetes are multiple but strongly correlated with genetic, lifestyle, and/or environmental factors.

Gestational diabetes is a specific type that is diagnosed in females during pregnancy and is associated with an increased risk for developing type 2 diabetes (T2DM) in the future.

Diabetes mellitus may be more generally defined as an inability to effectively produce or use insulin; thus, leading to increased levels of glucose in the blood.5 The human body requires insulin to metabolize glucose as an energy source; without the proper ability to complete this process, our bodies are inherently at risk for developing other serious complications that are either microvascular or macrovascular in nature.5 For example, hyperglycemic-specific cardiovascular disease (CVD) has been identified as the highest cause of mortality in people with diabetes mellitus. 6 In Canada, CVD (heart disease or stroke) accounts for approximately 80% of deaths in individuals with diabetes; roughly 41,500 diabetes related deaths occur in Canada each year. 7 However, attempting to control one’s glucose levels via lifestyle modification, such as increased physical activity and/or improved diet, may help to reduce the risk of diabetes and its associated complications. For instance, the Diabetes Control and Complications Trial (DCCT) found that significant glucose control reduced the risk of eye disease (retinopathy) by ~76%, kidney disease (nephropathy) by ~50%, and nerve disease (neuropathy) by ~60%.8 The follow-up study, aptly named the Epidemiology of Diabetes Interventions and Complications (EDIC), found that intensive plasma glucose control significantly reduced the risk of a CVD occurrence by 42%. 8

In order to effectively manage diabetic complications, glucose control targets have been established to delay progression and to minimize the level of morbidity and mortality among individuals living with diabetes. Given the gradual onset of T2DM, diagnoses may not occur until after glucose levels have risen; long-term complications are therefore developed prior to these diabetes diagnoses. 5,8 Clinically, this issue may be of even greater concern for higher-risk Canadian subpopulations. This thesis will limit discussion of diabetes mellitus specifically to T2DM, high-risk Canadian subpopulations, and the differences that exist between them.

Previous longitudinal research has examined the powerful impact that lifestyle choices may have on the neuroendocrine and inflammatory responses which lay the foundation for T2DM within the body; these risk factors are typically modifiable in nature.9,10

Individuals who do not engage in regular physical activity, do not consistently follow a healthful diet, smoke heavily, or abuse alcohol are at much higher risks for obesity and the subsequent physiological consequences that may lead to T2DM, particularly insulin resistance: cells within the body become resistant to the metabolizing effects of insulin .11,12, Once insulin resistance and the successive impairment of glucose uptake have occurred, the risk of developing T2DM is then increased.13,14

1.1.1 Impact as a Chronic Disease

Depending on the variable combination of one’s genetics and personal lifestyle choices, chronic illness may be dealt with at any stage of the life cycle. Whether it is an arthritic condition, hypertension, or diabetes mellitus, by the age of 65 at least 77% of men and 85% of women will be managing at least one chronic health condition. 15 Several chronic illnesses are found to be co-morbid in nature, with one condition frequently being indicative of the presence of another. 16 The reality for most diabetic patients is that if left untreated or unmanaged, their diabetes may lead to any number of the possible complications associated with the condition; CVD, retinopathy, nephropathy, and neuropathy have been previously mentioned. Other risks for type 2 diabetics include oral disease, arthritic conditions, cancer, and even a lowered perception of ones’ own quality of life. 17,18,19,20 Moreover, diabetes has even been linked with higher rates of certain mental illnesses, such as depression (this disease relationship has been noted as bidirectional in nature).21,22

While modifiable risks for T2DM may be clinically reduced through the adjustment of lifestyle behaviours and/or therapeutic treatments, other risk factors for T2DM are not as easily altered. This thesis will focus not only on the modifiable risk factors previously mentioned, but on non-modifiable risk factors as well; for example, racial/cultural origin, country origin, and intrinsically, ethnicity. Genetic predispositions, behavioural choices, cultural differences, or diverse combinations of all of the above may explain the observable disparities among high-risk Canadian subpopulations.

Overall, certain subpopulations observe increased risks for T2DM. Individuals with a family history of the disease, Aboriginals (First Nations, Inuit, and Métis), individuals of

South or East Asian, Latin American, or African descent, immigrants, and refugees are examples of groups with unique characteristics that lead to higher susceptibilities. 23,24 Of specific interest to this study are Canadian-immigrants and Aboriginals (First Nations, Inuit, and Métis).

1.2 Immigrant Health

Immigrant health is of a particular relevance to the Canadian healthcare system. Canada has a uniquely high rate of immigration that accounts for over 60% of the country’s population growth and about 20% of the overall population.25 Canadian-immigrants are predominantly unique in that many are of ethnicities that observe higher risks for T2DM. For instance, previous research has shown that long-term Canadian immigrants from South Asia, the Caribbean, sub-Saharan Africa, North Africa, the Middle East, and Latin America have experienced significantly higher rates of T2DM when compared to long- term Canada-born residents. 26 These individuals also experience the often-taxing process of immigration.

Migration can be quite demanding, and poor social or economic adjustment may result in adverse health outcomes (both physically and mentally).27 In the past, researchers have argued that immigrants are generally healthier than the Canada-born population upon arrival; however, rates of diabetes, hypertension, and most cancers are found to be quite similar after years of settlement.28,29 The concept of “acculturation” (an adaptation of Canadian socio-health related behaviours) has been proposed as a mediating factor that contributes to these eventual similarities.30

While immigrants may have better health than those born in Canada upon arrival, the longer that they choose to live and settle in Canada, the more their overall health begins to resemble that of the Canada-born population. Researchers refer to their initial benefits in health as the “healthy immigrant effect”.31 In summary, immigrants may be additionally prone (in comparison to Canada-born individuals) to a number of chronic conditions, especially T2DM, as a result of a “negative acculturation effect”.32 The stressful process of migration, social or economic maladjustment, and acculturation over time could synergistically contribute to this excess risk.

Immigrant communities also face unique barriers to their health when compared to the Canada-born population.33,34 Lower levels of either national language (English or French) competency or general health literacy may cause certain racial/cultural groups to be inherently marginalized, especially if they are recent immigrants;35,36 it has been suggested that this marginalization may cause wide variations in subsequent socio-health behaviours and outcomes following settlement.35,36 Moreover, certain migrant populations may innately differ in their approach to, and utilization of healthcare services and/or disease treatments.33,34,36 Inequities in healthcare usage can have substantial implications for preventative care programs and how they tailor to the specialized needs of various subpopulations in Canada.37

Immigrant health is becoming increasingly important to the overall wellbeing of Canadians; 37 thus, their risks for diseases like T2DM require increasingly scrutinized and comparative research.

1.3 Aboriginal Health

In Canada, Aboriginal groups consist of three unique identifications that are recognized constitutionally: First Nations, Métis, and Inuit. 38 The distinct cultural practices and historical beliefs of each group not only differentiate them from the non-Aboriginal population, but from each other as well. 39 For instance, languages and geographic roots will differ greatly among these Aboriginal groups. 38

National prevalence statistics tend to underline significant disparities in health between Aboriginal and non-Aboriginal Canadian populations. When examining diabetes mellitus estimates in particular, the crude and age-adjusted prevalence rates among Aboriginals well exceeds that of non-Aboriginals; previous research has shown that First Nations, Inuit, and Métis individuals typically have age-adjusted diabetes rates that are three to five times that of non-Aboriginal individuals in Canada.40,41,42,43

Aboriginals are also known to have the highest risk of diabetic complications and comorbidities. 44 For example, notably high prevalence rates of chronic kidney disease (CKD), 45 amputation due to lower limb neuropathy, 46 retinopathy, 47 and both micro-45

6 and macro-vascular disease have been observed in this subpopulation. 48 One theory suggests that the reason for these increased complications from diabetes stems from Aboriginal Canadians receiving more frequent diagnoses at younger ages than any other Canadian subpopulation. 40,49,50 Aboriginal youth have also been noted as having higher risks for developing T2DM when compared to youth of all other racial/cultural or ethnic origins in Canada.50

Important sex differences also exist between Aboriginals and non-Aboriginal Canadians. In particular, Aboriginal females are known to have significantly higher rates of gestational diabetes than non-Aboriginal females.51 Furthermore, Aboriginal sex differences in the rates of diabetes are uncharacteristically the reverse of what is seen in the non-Aboriginal Canadian population. Aboriginal females tend to have greater diabetes rates than males, and the opposite trend is seen in non-Aboriginals. 52

It has been posited that Aboriginals possess a unique genetic risk factor that may be responsible for increased diabetes diagnoses, via physiology. The “thrifty gene effect” theory states that as a protective mechanism, individuals of Aboriginal descent will naturally conserve more calories (leading to greater chances for obesity) because of their historic hunting-and-gathering lifestyles. 53 This hypothesis is based on the fact that culturally, food security was once considered to be extremely minimal. While the theory itself is contentious, if true, it would most likely be (etiologically) a fragment of a larger combination of behavioural, cultural, and genetic risk factors. 54

Geographic inequities also exist among Aboriginal communities and are of particular interest when evaluating diabetes in this high-risk subpopulation. Healthcare systems and living conditions will differ considerably depending on whether or not an Aboriginal individual is on-reserve or off-reserve, living in a large city, or living in a small rural community. Reserve communities are often located in remote areas where access to affordable foods, quality education, or even general healthcare services may be increasingly scarce or expensive; these factors may impede overall Aboriginal healthcare utilization. 55 Poor health surveillance, substandard chronic disease management systems, and low retention of healthcare service staff have been identified as unique barriers to healthcare systems that are tailored to Aboriginals. 55 Studies have also shown that

7 culturally catered healthcare services are few and far between for the Aboriginal population. 56,57

Comparable to immigrants, Aboriginals have unique risk factors for chronic diseases, specifically T2DM. Further exploration into this subpopulation’s distinctive diabetes burden will further enrich our knowledge concerning national surveillance, comparable to that of other high-risk Canadians.

1.4 Aim of the Present Study

Prevalence rates for diabetes tend to vary widely depending upon the data source and methodologies that are used in Canadian research. Self-reported data that is provided by single iterations of the Canadian Community Health Survey (CCHS) differ significantly when compared to the information collected and produced by the Canadian Chronic Disease Surveillance System (CCDSS), for instance. 1,3 The Public Health Agency of Canada states that while each source (and methodology) are valid and reliable, any single source may under- or over-estimate the actual rates and recent trends of diagnosed diabetes in Canada; 1 this is of particular importance when attempting to gauge the rising epidemic of diabetes, and the differential impact of certain risk factors for diabetes among Canadian subpopulations. In particular, Canadian-immigrants and Canadian-Aboriginals experience unique risks for diabetes that require further investigation.

Self-reported health survey data tends to underestimate the actual rate of diabetes mellitus and diabetes related complications in the population.1 Nevertheless, large-scale health surveys (most of which rely on self-report) such as the National Population Health Survey (NPHS), CCHS, Aboriginal Peoples’ Survey (APS), First Nations Regional Longitudinal Health Survey (FNRLHS), provide useful and accessible health data at several different points in time; the CCHS is particularly useful because of its national-level representation of the Canadian population, inclusivity of both immigrants and Aboriginals (off-reserve), and its cross sectional iterations that are available from 2001 to 2010 (when diabetes prevalence rates have reportedly increased considerably 2).

This thesis seeks to add to the current body of literature by using pooled methodology with several cycles of the CCHS. By combining all available iterations of the CCHS, this thesis will characterize Canadians with diabetes from 2001 to 2010, and compare Canadian subpopulations at higher risks for diabetes. Diagnosed diabetes trends for immigrants and Aboriginals will be assessed from 2001 to 2010, and the differences between these subpopulations, according to a relevant set of modifiable and non- modifiable risk factors, will be interpreted.

Chapter 2 2 Literature Review

This thesis examined Canadian subpopulations at higher risks for diabetes mellitus; immigrants and Aboriginals (First Nations, Métis, Inuit) were of pertinent interest to the literature search. Three separate literature reviews were conducted in the areas of type 2 diabetes mellitus (T2DM) risk factors, immigrant health, and Aboriginal health. The key words identified for this study include “Type 2 Diabetes”, “Risk Factor”, “Immigrant”, and “Aboriginal”. Each key word allowed for various inclusions of medical subject heading (MeSH) terms that summarized the pertinent literature within each relevant area of the study. PubMed and Social Sciences Abstract were used as primary and secondary literature sources, respectively.

The following table lists the individual MeSH term strategies used for each review:

Table 1: MeSH Term Search Strategies

T2DM Immigrants Aboriginals Risk Factors Risk Factor Immigrant Aboriginal Broad Search Diabetes Mellitus, Diabetes Mellitus, Diabetes Mellitus, Type 2 Type 2 Type 2 Non-Modifiable Healthy Immigrant First Nations Narrow Risk Factor Effect Search Race Acculturation Métis Ethnicity Barrier Inuit *All terms were inclusive for each keyword and alternate suffixes.

The use of both broad and narrow search strategies allowed for an exhaustive and comparative investigation of the significant literature; six different variations of searches were conducted in total. The broad searches initially generated literature that was specific to T2DM risk factors, immigrant T2DM, and Aboriginal T2DM. The narrow searches

10 were useful in exploring the applicable topic discussions identified by the broad T2DM literature.

Broad Searches : 1. [Diabetes Mellitus, Type 2] AND [Risk Factor] 2. [Diabetes Mellitus, Type 2] AND [Immigrant] 3. [Diabetes Mellitus, Type 2] AND {[Aboriginal] OR [Indigenous]}

Narrow Searches: 1. [Diabetes Mellitus, Type 2] AND {{[Risk Factor] AND [Non-Modifiable]} AND {[Race] AND [Ethnicity]}} 2. [Diabetes Mellitus, Type 2] AND [Immigrant] AND {[Health Immigrant Effect] AND [Acculturation] AND [Barrier]} 3. [Diabetes Mellitus, Type 2] AND {[Aboriginal] OR [Indigenous]} AND {[First Nations] AND [Métis] AND [Inuit]}

Broad searches revealed a wealth of literature for search 1 (12574 items), search 2 (102 items), and search 3 (12652 items). Narrow searches also revealed a great deal of literature for search 1 (1941 items without including “Risk Factor AND Non-Modifiable” (the inclusion of this term resulted in an additional 9 items)), search 2 (138 items), and search 3 (67 items). All literature searches were additionally filtered for items with full free text and abstracts available, as well as by date (January 1990 to Present). Relevancy was initially determined by scanning item titles, then abstracts, followed by an in-depth analysis of each appropriate article and its reference list.

2.1 Modifiable Risk Factors for T2DM and Diabetes Related Complications

The rising epidemic of diabetes mellitus is multi-faceted and complex in nature. 58 Several different causal pathways have been proposed in numerous patient populations that differ by risk factor type, and diagnosis; 58 a significant proportion of the at-risk population (worldwide) is believed to be undiagnosed for prediabetes and T2DM.58

Assessment instruments for detecting T2DM risk in the population have been extensively developed for standard use in primary health care.59 The complexity of the cumulative risk factors that are considered by these instruments is generally considered two-fold: assessments will make use of multi-variable models that operationalize risk factors as either modifiable (lifestyle), or non-modifiable (genetic).59 Common measurements will combine both types of risk factors into an aggregate absolute risk score that is used to predict T2DM in individuals and/or population subgroups.60,61,62,63,64 These scores have been comprehensively validated in the literature using prospective cohorts. The Framingham clinical model,65 the German Diabetes Risk Score,66 the Cambridge Diabetes Risk Score,67 the Finnish Diabetes Risk Score,68,69,70,71 and many others 72 are examples of measurements that accurately assess propensities for T2DM using similar and standardized variables. One of the most recently validated assessment tools available in Canada is the Canadian Diabetes Risk Assessment Questionnaire (CANRISK), one that is publically available on the Internet for anyone to determine if they have a higher risk of developing prediabetes or T2DM. 73,74 Higher risk populations typically flagged using the CANRISK assessment tool include those of Aboriginal, African, Asian, Latino, or South Asian descent, having a first-degree relative with T2DM, having a history of gestational diabetes mellitus, being 40 years of age or older, being overweight, having high cholesterol, and many more. 75

Modifiable variables generally include measures of obesity (body mass index (BMI), waist circumference, and waist-to-hip ratio), physical activity level, diet, smoking status, and alcohol use. 60,76,77 More specifically, the presence of high BMI, physical inactivity, poor nutrition, daily cigarette smoking, and heavy alcohol usage have all been associated with increased risks for T2DM.78,79,80,81,82

Indicative measures for obesity such as BMI, waist circumference, and waist-to-hip ratios have been extensively studied in relation to T2DM. 83 Larger waist-to-hip ratios and increased amounts of adipose fat located around the abdominal area have both been particularly linked to increased risks for T2DM in both men and women.80,83,84 Cassano et al. reported that within their prospective cohort, men in the top tertile for waist-to-hip ratios had a 2.4 greater risk of developing T2DM than men in the lowest category.85

Moreover, Snijder et al. indicated that women in their own prospective cohort study had a 2.66 greater risk of T2DM associated with each standard deviation increase in waist circumstance. 86 According to currrent literature, increased modifiable risks for T2DM that are linked with being obese (excess body weight, clinically defined as a BMI > 30kg/m 2) are traditionally mediated by insulin resistance (or impaired glucose tolerance).83,85,86,87 While the precise causes for insulin resistance are not completely understood, it is believed that this physiological state is brought on and exacerbated by the aforementioned modifiable risk factors. 88 Insulin resistance occurs when the body produces sufficient amounts of insulin but does not efficiently use it to absorb plasma glucose.88 Plasma glucose levels will then begin to rise, while pancreatic β-cells simultaneously compensate by secreting increased levels of insulin (to combat the developing insulin resistance).89,90 Eventually, β-cells will fail to meet the body’s increased needs for insulin, and plasma glucose levels will heighten to a detectable level; 88-90 prediabetes and/or T2DM diagnoses will typically follow, dependent upon the severity and duration of insulin resistance.88-90

Individuals that are clinically obese, overweight, or those that carry excess abdominal fat tend to pass through the initial stage of prediabetes via insulin resistance (referred to as a “reference range”).91 Prediabetes is frequently used as a clinical indicator for an increased modifiable risk of developing T2DM. 88,92 The growing rate of prediabetes (as well as T2DM) in Canadian children as a result of modifiable risk factors has become increasingly alarming; 92 the childhood rates of hyperglycemic disorders (like prediabetes) in Canada is greater today than ever before. 89,92 One study notes that children and adolescents experiencing insulin resistance and prediabetes are extremely vulnerable during times of pubertal growth. 93 Increased insulin needs will subsequently lead to more T2DM diagnoses in individuals under the age of 18, leading to longer durations of these children living with the disease and increased risks for diabetes related complications.89,92,93 T2DM is the only type of nonautoimmune diabetes mellitus that is increasing in the pediatric Canadian population. 94

T2DM is the leading co-morbidity associated with obesity next to CVD.87,95 Lengthy durations of either (diagnosed or undiagnosed) obesity or T2DM also leave individuals particularly vulnerable to coronary heart failure.78-82,88 Research has shown that the

13 detection and prevention of obesity and T2DM with lifestyle modifications has led to improved overall health outcomes, as well as a decreased economic burden for the individual and the healthcare system.88,96 In 2011, approximately 34% and 11% of adults (over the age of 19) in the United Sates were clinically obese or had diabetes, respectively. 97 In Canada, self-reported rates of adult obesity and diabetes have risen in the past 13 years; 98 the Public Health Agency of Canada reports that the actual rate of adult obesity and diabetes in Canada is roughly 25% and 6.8%, respectively. 2,98 Of particular note when discussing obesity and T2DM is that they have the potential for mutual exclusivity, but more significantly, higher rates of potential inclusivity; 87-90 not all individuals with T2DM are obese, nor are all obese individuals diabetic. However, the potential insulin resistance that accompanies either condition may facilitate compounded risks that are both modifiable and non-modifiable in nature.99

Eckel et al. summarizes three known mechanisms that have linked obesity to T2DM via insulin resistance (the mediating condition).100,101,102 Firstly, increased adipokines or ‘cytokines’ within the body (as a result of obesity) will induce physiological insulin resistance; these cytokines include tumour necrosis factor-α and reduced adiponectin. 100,103 Secondly, organ fat deposition within the liver, musculoskeletal system, and the “dysmetabolic sequelae” has also been shown to facilitate a predisposition to T2DM.100,104,105,106,107 Thirdly, decreased functional activity of an individual’s mitochondria has been show to negatively impact insulin sensitivity and subsequently degrade the productivity of β-cells in obese individuals. 100,108 The potentially deleterious effects of obesity and insulin resistance are often difficult to reverse;108 however, regular physical activity has been shown to provide therapeutic effects for obese individuals and/or those who are becoming insulin resistant at increased risks of developing T2DM.100,109,110 Studies have shown that a reduction in the circumference of one’s waist has been linked to improved adipose tissue secretion factors that circulate within the blood, lowering the risk of successive β-cell malfunction.88,100,108- 110

Previous research has demonstrated a synergistic interaction concerning the combination of obesity and physical inactivity related to risks for T2DM.111 Physical inactivity

(coupled with obesity) may accelerate harmful pathophysiology towards co-morbidity and early mortality. 112 However, much like the relationship between obesity and TD2M, the co-existence of both physical inactivity and obesity do not always predict T2DM. 100 Physical inactivity is independently associated with insulin resistance, obesity, and T2DM. 87-90,100 Interestingly, even though regular physical activity has been proven to minimize the modifiable risk for T2DM in most individuals, the beneficial effects have also been shown to vary widely by BMI category. 113,114,115,116 Several studies have indicated little therapeutic benefits of physical activity in obese individuals, relative to a reduction in insulin resistance or T2DM risk. 117,118,119,120 Preventative benefits are most observed in those who are in the normal weight range. 119 This research demonstrates that while exercise may be helpful for improving insulin sensitivity, without a great deal of initial weight loss or reduced waist circumference, plasma glucose regulation in the short term is minimal.121,122 These findings also highlight the importance of long-term weight loss interventions that aim to modify T2DM risks that are generally associated with increased BMI, waist-to-hip ratios, and waist circumference.

In addition to physical inactivity, poor diet (high intakes of saturated fat, refined carbohydrates, and/or overall calories) has also been shown to increase one’s risk of T2DM. 123,124,125,126 The healthful combination of dietary changes (increased protein and dietary fibre, decreased fat and refined carbohydrates, decreased overall caloric intake, etc.) and increased amounts of exercise have proven to be successful as preventative and remedial measures for obese and overweight adults and children at risk for T2DM.123-126 In some cases, bariatric surgery may be required to help obese individuals surmount an initial significant weight loss. 127

Nicotine, the highly addictive ingredient used in cigarettes, has been proven to reinforce negative behaviours that are relevant to compulsive drug use, diet, and physical activity;128 cigarette smoking has been linked to T2DM via these mechanisms. 60,72,77,129 Several different prospective cohort studies have observed that the presence of daily cigarette smoking increases the modifiable risk for T2DM.130,131,132,133 After controlling for the presence of other relevant risk factors among cohorts ranging from 2312 men 133 to 114,247 women,130 those who smoked experienced greater relative risks (1.42-1.94) for

T2DM in comparison to individuals who had never smoked, over time. 130-133 These longitudinal studies have suggested that cigarette smoking aggravates insulin resistance, reduces metabolic control, and subsequently raises plasma glucose levels via increased abdominal adipose fat distribution.130-134 One drug that is frequently and compulsively associated with addictive smoking behaviour is alcohol. 129 Independently, heavy alcohol use (or abuse) has been associated with an increased risk for T2DM.135 Moderate use of alcohol is known to decrease the risk for T2DM via cardiovascular benefit;136 however, exceeding the daily or weekly recommended servings for low-risk consumption have been shown to negatively impact insulin resistance in the same manner as cigarette smoking. 135,136 Alcohol has also been shown to raise triglyceride levels and even cause weight gain when consumed in excessive amounts; 137,138,139 thus, provoking and/or aggravating insulin resistance, and predisposing the body to T2DM.

Socioeconomic status (SES) has also been linked to the development of T2DM; this economic and sociological measure typically constitutes income earnings, previous education, and/or occupation.140,141,142 Nevertheless, the increased risk of TD2M that is associated with SES is mediated by complexities involved with other modifiable risk factors. 140-142 In wealthier or more developed countries, low SES (age- and sex-adjusted) has been correlated with limited accessibility to healthful dietary choices, locations for physical exercise, healthcare services, and prosperous economic opportunity. 141,142,143,144,145,146 Concurrently, in poorer or less developed countries, higher SES has been associated with decreased levels of physical activity, obesity, and sedentary lifestyles.146,147,148 SES is commonly measured by a proxy variables (level of education, occupation status, type of diabetes, and address/postal code); 141-146 this tends to complicate its consistency across the research literature. 140-142,145 Moreover, the psychosocial factors that are negatively associated with low SES are inherently quite difficult to measure and to investigate in relation to T2DM. 149,150 Research has shown that depression (independent of, as well as in combination with low SES) in adults has been associated with an increased risk for T2DM;150,151 the relevant literature also indicates that this relationship is bi-directional. 152,153,154 An extensive review of several longitudinal cohort studies has specified that adults who are suffering from depression are at a 37% greater risk for developing T2DM. 152-155 This relationship is typically mediated by the

16 behaviours brought on by depression, or associated with T2DM.152-156 Sedentary activity, poor diet, cigarette and/or alcohol abuse, and more specifically, insulin resistance, have all been linked to either condition. 152-156

Several of the aforementioned modifiable risk factors also independently increase risks for diabetes related complications in individuals already diagnosed with T2DM.157 If left unmanaged or untreated, T2DM (and its modifiable risks) may lead to a whole host of related health difficulties.158,159 As a leading cause of CVD, mortality due to CVD among diabetics is roughly 60-80%. 6,7 Additionally, diabetic retinopathy has been noted as a major cause of blindness in the Canadian population, in addition to a reduction in the overall quality of life for those with T2DM.157,160 Approximately 27% of US adults over the age of 75 with T2DM were burdened with varying levels of diabetic retinopathy in 2005. 161 Diabetic nephropathy, such as CKD, is also frequent in individuals with T2DM and is correlated with an increased risk for renal failure and end-stage renal disease (ESRD);157-159,162,163 diabetes mellitus is the leading cause of ESRD. 162,163,164 One of the most common complications associated with all types of diabetes mellitus is neuropathy of the central and peripheral nervous system;165,166,167,168 damage to the brain, numbness, or amputation of the extremities occurs in 30-50% of individuals with diabetes mellitus.142-145 While the primary risk factor for diabetic neuropathy is raised plasma glucose levels (hyperglycemia), which typically accompany prediabetes and T2DM disease states, lengthier disease durations (diagnosed and undiagnosed), daily cigarette smoking, elevated triglycerides from poor diet, increased BMI, and excessive alcohol consumption have all been independently found to incite this complication. 157,165-166

2.2 Non-Modifiable Risk Factors for T2DM and Diabetes Related Complications

The literature concerning variables that are virtually unchangeable or genetic is vast, and relative to a wide range of Canadian subpopulations. The diversity among non-modifiable risk factors is just as complex and multiplicative as those that are modifiable. 40-46 One of the most intriguing issues that surfaces when examining the non-modifiable risk factor

17 literature is the inevitable genetic hypotheses concerning T2DM predispositions in specific population groups. 41,42

Common risk factors include older age, a family history of diabetes mellitus (type 1 or 2), sex (male or female, depending upon the population being examined), a history of gestational diabetes (in women), and certain racial/cultural origins and/or ethnicities (those of Aboriginal, African, Asian, Latino, or South Asian descent, for example).41,169,170 One of the most established predictors of future T2DM in individuals is a family history of the disease; 148 landmark studies such as the Framingham Offspring Study,148,171,172 the Health Professionals Follow-up Study,173 and the Rotterdam Study 174 are but a few prospective cohort examples that have demonstrated this link. Additionally, organ transplantation (most frequently kidney transplantation) has also been linked to the sudden and/or gradual development of T2DM (in small numbers, however).175,176 Following a transplantation procedure, some individuals may experience a sudden or gradual onset of severe insulin resistance (and eventually T2DM), usually in the first few months. 175-177 However, risk factors for this unusual development are similar for individuals who do not undergo organ transplantation (mostly non-modifiable): older age, certain ethnicities (African American and Hispanic kidney recipients observe much higher risks), daily cigarette smokers, increased BMI, and a family history of T2DM as well. 177,178,179

Sex differences for diabetes mellitus (inclusive of type 1 and type 2) rates have been noted in the Canadian population.180 According to 2008/2009 data provided by the Canadian Chronic Disease Surveillance System (CCDSS), approximately 6.4% of all Canadian females over the age of one year have diabetes mellitus, while 7.2% of all Canadian males over the age of one year have diabetes mellitus.180 In addition to higher prevalence rates, Canadian males also experience higher incident rates for diabetes. From 2008 to 2009, among the average 6.3 new cases per 1000 individuals diagnosed (approximately 200,000), males constituted 6.8 cases per 1000, while females constituted 5.7 cases per 1000. 180 The tendency for males to be more prone to diabetes has been noted in the literature; 41,42,169,170 however, certain ethnicities experience a reversal of this trend, and will be discussed later on in this review. 41,169,170

Older age has long been associated with a higher risk for developing T2DM.40-46,171,172 As we age, the healthy function of pancreatic β-cells deteriorates, and so does our body’s ability to metabolize glucose. 104-107,171,172 Using the same dataset from the CCDSS described above, the Public Health Agency of Canada reports that the proportion of diagnosed diabetes mellitus (inclusive of type 1 and 2) sharply rises with increasing age, most specifically after the age of 40. 180,181 The highest Canadian proportions are seen between the ages of 75 and 79, with approximately 23.1% of females having diabetes, and 28.5% of males. 180,181 Age (in general) is also of particular concern as a non- modifiable risk factor for T2DM because of the associated propensity for all-cause mortalities. 181 Individuals that have been diagnosed in young adulthood (ages 20 to 39 years old) observe all-cause mortality rates that are 4.2 to 5.8 times higher than those in the same age group who do not have T2DM.181 Similarly, individuals diagnosed in older adulthood (ages 40 to 74 years old) observe all-cause mortality rates that are 2 to 3 times higher than those in the same age group who do not have T2DM.181

Risk factors, both modifiable and non-modifiable, are applicable to every ethnicity; 40-42 yet, there are disparities between certain ethnicities that are notable when examining obesity, T2DM, and diabetes related complication rates.40-46 The following portions of the literature review will focus primarily on the observed differences in risks associated with racial origin and ethnicity.

2.2.1 The Definition of Ethnicity in Epidemiological Research

At best, ethnicity does not have a standardized definition in the T2DM literature; this due to the conflicting interests of researchers that are driven by the operationalization of context surrounding various ethnicities, and those that are focused on the conceptualization of the definition.182 ,183,184 Clarke et al. reviews the definition of ethnicity by comparing two different perspectives that are typically observed in the literature: the “fixed primordialist ” view,185,186 and the “social constructivist ” view.187 Clarke et al. speak in favour of the social constructivist view of ethnicity by saying that ethnicity itself not only involves the “sharing of a common culture, but also language, religion, national identity, social and/or political position within a country’s social

19 system.”188,189 Two additional features of ethnicity are proposed by Clarke et al.: “identity ,” which refers to one’s self identification in cultural groups, and “ origin ,” which refers to one’s ancestral classification of ethnic or cultural populations. 161,188,189

Relevant to the present study, Clarke et al. states that Canada’s national health surveys (such as the CCHS) characteristically use racial, cultural, and ethnic origin variables that attempt to capture both identity and origin via multiple choice categorization.188-190,191,192 Subsequently, the groups that are identified by these self-report variables acknowledge multiethnic groups, self-identified cultural origins, and the various ancestries of respondents (integral to ethnic comparisons of health outcomes in epidemiology). 193 Based on Clare et al.’s review, the racial, cultural, and ethnicity variables formulated by the CCHS (and used this thesis) will be deemed as appropriate proxies and synonyms for ethnicity and/or ethnic origin.

2.2.2 Ethnic Origin and Obesity

Obesity is an increasingly important modifiable risk factor for T2DM, as discussed previously. Concerning non-modifiable risk factors, ethnic differences in the propensity for obesity (and subsequently, T2DM) have been recorded in the literature. 40-45 Relative to adiposity distribution and abdominal obesity, ethnic/racial differences accounted for a 47.8% excess risk of obesity observed in African-American women in one United States study,194 and a 39.9% excess risk reported by the American Diabetes Association.195 Other United States investigations have noted a greater prevalence of obesity in Hispanic- Americans 196,197 (as well as African-Americans) at the time of T2DM diagnoses. 198,199 In the United Kingdom, non-Caucasian British-South Asians, British-Africans, and British- Caribbean individuals have a greater propensity for abdominal obesity,200,201,202,203 and significantly high rates of self-reported sedentary lifestyles. 204 Studies in both the United States 205 and the United Kingdom 206 have additionally confirmed that the Indian ancestry population (a large majority of those whose ancestry is from the South Asian continent) reports higher levels of obesity (measured by abdominal fat, waist-circumference, or waist-to-hip ratio) at every BMI category when compared to the Caucasian and African (ancestral) population in either country.

2.2.3 Ethnic Origin and Insulin Resistance

T2DM pathophysiology involves the variable combination of β-cell malfunction and insulin resistance. 88-90,207 Studies in the United States have confirmed an interesting paradox in the African-American population: African-American children show increased rates of insulin resistance when compared to Caucasian children (this effect is even greater for female children),208,209 while rates of insulin resistance are fairly similar in the adult population. 210 Moreover, the differences seen in the pediatric population were observed even after controlling for dietary intake and physical activity level.187,211 Additional studies have shown that West African descendants (African Americans and Ghanaian Americans, in particular) have higher rates of insulin resistance when compared to Caucasians. 212,213 The Insulin Resistance Atherosclerosis Study suggests that the physiological mechanism responsible for the higher rates of insulin resistance observed in African Americans and Hispanic Americans is probably due to an altered hepatic insulin clearance.214 It is important to note that their results also suggest genetic linkages for specific ethnicities even after adjusting for age, sex, BMI, physical activity level, and waist-to-hip ratios. 214 Studies in the United Kingdom have similarly suggested genetic propensities for insulin resistance in the British-South Asian population. 200-203,206

2.2.4 Findings from Canada: Ethnic Origin and T2DM

Canadian research has previously highlighted important ethnicity differences in modifiable risk factors (such as obesity, physical activity level, and cigarette smoking) for T2DM. 217 Based on data from a single iteration of the Canadian Community Health Survey (2010), some studies found that Chinese and South Asian Canadians had the smallest rates of obesity (3.3% and 9.5% prevalence rates, respectively) when compared to Caucasian, African, Filipino, and Latin American Canadians (19.7%, 18.7%, 12.5%, 15.7%, respectively). 180,181,215 Concurrently, these studies also found that Chinese and South Asian Canadians had the highest rates of physical inactivity (61.9% and 60.5%, respectively) when compared to Caucasian, African, Filipino, and Latin American Canadians (45.8%, 57.5%, 54.8%, 54.9%, respectively). 217 Inversely, Caucasian

Canadians were found to have higher rates of daily cigarette smoking than any other ethnic origin group (17.5%; no other ethnic origin’s prevalence rate was above 9.5%). 217

Currently, the Public Health Agency of Canada reports that certain ethnic subgroups have higher T2DM prevalence rates in Canada. 180,181 Compared to Canadians of European descent, individuals of South Asian, Hispanic or Latino American, Aboriginal, Chinese, and African descent are at much higher risks for developing T2DM.180,181 Furthermore, these groups experience diagnoses at younger ages, and at lower BMI values than those of European descent. 216,217 The majority of these ethnicity findings, however, was gathered either from research conducted in the United Kingdom, United States, or provincially-specific surveillance projects.180,217 Reliance on data from the United Kingdom and the United States is most likely due to the similar ethnicity distributions that exist among the three nations. 203,206 Nevertheless, T2DM research that explicitly investigates ethnic-origin differences at the national level in Canada, is currently lacking.

2.2.5 Findings from the United Kingdom: Ethnic Origin and T2DM

Comparable with Canada, the United Kingdom has an extremely diverse population with similar chronic disease burdens.218 Since the mid-1990’s, the United Kingdom has notably observed extremely high rates of T2DM, hypertension, and renal failure in the South Asian, African, and Caribbean descendant subpopulations, especially when compared to Caucasians.219,220 In particular, these ethnicities (collectively) have T2DM rates almost five times greater than Caucasians adults aged 40 to 69 (20% and 5%, respectively).221,222 Issues of early onset diagnoses and lengthier disease durations tend to appear for various ethnicity subpopulations in the United Kingdom as well. 223,224 Findings from several cardiovascular risk factor investigations have determined that British-South Asians are particularly prone to reporting habitual physical inactivity;225 this may also contribute to British-South Asians’ non-modifiable risk for T2DM. 225

One study in the United Kingdom attempts to explain ethnic differences in T2DM prevalence rates by examining specific components of insulin resistance pathophysiology; 226 in particular, insulin-like growth factor 1 (IGF1). 226 IGF1, integral to glucose metabolism, is comprised of various IGF binding proteins. 226 In low plasma

22 concentrations, IGF binding protein 1 (IGFBP1) has been associated with insulin resistance; 226,227,228 these concentrations reportedly vary by ethnic origin. 206,227,228 Even though dietary intake was found to be strongly linked to these ethnic differences in IGF binding protein concentration, ethnicity alone still significantly influenced concentrations after diet was adjusted for.229

Ethnic differences in diabetes related complications have also been extensively researched in the United Kingdom. 230-233 General nephropathy and renal failure has been observed in much higher frequencies among the South Asian, African, and Caribbean descendent diabetes mellitus population when compared to Caucasian individuals with diabetes.230 Moreover, mortality from diabetes mellitus (both type 1 and 2) is nearly 3.5 times the overall rate in the United Kingdom for South Asians (both men and women) and for those of African or Caribbean descent (men). 231 One study has noted a decreased risk of amputation (as a result of peripheral neuropathy) in African and Caribbean descendent men,232 suggesting a possibly protective interaction effect between these ethnicities and sex, relevant to this complication.232 Concurrently, one study concerning cardiovascular risk factors found that mortality from CVD among individuals with T2DM was 3 times greater for South Asian descendants, and that mortality from stroke was 3.5-4 times greater for African and Caribbean descendants.233

2.2.6 Findings from the United States: Ethnic Origin and T2DM

In the United States, Hispanic-234 and African-American 235 subpopulations consistently observe higher rates of T2DM when compared to other ethnicities within the country.

Hispanic Latinos experience remarkably high rates of poverty in addition to high T2DM rates when compared to other American ethnicities. 236,237 In the year 2000, the Centers for Disease Control and Prevention reported that the Hispanic Latino population in the United States had a 100% increased risk (age- and sex-adjusted) of developing T2DM when compared to non-Hispanic or Latino Caucasians (prevalence rates of 11.7% and 4.8%, respectively), and were twice as likely to be hospitalized for diabetes related complications.238,239 Hispanic Latinos also report the lowest rates of healthcare utilization

(relative to the management of T2DM) when compared to any other ethnicity in the United States.240

African Americans (of black racial and ethnic origin) have T2DM diagnosis rates well above the rates observed by Caucasians.241 The prevalence of T2DM in African Americans is approximately 1.4 to 2.3 times higher than Caucasians in the United States. 241 Moreover, this ethnic subpopulation has higher rates of peripheral neuropathic amputation,242 diabetic retinopathy,243 fatal diabetic nephropathy,244 and diabetes related morality 235 in comparison to Caucasians. Interestingly, studies have shown that low SES impacts the black African American population more so than any other ethnicity in the United States; 245,246 even after adjusting for SES, these studies have demonstrated that the interethnic differences between African Americans and other ethnicities concerning T2DM prevalence remain.245,246 It has been suggested that increased African American and Hispanic Latino American propensities for T2DM are the combined result of a G6PD genetic deficiency and the typical “American” or “Western” diet (otherwise classified as a poor diet); the basis for this association stems from the theoretical “thrifty gene ” effect.247,248

2.2.7 The Thrifty Gene Effect

Once accepted as a legitimate pathophysiological explanation for obesity in certain ethnicities, the thrifty gene effect theory has been suggested as a causal hypothesis for linking a ‘thrifty gene’ to the development of T2DM;247,248,249 this has since been contentiously debated in the literature. 247-249 The theory proposes that, based on historical interpretations of human survival patterns, certain ethnicities have been genetically programmed with specific traits relevant to their metabolism. 247-249 When food was previously considered to be unavailable, these genes provided a ‘thrifty’ metabolic advantage that led to the efficient storage of energy (and subsequently, fat). 249 In times of hunting and gathering, these genes would have provided an advantage; 247-249 however, the typical “ Western ” diet and lifestyle (in the present era) have apparently negated the historically ‘thrifty’ benefits.234 While the theory may be considered as a plausible mechanism for increased propensities for obesity, the legitimacy of its argument as an

24 explanation for T2DM is flawed. 250 Dr. Samuel Dagogo-Jack from the University of Tennessee argues, “Obesity ensured survival by making available stored energy during starvation; diabetes, on the other hand, reduces survival by a full decade and compromises quality of life through its numerous complications.” 250 The complete comparative logic behind the theory may be found in the following figure:250

Figure 1: Thrifty Hypothesis Schema

Two theoretical alternatives for explaining ethnic variations in T2DM risk are the “Antagonistic Pleiotropic” theory 251 and the “Genetic Trash Can” theory. 252 First, antagonistic pleiotropy involves a genetic tradeoff between younger age survival security, and older age T2DM predisposition. 250,251 Genetically, pleiotropy involves one gene controlling for the expression of more than one trait in an individual; antagonistic pleiotropy occurs when one gene controls for a trait that may be beneficial (ensuring survival security at younger ages, for instance) and another trait that may be detrimental (in this case, higher risks for T2DM at older ages). 250,251 The latter theory posits that multiple gene mutations (individually considered neutral) will confer adverse risks when condensed in a given ethnic subgroup. 250,252,253 The theory also suggests that these “diabetogenes” would recessively be retained in ethnic populations where frequent migration, interbreeding, and other ethnic dilution does not commonly take place. 250,254

The origins of ethnicity differences in T2DM are evidently modulated by certain non- modifiable (possibly pathophysiological) risk factors. Social, cultural, and modifiable

25 lifestyle risk factors are also present in this association by varying degrees (as a modulation of the possible T2DM expression). Of particular importance to several ethnic subpopulations’ health in Canada, the United States, and the United Kingdom, are the unique risk factors that are associated with migration; 25 a large portion of the Canadian population (20%) are immigrants.25,26 Understanding the complexities and impact that immigration may have on immigrants’ health is integral to the rationale of this thesis, and our understanding of T2DM in various ethnic origin groups. 25

2.3 Migration, Health, and Birth Origin

The historical study of migration and population health was previously rooted in attempts to discover and control epidemic threats, believed to be migrant-imported. 255 Infectious diseases were initially associated with the arrival of foreigners; 255 investigations of immigrant health were thought of as safety precautions for the health and well-being of non-immigrant citizens.255 Nevertheless, as health policy began to change, so did the understanding that immigrants were typically burdened with more noninfectious/chronic disease conditions than infectious ones. 255 As a result, research interests concerning migration became much more diverse.25,255

As cited previously in this thesis, nearly 20% of Canadian citizens were born outside of Canada. 25,26,256 Concurrently, approximately 12% of Americans were born outside of the United States. 257 Migration and birth origin have become increasingly important when investigating ethnic differences in health. 254-256 Unfortunately, insufficient sample size has been linked to numerous immigrant health investigations that are chiefly relevant to the Canadian population;258,259 these studies were unable to effectively assess the effect of birth origin on health and disease outcomes at the national level. Rather, they were limited to considerably general comparisons of immigrants vs. non-immigrants due to a lack of sufficient subpopulation granularity. 258,259 However, these types of studies are useful for generating literature that focuses on the systematic barriers to immigrants and their health.

Immigrant health may vary by global region of birth, and more specifically, by the duration of time that immigrants spend in country.260 The growing diversity of the North

American population necessitates the need for an immigrant perspective when surveying both health and disease outcomes. 258,259

A 2011 report from Statistics Canada summarizes the observed immigrant paradoxes concerning overall health.260 In some instances, immigrants have been found to report poorer overall health in comparison to Canada-born citizens.254 Other studies have suggested the existence of a “ healthy immigrant effect ”: immigrants observing and reporting better health upon arrival in Canada when compared to Canada-born citizens, especially regarding chronic disease.261,262,263,264 Research into the healthy immigrant effect, acculturation effects, and the time duration that immigrants spend in-country will allow for a more complete assessment of the potential disparities that exist between immigrant Canadian citizens and Canada-born citizens relevant to healthcare utilization barriers, and overall health. 264

2.3.1 The Healthy Immigrant Effect and Acculturation in North America

The health of newly arrived immigrants is a product of social, cultural, economic, and genetic determinants. 261-264 Immigrant health is also heavily influenced by time. 264

Overall, immigrants appear to report better health upon arrival in Canada. 265 Above 90% of Canadian immigrants have been shown to self-report ‘good’ to ‘excellent’ overall health, even more so than Canada-born citizens (in the first to fifth years following arrival); 266,267,268,269,270,271,272 this observation has been confirmed across studies using data from the National Population Health Survey, the Canadian Community Health Survey, the Longitudinal Survey of Immigrants linked with health service administrative data, and with Canadian landed immigrant databases. 266-272 The capability of movement to another country (and preparing oneself physically and mentally) may be an indicator of the already pervasive strength in health that this subpopulation observes. 262 Gushulak et al. believe that while this data proposes positive scenarios for Canadian immigrants, these findings only capture one-third of the immigrant health product. 265 Inline with the idea that immigrants’ well-being is determined by a multitude of factors, Gushulak et al. summarizes these factors in a three-stage, timeline sensitive perspective: pre-migration,

27 migration, and post-migration. 265 Gushulak et al.’s perspective suggests that the initial health benefits that we see in the immigrant population are only representative of the brief amount of time immediately following migration. 265,272

The potential reasons for the healthy immigrant effect have been discussed in the literature. 261-268 Social and cultural aspects relevant to modifiable risk factors for chronic diseases may be very different from what is expected or considered standard in Canada and the United States.272,273 For example, immigrants from less developed regions of the world typically do not demonstrate lifestyle behaviours that are associated with obesity or T2DM.255,256 Likewise, attitudes towards various food products, alcohol, and/or cigarettes will differ not just in the various social/cultural contexts of different countries, but also by the health policies and governance regarding marketing, distribution and availability of said products in these different countries.274,275 Additionally, Canada’s immigration policies are known to be stringent. 265,274,275

Certain aspects of immigration law prioritize young, educated individuals for admittance, and frequently deny unhealthy individuals entrance into the country.276,277 Immigrant legislation in Canada and other developed countries requires a complete medical examination upon arrival.242 These examinations are in place to determine if one’s health status meets the standard for admittance into the country; 265 an unhealthy migrant will not be considered a productive or contributing member of society. 265 This type of selective entry practice will bias the health of the entering immigrant population towards a much healthier overall group upon arrival. 265,276,277 Still, some studies have found that certain immigrant populations do not experience the initial healthy immigrant effect in Canada.278 A 2011 health report from Statistics Canada found that immigrants from the United States (male and female) who chose to settle in one of Canada’s three largest metropolitan city areas (Vancouver, Toronto, and Montreal) did not experience any beneficial immigrant effects relative to most health outcome measures.278 A similar, but less significant rate was observed in Sub-Saharan African immigrants who settled in the same metropolitan areas. 278 Nevertheless, this study did confirm a national-level healthy immigrant effect for all other immigrant groups concerning all-cause mortality. 278

Unfortunately, the healthy immigrant effect is not pervasive over time. 265,275-277 Several studies have found that after lengthier periods of residency in North America, immigrants’ health typically declines toward similar levels seen in the North American- born population, if not worse. 265,275-277,279 The combination of behavioural, environmental, and cultural changes have been proposed as mechanisms for this degradation, being summarized as negative “ acculturation ” effects. 279,280 Over time, changes in physical activity level, dietary intake, alcohol, tobacco, or drug use, as well as healthcare service utilization, may lead to increased rates of obesity, insulin resistance, T2DM, and other adverse health conditions in immigrants. 265,275-277,279,280 Previous studies have also investigated the aforementioned modifiable variables in relation to increased risk in immigrants.281,282 Overall, the effects of acculturation differ by the country of investigation (emigrant destination), birth origins of the immigrants being studied, and the varying levels of change associated with each modifiable risk factor.283,284 Other studies have attempted to investigate the specific cultural impact that living in North America has on new immigrants over time (both the United States and Canada).285-286 This research has shown that components of Western culture are correlated with increased risks for the development of hypertension, 285 obesity, 286 and T2DM.287 Similar research has also found that traditional or culturally specific health behaviours that accompany initial (positive) health benefits upon arrival predictably degrade over time with increased acculturation. 288,289

The unique concept of acculturation has been questioned in the literature concerning its operationalization and measurement. 288,289 Proxy measures are frequently used across datasets in order to decipher the relationship between what is believed to be acculturation, and the health status of immigrant individuals. 279-289 While a crude definition of acculturation has been generally agreed upon as “an indication of the cultural change of minority individuals to the majority culture,”290 the applicable construct is consistently more contentious. Primarily, it is a “process of adjustment and adaptation to a new culture involving instances of cultural learning, maintenance, and synthesis.”291,292,293 Given the inherent sampling design limitations of many health surveillance systems and national health survey questionnaires, proxy variables are not only useful, but are often used as the only feasible method for assessing acculturation in the immigration population.294

Immigrant generation status, language proficiency and/or preference, time since immigration, and age at immigration are just a few examples of suitable proxy variables for measuring acculturation.294,295,296,297,298 However, Chun et al. (2011) details a slight problem with attempting to research acculturation either by proxy variables or self-report: sociopolitical and family contexts that shape changes in both cultural and health behaviours are unfortunately lost in the course of research, and are frequently understudied. 291 The entire process of acculturating is inherently dynamic and will vary in degree, dependent upon the acculturative demands of the destination country over time. 291,294

The various demands of immigrants subgroups are what lead to the disparities we see in adverse health outcomes that are relative to the present study. 294,298 Certain immigrant ethnic origins will be more easily acculturated than others. 299,300,301 For instance, some studies have shown that Japanese and Chinese immigrants to the United States have a more consistent relationship between acculturation and eventual T2DM diagnoses when compared to other United States immigrant populations. 299-301

2.3.2 Potential Barriers to Immigrant Health, Diabetes Management, and Healthcare Utilization

The process of migration is known to be difficult; 291 for some immigrants, the transition to North America may lead to social and economic maladjustment. 291,294,295

The influence of SES on increased risks for adverse health outcomes (like T2DM) has been well established in the North American and Australian immigrant populations. 276- 302,303,304 Among immigrant subgroups, SES indices will vary dependent upon pre- migration education level, occupation, and a host of other SES factors. 291,301-304 Many immigrants find themselves in a vulnerable SES position upon arrival in a new country; 305 occupational instability and cultural unawareness has been associated with high levels of precarious employment for many newly arrived individuals.305

One of the largest potential barriers to immigrants, their health, and their use of healthcare services (both preventative and remedial) is inadequate language competency

(occasionally associated with a lack of acculturation).294-298 In 2005, approximately 36% of newly arrived immigrants to Canada had reported no competency in, or knowledge of English or French (our country’s national languages). 261,306 Less acculturated immigrants have been shown to report higher levels of communication difficulties with healthcare service providers when compared to more acculturated immigrants or North American- born citizens. 307,308,309 Subsequently, this communication barrier has been associated with a lack of reported healthcare service use. 306-308 One study in particular found that relatively acculturated immigrants often report high levels of perceived health, chronic disease screening, and access to healthcare services. 310 Individuals who tend to underuse healthcare services typically have higher rates of adverse health outcomes, including T2DM.295,309

Other possible barriers to immigrant health, as well as T2DM management, include negative perceptions of service usage (cultural or social), gender expectations, perceived isolation, abandonment, or even loneliness from leaving home-country communities behind.311,312,313,314 Various approaches to diabetes management and treatment in the immigrant population must incorporate immigrants’ cultural views concerning “autonomy, authority, and self-efficacy” 315,316,317 in addition to considering the stressful psychosocial factors that are attributed to migration. 318 Each of these aspects will contribute to immigrants’ attitudes and receptiveness towards the healthcare system.264,218

Thabit et al. illuminates the importance of health literacy in immigrants with T2DM.318 If immigrant individuals have poor English language competency or specific cultural beliefs that do not coincide with traditional Western medicine, their ability to understand, accept, or interpret warnings and/or instructions from healthcare providers (concerning plasma glucose levels, treatment dosages, etc.) may be impeded, hence increasing the potential for unfavorable patient outcomes.318 Effective healthcare and diabetes management is challenging for both care providers and North American immigrants.319,320 The diversity and complexity of this patient population is extremely relevant in Canada.

2.4 T2DM in Immigrants and Migration in Canada

Inline with the acculturation findings detailed previously, research has shown that incident T2DM in immigrants increases over time, since immigration;288 this has been specifically observed in immigrants following 15 years of Canadian residency.30 In a 2005 Ontario study, long-term immigrants from South Asia, the Caribbean, sub-Saharan Africa, North Africa, the Middle East, and Latin America showed significantly higher rates of T2DM when compared to long-term Ontario residents. 321 After stratifying by time since immigration, it was also confirmed that these Ontario immigrants experienced the highest prevalence rates of T2DM following 15 years of residency when compared to immigrants who had been in Canada for 5 to 9 years (odds ratios of 1.40 (95% CI 1.36– 1.44) for females, and 1.52 (95% CI 1.48–1.56) for males).321 Furthermore, when compared to immigrants from Western Europe and the United States, long-term Ontario immigrants from South Asia reported triple the (age, sex, and time since immigration adjusted) risk of T2DM, while immigrants from the Caribbean and sub-Saharan Africa reported twice the risk of T2DM.318 Specific research concerning ethnic or racial origins, T2DM, and their associated rates by immigration status is limited at the national-level. Broadly speaking, however, current evidence has shown that recent immigrants (1 to 5 years) tend to be limited in their access to healthcare services and have lower overall employment income than the general Canadian population.322 These issues may contribute the (eventual) longer-term rises in T2DM cases that are diagnosed in the immigrant subpopulation.

Canada’s immigrant population grows by approximately 225,000 to 250,000 per year. 265 Gushulak et al. compares the proportionate differences between Canada’s average immigration rate to similarly developed countries like the United States (1,000,000 legal permanent residents per year for a population that is approximately 10 times the size of Canada) and Australia (about 192,000 new immigrants each year for a population of approximately 22 million), citing the importance of Canada’s comparative diversity. 265 Yet, even with remarkable rates of accepted immigrants, some 200,000 are living in the Canada without formal citizenship, and are subsequently limited in their access to healthcare services. 323

Immigration to Canada has changed dramatically over the past 30 years. 265 Changes to the national immigration policy (away from country preferences and towards a universal point system in 1967)324 began to alter the diversity of immigrants arriving in Canada each year, beginning in the 1970’s. 324 Prior to 1986, 95% of the Canada’s incoming immigrants were from Britain, Europe, and the United States. 324,325 Since the 1980’s, representation from Africa, Asia, the Caribbean, Latin America, and the Middle East has surpassed any other country origin group (as seen in Table 2324,325 ). 265,325 Approximately 78% of recent immigrants (the large majority of which come from Asia, Africa, the Caribbean, Latin America and the Middle East) choose to settle in either Vancouver, Toronto, or Montreal;326,327 occupational opportunities in these cities are perceived prosperously in comparison to other locations in Canada. 326,327

Table 2: Country Origin of Recent Immigrants by Canadian Census Year RANK 2006 2001 1996 1991 1981 People’s People’s United 1 Republic of Republic of Hong Kong Hong Kong Kingdom China China People’s 2 India India Republic of Poland Vietnam China People’s United 3 Philippines Philippines India Republic of States of China America 4 Pakistan Pakistan Philippines India India United 5 States of Hong Kong Sri Lanka Philippines Philippines America United 6 South Korea Iran Poland Jamaica Kingdom 7 Romania Taiwan Taiwan Vietnam Hong Kong United United 8 Iran States of Vietnam States of Portugal America America United United 9 South Korea States of Lebanon Taiwan Kingdom America People’s United 10 Colombia Sri Lanka Portugal Republic of Kingdom China

While the growing diversity of Canada’s population necessitates an immigrant perspective when discussing ethnicity disparities, it is important to review the unique considerations that are relevant to native, Canada-born ethnicities as well. The next portion of this review will describe Canadian Aboriginals (First Nations, Métis, and Inuit) and the distinctive risk factors that are applicable to their health.

2.5 Canadian Aboriginals: Unique Characteristics

The Public Health Agency of Canada reports that according to 2006 Census data, Aboriginals constituted approximately 3.8% of the Canadian population (59.5% First Nations, 33.2% Métis, and 4.3% Inuit).328 In particular, First Nations encompass approximately 615 different communities and speak 60 different languages.328 Using the same dataset, it has been reported that Canadian Aboriginals are also the youngest subpopulation in Canada with 47.8% under the age of 25 (comparable with 31.7% for non-Aboriginal Canadians). 329

When examining health outcomes associated with Aboriginals in Canada, important historical influences must be taken into consideration. Centuries ago, Aboriginal populations were subject to European colonization. 330 Culturally traditional lands were dispossessed, assimilation policies like the Indian Act were put into place, and Aboriginal reserve communities were eventually formed. 331,332 Policies, rights, and government practices have since changed; 331-332 yet, economic, social, and health disparities between the Aboriginal population and the non-Aboriginal population persistently exist today. 333 Cultural barriers, environmental conditions, jurisdictional issues, and community determination are just a few examples of the many determinants that differentiate Aboriginals from other subpopulations concerning their health outcomes. 331-334

Currently, federal and provincial governments are responsible for Aboriginal health, as well as some community-based bodies. 335 Nevertheless, there has been a growing effort from reserve communities that demands the direction of physician and hospital services be allocated to more culturally receptive community-based authority, especially in remote or isolated regions of Canada. 335 Potential barriers to health are exceptionally pronounced

34 in this subpopulation due to the geographic isolation of many Aboriginal communities, and the subsequent complexities surrounding their healthcare systems and services. 336,337

Similar to many immigrant communities, the cultural and language differences that exist in this population have also been linked to communication difficulties between healthcare providers and Aboriginal patients (both on- and off-reserve). 338,339 As a result, First Nations, Métis, and Inuit populations have habitually low rates of healthcare utilization when compared to the non-Aboriginal Canadian population. 340 Moreover, high rates of poverty, substandard housing, and unemployment are also reported for Aboriginals. 341,342 Many of the unique SES barriers to Aboriginals are systemic, typically leading towards poorer access to post-secondary education and non-precarious or gainful employment.343 As mentioned previously, low SES has often been linked to adverse health outcomes. 141- 143

2.5.1 Modifiable Risk Factors for T2DM in Canadian Aboriginals

The complex risk factors (both modifiable and non-modifiable) for T2DM in the Aboriginal population are cumulative, stemming from rapid social/cultural and environmental changes, as well as adaptation.344,345,346 Even though risk factors are similar to those in the overall Canadian population, Aboriginals experience these risks in greater magnitude.344-346 Across Canada, many potential detriments to Aboriginal health have been linked to high rates of obesity, insulin resistance, and T2DM;344 harsh environmental conditions, widespread physical inactivity and unhealthy eating habits, high rates of obesity, and high rates of daily cigarette smoking have all been noted in the literature. 344- 346

Certain Aboriginal communities face environmental factors that prevent regular physical activity; 347 an absence of recreation facilities, loose dogs in many reserve communities, poor roads, and a lack of overall safety have been identified as detriments to positive health behaviours in reserve Aboriginals.347,348 According to 2009 data analyzed from the Aboriginal Peoples Survey and the Canadian Community Health Survey, only 26.0% (95% CI: 24.5-27.5%) of on-reserve First Nations adults reported sufficient physical activity during leisure time, and 51.8% (95% CI: 48.0-55.5%) of off-reserve First Nations

35 adults reported physical inactivity during leisure time, comparable to non-Aboriginal adults (49.7% (95% CI: 49.2-50.3%)).349 These findings suggest very different tendencies for physical activity dependent upon geographic residency (associated with the aforementioned barriers on reserve communities). Métis adults similarly reported high levels of physical inactivity during leisure time (46.7% (95%CI: 42.8-50.6%)), while Inuit adults reported the highest rates of physical inactivity during leisure time (59.6% (95% CI: 50.5-68.6%)). One study notes that in 2004, approximately 82% of Inuit adults did not meet the recommended daily or weekly levels of physical activity.350

Aboriginals residing on reserves are also subject to incredibly high food costs, which severely limits their choices of healthful foods (this barrier is compounded by pervasively low rates of SES in these communities).343,351 In particular, one study found that the majority of Aboriginals living on the Six Nations reserve (the largest First Nations community in Canada, found in Ontario) purchased groceries off of the reserve due to decreased food availability and high food costs. 352 Moreover, Aboriginals also reported spending approximately $151.00 per week on food, in comparison to $140.00 per week reported by the general Ontario population. 339,353

Aboriginals (on-reserve in particular) have grown increasingly dependent on prepared foods. 344-346,352,353 Consistent consumption of these foods typically results in higher intakes of saturated fat, low dietary fiber, refined carbohydrates (that are high on the glycemic index), and overall calories (a diet that has been proven to cause weight gain and insulin resistance).123-126,354 Traditional Aboriginal diets characteristically consist of a high amounts of protein and low amounts of carbohydrates; 355,356,357 research has found that this type of diet reduces the risk of obesity and T2DM. 358 Acculturation of Aboriginal communities, their diets, and less frequent dependence on hunting and fishing has reportedly led to an overall decrease in the amount of energy expenditure observed in these individuals (as described by the physical inactivity rates stated previously);355-358 this results in a surplus of calories that was previously metabolized.355 Data gathered from the Aboriginal Peoples Survey and the Canadian Community Health Survey have observed this transitional trend for First Nations, Métis, and Inuit populations across Canada. 345,359

Rates of obesity in Canadian Aboriginals are significant higher than in non- Aboriginals. 344 According to self-reported BMI data from the Aboriginal Peoples Survey and the Canadian Community Health Survey, 74.4% of on-reserve First Nations adults, 62.5% off-reserve First Nations adults, 60.8% of Métis, and 58.3% of Inuit were either overweight or obese (in comparison to 51.9% of the non-Aboriginal population). 344,360,361

Tobacco use has been traditionally practiced by many Aboriginal cultures.362 Across Canada in 2009, Aboriginal adults (on-reserve First Nations, off-reserve First Nations, Métis, and Inuit) reported prevalence rates of daily smoking that ranged from 2.2 (Métis) to 2.8 (on-reserve First Nations) times the non-Aboriginal daily smoking rate for adults. 344,359 Nevertheless, the lack of tobacco regulation policies on Aboriginal reserves, the expansion of tobacco production in these areas, and the increased ease of access to cigarettes on reserves has led to amplified numbers of Aboriginal adults becoming addicted to cigarettes.362

2.5.2 Non-Modifiable Risk Factors for T2DM in Canadian Aboriginals

Non-modifiable risk factors for T2DM, such as sex and age, differ in Aboriginals when compared to the non-Aboriginal population in Canada. 359

The literature concerning the health of Aboriginal women has suggested a reversal of the T2DM sex trend that is normally observed in the non-Aboriginal Canadian population. 52,363 Typically, females have lower rates of T2DM than men; 52 however, in Aboriginals, females often report higher rates of T2DM than males. 52,363 This sex trend has been identified by studies conducted in Ontario, 356 Alberta,364 and in Saskatchewan. 341 For example, Albertan Métis females were found have a T2DM prevalence of 7.8% while Albertan Métis males reported a T2DM prevalence of 6.1%. 364 In addition to the peculiar T2DM risk that Aboriginal females evidently possess, they also report the highest rates of gestational diabetes when compared to all other Canadian females.365,366,367,368,369,370,371,372 Some studies have observed Aboriginal gestational diabetes prevalence rates that are nearly triple that of the gestational diabetes rates seen in non-Aboriginals.361,365 One recent study on First Nations women in Ontario found that

37 prevalence rates for gestational diabetes (6.5% compared to 4.2% of non-Aboriginal females) and pre-pregnancy diabetes mellitus (3.9% compared to 1.8% of non-Aboriginal females) were higher for Aboriginal females, and that the Aboriginal females were generally younger at the time of pregnancy and diagnosis (average age of 28.76 years old compared to 32.79 years old for non-Aboriginal females).371

Overall, Aboriginals are younger than the non-Aboriginal Canadian population and characteristically report the highest rates of T2DM among youth (under the age of 18);373,374 all three identification groups are regularly diagnosed at younger ages when compared to the non-Aboriginal population.375 A 2010 study of children under the age of 18 that were recently diagnosed with T2DM found that approximately 44% cases were Aboriginal, far surpassing the proportions of other ethnically identified groups. 376 While some researchers have commonly cited the thrifty gene effect (Section 2.2.7) as a causal explanation for the high rates of obesity and T2DM seen in Aboriginals, 248,249 this claim remains contentious. 377 Though, researchers have discovered a polymorphism of the hepatic nuclear factor (HNF), called 1-α transcription factor, that has not only been linked to reduced insulin secretion in the First Nations people of Oji-Cree ancestry (of Manitoba and Ontario), but has also been proposed as an underlying cause of earlier-onset T2DM in Aboriginals; 378,379,380 the relative importance of these isolated communities may play a role in future investigations of gene-dose dependent effects. 378-380 The presence of this transcription factor has been associated with T2DM development following only mild insulin resistance, indicating a decreased ability to secrete insulin in those that carry the polymorphism.378

2.5.3 Canadian Aboriginals: Diabetes Burden and Diabetes Related Complications

Prior to the 1950’s, diabetes mellitus was rarely, if ever observed in Canadian Aboriginal populations; 344 however, Aboriginal T2DM prevalence rates have increased rapidly in the past 60 years toward near epidemic proportions. 349,359

Prevalence rates of T2DM in some Canadian Aboriginal groups well exceeds the rates seen in the non-Aboriginal population. 349,359,371 Based on Public Health Agency of

Canada data (2008/2009) collected from single cycles of the Canadian Community Healthy Survey, the Aboriginal Peoples Survey (APS), and the First Nations Regional Longitudinal Health Survey (FNRLHS), on-reserve First Nations adults aged 18 years or older reported a 15.3% crude prevalence rate of diabetes diagnoses, followed by First Nations off-reserve individuals aged 12 years or older at 8.7%, Métis individuals aged 12 years or older at 5.8%, and Inuit individuals aged 15 years or older at 4.3%. 381 Age- adjusted results showed similar patterns at 17.2%, 10.3%, 7.3%, and approximately 6.0%, respectively. 381,382 It had been previously believed that the Inuit ethnicity was protective against T2DM (more so than any other Canadian ethnicity).383,384 Nevertheless, recent age-adjusted data has indicated that diagnoses in Inuit are relatively comparable to the non-Aboriginal population. 383,384 The majority of these figures are considerably larger than the overall national diabetes rate of 6.8%,1 underlining the importance of investigating the T2DM epidemic in the overall Aboriginal population.

Other research has illuminated increasingly high propensities for diabetes related complications 385,386,387 and mortality 388 in Canadian Aboriginals. The tendency for Aboriginals to be diagnosed at younger ages is thought to be the reason why diabetes related complications (in particular, nephropathy such as CKD and ESRD) are quite common in the Aboriginal community.389,390 Younger individuals with T2DM will inherently experience longer durations of exposure to the physiological and metabolic burden of diabetes, leaving them with increased risks for complications.386,391,392 For instance, retinopathy is also exceedingly prevalent in certain Aboriginal communities with T2DM.393

Overall, mortality rates are significantly higher for Aboriginals than non- Aboriginals. 383,384 More specifically, Aboriginals are 2 to 4 times more likely to die from diabetes related complications than the general Canadian population.384 Aboriginal mortality from CVD is also twice the rate seen in the non-Aboriginal population. 394

2.6 Summary

Prevalence rates for diabetes vary widely depending upon the data source and methodology; moreover, rates of diabetes tend to fluctuate with varying racial, cultural,

39 ethnic, and country origins. A lack of consensus concerning diabetes prevalence is observed when comparing one or two iterations of the Canadian Community Health Survey (CCHS), information collected and produced by the Canadian Chronic Disease Surveillance System (CCDSS), previous cohort investigations, and racial, cultural, ethnic, or country origin data from the United Kingdom and the United States. While each source (and methodology) are valid and reliable, any single source may under- or over-estimate the actual prevalence rates and most recent trends of diagnosed diabetes in certain Canadian subpopulations. In order to comparatively gauge the rising epidemic of diabetes in Canada and the differential impact of certain risk factors for diabetes among Canadian subpopulations, holistic data sources must be used.

Cross-sectional Canadian health surveys provide the necessary time-relevant data for an investigation of diabetes time trends and risk factors for diabetes, especially in immigrants and Aboriginals; recent longitudinal health data for these subpopulations are either unobtainable or not conducive to comparing risk factor differences for diabetes across subgroups.

Recent Canadian research has found exceedingly high prevalence rates for diabetes,349,359,371 diabetes related complications 395,396,397 and mortality 398 in Aboriginals; however, no previous study has examined these rates using all available iterations of the CCHS (spanning 2001 to 2010). The national representativeness of the CCHS, especially for off-reserve Aboriginals, will help to address the increasing need for knowledge regarding Aboriginal health and the extent to which this subpopulation is experiencing the rising epidemic of diabetes. 1,2,349,359 Conversely, immigrants possess unique health determinants that necessitate a consideration of immigrant status when investigating ethnicity disparities and diabetes. A gap currently exists in the literature concerning Canada’s growing immigrant population and the extent to which they are experiencing the rising epidemic of diabetes, especially at the national-level. 30,288,321

An additional 1.2 million people are expected to be diagnosed with diabetes in Canada over the next seven years, raising the total prevalence of the disease from 4.2% (in 2000) to approximately 9.9% (in 2020).2 By assessing diabetes trends and period prevalence rates of diabetes from 2001 to 2010 (using a comprehensive dataset), population health

40 surveillance systems may be better able to assess the validity of this prediction for the overall Canadian population, as well as high-risk Canadian subpopulations. As described in this review, immigrants and Aboriginals possess unique risks for diabetes that require increased scrutiny and research at the national-level; differential increases in diabetes trends across subpopulations may indicate informational or systemic barriers to diabetes care. These potential disparities have important policy implications for prevention and health promotion systems that may cater to the complex determinants of health in Canadian subpopulations.

This thesis will add to the current body of diabetes surveillance literature by using pooling methodology with several cycles of the CCHS. By combining seven different iterations of the CCHS, this thesis will characterize Canadians with diabetes from 2001 to 2010, and compare high-risk Canadian subpopulations using an extremely large representative sample. Diabetes period prevalence rates and time trends for all Canadians, Canada-born individuals, immigrants, and Aboriginals will be presented from 2001 to 2010. As well, the modifiable and non-modifiable risk factor differences between these subpopulations (from 2001 to 2010) will be interpreted. Due to CCHS survey limitations (a lack of explicit distinction between T1DM and T2DM in the survey questionnaire), T2DM will now be referred to, analyzed as, and discussed more generally as diabetes (mellitus).

Chapter 3 3 Objectives

Chapter 2 reviewed the current state of diabetes research and identified certain gaps and inconsistencies in the literature that are relative to data methodology and investigations of certain subpopulations’ health. Concerning all Canadians, there is a need for surveillance research at the national-level which evaluates the predicted rise in diabetes prevalence rates from the year 2000 onward. 2 Furthermore, immigrant and Aboriginal subpopulations have unique health determinants, as well as distinctive barriers to their health; there is currently a need for comparative surveillance research to specifically assess diabetes prevalence rates and recent diabetes trends among these subpopulations. Using all available iterations of the CCHS (seven in total, spanning the years 2001 to 2010), the present study aims to address these gaps.

There are two primary objectives for this thesis. Firstly, this thesis will characterize Canadian diabetics and examine the period prevalence rates of diabetes, as well as the diabetes time trends among all Canadians, Canadian-immigrants, Canadian-Aboriginals (First Nations, Inuit, and Métis), and Canada-born (non-immigrant, non-Aboriginal population) individuals from the year 2001 to 2010. Secondly, this thesis will assess the differential impact of modifiable and non-modifiable risk factors for diabetes among Canadian-immigrants, Canadian-Aboriginals (First Nations, Inuit, and Métis), and Canada-born (non-immigrant, non-Aboriginal population) individuals from 2001 to 2010.

Differences between the subpopulations of main interest will then be discussed using comparable and current statistics provided by national-level population health surveillance programs.

Objective One 1. Characterize the distribution of modifiable and non-modifiable risk factor variables for the total sample of Canadians (aged 12 and above, as sampled by the CCHS) with diabetes from 2001 to 2010;

2. Determine the period prevalence rates of diabetes among all Canadians, Canadian- immigrants, Canadian-Aboriginals (First Nations, Inuit, and Métis), and Canada- born (non-immigrant, non-Aboriginal population) individuals from 2001 to 2010; 3. Determine the period time trends for diabetes among all Canadians, Canadian- immigrants, Canadian-Aboriginals (First Nations, Inuit, and Métis), and Canada- born (non-immigrant, non-Aboriginal population) individuals from 2001 to 2010

Objective Two 1. Assess the differential impact of modifiable risk factors for diabetes among Canadian-immigrants, Canadian-Aboriginals (First Nations, Inuit, and Métis), and Canada-born (non-immigrant, non-Aboriginal population). These risk factors include: smoking status, alcohol consumption frequency, BMI, socioeconomic status (income and education level), diet, and national language competency; 2. Assess the differential impact of non-modifiable risk factors for diabetes among Canadian-immigrants, Canadian-Aboriginals (First Nations, Inuit, and Métis), and Canada-born (non-immigrant, non-Aboriginal population). These risk factors include: age, sex, racial/cultural origin, and ethnic origin. Risk factors that are specific to immigrants include: time since immigration, age at immigration, and country origin.

Chapter 4 4 Methodology

The present study compiled and analyzed data from all available cycles of the Canadian Community Health Survey (CCHS); this incorporated seven different iterations from the year 2001 to 2010. The CCHS is a cross-sectional national population health survey that was previously conducted biennially (2001 to 2005), and is now conducted annually (2007 to present). While some of the CCHS questionnaire’s content modules have changed over time, the modules, outcome measure, and explanatory variables used in this study have remained consistent from 2001 to 2010. Detailed materials concerning sampling design and weighting procedures have been provided by Statistics Canada (the survey’s administrator) in order to ensure that valid and reliable health data is available for national population surveillance research. This chapter will specifically describe the dataset, outcome measures, variables, and methods used for analyses. Information regarding survey design and methodology has been collected from a combination of user guides, provided by Statistics Canada.

4.1 Data Source and Sampling Design

The CCHS aims to collect pertinent information that is relative to Canadians’ “health status, healthcare utilization and health determinants…” and has done so using extremely large samples from the population; on average, approximately 65,000 are surveyed each year. 399,400,401,402,403,404,405 As its main use, the CCHS supports population surveillance projects and programs that wish to use multiple levels of health data (national, provincial, or smaller health regions). The survey has been formulated to be a flexible tool that can be catered to emergent or developing issues relative to the Canadian population at the time of administration; as the population changes, so do some parts of the CCHS content.

Since 2001, the CCHS target population has included all Canadians aged 12 years old and older, and has maintained geographic sampling consistency (all provinces and territories are included in each iteration). Exclusions include full-time members of the Canadian Armed Forces, individuals living on-reserve or Aboriginal settlements, individuals living

44 on Crown Lands, members of institutionalized groups, and certain (extremely) isolated regions in Nunavut and Quebec. In total, these excluded groups encompass less than approximately 3% of the total Canadian target population. 374-380 Given the nature of the CCHS exclusion criteria, any conclusions made pertaining to identified Canadian- Aboriginals (First Nations, Inuit, and Métis) will be specific to off-reserve community populations.

Participation in the cross-sectional CCHS is completely voluntary, and all responses are recorded via self-report during facilitated interviews (details below). CCHS cycle data collection typically runs from January to December of the iteration year (for instance, January 2010 to December 2010 for the 2010 cycle); however, certain cycles have varied in the timing of administration prior to 2007. Cycle 1.1 (2001) data collection ran from September 2000 – November 2001, Cycle 2.1 (2003) data collection ran from January 2002 – December 2003, Cycle 3.1 (2005) data collection ran from January 2004 – December 2005, and Cycle 4.1 (2007) data collection ran from January 2006 – December 2007.

Content structure for the CCHS includes three components. First, common content includes questions that are asked of respondents regardless of province or territory, and is separated into two types: annual (questions that are asked of all respondents, regardless of year or cycle) and one or two-year content (previously known as ‘theme content’) involving queries that are periodically presented. Second, the CCHS survey contains optional content that addresses issues relevant to specific provinces or health regions; these questions are normally requested based on regional public health priorities. Last, the rapid response component includes inquiries made by organizations that wish to obtain explicit national population health estimates. These parameters are measured using distinctive topic or content areas, such as nutrition or mental health. In order to ensure geographic and temporal consistency across the cycles used, only annual common content (typically including demographics, height, weight, overall health, healthcare utilization, and other general health information) used for this study; thus, allowing for valid conclusions to be made about Canadians at the national level, across the designated time period (2001 to 2010).

Given that this thesis uses all available iterations of the CCHS, it is important to take into account any major changes to the survey, its administration, its content, or sampling design. Below is a summary of annual content modifications over time (Table 3406 ), taken directly from the Statistics Canada website. It should be noted that across time, no major changes have been made to the annual common content; any variables or outcome measures used for this study were unaffected by CCHS changes.

Table 3: Summary of CCHS Annual Content Modifications from Cycle to Cycle Since 2000/01, this survey includes the National Population Health Survey - Household Component (number 3236), which contains 2000-2001, data at the provincial level for 1994/95 to 1998/99. Cycle 1.1 Since 2000/01, this survey includes the National Population Health Survey - North Component (number 5004), which contains data at the territorial level for 1994/95 to 1998/99. Since 2003, this survey includes the Health Services Access Survey (number 5002), which contains data at the provincial level for 2003, Cycle 2.1 2001. Please refer to its description under the Documentation section. The data release of December 21, 2005 covers data collected over the first 6 months (January to June 2005) of the CCHS Cycle 3.1. At that time, the survey had collected information from about 69,000 individuals, aged 12 and older. 2005, Cycle 3.1 Only part of the data collected with the CCHS Cycle 3.1 questionnaire has been processed and finalized for this release. Data covering the entire 12 months collection period (January to December 2005) were released on June 13, 2006. Until 2005, the CCHS data were collected every two years over a one year period and released every two years, about six months after the end of the collection period. In 2007, the CCHS was redesigned to address two main points: the needs of partners to 2007, Cycle 4.1 increase the survey's content and the frequency of data releases, and to ensure better use of operational resources. For these reasons, the proposed changes to the CCHS design focused on improving the survey's efficiency and flexibility through ongoing data collection. Instrument design - As of the 2009 reference period, the theme 2009 content was removed from the questionnaire. Imputation - Beginning with the 2011 reference year, the household 2011 income variable is imputed.

As mentioned previously, the CCHS divides the target population into smaller geographic regions: firstly by province, then by health region (see Figure 2 407 ). In order to obtain a large enough sample with sufficient statistical power to generate reliable estimates for

46 each health region, a minimum required sample size is calculated for each cycle. Given that the number of health regions has slightly changed from year to year, the minimum sample size has also varied from year to year (especially once survey administration began annually in 2007) (see Table 4395-401 ).

The CCHS uses three sampling frames to derive the required sample of households: an Area frame, a List frame (of telephone numbers), and a Random Digit Dialing (RDD) frame. The amount of households per frame varies from cycle to cycle (proportions by cycle are listed in Table 4395-401 ). Proportional geographic allocation is ensured by initially assigning provincial sample quotas based on population size and number of health regions (within each province). Next, health regions are allocated their own sample quotas based on the square root of the population in their province. Lastly, clusters within each health region are then allocated sample quotas based on the number of households within their health region. While sampling frame proportions remain fairly consistent across cycles, certain health regions (a very small amount) exclusively use one or two of the three sampling frames. As well, the sampling frame proportion used in the 2000-2001 cycle initially resembled that of the National Population Health Survey (NPHS); during this year, the NPHS ceased cross-sectional data collection in favour of the CCHS collecting health survey data in a similar fashion. 395-401 Sampling frame proportions and health region numbers were reallocated following the initial 1.1 cycle in order to equalize the probably of sampling from all possible health regions (Table 4).

Figure 2: Example of Southern Ontario Health Region Divisions for Cycle 1.1 (2001)

Table 4: Number of Health Regions and Sample Sizes Required by CCHS Cycle Number Amount of Amount of Amount of Minimum of Sample Sample Sample Sample CCHS Cycle Health Derived Derived Derived Size Regions from Area from RDD from List Required (HRs) Frame Frame Frame 2001, Cycle 1.1 136 133,300 83% 7% 10% 2003, Cycle 2.1 133 130,000 48% 2% 50% 2005, Cycle 3.1 122 130,000 49% 1% 50% 2007, Cycle 4.1 121 65,000 49% 1% 50% 2008 121 65,000 49% 1% 50% 2009 121 65,000 49% 1% 50% 2010 117 65,000 49.5% 1% 49.5%

Within each of the individual sampling frames, different forms of stratification enable proportionate household numbers across the provinces and territories. For instance, both the telephone list and RDD frames use health region stratification, then randomly sample within each health region stratum. The area frame is stratified primarily by population density, and then by geography; median household income is secondarily taken into account. Within each stratum of the area frame, clusters are made and households are selected within each cluster. 395-401

4.2 Data Collection

CCHS data collection is comprised of numerous computer-assisted interviews (CAI); each of the required respondents in a selected household is interviewed. CAI is administered in two separate fashions, dependent upon the type of sampling frame. Computer-assisted personal interviewing (CAPI) (approximately half of all interviews) is conducted in person, and households are generally selected via the area frame. Field interviewers are trained to effectively establish initial contact with each potential dwelling. Computer-assisted telephone interviewing (CATI) (approximately half of all interviews) is conducted via the telephone, and households are selected from the telephone and RDD frames. Centralized call centers are used to train interviewers and to facilitate data collection using the CATI method. In all household dwellings that are chosen for interview (regardless of the method), a single household member is asked to submit basic demographic information for all household residents. As well, an additional individual (who could also be the original household member) is asked to take part in an in-depth, health-specific interview. An advantage of using CAI data collection methods is that interviewers are able to conduct custom interviews according to respondents’ individual characteristics, survey answers, and demographics. As well, interviewers are immediately notified if responses are considered to be out-of-range or invalid during survey completion. Respondents and interviewers must correct these inconsistencies in order to proceed with the questionnaire; thus, helping to ensure the validity of respondents’ self-report by accounting for consistent responses throughout the entire interview.

In order to minimize household nonresponse, CCHS interviewers are instructed (and properly trained concerning how) to initially send out introductory letters, bridge early contact with households prior to scheduling an interview, encourage participation from reluctant respondents, minimize refusal, cater to language barriers, offer youth interviews, and even proxy interviews.

4.2.1 Nonresponse and Data Quality

Partial nonresponse (a respondent choosing not to answer one or even several questions) has not been a major issue during CCHS data collection; however, imputation methods are used for certain modules included in the annual content in order to ensure data quality. Data is typically imputed from the ‘nearest neighbour’, based on a distance function that incorporates surrounding neighbourhood demographics for both proxy and non-proxy interviews. For modules that are strictly self-report in nature, or where imputation is not possible, partial nonresponses are coded as missing values. Missing data for any of the variables used in the analytical models were included and tested for significance through the use of an additional dummy variable.

Complete nonresponse is adjusted for with proper weighting (strategies are provided by Statistics Canada) and by increasing the desired sample size during the data collection process (ensuring the representativeness of the sample by offsetting the amount of non- responders). Without weighting, CCHS data would be subject to nonresponse bias (those who choose to respond to the survey could be considered statistically or clinically different from those who chose to not respond to the survey).

Minimizing the number of proxy interviews conducted during data collection has also ensured data quality. Across all cycles, no more than approximately 7% (on average) of all interviews were conducted using a proxy. 395-401

4.2.2 Self-Reporting Bias and Interview Mode Effects

Concerns for reporting biases inherently result from CCHS data collection methods. A pertinent limitation of the survey is the potential for self-reporting bias and interview mode effects. Certain measures that are calculated using self-reported answers involve a slight risk of under- or over-estimating the actual magnitude of the variables used. 408 Health-related estimates that have been found to be most prone to this risk are smoking status, BMI (height and weight), and certain physical activity indices. 409,410 Unfortunately, some CCHS respondents may be prone to under- or over-estimating their smoking behaviour, height, weight, and/or level of physical activity because of social desirability during in-person interviews; thus, suggesting the possibility of interview mode effects. 411

The influence of social desirability has been found to be particularly relevant for in- person interviews. 404 For instance, smoking is generally perceived to be an example of negative health behaviour; perceiving potential judgment from interviewers, certain respondents may be more likely to answer dishonestly about their smoking status when they are not provided with anonymity (such as during telephone interviews). Equally, interviewer variability may also lead to some variation in responses across interview modes.407 It has also been found that certain chronic disease diagnoses, like diabetes mellitus, tend to be under-reported by older respondents in health-related surveys. 412 However, despite the possibility (or presence) of under- or over-reporting or interview mode effects in CCHS data, the present study is modestly robust in three ways.

First, even though BMI and smoking status are included as modifiable risk factor variables in the analyses for this thesis, their relevant inclusion is based on well- documented co-variation (and are not pertinent to the exclusively comparative nature of the objectives); it is assumed that any under- or over- estimation of smoking status or behaviour, height, and weight will be consistent across the subpopulations of interest. Given the difficulty of gauging various types and levels of physical activity, in combination with the lack of measurement consistency across iterations of the CCHS, physical activity (as a risk factor variable) is excluded from the analyses in this thesis. Secondly, the magnitude of any possible self-report bias that is relative to diabetes mellitus diagnoses is considered constant across time. Essentially, randomized sampling within CCHS frames provides the same level of potential self-reporting bias across cycles (since the method of random sampling has remained consistent from CCHS cycle to CCHS cycle within the time period used for analyses). Therefore, no single year or iteration will be subject to more or less bias than any other year or iteration. Lastly, a dummy variable for interview mode was created and used in all analytic regression models (using phone interviews as the reference group). By controlling for this dummy variable, all parameter estimates reflect an average interview mode; thus, the possibility of parameters being representative of CCHS respondents who were interviewed using one method or another is minimized.

4.3 Weighting

In order to ensure that each CCHS sample is representative of the actual Canadian population, respondents must ‘represent’ several other individuals that are not in the sample. This ‘weight’ essentially corresponds to the number of persons in the target population that are represented by each individual respondent. In a simple random 2% sample of the Canadian population, the CCHS strives for each person in that sample to represent approximately 50 persons, overall. Using this terminology, it can be said that each person in the CCHS sample has a weight of 50. Effective weighting strategies account for the rather complex sampling design used by the CCHS, non-response rates, and unequal probabilities of selection across iterations. 395-401

In order to generate appropriate person-level weights for each cycle, a number of steps are completed. First, the telephone list and RDD frames are combined, leaving only 2 sampling frames to calculate weights for (Telephone and Area). The CCHS weighting strategy treats each frame independently, generating separate household-level weights for each. After household weights are created, they are combined via integration. Following integration, the weights are adjusted using person-level selection formulae; thus, creating a final person-level weight.

4.3.1 Area Frame Weight

Area frame calculations are comprised of five different levels, beginning with an initial weight that is obtained from the Labour Force Survey (LFS). 413 LFS sampling design consists of selecting household dwellings within clusters from area strata similar to those that are used in the CCHS (CCHS area frame sampling is based on LFS design). The initial weight is then adjusted in order to account for health regions, sub-sampling within clusters, and eventually stabilization. Stabilization corrects for some health regions having larger sample sizes than others, and will attenuate the sample size to a desired level by randomly sub-sampling dwellings within clusters; this produces an adjustment factor that is multiplied by the initial weight. The adjusted weight is then altered to account for the removal of out-of-scope units (households that are demolished or vacant, institutions, etc.) and for household nonresponses (the weights of these households are

52 redistributed to responding households), creating a final (household-level) area frame weight.

4.3.2 Telephone Frame Weight

Telephone frame calculations are also comprised of five different levels, beginning with an initial weight that is derived from the inverse of the probability of selection within both the telephone list and RDD frame. The value of this initial weight is essentially a ratio. For the telephone list frame: the number of sampled units, to the number of telephone numbers on the list within each health region. For the RDD frame: the ratio between the number of sampled units and one hundred times the number of random digit banks within each RDD stratum. 395-401 It is important to note that for the telephone list frame, only listed telephone numbers are used and cell phone numbers are not included. The CCHS attempts to adjust for changes in listed numbers by using an external administrative telephone list frame that is updated twice per year. 395-401

Next, the timing of CCHS data collection must be considered. Area frame households are sampled entirely at the beginning of each cycle year, and telephone frame households are sampled every two months. Therefore, each two-month telephone frame sample carries its own initial weight. Each of these weights is first reduced to ensure that the total sample is counted only once, and then multiplied by an adjustment factor. Then, like the area frame, out-of-scope numbers and household nonresponses are corrected for, producing a final (household-level) telephone frame weight.

4.3.3 Final Weight Integration

During integration, household-level weights (generated from each individual frame) are converted to person-level weights. First, person-level nonresponses are corrected for using the same method as household nonresponses, followed by “ winsorization ”. 395- 401 After multiple adjustments, some units may be generated with extreme weights; winsorization involves attenuating outlier weights downward (as a trimming approach) so that they may be calibrated for final CCHS person-level weights. The calibrated master weights are found in the data file associated with each cycle under specific variable names (WTSA_M for 2001, WTSC_M for 2003, WTSE_M for 2005, and WTS_M for

2007 and onward). Final calibrated weights are provided in order to scale estimates that are generated from the sample level, towards the population level. As a result, each of these weights were rescaled (by a factor of αi=1/k) prior to computing any analyses; individual sampling weights were divided by mean sampling weights for each CCHS iteration, then combined for the total sample (thus, ensuring that the total weighted sample was equal to the actual total sample from 2001 to 2010). All analyses were conducted using these rescaled weights.

4.4 Combining CCHS Cycles

When investigating a rare population or disease, a single cycle of the CCHS may not provide enough data to generate the sample size or disease cases required for valid and reliable parameter estimates. Statistics Canada states “combining cycles yields larger sample sizes for analysis, and the resulting estimates are of higher quality than those from one cycle alone.”414 For the present study, this was indeed the case. Multiple cycles were required in order attain a sufficient sample for each subpopulation, and subsequent stratums within each covariate of interest. Nevertheless, certain aspects of the survey, its data collection, and the parameter estimates that have been generated must be taken into consideration. Essentially, changes to the survey and its administration over time (questionnaire content, geographic coverage, etc.) must be acknowledged and accounted for. Fortunately, all variables included in this thesis are taken strictly from annual modules that have remained unchanged in terms of content wording and coverage across all iterations used. Any potential for biased responses due to changes in interview mode (method of data collection) have been acknowledged in Section 4.2.2 .

A key concern for researchers when combining CCHS cycles is that the Canadian population is naturally evolving over time. Parameter estimates that are computed based on a combined sample are therefore not directly representative of the Canadian population at each individual cycle year; rather, they are representative of an average population garnered over the period of time used for the study. 415 Statistics Canada warns that changes in the population occurring between cycles may actually reflect variations in the outcome variables (parameters) used in one’s study. Statistics Canada also notes that interpretations of time trends are slightly obscured when combining multiple cycles;

54 single year estimates must be considered as part of an overall period estimate when describing parameters. 411 Nevertheless, certain methodological approaches are offered in order to account for any discrepancies that may emerge in the population from year to year. For instance, by incorporating a distinct ‘time’ variable into trend analyses, regression models may be used to assess fluctuations in parameter estimates regardless of the contextual change that may have occurred. Statistics Canada recommends that if the time variable is found to be significant, it may be controlled for; thus, creating comparable trend results across multiple cycles.

Two techniques are presented when combining multiple CCHS cycles: the “separate” approach, and the “pooled” approach. 410 The separate approach involves conducting analyses and generating estimates for each individual cycle independently, followed by combination. The pooled approach differs in that the combination step is done prior to any analyses being conducted. The combined data set is then treated as a new total sample that may be analyzed. Both methods are considered to be statistically valid. Ultimately, the choice between the two techniques depends on the objectives of the study. The two approaches do not always produce identical parameter estimates, however. The separate approach is considered to be a simple average of two ratios (a/b and c/d, for instance), while the pooled approach generates period estimates ((a/b + c/d) ≠ (a + c)/(b + d)). 410 Statistics Canada states “using a Canadian estimate as an example, some researchers may choose to study the average of provincial estimates, which gives equal weight to each province (separate approach), while others are interested in the national estimate (pooled approach), which is influenced more by larger provinces.” Essentially, the separate approach provides researchers with an average of estimates using multiple CCHS cycle iterations. These average estimates can be time consuming to compute and often quite difficult to interpret. Given the aim of the present study (to generate national level estimates) and the planned analytical framework, the pooled approach was employed.

A large advantage of using the pooled approach is that it lends itself very well to complex and/or detailed regression models. 401 When regression parameters (meaningful proportions or probabilities) are sought as nationally representative estimates of the outcome measure, the pooled approach has been shown to be the more appropriate

55 alternative. 389 The primary goal of this study was to describe nationally representative differences that are observed between large subpopulations that are spread out across Canada, and in variable numbers (depending on the province or territory). Inline with the pooled approach example provided in the previous paragraph, this technique was much more conducive to this goal and subsequent framework.

Pooling methodology not only provides an increase in statistical power, but it also allows for simpler interpretation of the parameter estimates. By rescaling master weights (in addition to the inclusion of a time variable), resultant odds ratios, probability proportions, and distribution frequencies may be described as meaningful period estimates as well as year-specific trend statistics.

4.5 CCHS Data Access

Following the completion of a confidentially oath (required by Statistics Canada), all analyses were computed using master data files and screened by an appointed analyst in the Research Data Centre located at the University of Western Ontario.

4.6 Outcome Variable and Total Sample

In order to effectively assess the probability of diabetes mellitus among all CCHS respondents according to the subpopulations of interest, individual cycles (2001, 2003, 2005, 2007, 2008, 2009, and 2010) were pooled, generating a total combined sample of N = 656,884. This extremely large sample provided more than enough power to generate (valid) odds for the disease (from 2001 to 2010).

The dependent (outcome) variable for the present study was simply whether or not a CCHS respondent had indicated that they have diabetes at the time of survey completion. When completing the CCHS annual content questionnaire, respondents answered the following question related to diabetes mellitus: 416 ([Do/Does] [you/FNAME] have:) Diabetes? Yes or No . The variable was therefore binary, and coded as 0 = No Diabetes, and 1 = Has Diabetes. A total of n = 41, 807 diabetes mellitus cases were found within the total combined sample, and n = 40, 596 diabetes cases were found excluding gestational diabetes (n = 1211). Approximately n = 7, 031 respondents with diabetes

56 mellitus did not provide ethnic/racial/cultural origin or citizenship status data, leaving n = 33, 565 diabetes mellitus cases within the total sample used for the analyses.

Unfortunately, the CCHS does not make the distinction between whether or not a respondent has T1DM or T2DM. As well, the CCHS uses non-specific diagnosis during pregnancy as a proxy for determining whether or not a respondent’s diabetes mellitus was explicitly gestational. This eliminated the possibility of calculating exact frequencies for either T1DM or T2DM within the sample; however, the number of T2DM cases will be inferred based on previous statistics concerning its proportion in population (approximately 90%).1 Moreover, gestational diabetes mellitus was effectively controlled for using the provided CCHS proxy question.

4.7 Conceptual Framework

Various conceptual frameworks have been proposed in the literature relative to the outcome of diabetes mellitus, especially within the Canadian context. Previous research has pinpointed a number of influential factors that contribute to an increased risk for T2DM (several modifiable and non-modifiable risk factors, as detailed previously in this thesis). In order to effectively formulate a conceptual framework that encompassed each area of the present study’s objectives, as well as the current literature, diverse models and studies were taken into consideration. Much of this section provides a brief elaboration of the rationale for pertinent factors derived from the literature review (Chapter 2), and detailed in the study objectives (Chapter 3).

Previous research into T2DM risks has been inclusive of both modifiable and non- modifiable risk factors. 41 Risk score models that have monitored this dichotomy frequently incorporate elements of both risk areas when examining T2DM. 42,59-64,148,171-172 In addition to these considerations, immigrant-specific literature has detailed the importance of acculturation (time), and racial/cultural/ethnic origin,264,417 while Aboriginal-specific literature highlights the distinctive influences that are understood in reserve communities. By using CCHS data, however, the present study’s framework is limited to analyses of off-reserve Canadian-Aboriginals (First Nations, Inuit, and Métis). Still, according to Statistics Canada approximately 60% of First Nations people lived off-

57 reserve in 2006 (increasing from 58% in 1996); 418 therefore, it is assumed that analyses of this population using the CCHS are still fairly representative.

Relative to immigrants, the variation in immigrant country origins across the study period (referenced in Table 2) must also be taken into consideration. Over time (from 2001 to 2010), the picture of immigration has changed. In order to assess the relative changes in immigrant country origins among CCHS participants across time, frequency distributions for immigrants with diabetes across the survey iterations are provided (Appendix A).

Two particular Canadian studies have provided useful suggestions for formulating the present study’s conceptual model, especially concerning immigrant-specific factors for inclusion. Creatore et al. modeled their investigation of diabetes mellitus in Ontario immigrants by combining demographic variables, immigrant-specific variables (time in- country), modifiable risk factor variables (physical activity level, BMI, etc.), and non- modifiable risk factor variables (specifically, world region of birth) from provincial administrative health databases.321 Similarly, Liu et al. also examined several of these influences (excluding immigrant-specific factors, however) in their investigation of cardiovascular risk factors by ethnicity (at the national level) using three cycles of the CCHS (2001, 2003, and 2005); nevertheless, these researchers chose to exclude comparisons of specific ethnic origins. 419

The present study attempts to combine these previous models by incorporating comparisons of racial/cultural and ethnic origins, country origins, immigrant-specific factors, and variables that have been cumulatively designed to assess diabetes mellitus risks, as well as differences between subpopulations. Variables included in the conceptual framework were organized into two main groups according to objective two , as described in Chapter 3: modifiable and non-modifiable (inclusive of immigrant-specific health determinants) risk factor variables (see Figure 3). All covariates provided an appropriate structure for the descriptive frequencies and regression analyses computed in this thesis, and also provided a framework for the comparative discussions detailed in Chapter 6. Specific coding for each variable is found in Table 5.

MODIFIABLE RISK FACTORS: NON - Smoking Status MODIFIABLE RISK FACTORS: Alcohol Consumption Frequency Age BMI Sex Diet Racial/Cultural Origin Household Income ( SES ) Ethnicity (Ethnic Personal Income ( SES ) Origin) Household Education ( SES ) Time Since Immigration Personal Education ( SES ) Age at Immigration National Language Competency (Conversational) Country Origin

DIABETES MELLITUS (T2DM)

Figur e 3: Conceptual Framework

Table 5: Variable Coding

VARIABLE CODING FOR ANALYSES

MODIFIABLE RISK FACTORS Never Smoked Daily Smoker Former Daily Smoker Smoking Status Occasional Smoker Always an Occasional Smoker Former Occasional Smoker Less Than Once a Month Once a Month Alcohol 2 to 3 Times a Month Consumption Once a Week Frequency 2 to 3 Times a Week 4 to 6 times a Week Everyday Normal Weight Body Mass Underweight Index Overweight Obese No income Less than $5000 $5000 to $9999 $10000 to $14000 $15000 to $19999 Household $20000 to $29999 Income * $30000 to $39999 $40000 to $49999 $50000 to $59999 $60000 to $79999 $80000 or More No income Less than $5000 $5000 to $9999 $10000 to $14000 $15000 to $19999 Personal $20000 to $29999 Income $30000 to $39999 $40000 to $49999 $50000 to $59999 $60000 to $79999 $80000 or More Household Less than Secondary School Education

Education * Secondary School Graduation Some Post-Secondary Education Post-Secondary Education Less than Secondary School Education Personal Secondary School Graduation Education Some Post-Secondary Education Post-Secondary Education Diet (Meets full Less than 5 Servings per Day servings of 5 to 10 Servings per Day Fruits and 10 or More Servings per Day Vegetables) English Only Neither English or French (Other) National French Only Language English and French Only Competency English and French and Other (Conversational) English and Other (Not French) French and Other (Not English) NON-MODIFIABLE FACTORS 12-19 years old 20-29 years old 30-39 years old 40-49 years old Age 50-59 years old 60-69 years old 70-79 years old 80 years old and above Male Sex Female White Black Korean Filipinos Japanese Chinese Racial/Cultural South Asians Origin Southeast Asians Arabs West Asians Latin Americans Other Racial Origins Multiple Racial Origins Canadian Ethnicity French (Ethnic Origin)* English German

Scottish Irish Italian Ukrainian Dutch (Netherlands) Chinese Jewish Polish Portuguese South Asian Aboriginal Non-Aboriginal (non-Immigrant) Citizen Status Aboriginal (First Nations, Inuit, and Métis) Citizen Other North America (United States) South, Central America and Caribbean Europe Country Origin Africa Asia Oceania IMMIGRANT-SPECIFIC RISK FACTORS Immigration Non-Immigrant (non-Aboriginal) Citizen Status Immigrant Citizen 0-9 years 10-19 years Time Since 20-29 years Immigration 30-49 years 40 years or more 0-19 years old 20-29 years old Age at 30-39 years old Immigration 40-49 years old 50-69 years old 70 years old or above Note s: (1) Reference groups are bolded for each variable. (2) Given the luxury of a large sample size, variables noted with an asterisk (*) were included to capture similar, but separate domains of risk factor variables encompassed in the conceptual framework. (a) All regression analyses controlled for the linear effect of time.

4.8 Statistical Analyses

All statistical analyses were carried out using SAS version 9.3, and were conducted at the University of Western Ontario’s Research Data Centre. In order to ensure that the computed parameter estimates and trends were appropriate for discussion at the national level (inclusive of all provinces and territories), variables were limited specifically to the data collected from the annual component questionnaire (as mentioned previously), the

CCHS datasets (from 2001 to 2010) were pooled, and all analyses were weighted appropriately (accounting for the unequal probability of selection inherent in the CCHS sampling design). Statistical significance was determined for all (logistic) regressions at an alpha ( α) type 1 error probability rate of 0.05, and parameter estimates were interpreted as crude and adjusted odds ratios.

4.8.1 Objective One

Firstly, the total sample of Canadians with diabetes from 2001 to 2010 was calculated using cross tabulations of the diabetes outcome variable and the time variable. Next, in order to determine the number of diabetes cases by each subpopulation, the sample was stratified into the desired groups. In order to stratify the sample, two binary variables were created. For immigrant status, the original CCHS variable was recoded to encompass the entire weighted portion of the sample, excluding Canadian-Aboriginals (First Nations, Inuit, and Métis). To account for this exclusion, non-immigrant/non- Aboriginals were coded as the reference group for this variable, and Canadian- Aboriginals were re-coded into a separate binary variable.

Prior to the 2005 cycle of the CCHS, Canadian-Aboriginals were able to select their descendent origin as an option from the racial/cultural origin question: “ People living in Canada come from many different cultural and racial backgrounds. Are you:…?” 395-397 However, since the 2005 cycle iteration, Canadian-Aboriginals have been identified by a specific, separate question: “ Are you an Aboriginal person that is First Nations, Inuit, and Métis? A binary variable for Aboriginal status was created and coded to encompass not only the entire weighted portion of the sample that excluded Canadian-immigrants, but also to adjust for the changing method of identification. The creation of these recoded binary variables allowed for Canada-born, non-immigrant, non-Aboriginal citizens to be used as the reference group for frequency calculations and regression analyses concerning both Canadian-immigrants, and Canadian-Aboriginals. Sensitivity analyses were conducted using all three subpopulations as the reference group; these options yielded nearly identical statistical results. Stratified sample totals and diabetes cases were then calculated for each subpopulation by using cross tabulations of the diabetes outcome variable and the time variable. Period prevalence rates were found by dividing the total

63 number of diabetes cases within each subpopulation (from 2001 to 2010) by the associated total frequency (of each respective subpopulation).

In order to characterize the distribution of modifiable and non-modifiable risk factor variables for the total sample of Canadians with diabetes from 2001 to 2010, cross tabulations of the diabetes outcome variable and each risk factor variable were computed. All variables used for the analyses were categorical; age, time since immigration, and age at immigration were initially continuous but were recoded categorically in order to capture a great deal of data spread, and to better depict the effect of older age on diabetes outcomes. Period prevalence rates were calculated for each category of every variable, depicting the relative distribution of characteristics among the total (pooled) diabetic sample. Period prevalence rates were found by dividing the number of diabetes cases within each single category (of all variables) by the associated total frequency (from the total sample).

Period time trends for Canadian-immigrants, Canadian-Aboriginals (First Nations, Inuit, and Métis), and Canada-born (non-immigrant, non-Aboriginal population) individuals from 2001 to 2010 were calculated by dividing the number of diabetes cases within each subpopulation by the total subpopulation size per CCHS cycle year; these individual prevalence rates were then depicted collectively over time (Figure 4). A period time trend for all Canadians was calculated in the same manner using the total number of diabetes cases and the total sample size per CCHS cycle year.

4.8.2 Objective Two

In order to assess the differential impact of modifiable and non-modifiable risk factors for each subpopulation, bivariate logistic regressions (providing crude results) were conducted for each variable, and for each subpopulation. Multivariable logistic regressions (providing adjusting results) were then conducted for each subpopulation, inclusive of all risk factors. For immigrants, additional variables were included (for both bivariate and multivariable regressions) in order to account for immigrant-specific health determinants such as time since immigration, age at immigration, and country origin. As stated previously, statistical significance was determined for all regressions at an alpha

(α) type 1 error probability rate of 0.05, and parameter estimates were interpreted as crude and adjusted odds ratios.

Chapter 5 5 Results

This chapter begins with a description of the overall diabetes mellitus sample characteristics (from 2001 to 2010), focusing on the relative distributions of modifiable risk factors, non-modifiable risk factors, and immigrant-specific risk factor variables (Section 5.1). The diabetes period prevalence rates (as well as diabetes cases by CCHS survey year) among all Canadians (ages 12 and above), Canadian-immigrants, Canadian- Aboriginals (First Nations, Inuit, and Métis) and Canada-born (non-immigrant, non- Aboriginal) individuals from 2001 to 2010 will then be presented. Prevalence rate time trends will then be depicted graphically in order to assess the differences associated with each subpopulation from 2001 to 2010. The comparative impact of each risk factor for diabetes among Canadian-immigrants, Canadian-Aboriginals (First Nations, Inuit, and Métis) and Canada-born (non-immigrant, non-Aboriginal) individuals from 2001 to 2010 will then be described using the crude bivariate (Section 5.3, Table 8) and adjusted multivariable logistic regression results (Section 5.4, Table 9).

5.1 Diabetes Mellitus Sample Characteristics

The following data descriptions pertain specifically to the period prevalence rates within variable categories (for diabetics), as well as overall period prevalence rates (for the total sample, inclusive of both diabetics and non-diabetics); these results are found in Table 6. A distribution of variables that is broken down by CCHS survey year (2001 to 2010) is provided in the Appendix (Appendix 1).

To begin, diabetics most frequently indicated that they were either non-smokers (31.8%) or former daily smokers (38.2%); yet, diabetic non-smokers comprised only 4.1% of all non-smokers, while diabetic former daily smokers comprised 8.5% of all former daily smokers.

Moderate alcohol consumption from as low as once per month to as high as four-to-six times per week ranged from 2.7% to 8.6% among diabetics, while low alcohol

66 consumption of less than once per month was considerably more prevalent at 21.1%. Those with diabetes who consumed alcohol less than once per month comprised 6.3% of all individuals who reportedly consumed the same amount, while those with diabetes who consumed alcohol everyday comprised 5.3% of all individuals who consumed alcohol in the same frequency; any reported alcohol consumption of more than ‘less than once per month’ and ‘less than everyday’ ranged from of 2.9% to 4.5% period prevalence rates within the associated variable categories (see Table 6).

Among those who provided their height and weight, calculated BMIs in the normal (20.8%), overweight (30.0%), and obese (36%) categories comprised the majority of the diabetic sample; these figures represented 2.7%, 5.9%, and 11.4% of their associated variable categories (period prevalence rates).

Concerning both household and personal income, individuals with diabetes who reported earnings for either variable between $5,000 and $9,999 upwards of $20,000 and $29,999 had the highest period prevalence rates among their respective earnings categories (ranging from 8.4% to 9.8% for household income, and 5.9% to 8.6% for personal income, respectively).

Approximately 55.2% of diabetics indicated that the highest level of education obtained by an individual in their household was post-secondary, while 41.3% indicated that they had personally completed post-secondary education. However, each of these variable categories comprised only 4.2% and 4.3% of their respective education category. Conversely, 19.2% of diabetics had indicated that the highest level of education obtained by an individual in their household was less than secondary school, while 33.9% had indicated that they had personally not completed secondary school; each comprised 11.7% and 7.1% of their respective education categories (the highest period prevalence rates between the two variables).

Among those with diabetes, 50% consumed less than 5 servings per day of fruits and vegetables, 32.1% consumed 5-10 servings per day, and 2.8% consumed more than 10 servings per day; these constituted 5.0%, 5.0%, and 3.8% of their associated diet categories, respectively (period prevalence rates).

The majority of diabetics indicated that they were only able to converse in English (47.6%), while 3.8% reported that were not able to converse in English or in French at all (3.8%); however, English-only speaking diabetics constituted 5.2% of all English-only speaking participants while diabetics who spoke neither English or French comprised 6.9% of all participants that could not converse in either language.

The majority of the diabetic sample were aged 50 to 59 (23.5%), 60 to 69 (26.4%) and 70 to 79 (21%); the age categories with the largest period prevalence rates were those aged 50 to 59 (7.6%), 60 to 69 (13.0%), 70 to 79 (15.7%), and 80 or Above (14.3%). Males and females comprised 54% and 46% of all diabetes cases, with period prevalence rates of 5.6% and 4.6%, respectively.

The vast majority of the diabetics from 2001 to 2010 were White (79.1%); this accounted for 5.1% of all Whites in the total sample. South Asians (7.2%), Latin Americans (6.3%), and Japanese (6.12%) individuals were found to have the highest representative period prevalence rates of diabetes in their respective racial/origin category. Inversely, individuals who were Arab (2.9%) or Chinese (3.6%) had the lowest representative period prevalence rates of diabetes. Concerning ethnic origin, South Asians were (again) found to have the highest prevalence rate of diabetes when compared to any other ethnic origin category (7.0%). Polish (3.7%) and Chinese (3.8%) individuals had the lowest period prevalence rates of diabetes.

Most diabetics who provided country origin data were originally from Europe (38.2%) or Asia (28.6%). Individuals from South, Central America, the Caribbean (7.4%), and Europe (7.2%) had the highest period prevalence rates among immigrants.

Immigrants who had been in Canada from 0 to 9 years had remarkably lower rates of diabetes (2.3%) when compared to those who had been in Canada for 30 to 39 years (10.0%) or 40 or more years (11.3%). Conversely, those who were 12 to 19 years of age upon arrival in Canada had the lowest period prevalence rate of diabetes (3.6%) among immigrants. Those who immigrated to Canada at 50 to 69 years of age (16.4%) or 70 and Above years of age (16%) had the highest period prevalence rates of diabetes among immigrants.

Table 6: Distribution of Variables for the Total Sample (2001 to 2010 Inclusive) Period Number of Prevalence Diabetes Cases (number of (Percentage of Total Sample diabetes cases Variable divided by the Column Total) total sample ) 1. MODIFIABLE RISK FACTOR VARIABLES Smoke Status Is Not a Smoker 10688 258783 4.1% (31.8%) Occasional Smoker 4088 96968 4.2% (12.3%) Former Daily Smoker 12826 150232 8.5% (38.2%) Always an Occasional Smoker 239 13120 1.8% (0.7%) Former Occasional Smoker 622 18565 3.4% (1.9%) Daily Smoker 4895 116136 4.2% (14.6%) Not Available/Applicable 205 3080 6.7% (0.6%)

Alcohol Consumption Frequency Less than Once per Month 7089 112278 6.3% (21.1%) Once per Month 2496 55021 4.5% (7.4%) Two to Three Times per 2247 69940 3.2% Month (6.7%) Once per Week 2811 89419 3.1% (8.4%) Two to Three Times per 2885 99156 2.9% Week (8.6%) Four to Six Times per Week 906 28474 3.2% (2.7%) Everyday 2363 44262 5.3% (7.0%) Not Available/Applicable 12766 158335 8.1% (38.0%)

Body Mass Index Underweight 322 19839 1.6% (1.0%) Normal 6988 254610 2.7% (20.8%) Overweight 10058 170208 5.9% (30.0%) Obese 12074 105653 11.4% (36.0%) Not Available/Applicable 4122 106573 3.9% (12.3%)

Household Income No Income 71 1715 3.9% (0.2%) Less than $5,000 167 3195 5.2% (0.5%) $5,000 to $9,999 703 8347 8.4% (2.1%) $10,000 to $14,999 2129 21801 9.8% (6.3%) $15,000 to $19,999 2168 22224 9.8% (6.5%) $20,000 to $29,999 4517 51994 8.7% (13.5%) $30,000 to $39,999 3721 57095 6.5% (11.1%) $40,000 to $49,999 2892 54226 5.3% (8.6%) $50,000 to $59,999 2463 53908 4.6% (7.3%) $60,000 to $80,000 3252 85355 3.8% (9.7%) $80,000 and Above 2859 160271 2.9% (8.5%) Not Available/Applicable 8625 199236 4.3% (25.7%)

Personal Income No Income 1243 30001 2.9% (3.7%) Less than $5,000 864 31310 2.8% (2.6%) $5,000 to $9,999 2698 40329 6.7%

(8.1%) $10,000 to $14,999 4499 52473 8.6% (13.5%) $15,000 to $19,999 3076 40728 7.6% (9.3%) $20,000 to $29,999 4627 78949 5.9% (13.9%) $30,000 to $39,999 3607 73205 4.9% (10.9%) $40,000 to $49,999 2285 54893 4.2% (6.9%) $50,000 to $59,999 1736 40227 4.3% (5.2%) $60,000 to $80,000 1543 43906 3.5% (4.6%) $80,000 and Above 1291 39147 3.3% (3.9%) Not Available/Applicable 5764 123328 4.7% (17.3%)

Household Education Less than Secondary School 6458 55328 11.7% Education (19.2%) Secondary School Graduation 4099 71089 5.8% (12.2%) Some Post-Secondary 1805 38450 4.7% Education (5.4%) Post-Secondary Education 18529 445045 4.2% (55.2%) Not Available/Applicable 2674 46971 5.6% (8.0%)

Personal Education Less than Secondary School 11386 160248 7.1% Education (33.9%) Secondary School Graduation 5126 106989 4.8% (15.3%) Some Post-Secondary 1935 52929 3.7% Education (5.8%) Post-Secondary Education 13846 321543 4.3% (41.3%)

Not Available/Applicable 1271 15174 8.4% (3.8%)

Dietary Consumption of Fruit and Vegetables Consumes Less than 5 16766 336736 5.0% Servings per Day (50.0%) Consumes 5 to 10 Servings 10775 217501 5.0% per Day (32.1%) Consumes More than 10 933 24310 3.8% Servings per Day (2.8%) Not Available/Applicable 5091 78337 6.5% (15.2%)

National Language Competency (Conversational) English Only 15962 306169 5.2% (47.6%) French Only 4010 70311 5.7% (11.9%) English and French Only 4123 107951 3.8% (12.3%) English and French and 1281 34777 3.8% Other (3.8%) English and Other (Not 5646 106742 3.7% French) (16.8%) French and Other (Not 306 4416 5.3% English) (0.9%) Neither English or French 1272 12848 6.9% (Other) (3.8%) Not Available/Applicable 965 13670 7.1% (2.9%)

2. NON-MODIFIABLE RISK FACTOR VARIABLES Age 12 to 19 282 80381 0.4% (0.8%) 20 to 29 748 104975 0.7%

(2.2%) 30 to 39 1697 108978 1.6% (5.1%) 40 to 49 4050 125712 3.2% (12.1%) 50 to 59 7873 103001 7.6% (23.5%) 60 to 69 8866 68028 13.0% (26.4%) 70 to 79 7054 44815 15.7% (21.0%) 80 and Above 2994 20994 14.3% (8.9%) Not Available/Applicable N/A N/A N/A

Sex Male 18121 323700 5.6% (54.0%) Female 15443 333184 4.6% (46.0%) Not Available/Applicable N/A N/A N/A

Racial/Cultural Origin White 26561 529761 5.0% (79.1%) Black 766 12766 6.0% (2.3%) Korean 127 2444 5.2% (0.4%) Filipino 354 7178 4.9% (1.1%) Japanese 78 1275 6.1% (0.2%) Chinese 791 21775 3.6% (2.4%) South Asian 1457 20181 7.2% (4.3%) Southeast Asian 256 5330 4.8% (0.8%) Arab 133 4554 2.9% (0.4%) West Asian 129 2920 3.8% (0.4%) Latin American 240 6329 6.3% (0.7%)

Other 421 6690 4.4% (1.3%) Multiple 250 5673 4.4% (0.7%) Not Available/Applicable 2001 30007 6.7% (6.0%)

Ethnicity (Ethnic Origin) Canadian 6988 148378 4.7% (20.6%) French 4400 89396 5.5% (13.0%) English 6321 114919 4.3% (18.6%) German 2240 51834 4.3% (6.6%) Scottish 4037 82193 4.9% (11.9%) Irish 3481 71242 4.9% (10.3%) Italian 1446 24936 5.8% (4.3%) Ukrainian 841 20564 4.1% (2.5%) Dutch (Netherlands) 782 17662 4.4% (2.3%) Chinese 872 22994 3.8% (2.6%) Jewish 257 4643 5.5% (0.8%) Polish 556 15108 3.7% (1.6%) Portuguese 310 7040 4.4% (0.9%) South Asian 1404 20005 7.0% (4.1%) Not Available/Applicable 34030 a

3. IMMIGRANT-SPECIFIC RISK FACTOR VARIABLES Country Origin Other North America 362 8001 4.5% (3.6%) South, Central America and 1129 15241 7.4% the Caribbean (11.2%)

Europe 3846 53555 7.2% (38.2%) Africa 440 8213 5.4% (4.4%) Asia 2877 51315 5.6% (28.6%) Oceania 71 1105 6.4% (0.7%) Not Available/Applicable 1338 20219 6.6% (13.3%)

Time Since Immigration to Canada 0 to 9 Years Since 904 38754 2.3% Immigration (2.7%) 10 to 19 Years Since 1466 31605 4.6% Immigration (4.4%) 20 to 29 Years Since 1341 20533 6.5% Immigration (4.0%) 30 to 39 Years Since 2022 20289 10.0% Immigration (6.0%) 40 or More Years Since 3064 27016 11.3% Immigration (9.1%) Not Available/Applicable 24767 518687 4.8% (73.8%)

Age at Immigration to Canada 12 to 19 1804 49835 3.6% (5.4%) 20 to 29 2952 44358 6.7% (8.8%) 30 to 39 2035 27444 7.4% (6.1%) 40 to 49 999 10394 9.6% (3.0%) 50 to 69 915 5591 16.4% (2.7%) 70 and Above 92 575 16.0% (0.3%) Not Available/Applicable 24651 518867 4.8%

(73.8%) Notes : (1) All results were calculated using rescaled master weights (see Section 4.3.3 ); (2) aMultiple options available; number represents total respondents with more than one ethnicity option selected.

5.2 Period Prevalence of Diabetes Mellitus (2001 to 2010)

Overall, the period prevalence rate of diabetes for all Canadians from 2001 to 2010 was 5.3%. The period prevalence of diabetes was found to be lowest among Canada-born (non-immigrant, non-Aboriginal) individuals (5.0%), followed by Aboriginals (First Nations, Inuit, and Métis) (6.5%). Immigrants were found to have the highest period prevalence rate of diabetes among the three Canadian subpopulations at 6.7% (see Table 7).

The prevalence of diabetes among all Canadians ranged from 4.1% in 2001 (2000/2001 inclusive) to 6.4% in 2010 (2009/2010 inclusive).

The prevalence of diabetes in Canada-born individuals ranged from 3.9% in 2001 (2000/2001 inclusive) to 5.7% in 2010 (2009/2010 inclusive).

The prevalence of diabetes in Aboriginal individuals ranged from 5.4% in 2001 (2000/2001 inclusive) to 7.4% in 2010 (2009/2010 inclusive).

The prevalence of diabetes in immigrant individuals ranged from 5.0% in 2001 (2000/2001 inclusive) to 8.5% in 2010 (2009/2010 inclusive).

Overall, there were generally upward linear trends for diabetes among all subpopulations (see Figure 4). Both Aboriginals and immigrants remained consistently above the linear trend for Canada-born individuals from 2001 to 2010, while Canada-born individuals remained consistently below the linear trend for all Canadians. Immigrants had the highest prevalence rate of diabetes almost consistently across time from 2001 to 2010.

Table 7: Diabetes Mellitus Period Prevalence Rates Diabetes Cases by CCHS Survey Year Total Diabetes Total 2001 2003 2005 2007 2008 2009 2010 Cases Sample (Period Prev.) Canada- Born (Non- 23749 Immigrant, 4042 4428 4759 2672 2621 2552 2675 479659 (5.0%) Non- Aboriginal) 8797 Immigrant 1339 1573 1507 993 1123 1033 1229 129295 (6.7%) Aboriginal (First 910 Nations, 74 107 154 148 131 149 146 14164 (6.5%) Inuit, and Métis) 33565 Total 5423 6238 6493 3825 3876 3682 4028 656354 (5.3%) Notes : (1) All results were calculated using rescaled master weights (see Section 4.3.3 ); (2) Overall response rates (%) varied slightly across the years: 84.7, 80.7, 79.0, 77.6, 75.2, 73.2, 71.5 (from 2001 to 2010, respectively).

Figure 4: Diabetes Time Trends from 2001 to 2010

5.3 Comparative Impact of Diabetes Risk Factors - Crude Results

Statistical tests for differences between subpopulations according to each risk factor were not conducted in order to avoid multiple comparisons/testing across the groups of main interest. Instead, odds ratios (likelihood of diabetes) were generated for each subpopulation according to the outlined set of variables found in Table 5 (crude bivariate logistic regression results are found in Table 8).

5.3.1 Smoking Status

Smoking was found to be strongly associated with diabetes, but in varying degrees among the subpopulations of interest.

Crude results revealed that Canada-born individuals who were former-daily smokers were most likely to have diabetes when compared to Canada-born non-smokers (OR 2.58, 95% CI 2.50, 2.67).

Aboriginals who were daily smokers (OR 1.6, 95% CI 1.32-1.93), former daily smokers (OR 2.67, 95% CI 2.19-3.24), and former occasional smokers (OR 1.76, 95% CI 1.2- 2.48) were most likely to have diabetes when compared to Aboriginal non-smokers. Conversely, immigrants who were daily smokers (OR 0.84, 95% CI 0.78-0.91) or former occasional smokers (OR 0.54, 95% CI 0.44-0.66) were less likely to have diabetes when compared to immigrant non-smokers.

5.3.2 Alcohol Consumption Frequency

Alcohol consumption of once per month or more was not found to be significantly associated with diabetes among the subpopulations of interest. When compared to those who consumed alcohol less than once per month, all subpopulations were less likely to have diabetes at every increased level of alcohol consumption.

5.3.3 Body Mass Index

BMI was found to be strongly associated with diabetes in all three subpopulations. Overall, the likelihood of diabetes increased with BMIs higher than normal. On the other hand, the likelihood of diabetes decreased with BMIs lower than normal in the Canada- born and immigrant subpopulations.

5.3.4 Household and Personal Income

Household and Personal income were found to be fairly associated with diabetes. Crude results indicated that all three subpopulations were most likely to have diabetes if their reported household or personal income ranged from $5,000 to $29,999 when compared to individuals who reported household or personal income above $80,000.

5.3.5 Household and Personal Education

Household and Personal education were found to be slightly associated with diabetes in all three subpopulations. Overall, as education level increased (for either variable), the likelihood of diabetes decreased.

5.3.6 Diet

Dietary consumption of fruit and vegetables was not strongly associated with diabetes; however, significant results were found in the Canada-born and immigrant subpopulations. Crude results indicated that both Canada-born and immigrant individuals who consumed less than 5, or 5 to 10 servings of fruits and vegetables per day were marginally more likely to have diabetes than those who consumed more than 10 servings per day.

5.3.7 National Language Competency

National language competency (conversational) was found to be slightly associated with diabetes in certain subpopulations. When compared to individuals who were only able to converse in English, crude results indicated that Canada-born (OR 1.87, 95% CI 1.12- 3.14) and immigrant (OR 2.96, 95% CI 1.61-5.44) individuals who could not speak either English or French but could converse in another language were more likely to have

81 diabetes. Crude results also indicated that Aboriginal individuals who spoke English and another language other than French (OR 2.05, 95% CI 1.73-2.43), as well as Aboriginals who spoke French and another language other than English (OR 4.22 95% CI 1.71-10.43) were considerably more likely to have diabetes when compared to Aboriginals those who only spoke English.

5.3.8 Age

Age was found to be robustly associated with diabetes in all three subpopulations. The likelihood of diabetes significantly increased with older age; this was especially true for the immigrant subpopulation.

5.3.9 Sex

Sex was found to be strongly associated with diabetes in the Canada-born and immigrant subpopulations. Crude results showed that Canada-born and immigrant females were less likely to have diabetes when compared to males, respectively. However, while not statistically significant, Aboriginal females were found to be slightly more likely than Aboriginal males to have diabetes.

5.3.10 Racial/Cultural and Ethnic Origin

Racial/cultural origin was strongly associated with diabetes in the Canada-born and immigrant subpopulations.

For the Canada-born subpopulation, those who were Japanese, and those who did not provide racial/cultural origin data were most likely to have diabetes when compared to White Canada-born individuals.

For the immigrant subpopulation, Black immigrants, and immigrants with ‘Other’ racial/cultural origins were most likely to have diabetes when compared to White immigrants. Latin American immigrants were least likely to have diabetes when compared to White immigrants.

Concerning ethnic origin, Italian and South Asian immigrants had the highest likelihood of diabetes when compared to those who ethnically identified as Canadian. Conversely,

Portuguese Canada-born individuals had the smallest likelihood of diabetes when compared to those who identified ethnically as Canadian.

5.3.11 Immigrant-Specific Risk Factors

Country origin, time since immigration, and age at immigration were all found to be strongly associated with diabetes in the immigrant subpopulation.

Crude results showed that immigrant respondents from any country origin outside of North America had greater odds of reporting diabetes. Immigrants from South, Central America, and the Caribbean (OR 1.77, 95% CI 1.55-2.01), as well as Europe (OR 1.72, 95% CI 1.53, 1.94) had the highest likelihoods.

Crude results revealed that longer periods of time in Canada were linked to higher likelihoods for diabetes. When compared to immigrants who had been in Canada for 0 to 9 years, immigrants who had been in Canada for 10 years and above were more likely to have diabetes; this likelihood increased exponentially with lengthier periods of time in Canada.

Overall, when compared to immigrants who arrived in Canada at ages 70 and above, immigrants who arrived at younger ages had noticeably decreased likelihoods of having diabetes.

Table 8: Results from Bivariate Logistic Regression Models (Crude) Canada-Born (Non- Aboriginal (First Immigrant, Non- Immigrant likelihood of Nations, Inuit, and Aboriginal) likelihood of Diabetes Métis) likelihood of Diabetes Diabetes Odds Ratios Odds Ratios Odds Ratios (95% (95% (95% P-Value P-Value P-Value Confidence Confidence Confidence Intervals) Intervals) Intervals) 1. MODIFIABLE RISK FACTOR VARIABLES Smoke Status Is Not a Smoker Reference Group 1.01 0.89 1.03 Occasional Smoker 0.847 0.0009* 0.8574 (0.92, 1.11) (0.82, 0.95) (0.77, 1.37) 2.58 1.75 2.67 Former Daily Smoker <.0001* <.0001* <.0001* (2.50, 2.67) (1.66, 1.84) (2.19, 3.24) 0.45 0.46 0.67 Always an Occasional Smoker <.0001* <.0001* 0.1442 (0.38, 0.53) (0.37, 0.58) (0.40, 1.15) 1.17 0.54 1.76 Former Occasional Smoker <.0001* <.0001* 0.0013* (1.12, 1.22) (0.44, 0.66) (1.25, 2.48) 1.22 0.84 1.6 Daily Smoker <.0001* <.0001* <.0001* (1.17, 1.27) (0.78, 0.91) (1.32, 1.93) 1.89 0.96 4.08 Not Available/Applicable <.0001* 0.8867 0.0054* (1.43, 2.50) (0.53, 1.74) (1.52, 10.98)

Alcohol Consumption Frequency Less than Once per Month Reference Group 0.69 0.76 0.8 Once per Month <.0001* <.0001* 0.1311 (0.65, 0.73) (0.69, 0.83) (0.61, 1.07)

0.49 0.53 0.38 Two to Three Times per Month <.0001* <.0001* <.0001* (0.46, 0.51) (0.47, 0.59) (0.28, 0.52) 0.46 0.62 0.55 Once per Week <.0001* <.0001* <.0001* (0.43, 0.48) (0.57, 0.68) (0.42, 0.74) 0.44 0.52 0.61 Two to Three Times per Week <.0001* <.0001* 0.0005* (0.42, 0.46) (0.47, 0.57) (0.46, 0.81) 0.46 0.67 0.66 Four to Six Times per Week <.0001* <.0001* 0.0879 (0.43, 0.50) (0.58, 0.78) (0.41, 1.06) 0.77 1.03 0.74 Everyday <.0001* 0.521 0.1501 (0.72, 0.81) (0.94, 1.12) (0.49, 1.11) 1.34 1.2 1.6 Not Available/Applicable <.0001* <.0001* <.0001* (1.29, 1.38) (1.13, 1.28) (1.34, 1.91)

Body Mass Index Normal Reference Group 0.66 0.54 0.42 Underweight <.0001* <.0001* 0.0738 (0.57, 0.76) (0.45, 0.65) (0.16, 1.09) 2.42 1.83 1.89 Overweight <.0001* <.0001* 0.0013* (2.33, 2.52) (1.73, 1.94) (1.53, 2.34) 5.37 4.00 4.81 Obese <.0001* <.0001* <.0001* (5.18, 5.58) (3.77, 4.24) (3.96, 5.83) 1.58 1.72 0.77 Not Available/Applicable <.0001* <.0001* 0.0614 (1.51, 1.66) (1.59, 1.85) (0.58, 1.01)

Household Income 1.6 0.85 0.71 No Income 0.0045* 0.4141 0.7019 (1.16, 2.21) (0.58, 1.25) (0.12, 4.13) 1.84 1.47 0.9 Less than $5,000 <.0001* 0.0034* 0.81 (1.49, 2.26) (1.14, 1.90) (0.39, 2.08) $5,000 to $9,999 3.48 <.0001* 1.93 <.0001* 2.04 0.0009*

(3.16, 3.84) (1.61, 2.32) (1.34, 3.11) 4.1 2.27 2.21 $10,000 to $14,999 <.0001* <.0001* <.0001* (3.84, 4.39) (2.01, 2.56) (1.56, 3.12) 4.02 2.42 1.92 $15,000 to $19,999 <.0001* <.0001* 0.0004* (3.76, 4.30) (2.15, 2.71) (1.34, 2.75) 3.56 2.08 1.66 $20,000 to $29,999 <.0001* <.0001* 0.0033* (3.37, 3.77) (1.89, 2.28) (1.18, 2.32) 2.56 1.62 1.42 $30,000 to $39,999 <.0001* <.0001* 0.0451* (2.41, 2.71) (1.47, 1.79) (1.01, 2.00) 2 1.44 1.2 $40,000 to $49,999 <.0001* <.0001* 0.3471 (1.88, 2.13) (1.30, 1.60) (0.82, 1.73) 1.67 1.32 0.99 $50,000 to $59,999 <.0001* <.0001* 0.9574 (1.57, 1.78) (1.18, 1.47) (0.68, 1.45) 1.34 1.2 1.37 $60,000 to $80,000 <.0001* 0.0003* 0.075 (1.26, 1.42) (1.09, 1.33) (0.97, 1.93) $80,000 and Above Reference Group 1.4 1.43 0.99 Not Available/Applicable <.0001* <.0001* 0.9549 (1.33, 1.47) (1.31, 1.55) (0.73, 1.35)

Personal Income 1.11 1.38 0.93 No Income 0.0322* <.0001* 0.7564 (1.01, 1.22) (1.20, 1.58) (0.57, 1.51) 0.8 0.91 0.52 Less than $5,000 <.0001* 0.2985 0.0177* (0.72, 0.88) (0.77, 1.09) (0.31, 0.89) 2.03 2.21 1.79 $5,000 to $9,999 <.0001* <.0001* 0.0057* (1.88, 2.19) (1.93, 2.52) (1.19, 2.71) 2.66 2.83 1.7 $10,000 to $14,999 <.0001* <.0001* 0.0106* (2.48, 2.85) (2.50, 3.21) (1.13, 2.54) 2.4 2.2 2.34 $15,000 to $19,999 <.0001* <.0001* <.0001* (2.23, 2.58) (1.93, 2.51) (1.55, 3.53)

1.8 1.75 1.43 $20,000 to $29,999 <.0001* <.0001* 0.0862 (1.68, 1.93) (1.55, 1.98) (0.95, 2.15) 1.51 1.45 1.3 $30,000 to $39,999 <.0001* <.0001* 0.2238 (1.41, 1.62) (1.27, 1.64) (0.85, 1.97) 1.22 1.36 1.04 $40,000 to $49,999 <.0001* <.0001* 0.8857 (1.13, 1.32) (1.19, 1.56) (0.65, 1.64) 1.29 1.39 1.11 $50,000 to $59,999 <.0001* <.0001* 0.6704 (1.19, 1.40) (1.20, 1.61) (0.69, 1.80) 1.09 1.02 0.81 $60,000 to $80,000 0.0314* 0.8393 0.4158 (1.01, 1.19) (0.88, 1.18) (0.49, 1.34) $80,000 and Above Reference Group 1.25 1.55 0.81 Not Available/Applicable <.0001* <.0001* 0.2895 (1.17, 1.34) (1.37, 1.75) (0.54, 1.20)

Household Education Less than Secondary School 3.31 2.5 1.93 <.0001* <.0001* <.0001* Education (3.20, 3.43) (2.34, 2.66) (1.63, 2.29) 1.48 1.31 0.78 Secondary School Graduation <.0001* <.0001* 0.043* (1.42, 1.54) (1.22, 1.40) (0.62, 0.99) 1.21 0.97 0.78 Some Post-Secondary Education <.0001* 0.5597 0.0706 (1.14, 1.28) (0.87, 1.08) (0.59, 1.02) Post-Secondary Education Reference Group 1.23 1.25 0.72 Not Available/Applicable <.0001* <.0001* 0.0127* (1.16, 1.30) (1.15, 1.37) (0.56, 0.93)

Personal Education Less than Secondary School 1.67 2 1.03 <.0001* <.0001* 0.7178 Education (1.62, 1.72) (1.90, 2.10) (0.89, 1.20) 1.1 1.21 0.6 Secondary School Graduation <.0001* <.0001* <.0001* (1.06, 1.14) (1.14, 1.29) (0.47, 0.77)

0.86 0.84 0.65 Some Post-Secondary Education <.0001* 0.001* 0.0015* (0.82, 0.91) (0.75, 0.93) (0.49, 0.85) Post-Secondary Education Reference Group 2.02 2.12 1.95 Not Available/Applicable <.0001* <.0001* 0.0041* (1.81, 2.26) (1.82, 2.47) (1.23, 3.07)

Dietary Consumption of Fruit and

Vegetables Consumes Less than 5 Servings per 1.29 1.35 1.13 <.0001* <.0001* 0.5327 Day (1.20, 1.40) (1.19, 1.54) (0.77, 1.64) 1.29 1.29 1 Consumes 5 to 10 Servings per Day <.0001* 0.0002* 0.9832 (1.19, 1.40) (1.12, 1.47) (0.68, 1.48) Consumes More than 10 Servings Reference Group per Day 1.78 1.63 1.19 Not Available/Applicable <.0001* <.0001* 0.3996 (1.64, 1.94) (1.41, 1.88) (0.79, 1.78)

National Language Competency

(Conversational) English Only Reference Group 1.15 0.66 1.5 French Only <.0001* 0.002* 0.0279* (1.11, 1.20) (0.51, 0.86) (1.05, 2.14) 0.76 0.51 1.02 English and French Only <.0001* <.0001* 0.8771 (0.73, 0.78) (0.43, 0.59) (0.83, 1.25) 0.53 0.66 1.02 English and French and Other <.0001* <.0001* 0.9226 (0.48, 0.58) (0.61, 0.72) (0.66, 1.57) 0.81 0.78 2.05 English and Other (Not French) <.0001* <.0001* <.0001* (0.77, 0.86) (0.74, 0.83) (1.73, 2.43) 0.76 1.02 4.22 French and Other (Not English) 0.1792 0.7584 0.0018* (0.51, 1.13) (0.90, 1.16) (1.71, 10.43)

1.87 2.96 3.34 Neither English or French (Other) 0.0175* 0.0005* 0.2730 (1.12, 3.14) (1.61, 5.44) (0.39, 28.80) 1.41 1.48 0.4 Not Available/Applicable 0.0007* <.0001* 0.1426 (1.16, 1.73) (1.37, 1.60) (0.12, 1.36)

2. NON-MODIFIABLE RISK FACTOR VARIABLES Age 12 to 19 Reference Group 2.06 2.53 2.81 20 to 29 <.0001* 0.0019* 0.0002* (1.79, 2.37) (1.41, 4.55) (1.62, 4.85) 4.19 8.98 5.54 30 to 39 <.0001* <.0001* <.0001* (3.67, 4.79) (5.20, 15.48) (3.29, 9.34) 8.33 24.85 10.33 40 to 49 <.0001* <.0001* <.0001* (7.34, 9.44) (14.5, 42.6) (6.24, 17.1) 19.78 66.53 29.06 50 to 59 <.0001* <.0001* <.0001* (17.49, 22.38) (38.9, 113.8) (17.8, 47.6) 38.4 101.88 51.26 60 to 69 <.0001* <.0001* <.0001* (33.95, 43.42) (59.6, 174.3) (31.13, 84.4) 48.36 123.25 62.04 70 to 79 <.0001* <.0001* <.0001* (42.72, 54.73) (72.0, 210.9) (36.7, 104.9) 43.22 113.6 56.61 80 and Above <.0001* <.0001* <.0001* (38.00, 49.16) (66.2, 195.0) (30.3, 105.9) Not Available/Applicable N/A N/A N/A N/A N/A N/A

Sex Male Reference Group 0.82 0.81 1.13 Female <.0001* <.0001* 0.0873 (0.80, 0.84) (0.78, 0.85) (0.98, 1.29) Not Available/Applicable N/A N/A N/A N/A N/A N/A

Racial/Cultural Origin White Reference Group 0.46 1.1 Black <.0001* 0.0218* N/A N/A (0.36, 0.57) (1.01, 1.20) 0.62 0.86 Korean 0.0621 0.1166 N/A N/A (0.37, 1.03) (0.71, 1.04) 0.13 0.83 Filipino <.0001* 0.0012* N/A N/A (0.06, 0.28) (0.74, 0.93) 1.62 0.59 Japanese 0.0004* 0.0194* N/A N/A (1.24, 2.13) (0.38, 0.92) 0.34 0.59 Chinese <.0001* <.0001* N/A N/A (0.27, 0.44) (0.54, 0.64) 0.22 1.3 South Asian <.0001* <.0001* N/A N/A (0.16, 0.31) (1.22, 1.39) 0.41 0.81 Southeast Asian <.0001* 0.0025* N/A N/A (0.26, 0.63) (0.71, 0.93) 0.09 0.48 Arab <.0001* <.0001* N/A N/A (0.03, 0.27) (0.40, 0.58) 0.04 0.71 (0.59, West Asian 0.0128* 0.0002* N/A N/A (0.00, 0.51) 0.85) 0.22 0.61 Latin American <.0001* <.0001* N/A N/A (0.12, 0.42) (0.54, 0.70) 0.59 1.23 Other <.0001 0.0003* N/A N/A (0.46, 0.75) (1.10, 1.38) 0.61 1.06 Multiple <.0001* 0.5285 N/A N/A (0.51, 0.74) (0.89, 1.26) 1.35 1.07 Not Available/Applicable <.0001* 0.5923 N/A N/A (1.26, 1.44) (0.83, 1.39)

Ethnicity (Ethnic Origin)

Canadian Reference Group 1.05 0.96 French 0.0185* 0.7555 N/A N/A (1.01, 1.09) (0.72, 1.27) 1.16 1.26 English <.0001* 0.0665 N/A N/A (1.12, 1.20) (0.98, 1.62) 0.82 1.4 German <.0001* 0.009* N/A N/A (0.78, 0.87) (1.09, 1.81) 1.03 1.17 Scottish 0.1588 0.2446 N/A N/A (0.99, 1.07) (0.90, 1.52) 1.04 0.92 Irish 0.0532 0.5668 N/A N/A (1.00, 1.09) (0.70, 1.22) 0.55 2.71 Italian <.0001* <.0001* N/A N/A (0.50, 0.61) (2.12, 3.47) 0.84 1.04 Ukrainian <.0001* 0.7950 N/A N/A (0.78, 0.91) (0.76, 1.43) 0.76 1.54 Dutch (Netherlands) <.0001* 0.0016* N/A N/A (0.70, 0.84) (1.18, 2.01) 0.35 0.84 Chinese <.0001* 0.1668 N/A N/A (0.28, 0.44) (0.66, 1.08) 1 1.44 Jewish 0.9519 0.0195* N/A N/A (0.84, 1.18) (1.06, 1.94) 0.69 0.91 Polish <.0001* 0.4753 N/A N/A (0.62, 0.77) (0.69, 1.19) 0.23 1.38 Portuguese <.0001* 0.0172 N/A N/A (0.17, 0.33) (1.06, 1.81) 0.27 1.7 South Asian <.0001* <.0001* N/A N/A (0.20, 0.36) (1.33, 2.17) 1.56 1.24 Not Available/Applicable 0.9776 0.3477 N/A N/A (0.94, 5.66) (0.87, 3.34)

3. IMMIGRANT-SPECIFIC RISK FACTOR VARIABLES Country Origin Other North America Reference Group South, Central America and the 1.77 N/A N/A <.0001* N/A N/A Caribbean (1.55, 2.01) 1.72 Europe N/A N/A <.0001* N/A N/A (1.53, 1.94) 1.21 Africa N/A N/A <.0012* N/A N/A (1.04, 1.41) 1.30 Asia N/A N/A <.0001* N/A N/A (1.15, 1.46) 1.50 Oceania N/A N/A <.0004* N/A N/A (1.14, 1.98) 1.44 Not Available/Applicable N/A N/A <.0001* N/A N/A (1.36, 1.53)

Time Since Immigration to Canada 0 to 9 Years Since Immigration Reference Group 2.04 10 to 19 Years Since Immigration N/A N/A <.0001* N/A N/A (1.87, 2.22) 2.92 20 to 29 Years Since Immigration N/A N/A <.0001* N/A N/A (2.68, 3.19) 4.64 30 to 39 Years Since Immigration N/A N/A <.0001* N/A N/A (4.28, 5.02) 40 or More Years Since 5.36 N/A N/A <.0001* N/A N/A Immigration (4.96, 5.78) 2.1 Not Available/Applicable N/A N/A <.0001* N/A N/A (1.96, 2.25)

Age at Immigration to Canada

0.2 12 to 19 N/A N/A <.0001* N/A N/A (0.16, 0.25) 0.37 20 to 29 N/A N/A <.0001* N/A N/A (0.30, 0.47) 0.42 30 to 39 N/A N/A <.0001* N/A N/A (0.33, 0.53) 0.56 40 to 49 N/A N/A <.0001* N/A N/A (0.44, 0.70) 1.02 50 to 69 N/A N/A 0.8478 N/A N/A (0.81, 1.29) 70 and Above Reference Group 0.26 Not Available/Applicable N/A N/A <.0001* N/A N/A (0.21, 0.33) Notes : (1)*Result was statistically significant at the 0.05 level (based on a chi-square test statistic); (2) All models controlled for time/survey year and interview mode; (3) All results were calculated using rescaled master weights (see Section 4.3.3 ).

5.4 Comparative Impact of Diabetes Risk Factors - Adjusted Results

Odds ratios (likelihood of diabetes) were generated for each subpopulation according to the outlined set of variables found in Table 5 (adjusted multivariable logistic regression results are found in Table 9).

5.4.1 Smoking Status

Overall, results persisted in the adjusted multivariable model with some modest attenuation of the odds ratios. Mixed results were noted for the association between smoking status and reporting diabetes. Compared to non-smokers, former daily smokers had greater odds of having diabetes, an effect noted in all three subpopulations. However, for the Canada born and immigrant subpopulations, compared to non-smokers, all the other categories (former/occasional/daily smokers) had lower odds of diabetes; this effect was not seen in the Aboriginal subpopulation.

5.4.2 Alcohol Consumption Frequency

Following adjustment for modifiable and non-modifiable risk factors, the initial trend observed in the crude results continued. When compared to those who consumed alcohol less than once per month, those who consumed alcohol more frequently were less likely to have diabetes; this was true for each of the three subpopulations.

5.4.3 Body Mass Index

Overall, increasing BMI was still associated with increased likelihoods of diabetes for all three subpopulations following adjustment. For example, obese BMI Aboriginals had nearly 4.85 times greater odds of having diabetes when compared to normal BMI Aboriginals (OR 4.85, 95% 4.21-4.56). For the obese BMI category, this was the highest odds ratio amongst the subpopulations.

5.4.4 Household and Personal Income

Following adjustment for other covariates, associations between lower income levels and increased likelihoods of diabetes were seen only for the Household income variable.

Broadly speaking, for the Canada-born and immigrant subpopulations, higher household income was associated with lower odds of reporting diabetes. The difference was most stark between the high-income range ($80,000 and above) and the low-income range ($5,000 to $29,999), where there was a two-fold difference in odds. Household income was not associated with diabetes in the Aboriginal subpopulation.

For Personal income, adjusted results revealed that Canada-born individuals who reported incomes in the $50,000 to $59,999 range had the highest likelihood of diabetes (OR 1.21, 95% CI 1.04-1.40) when compared to Canada-born individuals who reported personal incomes of $80,000 or more; this result was similarly observed in the immigrant subpopulation (OR 1.6, 95% CI 1.22-2.09).

5.4.5 Household and Personal Education

A mixed picture was seen for the effects of Household and Personal education on diabetes. For the Canada-born population, a trend of lower household education being associated with greater odds of reporting diabetes was seen. For example, compared to those with post-secondary education, Canada-born respondents from households with less then a secondary school education had 29% greater odds of reporting diabetes (OR 1.29, 95% CI 1.15-1.42). On the other hand, for the immigrant subpopulation, respondents from households with some post-secondary education had 32% lower odds of reporting diabetes (OR 0.68, 95% CI 0.54-0.85) compared to respondents from households with post-secondary education. In contrast, no association was noted between household education and diabetes in the Aboriginal subpopulation.

For personal education, a similar trend (i.e. lower education being associated with greater odds of reporting diabetes) was noted for all three groups.

5.4.6 Diet

Lower consumption of dietary fruits and vegetables was associated with 12% lower odds of diabetes in the Canada-born subpopulation (OR 0.88, 95% CI 0.78-1.00), but was associated with greater odds of diabetes in the immigrant group (OR 1.36, 95% CI 1.07- 1.71).

5.4.7 National Language Competency

After controlling for other risk factors, language competency associations revealed by the crude results were no longer significant. Compared to English-only speakers, all other language groups/combinations had lower odds of reporting diabetes, both in the Canada- born and immigrant populations.

5.4.8 Age

Both crude and adjusted results showed a similar pattern of successively greater odds of reporting diabetes with increasing age; this was noted for all three subpopulations.

5.4.9 Sex

Both crude and adjusted results showed that females were less likely to report having diabetes when compared to males in both the Canada-born and immigrant subpopulation. Adjusted results demonstrated a slightly increased likelihood of diabetes in Aboriginal females when compared to Aboriginal males; however, this result was not statistically significant.

5.4.10 Racial/Cultural and Ethnic Origin

For the Canada-born subpopulation, Blacks, Southeast Asians and those with multiple racial/cultural origins had significantly greater odds of reporting diabetes when compared to Whites. For the immigrant subpopulation, all racial/cultural groups had greater odds of reporting diabetes, compared to Whites.

Concerning ethnic origin, adjusted results showed that respondents with German, Ukrainian, and Polish ethnicities had lesser odds of reporting diabetes than ethnically identified Canadians in the Canada-born subpopulation. Among immigrants, those reporting Italian, Jewish, South Asian, and Portuguese ethnicities had greater odds of reporting diabetes.

5.4.11 Immigrant-Specific Risk Factors

When compared to immigrants from other North American countries, immigrants from South, Central America, and the Caribbean (29% greater odds), Africa (20% greater odds), and Asia (19% greater odds) had increased likelihoods for reporting diabetes.

Following adjustment, greater length of stay in Canada was strongly associated with greater odds of reporting diabetes. For instance, immigrants who had been in Canada for 40 years or more were most likely to have diabetes when compared to immigrants who had been in Canada for 0 to 9 years (OR 11.97, 95% CI 11.01-13.02).

The association between age at immigration and diabetes was also found to be significant both before and after controlling for other risk factors; immigration at older ages was strongly linked to increased likelihoods for diabetes. For instance, immigrants who arrived in Canada between the ages of 12 and 19 were drastically less likely to have diabetes when compared to immigrants who arrived in Canada at ages over 70 (OR 0.05, 95% CI 0.04-0.06).

Table 9: Results from Multivariable Logistic Regressions (Adjusted) Canada-Born (Non- Aboriginal (First Immigrant, Non- Immigrant likelihood of Nations, Inuit, and Aboriginal) likelihood of Diabetes Métis) likelihood of Diabetes Diabetes Odds Ratios Odds Ratios Odds Ratios (95% (95% (95% P-Value P-Value P-Value Confidence Confidence Confidence Intervals) Intervals) Intervals) 1. MODIFIABLE RISK FACTOR VARIABLES Smoke Status Is Not a Smoker Reference Group 1.06 0.82 0.91 Occasional Smoker 0.1177 0.0017* 0.7167 (0.99, 1.15) (0.72, 0.93) (0.54, 1.52) 1.92 1.44 1.7 Former Daily Smoker <.0001* <.0001* 0.0065* (1.81, 2.03) (1.31, 1.58) (1.16, 2.49) 0.54 0.33 0.84 Always an Occasional Smoker <.0001* <.0001* 0.6806 (0.41, 0.70) (0.21, 0.52) (0.36, 1.95) 0.97 0.82 0.97 Former Occasional Smoker 0.7172 <.0001* 0.9235 (0.84, 1.13) (0.72, 0.93) (0.50, 1.89) 0.93 0.42 1.22 Daily Smoker 0.0422* <.0001* 0.3024 (0.87, 1.00) (0.29, 0.59) (0.84, 1.76) 1.6 0.87 3.39 Not Available/Applicable 0.0013* 0.6579 0.0279* (1.20, 2.14) (0.47, 1.60) (1.14, 10.04)

Alcohol Consumption Frequency Less than Once per Month Reference Group 0.74 0.8 0.68 Once per Month <.0001* 0.0007* 0.0419* (0.69, 0.80) (0.71, 0.91) (0.47, 0.99)

0.51 0.5 0.37 Two to Three Times per Month <.0001* <.0001* <.0001* (0.48, 0.55) (0.43, 0.58) (0.24, 0.55) 0.51 0.73 0.44 Once per Week <.0001* <.0001* <.0001* (0.47, 0.54) (0.65, 0.83) (0.30, 0.66) 0.51 0.52 0.37 Two to Three Times per Week <.0001* <.0001* 0.0020* (0.48, 0.55) (0.45, 0.59) (0.24, 0.55) 0.59 0.62 0.91 Four to Six Times per Week <.0001* <.0001* 0.7392 (0.53, 0.66) (0.50, 0.77) (0.51, 1.62) 0.74 0.82 0.4 Everyday <.0001* 0.0029* 0.0032* (0.68, 0.80) (0.73, 0.94) (0.21, 0.73) 1.48 1.23 1.87 Not Available/Applicable <.0001* <.0001* <.0001* (1.43, 1.54) (1.15, 1.31) (1.55, 2.27)

Body Mass Index Normal Reference Group 0.75 0.73 0.56 Underweight 0.0002* 0.0009* 0.2374 (0.65, 0.87) (0.59, 0.88) (0.21, 1.47) 1.85 1.42 1.46 Overweight <.0001* <.0001* 0.0011* (1.78, 1.92) (1.34, 1.51) (1.16, 1.82) 4.38 3.25 4.02 Obese <.0001* <.0001* <.0001* (4.21, 4.56) (3.04, 3.47) (3.26, 4.95) 2.02 1.58 1.97 Not Available/Applicable <.0001* <.0001* 0.0001* (1.91, 2.13) (1.45, 1.71) (1.39, 2.78)

Household Income 1.7 1.14 3.76 No Income 0.0483* 0.7285 0.2046 (1.00, 2.87) (0.55, 2.34) (0.49, 29.08) 1.76 2.09 0.7 Less than $5,000 0.0007* 0.0031* 0.6337 (1.27, 2.44) (1.28, 3.40) (0.16, 3.03) $5,000 to $9,999 2.76 <.0001** 3.1 <.0001* 1.45 0.3252

(2.30, 3.31) (2.25, 4.28) (0.69, 3.05) 2.14 1.99 1.49 $10,000 to $14,999 <.0001* <.0001* 0.2415 (1.87, 2.44) (1.58, 2.50) (0.77, 2.89) 2.61 2.25 0.93 $15,000 to $19,999 <.0001* <.0001* 0.8362 (2.29, 2.97) (1.79, 2.82) (0.48, 1.80) 2.35 2.05 1.11 $20,000 to $29,999 <.0001* <.0001* 0.7188 (2.13, 2.59) (1.73, 2.42) (0.62, 2.00) 1.91 1.55 1 $30,000 to $39,999 <.0001* <.0001* 0.9992 (1.74, 2.10) (1.31, 1.83) (0.57, 1.76) 1.69 1.11 1.25 $40,000 to $49,999 <.0001* 0.2188 0.4389 (1.54, 1.85) (0.94, 1.32) (0.71, 2.22) 1.32 1.15 0.69 $50,000 to $59,999 <.0001* 0.0858 0.2220 (1.21, 1.45) (0.98, 1.36) (0.38, 1.25) 1.18 1.24 1.25 $60,000 to $80,000 <.0001* 0.0027* 0.3955 (1.09, 1.28) (1.08, 1.43) (0.75, 2.08) $80,000 and Above Reference Group 1.36 1.47 1.19 Not Available/Applicable <.0001* <.0001* 0.3073 (1.28, 1.43) (1.34, 1.61) (0.85, 1.67)

Personal Income 0.74 0.87 2.6 No Income 0.0018* 0.3746 0.2768 (0.61, 0.89) (0.63, 1.19) (0.47, 14.54) 0.63 0.72 2.79 Less than $5,000 <.0001* 0.0652 0.2366 (0.52, 0.76) (0.51, 1.02) (0.51, 15.28) 0.8 0.84 3.41 $5,000 to $9,999 0.0089* 0.2710 0.1485 (0.68, 0.95) (0.62, 1.15) (0.65, 17.97) 0.99 1.21 2.42 $10,000 to $14,999 0.9207 0.1733 0.2948 (0.85, 1.16) (0.92, 1.61) (0.46, 12.71) 0.87 0.93 3.85 $15,000 to $19,999 0.0946 0.6052 0.1088 (0.75, 1.02) (0.69, 1.24) (0.74, 20.04)

100

0.92 1.3 2.66 $20,000 to $29,999 0.2589 0.0481* 0.2421 (0.80, 1.06) (1.00, 1.70) (0.52, 13.66) 0.97 1.1 2.54 $30,000 to $39,999 0.6536 0.4736 0.2655 (0.84, 1.12) (0.85, 1.43) (0.49, 13.03) 0.94 1.22 1.61 $40,000 to $49,999 0.4141 0.1367 0.5742 (0.81, 1.09) (0.94, 1.59) (0.31, 8.43) 1.21 1.6 2.67 $50,000 to $59,999 0.0130* 0.0006* 0.2474 (1.04, 1.40) (1.22, 2.09) (0.51, 14.09) 1.04 0.81 1.22 $60,000 to $80,000 0.6437 0.1447 0.8244 (0.89, 1.20) (0.61, 1.08) (0.22, 6.89) $80,000 and Above Reference Group 1.04 1.16 1.08 Not Available/Applicable 0.2843 0.0262* 0.7322 (0.97, 1.12) (1.02, 1.32) (0.70, 1.67)

Household Education Less than Secondary School 1.29 0.9 1.26 <.0001* 0.0797 0.3281 Education (1.17, 1.42) (0.76, 1.08) (0.79, 2.00) 1.06 0.96 1.31 Secondary School Graduation 0.1751 0.6571 0.2783 (0.97, 1.16) (0.82, 1.13) (0.80, 2.14) 1.01 0.68 1.09 Some Post-Secondary Education 0.8226 0.0008* 0.7757 (0.91, 1.13) (0.54, 0.85) (0.61, 1.96) Post-Secondary Education Reference Group 1.12 1 1.19 Not Available/Applicable 0.0019* 0.9977 0.0886 (1.04, 1.20) (0.90, 1.12) (0.85, 1.67)

Personal Education Less than Secondary School 1.37 1.81 1.24 <.0001* <.0001* 0.3107 Education (1.26, 1.49) (1.57, 2.10) (0.82, 1.89) 0.98 0.98 0.47 Secondary School Graduation 0.5776 0.7295 0.0054* (0.90, 1.06) (0.84, 1.13) (0.27, 0.80)

101

0.98 1.11 0.86 Some Post-Secondary Education 0.7128 0.3394 0.5979 (0.88, 1.09) (0.90, 1.36) (0.48, 1.52) Post-Secondary Education Reference Group 1.45 1.69 3.21 Not Available/Applicable <.0001* <.0001* <.0001* (1.26, 1.66) (1.39, 2.04) (1.80, 5.73)

Dietary Consumption of Fruit and

Vegetables Consumes Less than 5 Servings 0.88 1.23 1.16 0.0483* 0.0797 0.6831 per Day (0.78, 1.00) (0.98, 1.55) (0.58, 2.33) Consumes 5 to 10 Servings per 1.03 0.6013 1.36 1.3 0.0111* 0.4740 Day (0.91, 1.17) (1.07, 1.71) (0.64, 2.65) Consumes More than 10 Servings Reference Group per Day 1.36 1.41 1.01 Not Available/Applicable <.0001* <.0001* 0.9566 (1.24, 1.48) (1.22, 1.64) (0.66, 1.55)

National Language Competency

(Conversational) English Only Reference Group 0.93 0.47 1.12 French Only 0.0368* 0.0050* 0.7398 (0.88, 1.00) (0.28, 0.80) (0.58, 2.15) 0.93 0.74 1.21 English and French Only 0.0232* 0.0201* 0.2873 (0.88, 0.99) (0.58, 0.95) (0.86, 1.70) 1.03 0.65 1.66 English and French and Other 0.6491 <.0001* 0.1381 (0.90, 1.18) (0.56, 0.76) (0.85, 3.26) 0.83 0.88 0.9 English and Other (Not French) 0.0005* 0.0071* 0.5727 (0.75, 0.92) (0.80, 0.97) (0.62, 1.30) 0.89 0.76 1.78 French and Other (Not English) 0.7598 0.0293* 0.4355 (0.43, 1.86) (0.59, 0.97) (0.42, 7.65)

102

1.7 0.84 0.22 Neither English or French (Other) 0.0541 0.0539 0.2992 (0.99, 2.91) (0.70, 1.00) (0.01, 3.91) 0.94 2.19 2.11 Not Available/Applicable 0.7705 0.0153* 0.5136 (0.60, 1.46) (1.16, 4.13) (0.22, 19.84) 2. NON-MODIFIABLE RISK FACTOR VARIABLES Age 12 to 19 Reference Group 2.02 2.42 2.8 20 to 29 <.0001* 0.0032* 0.0002* (1.74, 2.34) (1.35, 4.37) (1.62, 4.84) 4.21 9.25 5.54 30 to 39 <.0001* <.0001* <.0001* (3.67, 4.83) (5.36, 15.95) (3.28, 9.33) 8.46 27.26 10.3 40 to 49 <.0001* <.0001* <.0001* (7.42, 9.64) (15.91, 46.72) (6.23, 17.05) 20.14 81.28 29.06 50 to 59 <.0001* <.0001* <.0001* (17.71, 22.90) (47.51, 139.06) (17.76, 47.56) 39.67 134.92 51.18 60 to 69 <.0001* <.0001* <.0001* (34.90, 45.09) (78.84, 230.90) (31.08, 84.26) 51.13 180.09 61.67 70 to 79 <.0001* <.0001* <.0001* (44.94, 58.17) (105.15, 308.43) (36.48, 104.24) 46.97 178.01 56.14 80 and Above <.0001* <.0001* <.0001* (41.09, 53.69) (103.60, 305.85) (30.03, 104.97) Not Available/Applicable N/A N/A N/A N/A N/A N/A

Sex Male Reference Group 0.73 0.78 1.11 Female <.0001* <.0001* 0.1676 (0.71, 0.75) (0.74, 0.81) (0.96, 1.27) Not Available/Applicable N/A N/A N/A N/A N/A N/A

Racial/Cultural Origin

103

White Reference Group 1.48 2.52 Black 0.0016* <.0001* N/A N/A (1.16, 1.88) (2.31, 2.76) 0.84 2.19 Korean 0.5044 <.0001* N/A N/A (0.50, 1.41) (1.78, 2.68) 0.76 1.93 Filipino 0.4994 <.0001* N/A N/A (0.35, 1.68) (1.71, 2.17) 1.08 1.11 Japanese 0.5810 0.6484 N/A N/A (0.82, 1.43) (0.71, 1.75) 0.9 1.05 Chinese 0.3855 0.2435 N/A N/A (0.70, 1.15) (0.97, 1.14) 1.18 2.86 South Asian 0.3011 <.0001* N/A N/A (0.86, 1.63) (2.67, 3.06) 1.94 1.94 Southeast Asian 0.0041* <.0001* N/A N/A (1.23, 3.05) (1.68, 2.23) 0.38 1.31 Arab 0.0778 0.0052* N/A N/A (0.13, 1.12) (1.08, 1.57) 0.24 1.82 West Asian 0.2597 <.0001* N/A N/A (0.02, 2.85) (1.51, 2.20) 1.09 1.74 Latin American 0.7969 <.0001* N/A N/A (0.57, 2.08) (1.51, 2.01) 1.21 2.48 Other 0.1496 <.0001* N/A N/A (0.93, 1.56) (2.20, 2.80) 1.4 2.29 Multiple 0.0009* <.0001* N/A N/A (1.15, 1.70) (1.90, 2.75) 2.07 1.68 Not Available/Applicable <.0001* 0.0002* N/A N/A (1.93, 2.21) (1.28, 2.21)

Ethnicity (Ethnic Origin) Canadian Reference Group

104

1.02 1.15 French 0.3648 0.3667 N/A N/A (0.98, 1.06) (0.85, 1.56) 1.01 1.03 English 0.5350 0.8155 N/A N/A (0.98, 1.05) (0.80, 1.34) 0.92 1.09 German 0.0027* 0.5576 N/A N/A (0.87, 0.97) (0.82, 1.43) 1.01 0.97 Scottish 0.7388 0.8424 N/A N/A (0.97, 1.05) (0.73, 1.29) 1.03 0.79 Irish 0.2774 0.1195 N/A N/A (0.98, 1.07) (0.59, 1.06) 0.96 2.34 Italian 0.4372 <.0001* N/A N/A (0.87, 1.06) (1.77, 3.09) 0.9 1.14 Ukrainian 0.0127* 0.4510 N/A N/A (0.83, 0.98) (0.81, 1.62) 1 1.15 Dutch (Netherlands) 0.9139 0.3426 N/A N/A (0.90, 1.09) (0.86, 1.54) 1.39 0.83 Chinese 0.1940 0.5075 N/A N/A (0.85, 2.28) (0.48, 1.44) 0.87 1.45 Jewish 0.1348 0.0287* N/A N/A (0.73, 1.04) (1.04, 2.01) 0.78 1.2 Polish <.0001* 0.2643 N/A N/A (0.70, 0.88) (0.87, 1.66) 0.77 1.73 Portuguese 0.1442 0.0003* N/A N/A (0.54, 1.09) (1.29, 2.32) 1.19 1.43 South Asian 0.5973 0.1839 N/A N/A (0.63, 2.22) (0.85, 2.41) 1.2 0.9 Not Available/Applicable 0.077 0.0455* N/A N/A (0.71, 1.72) (0.81, 1.02) 3. IMMIGRANT-SPECIFIC RISK FACTOR VARIABLES Country Origin

105

Other North America Reference Group South, Central America and the 1.23 N/A N/A <.0028* N/A N/A Caribbean (1.09, 1.52) 1.04 Europe N/A N/A <.5855 N/A N/A (0.91, 1.18) 1.20 Africa N/A N/A 0.0428* N/A N/A (1.01, 1.44) 1.19 Asia N/A N/A 0.0408* N/A N/A (1.01, 1.40) 1.24 Oceania N/A N/A 0.1625 N/A N/A (0.92, 1.66) 1.45 Not Available/Applicable N/A N/A <.0001* N/A N/A (1.21, 1.67)

Time Since Immigration to Canada 0 to 9 Years Since Immigration Reference Group 2.36 10 to 19 Years Since Immigration N/A N/A <.0001* N/A N/A (2.16, 2.57) 4.1 20 to 29 Years Since Immigration N/A N/A <.0001* N/A N/A (3.75, 4.48) 8.35 30 to 39 Years Since Immigration N/A N/A <.0001* N/A N/A (7.66, 9.09) 40 or More Years Since 11.97 N/A N/A <.0001* N/A N/A Immigration (11.01, 13.02) Not Available/Applicable N/A N/A N/A N/A N/A N/A

Age at Immigration to Canada 0.05 12 to 19 N/A N/A <.0001* N/A N/A (0.04, 0.06) 20 to 29 N/A N/A 0.11 <.0001* N/A N/A

106

(0.08, 0.13) 0.18 30 to 39 N/A N/A <.0001* N/A N/A (0.14, 0.23) 0.34 40 to 49 N/A N/A <.0001* N/A N/A (0.27, 0.43) 0.7 50 to 69 N/A N/A 0.0039* N/A N/A (0.56, 0.89) 70 and Above Reference Group Not Available/Applicable N/A N/A N/A N/A N/A N/A Notes : (1)*Result was statistically significant at the 0.05 level (based on a chi-square test statistic); (2) All models controlled for modifiable, non-modifiable risk factors, time/survey year, and interview mode; (3) All results were calculated using rescaled master weights (see Section 4.3.3 ).

107

Chapter 6 6 Discussion

This thesis aimed to characterize Canadian diabetics from 2001 to 2010, estimate the associated period prevalence of diabetes among all Canadians, Canadian-immigrants, Canadian-Aboriginals (First Nations, Inuit, and Métis), and Canada-born (non-immigrant, non-Aboriginal population) individuals, as well as assess the changes in diabetes prevalence rates in Canada from 2001 to 2010. This thesis also aimed to illuminate the comparative impact of modifiable, and non-modifiable (immigrant-specific inclusive) risk factors among Canadian subpopulations. This chapter begins with an overview of the study results and how they compare to the current body of diabetes surveillance research (Section 6.1). A discussion of the challenges associated with subpopulation research will follow (Section 6.2). Lastly, a summary of the present study’s strengths, limitations, implications and conclusions will then be described (Section 6.3, Section 6.4, and Section 6.5, respectively).

6.1 Overview of the Findings

6.1.1 Main Subpopulation Differences

6.4% of Canadians reported having diabetes in 2010 according to the CCHS; though, the overall period prevalence of diabetes was found to be 5.3%. This 2010 prevalence finding was slightly lower than the prevalence rate specified by the CCDSS, which in 2011 (using 2008/2009 health data) stated that individuals diagnosed with diabetes constituted 6.8% of the Canadian population. 1 However, the crude 5.3% period prevalence of diabetes found in this study was considerably higher than the age-adjusted period prevalence of diabetes found using 1998-2009 CCDSS data (~4.5%);416 crude period prevalence is not provided by the CCDSS.

Conversely, the National Diabetes Surveillance System (NDSS), a Health Canada initiative designed to monitor diabetes prevalence rates and adverse health outcomes (proposed in part by the Institute for Clinical Evaluative Sciences (ICES)), has reported findings that are akin to the present study. While the NDSS has not released data beyond

108 the year 2009 (using 2006/2007 health data), its most recent report estimated that approximately 6.2% of Canadians were diagnosed with diabetes. 420 The NDSS had also predicted that by 2012, rates of diabetes would be similar to what was most recently published by the CCDSS. 416 While 6.4% (and 5.3% for the overall study period) may seem unusually low for such a large sample, three things must be considered.

First, this rate is much more representative of the realistic (crude) population rate of T2DM (recall that 90% of diabetes mellitus cases are T2DM) as indicated by the Canadian Diabetes Association. 7 Second, diabetes rates that have been presented in this study demonstrate an average that is calculated from the years 2001 to 2010, which encompassed a generally upward trend in the rates of reported diabetes (seen in Figure 4); the prevalence of diabetes (among all three subpopulations) increased from 2001 to 2010 in accordance with CCDSS and NDSS predictions.1,416 Third, slightly lower rates of diabetes found in this study may also be the result of underestimation due to self-report. As mentioned previously in this thesis, large health surveys must surmount a great deal of self-reporting bias when sampling large proportions of the population, especially in relation to the reporting of chronic diseases like diabetes mellitus. 381 If correct, CCHS diabetes rates may actually be more indicative of self-perceived health status, and could be useful in arguing for more efficient methods of investigating chronic disease outcomes in large Canadian samples.

The period prevalence of diabetes from 2001 to 2010 was found to be lowest among Canada-born (non-immigrant, non-Aboriginal) individuals (5.0%), followed by Canadian- Aboriginals (First Nations, Inuit, and Métis) (6.5%). Canadian-immigrants had the highest period prevalence of diabetes among the three Canadian subpopulations at 6.7%. Immigrants were also found to the have the highest rates of diabetes almost consistently across the study period when compared to Canada-born and Aboriginal individuals (based on previous Canadian surveillance literature, Aboriginals were expected to have had the highest rates of diabetes from 2001 to 2010 349,359,371 ). The difficulty with comparing specific subpopulation data found in previous studies to the current study’s results is that the information typically provided by Statistics Canada reports, surveillance systems, and even ICES publications, make frequent use of provincial data (relative to immigrants) or

109 data collected from single cycles of the NPHS, CCHS, APS, or the FNRLHS (relative to Aboriginals). Nevertheless, comparisons to these findings are still discussed; in particular, immigrant findings will be later considered in contrast to ICES studies that have made use of electronic registries/health databases (in order to depict snapshots of recent Canadian diabetes rates).

The significant (upward) linear trends observed for all three subpopulations, and the differences seen in their magnitudes, provide an interesting point of discussion. Regardless of immigrant status or racial/cultural/ethnic origin, a rise in diabetes diagnoses from 2001 to 2010 was seen among all three groups. Particularly, the rise in diabetes prevalence rates from 4.1% in 2001 to 6.4% in 2010 among all Canadians in this study is fairly consistent with the Public Health Agency of Canada’s predicted increase in prevalence from 4.2% in 2000. Nevertheless, this study adds to the current literature by providing a distinct stratification of high-risk Canadian subpopulations.

By examining Canada-born, immigrant, and Aboriginal subpopulations independently, this study found that Canada-born individuals had the lowest prevalence rates of diabetes overall, ranging from 3.9% in 2001 to 5.7% in 2010 (a 46% increase); this was the only trend observed in this study that was lower than the overall Canadian prevalence rate trend. By far, immigrants had the highest prevalence rates of diabetes consistently across time, ranging from 5.0% in 2001 to 8.5% in 2010 (a 70% increase). The present study is the first of its kind to demonstrate a diabetes time trend specifically for Canadian immigrants at the national-level, especially one that is considerably above the overall Canadian diabetes prevalence trend. This finding inherently poses a unique challenge for the Canadian healthcare system; when considering future diabetes management strategies, as well as preventative tactics for dealing with the current diabetes epidemic in high-risk subpopulations, immigrants must be considered. As mentioned previously, immigrant health and diabetes will be further discussed in Section 6.1.4.

Concerning Aboriginals, the prevalence rates of diabetes ranged from 5.4% in 2001 to 7.4% in 2010 (a 37% increase). The Public Health Agency of Canada notes that data sources examining diabetes prevalence rates in First Nations, Métis, and Inuit populations provide varying estimates of the rise in rates across time periods between 2001 and

110

2010. 349 For instance, one study from British Columbia examining the rise in diabetes prevalence rates in the First Nations population from 2002 to 2007 found an increase of 15.5% (age-adjusted), while another Quebec study found a crude increase of 36% for a similar time period (2001 to 2005);421 however, both studies did not rely on self-report data and found prevalence rates that were considerably higher than the ones found in the present study. Using data from the Aboriginal Peoples Survey, two separate studies have found that the self-reported prevalence rate of diabetes among Métis Canadians increased from 5.9% in 2001 to 7.0% in 2006 (an increase of 19%); 422,423 these prevalence rates are much more comparable with the prevalence rate found in this thesis. Nevertheless, a reliance on self-reported data could account for slightly lower rates of diabetes in the Aboriginal population found in this study, as well as others.422,423

Unfortunately, the present study was unable to make any conclusions about specific Aboriginals populations; however, the results from this study are applicable to off-reserve First Nations, Métis, and Inuit Aboriginal populations. This thesis adds to the Canadian- Aboriginal diabetes literature by providing an off-reserve time trend for diabetes prevalence rates in Canada that is generalized to all three Aboriginal groups. The present study’s findings also help to identify unique risk factor associations that are particular to the off-reserve Aboriginal population in Canada, regardless of provincial geography or potential reserve influence.

6.1.2 Modifiable Risk Factor Differences 6.1.2.1 Smoking

The presence of former smoking behaviour was found to be associated with increased odds of diabetes in all three subpopulations. On the other hand, other smoking behaviour categories were associated with decreased odds of diabetes in the Canada-born and the immigrant subpopulations. Overall, the highest period prevalence rate of diabetes was seen in Canadians who were former daily smokers (8.5%).

The present study’s findings are somewhat consistent with previous prospective cohort study results that have determined the correlation between cigarette smoking and increased risks for diabetes; 60,72,77,129-133 it is assumed that the physiological mechanisms

111 which facilitate increased risks for diabetes (identified in these cohort studies), relative to smoking, are true for both current and former smokers. One prospective cohort study also identified an elevated risk associated with smoking cessation in the first 3 years of quitting; even after adjusting for age, race, sex, education, physical activity and a number of other health related measures, this elevated risk was significantly higher than non- smokers’ and current daily smokers’ risks. 424 Moreover, CCHS respondents who indicated that they had quit smoking were not explicitly asked to provide the length of time since quitting smoking. It is possible that many diabetic former smokers were within this 3-year window period at the time of their CCHS participation. It is also possible that many diabetic former smokers had been diagnosed with diabetes while being current daily smokers; their diabetes may or may not have been the reason for their smoking cessation.

From 2001 to 2010, Canada-born and immigrant individuals had relatively decreased odds associated with daily and occasional smoking. Some research has posited that the nicotine found in cigarettes is able to increase energy expenditure, and decrease appetite. 425,426 Moreover, other cross sectional studies have indicated that the average BMI of smokers tends to be lower than the average BMI of nonsmokers.427 Nicotine may be responsible for controlling body weight in some daily or occasional smokers, which could decrease the risk for diabetes associated with obesity or being overweight. Whether or not the present study’s findings could be explained by cigarette smoking behaviour and its effect on metabolic rate or appetite suppression is mostly speculative, however.

6.1.2.2 Alcohol Consumption

Consistent with previous research indicating that moderate alcohol use may decrease the risk of T2DM (via cardiovascular benefit),135,136 alcohol consumption of once per month or more was found to be mildly protective (concerning likelihood) in comparison with alcohol consumption of less than once per month.

There are certain limitations with attempting to gauge alcohol consumption using the CCHS survey questionnaire. CCHS data does not provide any granularity beyond consumption occurrences that are less frequent than once per month, or more frequent

112 than once per day. It may be assumed that the increased likelihoods seen for those with missing data was determined primarily by individuals who do not consume alcohol at all, or who consume more than one serving of alcohol per day; thus, not attaining or exceeding the usage that is recommended for cardiovascular benefit. 136

6.1.2.3 Body Mass Index (BMI) & Diet

As BMI increased, so did the probability of diabetes. Overweight and obese BMI categories represented the majority of the total diabetes sample; each of these categories had period prevalence rates of 5.9% and 11.4%, respectively.

Previous literature has stated that increased BMI (>30kg/m 2) is a valid indicator of an increased risk for T2DM. 78-82,88 Given the highly co-morbid relationship between obesity and T2DM, 98 it should be expected that rising trends in diabetes (as observed in this study) would also co-occur with rising trends in self-reported obesity; 2,98 the current study’s findings greatly support the presence of this relationship. However, due to the cross-sectional design of the CCHS, this study cannot make any claims about the bidirectional relationship between obesity and diabetes. While the crude and adjusted regression results showed that the two disease states were highly associated, this study could not decipher whether or not obesity led to diabetes, or vice versa for CCHS participants. Overall, each subpopulation observed this co-morbid trend with only minor differences.

Of course, rates of obesity are inherently related to diet. The present study attempted to gather dietary information by using the quantifiable consumption of fruit and vegetables (normally said to be indicators of a healthful diet 123 ); this was found to be poorly associated with diabetes. Nevertheless, mixed results were seen concerning the association between lower consumption of dietary fruits and vegetables and the odds of reporting diabetes in the Canada-born and immigrant subpopulations.

Due to the complex nature of the pooling methodology, this study was unfortunately limited by which dietary variables could be used for analyses. For the future, assessing holistic dietary influences that are relative to each subpopulation would be advantageous; the various cultural influences of Canadian subpopulations may considerably impact the

113 over- or under- consumption of various macronutrients (that could lead to increased risks for diabetes). Knowledge of dietary discrepancies, while noticeably difficult to measure in any health survey,283 will help to evaluate culturally-catered preventative programs that aim to reduce diabetes rates via recommended dietary adjustments.

6.1.2.4 Socioeconomic Status (SES) & National Language Competency

Individuals with diabetes who reported earnings for either income variable between $5,000 and $9,999 and upwards of $20,000 and $29,999 had the highest period prevalence rates among their respective earnings categories (ranging from 8.4% to 9.8% for household income, and 5.9% to 8.6% for personal income, respectively). Moreover, lower levels of household income were found to be most strongly associated with increased risks of diabetes in the Canada-born and immigrant subpopulations. On the other hand, higher levels of personal income were found to be most strongly associated with increased risks of diabetes in these subpopulations.

As previously noted in the literature, low SES (with income being used as a proxy variable) has been linked to the development of T2DM in developed nations like Canada.140-142 Lower levels of SES have been associated with decreased access to healthful food, healthcare services, and even economic prosperity. 141-146 Though lower income findings in this study are consistent with previous SES and diabetes research, increased odds that were observed in higher personal income ranges are not as easily explained.

Previously, research has shown that in less developed nations, high levels of SES may be an indicator of decreased physical activity and a sedentary lifestyle, which subsequently leads to increased risks for diabetes. 146-148 Individuals who report higher levels of personal income in Canada may either be experiencing increased odds of diabetes because of modifiable risk factor mechanisms, or because of other non-modifiable risk factors that compound the chances of diabetes.

From 2001 to 2010, 19.2% of diabetics indicated that the highest level of education obtained by an individual in their household was less than secondary school, while 33.9%

114 of diabetics had indicated that they had personally not completed a secondary school education; each comprised 11.7% and 7.1% of their respective education categories (period prevalence). These period prevalence rates were considerably high as indicators for SES and strongly associated with an increased risk for diabetes. For personal education in particular, both Canada-born and immigrant individuals had increased likelihoods of diabetes if they reported having less than a secondary school education (OR 1.37, 95% CI 1.26-1.49, and OR 1.81, 95% CI 1.57-2.10, respectively). For household education, this trend was seen only in the Canada-born subpopulation. For immigrants, respondents from households with only some post-secondary education had lower odds of reporting diabetes than those from post-secondary educated households.

For Canada-born and immigrant subpopulations, low SES (as indicated by low levels of personal education) was indeed associated with high risks for diabetes. These findings are consistent with the previous SES literature that agree upon SES as a mediating construct for diabetes outcomes. 141-146 In essence, low levels of education are thought to lead to less gainful employment, lower levels of income, and more precarious living situations (that may facilitate health inequalities).141-146

The immigrant finding concerning household education could be also explained by high SES (as indicated by high levels of household education) being associated with decreased physical activity and a sedentary lifestyle, which may lead to increased risks for diabetes (previously discussed relative to the personal income finding).

Braveman et al. points out that measuring SES can be quite difficult in health surveys because of the natural variations and perceptions of privacy that exist among participants. 418 Racial and ethnic differences tend to occur systematically for income and other SES proxies like education;418 by assessing several SES indices, these differences might be better acknowledged. 428 Whether or not SES disparities exist systematically in relation to diabetes outcomes is worth addressing in the future (especially for Aboriginal and immigrant populations in Canada). While the correlation between SES and diabetes is somewhat subjective, the basis of its facilitating mechanisms is rooted in the complexities of other modifiable risk factors. SES is a concept that encompasses systematic schemas, which are difficult, in any methodology, to effectively measure or quantify.

115

Immigrants face unique barriers to their health when compared to the Canada-born population, particularly due to acculturative and SES influences.35 Language competency has been previously established as a useful marker of both acculturation and SES in immigrant communities. 34 Compared to English-only speakers, all other language groups/combinations had lower odds of reporting diabetes, both in the Canada-born and immigrant populations.

Previous research has suggested that lower levels of national language (English or French) competency and/or general health literacy may lead certain racial/cultural groups to be inherently marginalized, especially if they are longer-term immigrants.429,430,439 It has also been suggested that marginalization may cause wide variations in subsequent socio-health behaviours and outcomes following settlement; this could essentially compound diabetes risks (modifiable and/or non-modifiable) over time.35,36 English-only speakers may be more inclined to see a physician regarding their health due to a lack of communication barriers. Increased healthcare service usage in this group may subsequently lead to more diabetes diagnoses (in comparison to those who may be less inclined to see a physician because of potential communication difficulties). However, it is difficult to discern causality regarding healthcare usage and disease outcomes, especially using cross-sectional data.

Loosely defined as the “language(s), thoughts, communications, actions, customs, beliefs, values, and institutions of racial, ethnic, religious, or social groups,” the concept of culture has become more and more prevalent when investigating health disparities in subpopulations.431,432 If immigrant or Canadian-born communities are gathered collectively by a shared belief that encourages or discourages (either passively or actively) health service usage via common tradition, either of these communities may inevitably observe high or low amounts of diabetes diagnoses in relation to each other.

Incongruities in healthcare usage can have substantial implications for preventative care programs and how they tailor to the specialized needs of subpopulations in Canada.37 Certain US investigations into perceived discrimination among cultural communities and patient satisfaction (with local healthcare systems) have determined that language barriers may deter individuals, and subsequently, close-knit cultural communities from seeking

116 medical care. 433 These considerations are highly relevant for Canadian research into health disparities that exist among certain subpopulations; in particular, immigrant or Aboriginal communities that may hold pervasive (negative or positive) cultural attitudes towards traditional Canadian healthcare.

6.1.3 Non-Modifiable Risk Factor Differences

6.1.3.1 Age

Those aged 50 to 59 (23.5% of all diabetics), 60 to 69 (26.4% of all diabetics) and 70 to 79 (21% of all diabetics) had the highest period prevalence rates of diabetes; as described by previous research, older age was strongly associated with an increased risk for diabetes across all subpopulations. Still, the period prevalence rates in each of the above age categories fell considerably below the previous (2006/2007/2008/2009) data provided by the NDSS and the CCDSS: ages 50 to 59 (7.6% compared with ~10.4%), 60 to 69 (13.0% compared with ~18.7%), 70 to 79 (15.7% compared with ~24.8%) and 80 or Above (14.3% compared with ~23.1%). These lower rates may be particularly due to an underestimation of diabetes due to self-report, combined with the inherent nature of the period prevalence estimates. Underestimation of chronic diseases has been observed in greater magnitude among older health survey participants, which could explain the attenuated estimates that are presented in this study. 381

Of particular importance was the substantial effect of age found in the immigrant population. By far, immigrants had the highest odds ratios associated with increasing age amongst the subpopulations. This result is consistent with the Creatore et al.’s (ICES) finding that Ontario immigrants experience increased likelihoods of diabetes earlier in life (more frequently in the 20-40 years of age range) when compared to non-immigrant Canadians. 321 Immigrant health researchers have typically suggested either SES disparities 265,288,289 or genetic variations 261-264 among clusters of immigrant communities as possible explanations for the amplified effect of age. Both perspectives offer useful insight into the diverse health barriers that exist for Canadian immigrant populations (financial, informational, etc.), but also into the distinct variants that exist among

117 constituent ethnicities. Issues of small sample size and limited ability to investigate age differences through expansive cohort studies typically arise, nonetheless.418

6.1.3.2 Sex

The present study found that males and females had 5.4% and 4.6% respective period prevalence rates (from 2001 to 2010). However, Appendix 1 reveals a slight upward trend in the year-to-year diabetes frequencies for males, and a slight downward trend in the year-to-year diabetes frequencies for females. This trend finding is somewhat consistent with 1998-2009 CCDSS data that found age-adjusted prevalence rates of diagnosed diabetes to be higher among males over time. 180 Nevertheless, period estimates and trends are better assessed with regression analyses. Adjusted results indicated that both Canada- born and immigrant females were less likely to have diabetes than Canada-born and immigrant males, respectively. The present study’s findings illuminate the pervasive nature of sex differences for diabetes, regardless of Canadian subpopulation.

Aboriginals were also found to demonstrate a small reversal of the sex trend observed in the other subpopulations (consistent with previous research 52,341,356,363,364 ); yet, this result was not found to be statistically significant. The significance of this finding may have been affected by the use of only off-reserve Aboriginals for analyses. Primarily, research into Aboriginal female diabetes has focused on females living on-reserve. 361,365,371 By exploring diabetes outcomes in off-reserve Aboriginals, this study may suggest small sex differences that are primarily intermediated by SES or acculturation as opposed to genetic mechanisms.

6.1.3.3 Racial, Cultural, and Ethnic Origin

Race, culture, and ethnicity were found to be strongly associated with diabetes for both Canada-born individuals and immigrants; however, increased risks were seen in greater magnitude among immigrants following adjustment for other risk factors. South Asian, Southeast Asian, Black, ‘Other’ and ‘Multiple’ racial/cultural/ethnic origin identified immigrants were most likely to report diabetes when compared to White immigrants. In particular, White immigrants were least likely to have diabetes when compared to any other racially or culturally identified immigrant group. These findings are comparable

118 with high-risk subpopulation data provided by the CCDSS,1,59 and the risk guidelines used by the CANRISK assessment tool. 73,74 However, previously available statistical data has frequently analyzed racial/cultural/ethnic origin without the stratification of immigrant status.

This study showed that each of the identified higher-risk immigrant subgroups had larger odds ratios than their Canada-born counterpart subgroups (Table 9). This comparison suggests the importance of scrutinizing the immigrant population in national-level diabetes surveillance systems. At the provincial-level, critically relevant immigrant research has been very recently generated from the University of Toronto and ICES. At the national-level, Canada’s health surveillance systems have not publically released an informational report since 2009 (NDSS) and 2011 (CCDSS), let alone statistical data that are specific to immigrants. Future Canadian surveillance research should not only focus on the impact of race, culture, and ethnicity on diabetes, but should also investigate the important barriers that are presented for immigrants of high-risk races, cultures, and ethnicities identified in this study. Moreover, it may be beneficial for the Canadian Diabetes Association’s clinical practice guidelines to consider immigrant-specific recommendations in the future.

6.1.4 Immigrant-Specific Risk Factors

Immigrants from countries originating in South, Central America, and the Caribbean were found to have the highest likelihoods of diabetes when compared to other North American immigrants from the United States. Similarly, immigrants from Africa and Asia had relatively high likelihoods of diabetes in comparison to other North American immigrants.

Prior studies have also identified immigrants from South, Central America, and the Caribbean as having increased risks for diabetes; 202,203,205 however, several of these studies have been conducted in the United States and United Kingdom, or have not examined all the potential spheres that may influence immigrant diabetes (such as age at immigration and time since immigration). Using the Ontario Diabetes Registry and the Longitudinal Immigrant Database, Creatore at al. addressed this gap in the Canadian

119 literature and found that immigrants from South Asia, Latin America, the Caribbean, and sub-Saharan Africa had, on average, a 2-3 times increased risk of diabetes when compared to Western European and North American immigrant populations. The similarity between the results of the present study and Creatore et al.’s are clear; this thesis adds to the growing body of Canadian-immigrant diabetes literature by offering cross-sectional results that mirror Creatore et al.’s findings.

The difficulty with investigating health differences by country origin is that there are several contributing dynamics that combine to produce variations in the immigrant subpopulation. Immigrants that are originally from the high-risk country origins (identified in this study) are typically of racial/cultural/ethnic origins that already observe increased risks for diabetes; South Asians, Latin Americans, and Blacks are descendent origins that constitute large proportions of the population in South Asian, Latin American, Caribbean, and sub-Saharan African countries.224 This raises an important question for investigations of health disparities that include race, culture, and ethnicity: are racial/cultural/ethnicity differences in health outcomes a result of race, ethnicity, culture, or a combination of these factors? Even more so, this presents an issue concerning these factors and specifically, immigration: how does one quantify the disparities among Canadians that are specific to immigrants and unique from disparities that are due to race, culture, or ethnicity?

Some research has shown that specifically relying on ethnicity or race as a valid genetic perspective for explaining adverse health outcomes (like diabetes) is driven by the assumption that racial or ethnic origin differences are strictly biological in nature.424,434 The trouble with relying on genetic differences is that both immigrants and Canada-born individuals are of increasingly mixed origins; 425 solely trusting a genetic model to explain health disparities is not reliable with increasingly changing birth patterns. Complicating matters even more so are changes in immigration patterns. Across the study period, 5 of the top-10 country origins of recent immigrants to Canada (according to census data) changed considerably from 2001 to 2006; data beyond 2006 is currently unavailable (Table 2). 326,327 The varying diversity of immigrants arriving to Canada each year has a large influence on the diversity of the CCHS sample from year to year as well.

120

Furthermore, health research that is conducted using large population health surveys and surveillance databases are still limited by small sample sizes when investigating particular country origins that are not seen in Table 2.

This study also addressed an important concern that is relevant to Canadian immigrants who participated in the CCHS: time since immigration, and age at immigration. The health of newly arrived and long-term immigrants is typically a product of social, cultural, economic, and genetic factors; 261-264 nevertheless, time spent in and out of Canada is particularly important. With increasing periods of residency in Canada, immigrants experienced exponentially higher likelihoods of diabetes. For instance, immigrants who had been in Canada for 40 years or more were much more likely to have diabetes when compared to immigrants who had been in Canada for 0 to 9 years. Inversely, the younger that an immigrant was upon immigration to Canada the less likely they were to have diabetes when compared to older immigrants. The concern for timing and age relates back to the concept of the healthy immigrant effect; overall, immigrants typically report better health upon arrival in Canada when compared to Canada-born individuals. 261-265 Using Gushulak et al.’s model for immigrant health in three stages, this initial superior health report characteristically relates to the few years immediately following migration. 265

The present study’s results are consistent with other research findings which indicate that after lengthier periods of residency in North America, immigrant health may deteriorate to even worse levels than individuals born in-country.265,275-277,279 Inline with these previous findings, it appears as though the healthy immigrant effect is not pervasive over time in Canadian-immigrants. Whether or not this is due to acculturation is still debatable, however. Previous research has typically blamed an adaptation of Western culture (mostly in the United States’) for facilitating increased health risks due to modifiable behaviours. 270,280 Likewise, similar research has attributed poorer health (over time) to a degradation of the once-positive culturally specific behaviours that originate in home- countries, coupled with general acculturation. 288,289 Given the current theories relative to acculturation, the present study’s finding (that individuals who immigrated at younger ages are less likely to have diabetes than older immigrants) is unusual.

121

At younger ages, immigrants would experience longer durations of exposure to Western culture when compared to immigrants who arrived at older ages. By way of acculturation theory, younger immigrants would therefore be expected to have higher likelihoods of diabetes. Lower diabetes likelihoods found in Canada-born individuals (compared to immigrants) might indicate that the acculturation influence (seen quite heavily in the United States) is not as applicable to studies of Canadian-immigrants. Nevertheless, this may only be true for diabetes, and may not be a pervasive finding for other chronic diseases that are facilitated by increased modifiable risk factors. Moreover, the increasing likelihood of older immigrants to have diabetes when compared to younger immigrants may simply be a matter of older age. Increasing age is substantially associated with diabetes in the immigrant population (as suggested by the present study’s results); even though odds ratios were exceptionally high at younger ages (when compared to other subpopulations), older aged immigrants saw an exponentially large (comparative) likelihood that could explain this finding.

One of the implicit aims of the present study was to accommodate the highest possible amount of risk factors and subpopulation-specific considerations that were available for analyses. The CCHS was undeniably a valuable health data resource that allowed for an incorporation of each of these factors in some way; nevertheless, its usage was accompanied by some limitations that will be discussed in Section 6.4 .

6.2 Challenges in Subpopulation Research 6.2.1 Challenges in Studying Immigrant Health

Comparing results among immigrant groups is particularly difficult. While already discussed in great detail, the health of immigrants (recent or long-term) is the complex product of time-relative social, cultural, economic, and genetic factors. 261-264 Comparisons of results found in previous studies to the present findings is complicated by the existence of both genetic 437 and acculturative 424 perspectives; both provide a necessary component for researching health outcomes in this population. A large challenge associated with investigating genetic variances according to racial/cultural/ethnic origins is surmounting the need for large enough sample sizes in specific origins that are worthy of study. Using

122 large-scale health survey data is useful, but comes with certain trade-offs. These surveys are typically cross-sectional and lack the cohort-specific temporality that is necessary to generate quality causal explanations of diabetes according to various origins. Concurrently, large-scale cohort research is quite expensive and unfeasible when exploring a singular chronic disease condition, such as diabetes, explicitly.

6.2.2 Challenges in Studying Aboriginal Health

Concerning Aboriginals, the difficulty of investigation also lies in obtaining large enough sample sizes and quality data for health outcomes research. Off-reserve vs. on-reserve Aboriginal community research is difficult to equate; both types of communities experience varying degrees of risk factors, making “reserve-membership” difficult to control for. While using large-scale surveys, like the CCHS, may provide comprehensive risk factor information, these surveys are typically limited to off-reserve or on-reserve Aboriginals, exclusively, and rarely comprise of both for comparison purposes. Systematic research into the disparities faced by Aboriginals must surmount the unique challenge of surveying health-specific information, but culturally relevant variables as well; inherently, a holistic approach to investigating health outcomes in this subpopulation is essential.

6.3 Strengths

A considerable strength of this research project was the use of a cross-sectional, nationally representative population health survey. CCHS data allowed not only for relatively simplified analyses for the proposed topic of research, but also provided an incredibly large sample to explore the possibility of exposure-effect variations among the Canadian population. Given that the CCHS aims to collect pertinent information relative to a wide range of health determinants, health service utilization, and overall demographics, the survey data was undoubtedly the most appropriate resource for investigating specific subpopulations that are typically challenging to study. Statistics Canada also ensured the representativeness of the survey via overall sampling design, weighting strategies, and methodologies relative to data pooling.

123

While numerous studies in the past have focused explicitly either on Canadian- immigrants or Canadian-Aboriginals, separately, this study aimed to present their diabetes risks simultaneously and comparatively using the same dataset. In a unique fashion, the scope of this study’s health surveillance presents a novel investigation of risk factors that addresses unique concerns for Canada’s highest-risk populations for diabetes. By cumulatively presenting these data, policy analysts may be better able to prioritize and tackle systemic issues related to diabetes.

6.4 Limitations

6.4.1 Self-Report and Limitations with Measures

Concerns for self-reporting biases arose in Chapter 4 when initially discussing CCHS data collection methods. Unfortunately, due to limitations pertaining to the CCHS questionnaire, certain variable measures that were used were not optimal. Self-reporting in health surveys is typically problematic because of over or under-estimation of certain measures. 404 Due to the negative stigmas surrounding smoking behaviour and typically poor accuracy concerning height and weight reports, the smoking status and BMI calculated variables might be conservative. When BMI is calculated using self-report height or weight, overweight and/or obesity categories are typically underestimated. 406 As well, the tendency for older respondents in health surveys, like the CCHS, to underreport their diagnoses of certain chronic diseases (diabetes mellitus in particular) may have also skewed the results; 407 however, significant differences were seen regardless of any potential for underestimation.

Important differences may exist between the subpopulations that are not fully captured by the use of CCHS survey data. For instance, highly significant results were seen for those who did not respond to SES proxy variables such as income and education level (both household and personal). Even though SES has been noted as an effective way to measure acculturation, these variables had the highest degrees of missing data. While research concerning the validity of self-reported SES data is limited, 435,436 it is more so the presence of non-response bias that may be of concern to the study; this issue was also true for the diet, smoking, and alcohol consumption frequency variables.

124

Another limitation of the CCHS questionnaire pertained to the outcome measure. Diabetes (both generally and for gestational diabetes) was measured quite ambiguously. A considerable limitation for the study would be the lack of distinction between types of diabetes (type 1 and 2), as well as a dedicated question for clarifying gestational diabetes. While type 1 and gestational diabetes constitute very small percentages of total diabetes mellitus cases, had they been included as possible responses to diabetes-related questions, they would have been controlled for; analyses, results, and conclusions would therefore have been specifically dedicated to T2DM as opposed to general diabetes mellitus.

6.4.2 Non-Response, Response Rate, and Interview Mode

Addressing the issue of non-response identified above, Table 7 presented the response rates for the CCHS across the study period. Overall, response rates decreased from 2001 to 2010 (84.7% to 71.5%, respectively). Even though sampling weights and imputation methods were used to ensure that non-response (both complete and partial) is adjusted for, differences seen across time in relation to subpopulation diabetes magnitudes may still have been due to decreasing rates in response. Lastly, while interview mode was effectively controlled for throughout the analyses, some residual bias due to variances in interview mode may still have persisted.

6.4.3 Pooling Methodology

While pooling methodology provided a sufficiently large enough sample to investigate diabetes in relatively specialized groups, this methodology came with certain trade-offs. Given that the CCHS survey questionnaire content changed from year to year, variables that were available for analysis were somewhat limited. In order to make valid conclusions about the probability estimates generated by the analyses, only variables that remained consistent in terms of content wording and coverage across all iterations could be used. Unfortunately, this led to the exclusion of certain variable constructs like physical activity, diabetes treatments and/or medications, quality of life measures, and also limited the relevance of certain measures used, like diet. This study attempted to capture as much of the risk factor data as possible given this limitation; still, using fewer iterations of the CCHS may have provided additional variables for examination in the

125 regression analyses (possibly altering the study results following their inclusion). Nevertheless, using fewer CCHS iterations could have raised concerns for small sample sizes according to the subpopulations of interest.

Another limitation concerning this methodology relates to the period prevalence estimates described in the results. Unfortunately, fluctuations in single year estimates had to be considered as part of an overall period estimate when describing odds ratios and changes in prevalence rates. 411 While including a distinct ‘time’ variable into regression analyses helped to curb the interpretation in a way that described both the period and year-specific prevalence rates, there were certain limitations with the final period results. The main findings concerning diabetes prevalence rates over the study period are in fact averages of the individual iteration year rates, and are better viewed in combination with the trend analyses so as to examine the differences in parameters from year to year.

6.4.4 Cross-sectional Design and Temporality

Due to the cross-sectional design of the CCHS, this study does not establish exposure- effect relationships; causation cannot be determined by using this data. This study cannot establish the subpopulation-specific causal factors that lead to the eventual development of diabetes (given that temporality is not present). Instead, this study identifies important trends and risk factor associations that are linked with the comparative diabetes rates found in Canadian subpopulations.

6.5 Implications and Conclusions

The results of the current study may be used in a number of ways, and can be targeted towards future subpopulation health surveillance or diabetes epidemiology research proposals; recommendations for clinical practice that are relative to immigrant health and diabetes may also be formulated based on the study results.

For the future, diabetes research concerning the impact of race, culture, and ethnicity in Canadian immigrants should be holistically coupled with explorations of education, skills training, mobility, migration history, demography, and labour market contributions. By doing so, immigrant health researchers may be better able to investigate health outcomes

126

(like diabetes) using complex information that comprises both genetic and SES perspectives. It may be of value to explore the possibility of specialized linkages between healthcare databases across the provinces, coordinated at the national-level. Electronic health databases and landing registries may afford national population health surveillance systems (like the CCDSS and the NDSS) with more accurate (and quality) data that could be available for exploring unique barriers to immigrant health (using the Longitudinal Immigrant Database, for instance).

All Canadians, regardless of subpopulation, have experienced and will continue to experience a rise in obesity, diabetes, and CVD risk factor rates over the next 10 years. 1,2 Compounded by low SES, future challenges for the healthcare system and Canadian health policy makers include formulating preventative and management strategies for dealing with the Canadian diabetes epidemic at lower levels of SES. For Aboriginals and immigrants, both with and without diabetes, SES marginalization is a useful indicator of current and future risk factor disparities and subsequent diabetes development. Whether or not the observed differences between subpopulations that are presented in this thesis occur systematically, is a matter worth addressing in future research.

Previous subpopulation literature has stressed the need for diabetes prevention strategies to continue addressing the social determinants of health. 437 Healthcare utilization is an important measure of determinant outcomes, and also inequities in health and access to care. An important area of future research would be an examination of healthcare service usage (despite the type or frequency) in Canadian-immigrant and Canadian-Aboriginal diabetics; differential usages (in comparison to other subpopulations) may speak to the unique financial, linguistic, and population-specific needs of these subpopulations, where community support is heavily relied upon. Health promotion strategies 435 and support systems that congregate at the community-level may be of vital importance for tackling diabetes in immigrants and Aboriginals; service delivery programs should continue to consider the pervasive nuances observed in Canadian subpopulations that thrive in racially, culturally, and ethnically close-knit communities.

127

Appendix 1: Distribution of Variables by CCHS Survey Year (2001 to 2010)

Number of Diabetes Cases by CCHS Survey Year (Percentage of Variable Column Total) 2001 2003 2005 2007 2008 2009 2010 1. MODIFIABLE RISK FACTOR VARIABLES Smoke Status Is Not a Smoker 1623 1890 1947 1255 1313 1208 1452 (30.0%) (30.6%) (30.2%) (33.0%) (34.2%) (32.9%) (36.3%) Occasional Smoker 660 765 827 494 396 494 453 (12.2%) (12.4%) (12.8%) (13.0%) (10.3%) (13.4%) (11.3%) Former Daily Smoker 2085 2513 2550 1410 1448 1385 1435 (38.5%) (40.7%) (39.5%) (37.1%) (37.7%) (37.7%) (35.9%) Always an Occasional Smoker 39 43 35 31 56 22 15 (0.7%) (0.7%) (0.5%) (0.8%) (1.5%) (0.6%) (0.4%) Former Occasional Smoker 97 765 133 51 76 59 94 (1.8%) (12.4%) (2.1%) (1.3%) (2.0%) (1.6%) (2.4%) Daily Smoker 912 859 958 558 552 507 550 (16.8%) (13.9%) (14.9%) (14.7%) (14.4%) (13.8%) (13.8%)

Alcohol Consumption Frequency Less than Once per Month 1243 1288 1463 744 765 734 853 (38.2%) (34.4%) (36.0%) (31.0%) (32.0%) (30.7%) (33.5%) Once per Month 372 438 484 321 326 276 279 (11.4%) (11.7%) (11.9%) (13.4%) (13.6%) (11.5%) (11.0%) Two to Three Times per Month 356 430 384 282 233 284 278 (10.9%) (11.5%) (9.5%) (11.7%) (9.7%) (11.9%) (10.9%) Once per Week 425 528 521 361 307 312 357

128

(13.1%) (14.1%) (12.8%) (15.0%) (12.8%) (13.0%) (14.0%) Two to Three Times per Week 407 476 549 360 344 347 402 (12.5%) (12.7%) (13.5%) (15.0%) (14.4%) (14.5%) (15.8%) Four to Six Times per Week 134 157 161 86 137 137 95 (4.1%) (4.2%) (4.0%) (3.6%) (5.7%) (5.7%) (3.7%) Everyday 319 430 498 248 280 304 282 (9.8%) (11.5%) (12.3%) (10.3%) (11.7%) (12.7%) (11.1%)

Body Mass Index Underweight 70 71 67 30 28 30 26 (2.4%) (1.0%) (1.1%) (0.9%) (0.8%) (0.9%) (0.7%) Normal 654 1475 1571 766 914 767 841 (22.2%) (20.9%) (25.0%) (21.9%) (25.4%) (22.7%) (22.5%) Overweight 391 2287 2307 1343 1169 1170 1392 (13.3%) (30.1%) (36.8%) (38.5%) (32.5%) (34.6%) (37.3%) Obese 1830 2185 2328 1353 1489 1417 1473 (62.1%) (36.1%) (37.1%) (38.8%) (41.4%) (41.9%) (39.5%)

Household Income No Income 24 6 20 7 3 5 6 (0.5%) (0.1%) (0.4%) (0.2%) (0.1%) (0.2%) (0.0%) Less than $5,000 22 31 31 15 29 16 22 (0.5%) (0.6%) (0.6%) (0.6%) (1.1%) (0.7%) (0.9%) $5,000 to $9,999 166 140 121 59 74 54 88 (3.5%) (2.7%) (2.5%) (2.2%) (2.7%) (2.4%) (3.6%) $10,000 to $14,999 535 410 429 217 198 190 150 (11.2%) (7.9%) (8.9%) (8.0%) (7.3%) (8.5%) (6.1%) $15,000 to $19,999 452 428 381 210 273 202 222 (9.4%) (8.3%) (7.9%) (7.9%) (7.8%) (10.0%) (9.0%) $20,000 to $29,999 870 952 942 459 425 399 469

129

(18.2%) (18.5%) (19.4%) (17.0%) (15.5%) (17.8%) (18.9%) $30,000 to $39,999 686 649 729 409 440 373 435 (14.3%) (12.6%) (15.0%) (15.1%) (16.1%) (16.6%) (17.6%) $40,000 to $49,999 460 604 538 335 300 342 312 (9.6%) (11.7%) (11.10 (12.4%) (10.9%) (15.3%) (15.3%) %) $50,000 to $59,999 390 450 531 309 257 253 273 (8.1%) (8.8%) (10.9%) (11.5%) (9.4%) (11.3%) (11.0%) $60,000 to $80,000 498 616 722 441 459 235 280 (10.4%) (11.9%) (14.9%) (16.4%) (16.8%) (10.5%) (11.3%) $80,000 and Above 689 858 403 238 278 175 219 (14.4%) (16.8%) (17.3%) (21.2%) (23.5%) (13.1%) (13.8%)

Personal Income No Income 231 227 258 127 209 109 82 (4.8%) (4.4%) (4.8%) (4.1%) (6.6%) (3.6%) (2.5%) Less than $5,000 189 169 152 93 71 80 111 (3.9%) (3.3%) (2.9%) (3.0%) (2.2%) (2.7%) (3.4%) $5,000 to $9,999 640 569 550 228 253 223 234 (13.2%) (11.1%) (10.3%) (7.4%) (8.0%) (7.5%) (7.2%) $10,000 to $14,999 928 967 837 452 492 442 382 (19.2%) (18.8%) (15.7%) (14.7%) (15.5%) (14.8%) (11.8%) $15,000 to $19,999 538 564 542 360 359 312 401 (11.1%) (10.1%) (10.2%) (11.7%) (11.3%) (10.4%) (12.4%) $20,000 to $29,999 798 821 974 475 490 506 564 (16.5%) (15.9%) (18.3%) (15.5%) (15.4%) (16.9%) (17.4%) $30,000 to $39,999 541 600 679 478 416 411 482 (11.2%) (11.7%) (12.7%) (15.5%) (13.1%) (13.7%) (14.9%) $40,000 to $49,999 297 408 479 248 265 268 320 (6.1%) (7.9%) (9.0%) (8.1%) (8.3%) (8.9%) (9.9%)

130

$50,000 to $59,999 271 318 308 212 230 174 223 (5.6%) (6.2%) (5.8%) (6.9%) (7.2%) (5.8%) (6.9%) $60,000 to $80,000 236 272 347 222 223 231 202 (4.9%) (5.3%) (6.5%) (7.2%) (7.0%) (7.7%) (6.2%) $80,000 and Above 168 227 211 180 173 238 237 (3.5%) (4.4%) (4.0%) (5.9%) (5.4%) (8.0%) (7.3%)

Household Education Less than Secondary School 1451 1409 1168 630 640 563 594 Education (27.4%) (19.3%) (20.1%) (18.6%) (18.3%) (16.7%) (16.1%) Secondary School Graduation 848 776 725 449 389 411 502 (16.0%) (12.3%) (12.5%) (13.3%) (11.1%) (12.2%) (13.6%) Some Post-Secondary 346 372 327 177 199 191 194 Education (6.5%) (5.4%) (5.6%) (5.2%) (5.7%) (5.7%) (5.3%) Post-Secondary Education 2648 3280 3593 2126 2274 2207 2400 (50.0%) (55.4%) (61.8%) (62.9%) (64.9%) (65.5%) (65.0%)

Personal Education Less than Secondary School 2313 2380 2086 1225 1138 1093 1150 Education (43.2%) (34.0%) (33.6%) (33.9%) (30.9%) (30.8%) (29.8%) Secondary School Graduation 874 994 909 552 574 545 680 (16.3%) (15.3%) (14.6%) (15.3%) (15.6%) (15.4%) (17.6%) Some Post-Secondary 320 349 369 207 251 227 211 Education (6.0%) (5.8%) (5.9%) (5.7%) (6.8%) (6.4%) (5.5%) Post-Secondary Education 1847 2299 2843 1628 1725 1684 1821 (34.5%) (41.4%) (45.8%) (45.1%) (46.8%) (47.5%) (47.2%)

131

Dietary Consumption of Fruit and Vegetables Consumes Less than 5 Servings 3194 3258 1931 2015 2093 1962 2312 per Day (59.7%) (57.8%) (55.0%) (58.2%) (59.7%) (58.6%) (63.2%) Consumes 5 to 10 Servings per 1957 2184 1491 1323 1289 1270 1262 Day (36.6%) (38.7%) (42.5%) (38.2%) (36.8%) (37.9%) (34.5%) Consumes More than 10 196 195 90 124 124 117 86 Servings per Day (3.7%) (3.5%) (2.6%) (3.6%) (3.5%) (3.5%) (2.4%)

National Language Competency (Conversational) English Only 2717 3004 3110 1755 1770 1758 1847 (50.4%) (49.9%) (49.6%) (47.8%) (47.2%) (49.1%) (47.2%) French Only 698 700 864 484 448 420 396 (13.0%) (11.6%) (13.8%) (13.2%) (11.9%) (11.7%) (10.1%) English and French Only 666 772 798 478 472 444 493 (12.4%) (12.8) (12.7%) (13.0%) (12.6%) (12.4%) (12.6%) English and French and Other 196 250 251 128 191 126 138 (3.6%) (4.1%) (4.0%) (3.5%) (5.1%) (3.5%) (3.5%) English and Other (Not French) 841 979 1062 642 692 576 854 (15.6%) (16.3%) (16.9%) (17.5%) (18.5%) (16.1%) (21.8%) French and Other (Not English) 62 46 48 44 39 36 30 (1.2%) (0.8%) (0.8%) (1.2%) (1.0%) (1.0%) (0.8%) Neither English or French 206 270 140 141 137 222 156 (Other) (3.8%) (4.5%) (2.2%) (3.8%) (3.7%) (6.2%) (4.0%)

132

2. NON-MODIFIABLE RISK FACTOR VARIABLES Age 12 to 19 55 57 47 38 23 34 28 (1.0%) (0.9%) (0.7%) (1.0%) (0.6%) (0.9%) (0.7%) 20 to 29 115 127 181 108 83 59 74 (2.1%) (2.0%) (2.8%) (2.8%) (2.1%) (1.6%) (1.8%) 30 to 39 354 344 290 173 166 164 206 (6.5%) (5.5%) (4.5%) (4.5%) (4.3%) (4.5%) (5.1%) 40 to 49 752 773 786 444 450 388 457 (13.9%) (12.4%) (12.1%) (11.6%) (11.6%) (10.5%) (11.4%) 50 to 59 1157 1477 1517 930 1006 828 958 (21.3%) (23.7%) (23.4%) (24.3%) (26.0%) (22.5%) (23.8%) 60 to 69 1316 1669 1692 1012 1031 1053 1093 (24.3%) (26.8%) (26.1%) (26.5%) (26.6%) (28.6%) (27.1%) 70 to 79 1245 1330 1405 797 737 731 810 (23.0%) (21.3%) (21.6%) (20.8%) (19.0%) (19.9%) (20.1%) 80 and Above 429 461 575 323 380 425 401 (7.9%) (7.8%) (8.9%) (8.4%) (9.8%) (11.5%) (10.0%)

Sex Male 2839 3292 3503 2089 2062 2022 2315 (52.3%) (52.7%) (53.9%) (54.6%) (53.2%) (54.9%) (57.5%) Female 2584 2947 2990 1736 1814 1660 1713 (47.7%) (47.3%) (46.1%) (45.4%) (46.8%) (45.1%) (42.5%)

Racial/Cultural Origin White 4640 5000 5340 2881 2908 2855 2938 (87.5%) (84.8%) (87.5%) (82.4%) (80.8%) (83.4%) (78.3%) Black 108 98 117 128 81 102 132

133

(2.0%) (1.7%) (1.9%) (3.7%) (2.3%) (3.0%) (3.5%) Korean 17 4 17 8 12 39 30 (0.3%) (0.1%) (0.3%) (0.2%) (0.3%) (1.1%) (0.8%) Filipino 50 44 66 42 34 49 69 (0.9%) (0.7%) (1.1%) (1.2%) (0.9%) (1.4%) (1.8%) Japanese 10 12 23 6 8 9 21 (0.2%) (0.2%) (0.4%) (0.2%) (0.2%) (0.3%) (0.6%) Chinese 123 187 65 84 116 96 121 (2.3%) (3.2%) (1.1%) (2.4%) (3.2%) (2.8%) (3.2%) South Asian 169 233 235 205 235 122 258 (3.2%) (3.9%) (3.9%) (5.9%) (6.5%) (3.6%) (6.9%) Southeast Asian 26 38 41 23 25 35 69 (0.5%) (0.6%) (0.7%) (0.7%) (0.7%) (1.0%) (1.8%) Arab 30 15 21 17 13 19 19 (0.6%) (0.3%) (0.3%) (0.5%) (0.4%) (0.6%) (0.5%) West Asian 5 9 11 6 64 18 16 (0.1%) (0.2%) (0.2%) (0.2%) (1.8%) (0.5%) (0.4%) Latin American 19 21 41 58 50 37 14 (0.4%) (0.4%) (0.7%) (1.7%) (1.4%) (1.1%) (0.4%) Other 75 152 89 12 37 24 32 (1.4%) (2.6%) (1.5%) (0.3%) (1.0%) (0.7%) (0.9%) Multiple 33 90 37 26 16 17 32 (0.6%) (1.5%) (0.6%) (0.7%) (0.4%) (0.5%) (0.9%)

Ethnicity (Ethnic Origin) Canadian 1382 1197 1318 859 747 748 737 (23.7%) (19.0%) (19.4%) (23.1%) (19.5%) (20.6%) (19.1%) French 698 852 967 459 435 502 486 (12.0%) (13.5%) (14.2%) (12.4%) (11.4%) (13.8%) (12.6%) English 1077 1193 1340 686 712 656 657

134

(18.5%) (18.9%) (19.7%) (18.5%) (18.6%) (18.1%) (17.0%) German 353 441 434 249 251 242 270 (6.1%) (7.0%) (6.4%) (6.7%) (6.6%) (6.7%) (7.0%) Scottish 688 770 831 426 481 407 434 (11.8%) (12.2%) (12.2%) (11.5%) (12.6%) (11.2%) (11.3%) Irish 591 630 739 341 397 408 375 (10.2%) (10.0%) (10.9%) (9.2%) (10.4%) (11.2%) (9.7%) Italian 235 309 254 160 166 147 175 (4.0%) (4.9%) (3.7%) (4.3%) (4.3%) (4.0%) (4.5%) Ukrainian 148 180 163 86 81 84 99 (2.5%) (2.8%) (2.4%) (2.3%) (2.1%) (2.3%) (2.6%) Dutch (Netherlands) 156 148 176 85 74 77 67 (2.7%) (2.4%) (2.6%) (2.3%) (1.95) (2.1%) (1.7%) Chinese 126 210 77 95 134 104 127 (2.2%) (3.3%) (1.1%) (2.6%) (3.5%) (2.9%) (3.3%) Jewish 40 30 54 22 49 17 45 (0.7%) (0.5%) (0.8%) (0.6%) (1.3%) (0.5%) (1.2%) Polish 80 94 111 54 60 74 82 (1.4%) (1.5%) (1.6%) (1.5%) (1.6%) (2.0%) (2.1%) Portuguese 53 40 82 23 33 45 34 (0.9%) (0.6%) (1.2%) (0.6%) (0.9%) (1.3%) (0.9%) South Asian 194 206 246 169 202 118 269 (3.3%) (3.3%) (3.6%) (4.6%) (5.3%) (3.3%) (7.0%)

3. IMMIGRANT-SPECIFIC RISK FACTOR VARIABLES Country Origin Other North America 78 47 59 49 33 69 27 (5.6%) (3.0%) (4.8%) (4.6%) (2.9%) (6.5%) (2.2%) South, Central America and the 178 156 133 217 162 132 151 Caribbean

135

(12.9%) (9.9%) (10.7%) (20.5%) (14.0%) (12.5%) (12.1%) Europe 670 792 685 396 441 442 421 (48.4%) (50.1%) (55.3%) (37.3%) (38.1%) (41.8%) (33.6%) Africa 65 105 19 56 68 45 82 (4.7%) (6.6%) (1.5%) (5.3%) (5.9%) (4.3%) (6.5%) Asia 386 459 329 337 446 358 561 (27.9%) (29.0%) (26.6%) (31.8%) (38.5%) (33.9%) (44.8%) Oceania 8 23 14 6 7 11 10 (0.6%) (1.5%) (1.1%) (0.6%) (0.6%) (1.0%) (0.8%)

Time Since Immigration to Canada 0 to 9 Years Since Immigration 108 147 137 134 127 130 121 (8.1%) (9.4%) (9.1%) (13.5%) (11.3%) (12.6%) (9.9%) 10 to 19 Years Since 184 238 246 172 192 149 285 Immigration (13.7%) (15.1%) (16.3%) (17.3%) (17.1%) (14.4%) (23.2%) 20 to 29 Years Since 266 244 152 182 116 138 242 Immigration (19.9%) (15.5%) (10.1%) (18.3%) (10.3%) (13.4%) (19.7%) 30 to 39 Years Since 279 417 400 192 350 187 197 Immigration (20.8%) (26.5%) (26.6%) (19.3%) (31.2%) (18.1%) (16.0%) 40 or More Years Since 502 527 571 314 338 428 383 Immigration (37.5%) (33.5%) (37.9%) (31.6%) (30.1%) (41.5%) (31.2%)

Age at Immigration to Canada 12 to 19 292 392 354 165 243 179 179 (21.8%) (24.9%) (23.5%) (16.6%) (21.5%) (17.3%) (14.6%)

136

20 to 29 445 555 507 311 425 332 377 (33.2%) (35.3%) (33.7%) (31.3%) (37.6%) (32.1%) (30.7%) 30 to 39 313 330 342 257 218 229 345 (23.4%) (21.0%) (22.7%) (25.9%) (19.3%) (22.2%) (28.1%) 40 to 49 163 151 148 103 120 158 155 (12.2%) (9.6%) (9.8%) (10.4%) (10.6%) (15.3%) (12.6%) 50 to 69 113 126 147 146 117 100 167 (8.4%) (8.0%) (9.8%) (14.7%) (10.4%) (9.7%) (13.6%) 70 and Above 13 19 8 11 7 35 6

(1.0%) (1.2%) (0.5%) (1.1%) (0.6%) (3.4%) (0.5%) a Notes : (1) All results were calculated using rescaled master weights (see Section 4.3.3 ); (2) Multiple options available; number represents total respondents with more than one ethnicity option selected.

137

References

1 Shaw JE, Sicree RA & Zimmet PZ. Global estimates of the prevalence of diabetes for 2010 and 2030. Diabetes Res Clin Pract 2010;87:4 –14.

2 Public Health Agency of Canada. Analysis of 2008/09 data from the Canadian Chronic Disease Surveillance System. Ottawa, ON: Statistics Canada; 2011.

3 Informetrica Limited. Economic Cost of Diabetes in Canada: An Overview. Toronto, ON: Canadian Diabetes Association; 2009.

4 Diabète Québec. Diabetes: Canada at the Tipping Point, Charting a New Path. Toronto, ON: Canadian Diabetes Association; 2011.

5 American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care . 2006;29:43-48.

6 Lund SS, Vaag AA. Intensive glycemic control and the prevention of cardiovascular events: implications of the ACCORD, ADVANCE, and VA Diabetes Trials: a position statement of the American Diabetes Association and a scientific statement of the American College of Cardiology Foundation and the American Heart Association. Diabetes Care . 2009;32(7):90-91.

7 Canadian Diabetes Association Web site. The Prevalence and Costs of Diabetes. http://www.diabetes.ca/diabetes-and-you/what/prevalence/. Accessed June 1, 2013.

8 National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). DCCT and EDIC: The Diabetes Control and Complications Trial and Follow-up Study. Bethesda, M.D.: National Diabetes Information Clearinghouse (NDIC), U.S. Department of Health and Human Service; 2008.

9 Golden SH, Lazo M, Carnethon M, et al. Examining a bidirectional association between depressive symptoms and diabetes. JAMA. 2008;299:2751-2759.

10 Golden SH. A review of the evidence for a neuroendocrine link between stress, depression and diabetes mellitus. Curr Diabetes Rev . 2007;3(4):252-259.

11 Carnethon MR, Kinder LS, Fair JM, Stafford RS, Fortmann SP. Symptoms of depression as a risk factor for incident diabetes: findings from the National Health and Nutrition Examination Epidemiologic follow-up study, 1971-1992. Am J Epidemiol . 2003;158(5):416-423.

12 El-Zayadi A. Insulin resistance. Arab Journal of Gastroenterology . 2010;11:66-69.

13 Winokur A, Maislin G, Phillips JL, Amsterdam JD. Insulin resistance after oral glucose tolerance testing in patients with major depression. Am J Psychiatry . 1988;145(3):325-330.

14 Maes M, Vandewoude M, Schotte C, Martin M,Blockx P. Positive relationship between the catecholaminergic turnover and the DST results in depression. Psychol Med . 1990;20(3):493-499.

138

15 Statistics Canada Web site. Dependency, chronic conditions and pain in seniors. http://www.statcan.gc.ca/pub/82-003-s/2005000/pdf/9087-eng.pdf. Accessed January 14, 2013.

16 Grumbach, Kevin. Chronic Illness, Comorbidities, and the Need for Medical Generalism. Annals of Family Medicine . 2003;1:4-7.

17 Rewers A. Epidemiology of acute complications: Diabetic ketoacidosis, hyperglycemic hyperosmolar state and hypoglycemia. In: Ekoe J-M, Rewers M, Williams R, Zimmet P, eds. The Epidemiology of Diabetes Mellitus . 2nd ed. West Sussex, UK: Wiley-Blackwell; 2008:577-602.

18 Public Health Agency of Canada. Life with arthritis in Canada: A personal and public health challenge. Ottawa, ON: Statistics Canada; 2010.

19 Ehrlich SF, Quesenberry CP Jr., Van Den Eeden SK, Shan K, Ferrara A. Patients diagnosed with diabetes are at increased risk for asthma, chronic obstructive pulmonary disease, pulmonary fibrosis, and pneumonia but not long cancer. Diabetes Care . 2010;33(1):55-60.

20 Bowker SL, Majumdar SR, Veugelers P, Johnson JA. Increased cancer-related mortality for patients with T2DM who use sulfonylureas or insulin. Diabetes Care . 2006;29(8):254-258.

21 Talbot F, Nouwen A. A review of the relationship between depression and diabetes in adults: is there a link? Diabetes Care . 2000;23(10):1556-1562

22 Kemp DE, Calabrese JR, Ismail-Beigi F. Depressive symptoms and diabetes. JAMA. 2008;300:2115- 2115.

23 Elbein S. Genetic Factors Contributing to Type 2 Diabetes across Ethnicities. J Diabetes Sci Technol . 2009;3(4);685-689.

24 Nwasuruba C, Khan M, Egede L. Racial/Ethnic differences in multiple self-care behaviours in adults with diabetes. J Gen Intern Med. 2007;22(1):115-120.

25 Gushulak BD, Pottie K, Hatcher Roberts J, et al. Migration and health in Canada: health in the global village. CMAJ . 2011;183:952-958.

26 Manuel DG. Diabetes health status and risk factors. In: Hux JE, Booth GL, Slaughter PM, et al., editors. Diabetes in Ontario. Toronto, ON: Institute for Clinical Evaluative Sciences; 2003, p. 77–94.

27 Lam J, Yip T, Gee G. The Physical and Mental Health Effects of Age of Immigration, Age, and Perceived Difference in Social Status Among First Generation Asian Americans. Asian American Journal of Psychology . 2012;3:29-43.

28 Malenfant EC, Lebel A, Martel L. Projections of the Diversity of the Canadian Population, 2006- 2031. Ottawa: Statistics Canada, Catalogue 91-551-X; 2010.

29 Lou Y, Beaujot R. What Happens to the ‘Healthy Immigrant Effect’?: The Mental Health of Immigrants to Canada. London, ON: Population Studies Centre, University of Western Ontario; 2005.

30 Chen J, Wilkins R, Ng E. Health expectancy by immigrant status. Health Reports. 1996;8(3):29-37.

139

31 Newbold K. Chronic conditions and the healthy immigrant effect: Evidence from Canadian immigrants. Journal of Ethnic and Migration Studies . 2006;32:765-784.

32 Chiu M, Austin PC, Manuel DG, Tu JV. Cardiovascular risk factor profiles of recent immigrants vs. long-term residents of Ontario: A multi-ethnic study. Can J Cardiol. 201228(1):20-26.

33 Statistics Canada. Access to Health Care Services in Canada, January to December 2005. Ottawa, ON: Statistics Canada; 2006, Catalogue No. 82-575-XIE2006002, p. 9.

34 Gagnon AJ. The Responsiveness of the Canadian Health Care System Towards Newcomers. Ottawa, ON: Royal Commission on the Future of Health Care in Canada (Romanow Commission), Health Canada; 2002 .

35 Chalabian J, Dunnington G. Impact of language barrier on quality of patient care, resident stress and teaching. Teaching and Learning in Medicine. 1997;9:84-90.

36 Chugh U, Dillmann E, Kurtz SM, Lockyer J, Parboosingh J. Multicultural issues in medical curriculum: implications for Canadian physicians. Med. Teach. 1993;15:83-91.

37 Bowen S. Access to health services for underserved populations in Canada. In: Certain Circumstances: Equity in and Responsiveness of the Health Care System to the Needs of Minority and Marginalized Populations. Ottawa, ON: Health Canada; 2001, p. 1–60.

38 Indian and Northern Affairs Canada. Words first: An evolving terminology relating to Aboriginal Peoples in Canada. Ottawa, ON: Indian and Northern Affairs Canada; 2004.

39 Statistics Canada Web site. Aboriginal Affairs and Northern Development Canada. Terminology. http://www.aadnc-aandc.gc.ca/eng/1100100014642/1100100014643. Accessed June 2, 2013.

40 Dyck RF, Osgood N, Lin TH. Epidemiology of diabetes mellitus among First Nations and non-First Nations adults. CMAJ. 2010;182;3:249-256.

41 First Nations Centre. National Aboriginal Health Organization First Nations regional longitudinal health survey (RHS) 2002/03-results for adults, youth and children living in First Nations communities. Ottawa, ON: First Nations Centre, National Aboriginal Health Organization; 2005.

42 Shah B, Anand S, Zinman B, Duong-Hua M. Diabetes in Ontario: Diabetes and First Nations People. Practice Atlas . 2003;13:231.

43 Green C, Blanchard J, Young TK. The epidemiology of diabetes in the Manitoba-registered First Nation population: Current patterns and comparative trends. Diabetes Care. 2003;26:1993-1998.

44 Hanley AJ, Harris SB, Mamakeesick M, et al. Complications of T2DM among Aboriginal Canadians. Prevalence and associated risk factors. Diabetes Care . 2005;28(8):2054-2057.

45 Harris SB, Naqshbandi M, Bhattacharyya O. Major gaps in diabetes clinical care among Canada's First Nations: Results of the CIRCLE study. Diabetes Res Clin Pract. 2011;92(2):272-279.

140

46 Martens PJ, Martin BD, O'Neil JD. Diabetes and adverse outcomes in a First Nations population: Associations with healthcare access, and socioeconomic and geographical factors. Can J Diabetes. 2007;313:223-232.

47 Ross SA. McKenna A. Mozejko S. Diabetic retinopathy in Native and nonnative Canadians. Experimental Diabetes Research. 2007;76:271.

48 Naqshbandi M, Harris SB, Esler JG. Global complication rates of type 2 diabetes in Indigenous peoples: A comprehensive review. Diabetes Res Clin Pract. 2008;82(1):1-17.

49 Oster RT, Johnson JA, Hemmelgarn BR. Recent epidemiologic trends of diabetes mellitus among status Aboriginal adults. CMAJ. 2011;183:803-808.

50 Amed S, Dean HJ, Panagiotopoulos C et al. T2DM, medication-induced diabetes, and monogenic diabetes in Canadian children: A prospective national surveillance study. Diabetes Care. 2010;33(4):786-791.

51 Young K, Reading J, Elias B, O’Neil J. Type 2 diabetes mellitus in Canada’s First Nations: status of an epidemic in progress. CMAJ. 2000;163(5):561-566.

52 Health Canada. Diabetes Among Aboriginal People in Canada: The Evidence . Ottawa, ON: Health Canada; 2001.

53 Neel JV. Diabetes mellitus: A “thrifty” genotype rendered detrimental by “progress”? Am J Hum Genet. 1962;14:353-362.

54 Campbell LV. The thrifty gene hypothesis: maybe everyone is right? International Journal of Obesity . 2008;32:723-724.

55 Gracey M, King M. Indigenous health part 1: Determinants and disease patterns. Lancet. 2009;374:6575.

56 First Nations Information Governance Centre. First Nations Regional Longitudinal Health Survey (RHS) 2002/03: Results for adults, youth and children living in First Nations communities. Ottawa, ON: Assembly of First Nations, First Nations Information Governance Centre; 2007.

57 Shah BR, Gunraj N, Hux JE. Markers of access to and quality of primary care for Aboriginal people in Ontario, Canada. Am J Public Health. 2003;93(5)798-802.

58 Deshpande AD, Harris-Hayes M, Schootman M. Epidemiology of diabetes and diabetes-related complications. Physical Therapy. 2008;88(11):1254-1264.

59 Buijsse B, Simmons, RK, Griffin SJ, Schulze MB. Risk assessment tools for identifying individuals at risk of developing T2DM. Epidemiologic Reviews. 2011;33(1):46-62.

60 Schwarz PE, Schwarz J, Schuppenies A, et al. Development of a diabetes prevention management program for clinical practice. Public Health Rep . 2007;122(2):258–263.

61 Lindstrom J, Tuomilehto J. The diabetes risk score: a practical tool to predict T2DM risk. Diabetes Care . 2003;26(3):725–731.

141

62 Simmons RK, Harding AH, Wareham NJ, et al. Do simple questions about diet and physical activity help to identify those at risk of T2DM? Diabet Med . 2007;24(8):830–835.

63 Rahman M, Simmons RK, Harding AH, et al. A simple risk score identifies individuals at high risk of developing type 2 diabetes: a prospective cohort study. Fam Pract . 2008;25(3):191–196.

64 Hippisley-Cox J, Coupland C, Robson J, et al. Predicting risk of T2DM in England and Wales: prospective derivation and validation of QDScore. BMJ . 2009;338:b880.

65 Wilson PW, Meigs JB, Sullivan L, et al. Prediction of incident diabetes mellitus in middle-aged adults: the Framingham Offspring Study. Arch Intern Med . 2007;167(10):1068–1074.

66 Schulze MB, Hoffmann K, Boeing H, et al. An accurate risk score based on anthropometric, dietary, and lifestyle factors to predict the development of T2DM. Diabetes Care . 2007;30(3):510–515.

67 Stern MP, Williams K, Haffner SM. Identification of persons at high risk for T2DM: do we need the oral glucose tolerance test? Ann Intern Med . 2002;136(8):575–581.

68 Park PJ, Grifﬁn SJ, Sargeant L, Wareham NJ. The performance of a risk score in predicting undiagnosed hyperglycemia. Diabetes Care. 2002;25:984–988.

69 Balkau B, Lange C, Fezeu L, et al. Predicting diabetes: clinical, biological, and genetic approaches: data from the Epidemiological Study on the Insulin Resistance Syndrome (DESIR). Diabetes Care . 2008;31(10):2056–2061.

70 Alssema M, Feskens EJ, Bakker SJ, et al. Finnish questionnaire reasonably good predictor of the incidence of diabetes in The Netherlands [in Dutch]. Ned Tijdschr Geneeskd . 2008;152(44):2418– 2424.

71 Abdul-Ghani MA, Lyssenko V, Tuomi T, et al. Fasting versus post load plasma glucose concentration and the risk for future T2DM: results from the Botnia Study. Diabetes Care . 2009;32(2):281–286.

72 Herman WH. Predicting risk for diabetes: choosing (or building) the right model. Ann Intern Med . 2009;150(11):812–814.

73 Robinson CA, Agarwal G. Validating the CANRISK prognostic model for assessing diabetes risk in Canada’s multi-ethnic population. Chron Dis Inj Can. 2011;32:19-31.

74 Public Health Agency of Canada Web site. Diabetes, are you at risk? www.phac-aspc.gc.ca/cd- mc/diabetes-diabete/canrisk/index-eng.php. Accessed June 1 2013.

75 Canadian Diabetes Association Clinical Practice Guidelines Expert Committee. Canadian Diabetes Association 2013 Clinical Practice Guidelines for the Prevention and Management of Diabetes in Canada. Can J Diabetes. 2013;37(suppl 1):1-212.

76 Schulze MB, Hoffmann K, Boeing H, et al. An accurate risk score based on anthropometric, dietary, and lifestyle factors to predict the development of T2DM. Diabetes Care . 2007;30(3):510–515.

77 Sun F, Tao Q, Zhan S. An accurate risk score for estimation 5-year risk of T2DM based on a health

142

screening population in Taiwan. Diabetes Res Clin Pract . 2009;85(2):228–234.

78 Zimmet P, Alberti KG, Shaw J. Global and societal implications of the diabetes epidemic. Nature . 2001;414(6865):782-787.

79 Egede L, Dagogo-Jack S. Epidemiology of T2DM: focus on ethnic minorities. Med Clin North Amer 2005;89(5):949-975.

80 Hu FB, Manson JE, Stampfer MJ, et al. Diet, Lifestyle, and the Risk of Type 2 Diabetes Mellitus in Women. New England Journal of Medicine. 2001;345(11):790-797.

81 Rewers M. Hamman RF. Risk factors for non-insulin-dependent diabetes. In: Harris MI. Cowie CC. Stern MP, et al. eds. Diabetes in America. 2nd ed. Bethesda. M.D.: National Institutes of Health. National Institute of Diabetes and Digestive and Kidney Diseases; 1995:179-220.

82 Kaye SA. Folsom AR, Sprafka JM, et al. Increased incidence of diabetes mellitus in relation to abdominal adiposity in older women. J Clin Epidemiol . 1991;44:329-334.

83 Kissebah AH, Krakower GR. Regional adiposity and morbidity. Physiol Rev . 1994;74:761–809. 84 Schmidt MI, Duncan BB, Canani LH, Karohl C, Chambless L. Association of waist–hip ratio with diabetes mellitus. Strength and possible modifiers. Diabetes Care. 1992;15:912–914.

85 Cassano PA, Rosner B, Vokonas PS, Weiss ST. Obesity and body fat distribution in relation to the incidence of non-insulin dependent diabetes mellitus. A prospective cohort study of men in the normative aging study. Am J Epidemiol. 1992;136:1474–1486.

86 Snijder MB, Dekker JM, Visser M, et al. Associations of hip and thigh circumferences independent of waist circumference with the incidence of T2DM: the Hoorn Study. Am J Clin Nutr. 2003;77:1192– 1197.

87 Mokdad AH, Bowman BA, Ford ES, et al. The continuing epidemics of obesity and diabetes in the United States. JAMA. 2001;286:1195 –1200.

88 Soumaya K. Molecular mechanisms of insulin resistance in diabetes. Adv Exp Med Biol. 2012;771:240- 51.

89 Ginter E, Simko V. Type 2 diabetes mellitus, pandemic in 21st century. Adv Exp Med Biol . 2012;771:42- 50.

90 Sankhla M, Sharma TK, Gahlot S, Rathor JS, Vardey SK, Sinha M, et al. The ominous link between obesity and abdominal adiposity with diabetes and diabetic dyslipidemia in diabetic population of developing country. Clin Lab . 2013;59(1-2):155-161.

91 Tirosh A, Shai I, Tekes-Manova D, et al. Normal fasting plasma glucose levels and T2DM in young men. N Engl J Med. 2005;353:1454 –1462.

92 Martin MM, Martin ALA. Obesity, hyperinsulinism, and diabetes mellitus in childhood. J Pediatr. 1973;82:192–201.

143

93 Rosenbloom AL. Age-related plasma insulin response to glucose ingestion in children and adolescents. Metab Nutr Pediatr. 1974;2:1210.

94 Rosenbloom AL, Joe JR, Young, RS, Winter WE. Emerging epidemic of T2DM in youth. Diabetes Care. 1999;22(2);345-354.

95 Must A, Spadano J, Coakley EH, et al. The disease burden associated with overweight and obesity. JAMA . 1999;282:1523 –1529.

96 Kahn R, Alperin P, Eddy D, et al. Age at initiation and frequency of screening to detect T2DM: a cost- effectiveness analysis. Lancet. 2010;375:1365 –1374.

97 Eckel RH, Kahn SE, Ferrannini E, Goldfine AB, Nathan DM, Schwartz MW, et al. Obesity and T2DM: What can be unified and what needs to be individualized? Diabetes Care. 2011;34(6)1424-1430.

98 Public Health Agency of Canada Web site. Obesity in Canada – Snapshot. http://www.phac- aspc.gc.ca/publicat/2009/oc/index-eng.php. Accessed June 1 2013.

99 Fox CS, Sullivan L, D'Agostino RB Sr, et al. The significant effect of diabetes duration on coronary heart disease mortality: the Framingham Heart Study. Diabetes Care 2004;27:704 –708.

100 Eckel RH, Alberti KG, Grundy SM, Zimmet PZ. The metabolic syndrome. Lancet . 2010;375(9710):181- 183.

101 Björntorp P. Metabolic implications of body fat distribution. Diabetes Care .1991; 14:1132 –1143.

102 Freemantle N, Holmes J, Hockey A, Kumar S. How strong is the association between abdominal obesity and the incidence of T2DM? International Journal of Clinical Practice. 2008;62(9):1391-1396.

103 Deng Y, Scherer PE. Adipokines as novel biomarkers and regulators of the metabolic syndrome. Ann N Y Acad Sci . 2010;1212:1 –19.

104 Larson-Meyer DE, Newcomer BR, Ravussin E, et al. Intrahepatic and intramyocellular lipids are determinants of insulin resistance in prepubertal children. Diabetologia. 2011;54:869 –875.

105 Bjorntorp P. Metabolic implications of body fat distribution. Diabetes Care. 1991;14(12):1132–43.

106 Boden G. Role of fatty acids in the pathogenesis of insulin resistance and NIDDM. Diabetes . 1997;46(1):3–10.

107 McGarry JD. Banting lecture 2001: dysregulation of fatty acid metabolism in the etiology of T2DM. Diabetes . 2002;51(1):7–18.

108 Bournat JC, Brown CW. Mitochondrial dysfunction in obesity. Curr Opin Endocrinol Diabetes Obes . 2010;17:446 –452

109 Tuomilehto J, Lindstrom J, Eriksson JG, et al. Prevention of type 2 diabetes mellitus by changes in lifestyle among subjects with impaired glucose tolerance. N Engl J Med. 2001;344:1343–1350.

144

110 Knowler WC, Barrett-Connor E, Fowler SE, et al. Reduction in the incidence of T2DM with lifestyle intervention or metformin. N Engl J Med. 2002;346:393–403.

111 Qin L, Knol MJ, Corpeleijn E, Stolk RP. Does physical activity modify the risk of obesity for T2DM: A review of epidemiological data. European Journal of Epidemiology. 2010;25(1);5-12.

112 LaMonte MJ, Blair SN, Church TS. Physical activity and diabetes prevention. J Appl Physiol. 2005;99(3):1205–1213.

113 Hu FB, Sigal RJ, Rich-Edwards JW, et al. Walking compared with vigorous physical activity and risk of T2DM in women: a prospective study. JAMA .1999;282(15):1433–1439.

114 Hu FB, Leitzmann MF, Stampfer MJ, et al. Physical activity and television watching in relation to risk for T2DM in men. Arch Intern Med . 2001;161(12):1542–1548.

115 Hu G, Lindstrom J, Valle TT, et al. Physical activity, body mass index, and risk of T2DM in patients with normal or impaired glucose regulation. Arch Intern Med . 2004;164(8):892–896.

116 Rana JS, Li TY, Manson JE, et al. Adiposity compared with physical inactivity and risk of T2DM in women. Diabetes Care . 2007;30(1):53–8.

117 Kriska AM, Saremi A, Hanson RL, et al. Physical activity, obesity, and the incidence of T2DM in a high-risk population. Am J Epidemiol . 2003;158(7):669–675.

118 Weinstein AR, Sesso HD, Lee IM, et al. Relationship of physical activity vs body mass index with T2DM in women. JAMA . 2004;292(10):1188–94.

119 Hu FB, Manson JE, Stampfer MJ, et al. Diet, lifestyle, and the risk of T2DM in women. N Engl J Med . 2001;345(11):790–797.

120 Meisinger C, Lowel H, Thorand B, et al. Leisure time physical activity and the risk of T2DM in men and women from the general population. The MONICA/KORA Augsburg cohort study. Diabetologia . 2005;48(1):27–34.

121 Ross R, Dagnone D, Jones PJ, et al. Reduction in obesity and related comorbid conditions after diet- induced weight loss or exercise-induced weight loss in men. A randomized, controlled trial. Ann Intern Med . 2000;133(2):92–103.

122 Ross R, Janssen I, Dawson J, et al. Exercise-induced reduction in obesity and insulin resistance in women: a randomized controlled trial. Obes Res . 2004;12(5):789–798.

123 Larsen TM, Dalskov SM, van Baak M, et al. Diet, Obesity, and Genes (Diogenes) Project. Diets with high or low protein content and glycemic index for weight-loss maintenance. N Engl J Med. 2010;363:2102–2113

124 Papadaki A, Linardakis M, Larsen TM, et al., DiOGenes Study Group. The effect of protein and glycemic index on children ’s body composition: the DiOGenes randomized study. Pediatrics. 2010;126:1143 –1152.

145

125 Psaltopoulou T, Ilias I, Alevizaki M. The role of diet and lifestyle in primary, secondary, and tertiary diabetes prevention: A review of meta-analyses. The Review of Diabetic Studies. 2010;7(1):26-35.

126 Qi L, Liang J.Interactions between genetic factors that predict diabetes and dietary factors that ultimately impact on risk of diabetes. Current Opinion in Lipidology, 2010;21(1):31-37.

127 Sjöström L, Lindroos AK, Peltonen M, et al. Swedish Obese Subjects Study Scientific Group. Lifestyle, diabetes, and cardiovascular risk factors 10 years after bariatric surgery. N Engl J Med . 2004;351:2683 –2693.

128 Glasgow R, Mullooly J, Vogt T. Biochemical validation of smoking status in public health settings: pros, cons, and data fro m four low-intensity intervention trials. Addict Behav. 1993;18:511–527.

129 Haire-Joshu D, Glasgow RE, Tibbs TL. Smoking and diabetes. Diabetes Care . 1999;22(11):1887-1898.

130 Rimm E, Manson J, Stampfer M. Cigarette smoking and the risk of diabetes in women. Am J Public Health . 1993;83:211-214.

131 Rimm E, Chan J, Stampfer M, Colditz G, Willett W. Prospective study of cigarette smoking, alcohol use, and the risk of diabetes in men. BMJ. 1995;310:555–559.

132 Kawakami N, Takatsuka N, Shimizu H, Ishibashi H. Effects of smoking on incidence of non-insulin- dependent diabetes mellitus. Diabetes Care. 1997;16:103–109.

133 Targher G, Alberiche M, Zenere M, Bonadonna R, Muggeo M, Bonora E. Cigarette smoking and insulin resistance in patients with non-insulin-dependent diabetes mellitus . J Clin Endocrinol Metab. 1997;82:3619-3624.

134 Chan J, Rimm E, Colditz G, Stampfer M, Willett W. Obesity, fat distribution, and weight gain as risk factors for clinical diabetes in men. Diabetes Care . 1994;17:1–10.

135 Telner, A. Alcohol, diabetes and health: A review. Canadian Journal of Diabetes . 2002;26(3):378-381.

136 Centre for Addiction and Mental Health Web site. Canada’s Low-Risk Alcohol Drinking Guidelines. http://www.camh.ca/en/hospital/health_information/a_z_mental_health_and_addiction_information/al cohol/Pages/low_risk_drinking_guidelines.aspx. Accessed December 21, 2012.

137 Health Risks and Benefits of Alcohol Consumption. Alcohol Research & Health . 2000;24(1):5-11.

138 Zilkens RR , Puddley IB. Alcohol and cardiovascular disease - more thanone paradox to consider. Alcohol and T2DM - another paradox? J Cardiovasc Risk . 2003;10:25-30.

139 Howard AA, Arnsten JH, Gourevitch MN. Effect of alcohol consumption on diabetes mellitus: a systematic review. Ann Intern Med . 2004;140:211-219.

140 Agardh E, Allebeck P, Hallqvist J, Moradi T, Sidorchuk A. T2DM incidence and socio-economic position: A systematic review and meta-analysis. International Journal of Epidemiology, 2011;40(3):804-818.

141 Brown AF, Ettner SL, Piette J et al. Socioeconomic position and health among persons with diabetes

146

mellitus: a conceptual framework and review of the literature. Epidemiol Rev. 2004;26:63–77.

142 Agardh EE, Ahlbom A, Andersson T, et al. Explanations of socioeconomic differences in excess risk of T2DM in Swedish men and women. Diabetes Care. 2004;27:716–721.

143 Connolly V, Unwin N, Sherriff P, Bilous R, Kelly W. Diabetes prevalence and socioeconomic status: a population based study showing increased prevalence of type 2 diabetes mellitus in deprived areas. J Epidemiol Community Health. 2000;54:173–177.

144 Espelt A, Borrell C, Roskam AJ, et al. Socioeconomic inequalities in diabetes mellitus across Europe at the beginning of the 21st century. Diabetologia. 2008;51:1971–1979.

145 Evans JM, Newton RW, Ruta DA, MacDonald TM, Morris AD. Socio-economic status, obesity and prevalence of Type 1 and T2DM. Diabet Med . 2000;17:478–480.

146 Tang M, Chen Y, Krewski D. Gender-related differences in the association between socioeconomic status and self-reported diabetes. Int J Epidemiol. 2003;32:381–385.

147 Feinstein JS. The relationship between socioeconomic status and health: a review of the literature. Milbank Q. 1993;71:279–322.

148 Lantz PM, House JS, Lepkowski JM, Williams DR, Mero RP, Chen J. Socioeconomic factors, health behaviors, and mortality: results from a nationally representative prospective study of US adults. JAMA . 1998;279:1703–1708.

149 Wang Y, Beydoun MA. The obesity epidemic in the United States-gender, age, socioeconomic, racial/ethnic, and geographic characteristics: a systematic review and meta-regression analysis. Epidemiol Rev . 2007;29:6-28.

150 Daniels G. Present trends in the evaluation of psychic factors in diabetes mellitus. Psychosom Med . 1939;1:528–552.

151 Menninger W. The inter-relationships of mental disorders and diabetes mellitus. J Ment Sci . 1935;81:332–357.

152 Jacobson AM. Depression and diabetes. Diabetes Care . 1993;16:1621–1623.

153 Connolly V. ‘‘Sociotype’’: a key determinant of diabetes health. Br J Diab Vasc Dis . 2004;4:141–143.

154 Musselman DL, Betan E, Larsen H, Phillips LS. Relationship of depression to diabetes types 1 and 2: epidemiology, biology, and treatment. Biol Psychiatry. 2003;54:317–329.

155 Knol MJ, Twisk JWR, Beekman ATF, Heine RJ, Snoek FK, Pouwer F. Depression as a risk factor for the onset of T2DM. A meta-analysis. Diabetologia . 2006;49:837–845.

156 Cosgrove MP, Sargeant LA, Griffin SJ. Does depression increase the risk of developing T2DM? Occupational Medicine. 2008;58(1):7-14.

147

157 Engelgau MM, Narayan KM, Saaddine JB, Vinicor F. Addressing the burden of diabetes in the 21st century: Better care and primary prevention. Journal of the American Society of Nephrology. 2003;14(Suppl 2):s88-91.

158 Wetterhall SF, Olson DR, DeStefano F, Stevenson JM, Ford ES, German RR, Will JC, Newman JM, Sepe SJ, Vinicor F. Trends in diabetes and diabetic complications, 1980 –1987. Diabetes Care. 1992;15:960 –967.

159 Geiss LS, Engelgau M, Frazier E, Tierney E. Diabetes Surveillance, 1997. Atlanta, G.A.: Centers for Disease Control and Prevention, U.S. Department of Health and Human Services; 1997.

160 Klein R, Klein BEK, Moss SE. Visual impairment in diabetes. Ophthalmology . 1984;91:1 –9.

161 Centers for Disease Control and Prevention Website. Diabetes Data and Trends. http://www.cdc.nov/diabctes/statistics/newDataTrends.html. Accessed October 12, 2012.

162 Pirart J. Diabetes mellitus and its degenerative complications: a prospective study of 4,400 patients observed between 1947 and 1973 [in French]. Diabetes Metab . 1977;3:97-107.

163 Gregg EW. Sorlie P, Paulose-Ram R, et al. Prevalence of lower-extremity disease in the US adult population > 40 years of age with and without diabetes: 1990-2000 national health and nutrition examination survey. Diabetes Care : 2004;27:1591-1597.

164 Dick RF, Osgood ND, Lin TH, Gao A, Strang MR. End Stage Renal Disease Among People with Diabetes: A Comparison of First Nations People and Other Saskatchewan Residents from 1981 to 2005. Can J Diabetes . 2010;34(4):324-333.

165 Alder AI, Boyko EJ, Ahroni JH, et al. Risk factors for diabetic peripheral sensory neuropathy: results of the Seattle Prospective Diabetic Foot Study. Diabetes Care. 1997;20:1162-1167.

166 Shaw JH, Zimmet PZ. The epidemiology of diabetic neuropathy. Diabetes Review. 1999;7:245-253.

167 Boulton AJ, Vinik AI, Arezzo JC, et al. Diabetes neuropathies: a statement by the American Diabetes Association. Diabetes Care. 2005;28:956-962.

168 Perkins BA, Greene DA, Bril V. Glycemic control is related la the morphological severity of diabetic sensorimotor polyneuropathy. Diabetes Care . 2000;24(11):748-752.

169 Centers for Disease Control and Prevention. National Diabetes Fact Sheet: Information and National Estimates on Diabetes in the United States, 2005. Atlanta, G.A.: US Department of Health and Human Services, Centers for Disease Control and Prevention; 2005.

170 Meigs JB, Cupples LA & Wilson PW. Parental transmission of T2DM: the Framingham Offspring Study. Diabetes. 2000;49:2201 –2207.

171 Meigs JB, Shrader P, Sullivan LM, et al. Genotype score in addition to common risk factors for prediction of type 2 diabetes. N Engl J Med . 2008;359(21):2208–2219.

172 Fox CS. Cardiovascular disease risk factors, T2DM, and the framingham heart study. Trends in Cardiovascular Medicine. 2010;20(3), 90-95.

148

173 Cornelis MC, Qi L, Zhang C, et al. Joint effects of common genetic variants on the risk for T2DM in U.S. men and women of European ancestry. Ann Intern Med . 2009; 150(8):541–550.

174 Van Hoek M, Dehghan A,Witteman JC, et al. Predicting type 2 diabetes based on polymorphisms from genome-wide association studies: a population-based study. Diabetes . 2008;57(11):3122–3128.

175 Kasiske BL, Snyder JJ, Gilbertson D, Matas AJ. Diabetes mellitus after kidney transplantation in the United States. Am J Transplant . 2003;3(2):178-185.

176 Wilkinson A, Davidson J, Dotta F, et al. Guidelines for the treatment and management of new-onset diabetes after transplantation. Clin Transplant . 2005;19(3):291-298.

177 Markell M. New-onset diabetes mellitus in transplant patients: pathogenesis, complications, and management. Am J Kidney Dis . 2004;43(6):953-965.

178 Walczak DA, Calvert D, Jarzembowski TM, et al. Increased risk of post-transplant diabetes mellitus despite early steroid discontinuation in Hispanic kidney transplant recipients. Clin . 2005;19(4):527- 531.

179 Montori VM, Basu A, Erwin PJ, Velosa JA, Gabriel SE, Kudva YC. Posttransplantation diabetes: a systematic review of the literature. Diabetes Care. 2002;25(3):583-592.

180 Public Health Agency of Canada. Unpublished analysis using 2008/09 data from the Canadian Chronic Disease Surveillance System. Ottawa, ON: Statistics Canada; 2011.

181 Public Health Agency of Canada. Unpublished analysis using 2007-2009 data from the Canadian Health Measures Survey. Ottawa, ON: Statistics Canada; 2011.

182 Comstock RD, Castillo EM, Lindsay SP. Four-Year Review of the Use of Race and Ethnicity in Epidemiologic and Public Health Research. Am J Epidemiol . 2004;159:611-619.

183 Troper H, Weinfeld M. Ethnicity, Politics and Public Policy: Case studies in Canadian Diversity. Toronto, ON: University of Toronto Press. 1999:3-25.

184 Isajiw WW. Understanding diversity: Ethnicity and race in the Canadian Context. Toronto, ON: Thompson Educational Publishing Inc. 1999:17-36.

185 Isajiw WW. Definitions of Ethnicity. Ethnicity . 1974;1:111-124.

186 Connor W. Ethnonationalism: The Quest for Understanding. Princeton, N.J.: Princeton University Press. 1994:89-117.

187 Smith AD. The ethnic origins of nations. New York, N.Y.: Basil Blackwell. 1986:1-20.

188 Clarke DE, Colantonio A, Rhodes AE, Escobar M. Ethnicity and mental health: Conceptualization, definition and operationalization of ethnicity from a Canadian context. Chronic Diseases in Canada. 2008;28(4):128-147.

189 Allahar A. The social construction of primordial identities. In: Hier SP and Bolaria BS, eds. Identity and

149

Belonging: Rethinking race and ethnicity in Canadian society. Toronto: Canadian Scholars’ Press Inc. 2006;(2):31-42.

190 Statistics Canada Web site. Canadian Community Health Survey, Cycle 1.1. http://www.statcan.ca/english/concepts/health/. Accessed January 14, 2013.

191 Statistics Canada Web site. National Population Health Survey, Cycle 2. http://www.statcan.ca/cgibin/imdb/p2SV.pl?Function=getSurvey&SDDS=3225&lang=e n&db=IMDB&dbg=f&adm=8&dis=2. Accessed January 14, 2013.

192 Statistics Canada Web site. National Longitudinal Survey for Child and Youth, Cycle 1. http://www.statcan.ca/cgi-bin/imdb/p2SV.pl?Function=getSurvey&SurvId=44 50&SurvVer=0&InstaId=16044&InstaVer =1&SDDS=4450&lang=en&db=imdb&d bg=f&adm=8&dis=2. Accessed January 14, 2013.

193 Nazroo JY. Genetic, cultural or socioeconomic vulnerability? Explaining ethnic inequalities in health. Sociol Health Illn. 1998;20:710-730.

194 Brancati FL, Kao WHL, Folsom AR, Watson RL, Szklo M. Incident type 2 diabetes mellitus in African American and white adults. The Atherosclerosis Risk in Communities Study. JAMA. 2000;283:2253– 2259.

195 Narayan KM, Boyle JP, Geiss LS, et al. Impact of recent increase in incidence on future diabetes burden: U.S., 2005-2050. Diabetes Care. 2006;29:2114-2116.

196 Pawson IG, Martorell R, Mendoza FE. Prevalence of overweight and obesity in US Hispanic populations. Am J Clin Nutr . 1991;53(Suppl 9):s1522–1528.

197 Flegal KM, Ogden CL, Carroll MD. Prevalence and trends in overweight in Mexican-American adults and children. Nutr Rev . 2004;62(Suppl 2):s144–148.

198 Harris MI. Noninsulin-dependent diabetes mellitus in black and white Americans. Diabetes Metab Rev . 1990;6:71-90.

199 Stern MP, Gaskill SP, Hazuda HP, Gardner LI, Haffner SM. Does obesity explain excess prevalence of diabetes among Mexican Americans: Results of the San Antonio Heart Study. Diabetologia .1983;24:272-277.

200 Raleigh VS. Diabetes and hypertension in Britain's ethnic minorities: Implications for the future of renal services. BMJ (Clinical Research Ed.). 1997;314(7075);209-213.

201 Health Education Authority. Black and minority ethnic groups in England: health and lifestyles. London, UK: Health Education Authority; 1994.

202 Bhopal R, Unwin N, White M, et al. Heterogeneity of coronary heart disease risk factors in Indian, Pakistani, Bangladeshi, and European origin populations: cross sectional study. BMJ. 1999;319:215– 220.

203 Simmons D, Williams DR, Powell MJ. Prevalence of diabetes in different regional and religious south Asian communities in Coventry. Diabetes Med. 1992;9:428–431.

150

204 Cruickshank JK. Diabetes: contrasts between peoples of black (West African), Indian and white European origin. In: Cruickshank JK, Beevers DG, eds. Ethnic factors in health and disease. London, U.K.: Wright; 1989.

205 Banerji MA, Faridi N, Atluri R, Chaiken RL, Lebovitz HE. Body composition, visceral fat, leptin and insulin resistance in Asian Indian men. J Clin Endocrinol Metab . 1999;84:137 –144.

206 Deurenberg P, Deurenberg Y, Guricci S. Asians are different from Caucasians and from each other in their body mass index/body fat percent relationship. Obesity Rev . 2002;3:141 –146.

207 Polonsky KS, Sturis J, Bell Gl. Non-insulin-dependent diabetes mellitus: a genetically programmed failure of the beta cell to compensate for insulin resistance. N Eng J Med . 1996;334:777-783.

208 Svec F, Nastasi K, Hilton C, et al. Black-white contrasts in insulin levels during pubertal development. Diabetes. 1992;41:313–317.

209 Goran MI, Bergman RN, Cruz ML, et al. Insulin resistance and associated compensatory responses in African-American and Hispanic children. Diabetes Care. 2002;25:2184–2190.

210 Haffner SM, Howard G, Mayer C, et al. Insulin sensitivity and acute insulin response in African Americans, non-Hispanic whites, and Hispanics with NIDDM: the insulin resistance atherosclerosis study. Diabetes. 1997;46:63–69.

211 Lindquist CH, Gower BA, Goran MI. Role of dietary factors in ethnic differences in early risk of cardiovascular disease and T2DM. Am J Clin Nutr. 2000;71:725–732.

212 Osei K, Schuster DP, Owusu SK, Amoah AG. Race and ethnicity determine serum insulin and C- peptide concentrations and hepatic insulin extraction and insulin clearance: comparative studies of three populations of West African ancestry and white Americans. Metabolism . 1997;46:53-58.

213 Osei K, Gaillard T, Schuster DP. Pathogenetic mechanisms of impaired glucose tolerance and type II diabetes in African Americans: the significance of insulin secretion, insulin sensitivity, and glucose effectiveness. Diabetes Care . 1997;20:396-404.

214 Haffner SM, D'Agostino R, Saad MF, et al. Increased insulin resistance and insulin secretion in non- diabetic African Americans and Hispanics compared with non-Hispanic whites: the Insulin Resistance Atherosclerosis Study. Diabetes . 1996;45: 742-748.

215 Nwasuruba C, Khan M, Egede L. Racial/ethnic differences in multiple self-care behaviors in adults with diabete s. J Gen Intern Med. 2007;22(1):115-120.

216 Ekoe J-M, Rewers M, Williams R, Zimmet P, eds. The Epidemiology of Diabetes Mellitus. 2 nd ed. West Sussex, U.K.: Wiley-Blackwell; 2008.

217 Chiu M, Austin PC, Manuel DG, Shah BR, Tu JV. Deriving ethnic specific BMI cutoff points for assessing diabetes risk. Diabetes Care. 2011;34(8):1741-1748.

218 Riste L, Khan F, Cruickshank K. High prevalence of T2DM in all ethnic groups, including Europeans, in a British inner city: relative poverty, history, inactivity, or 21st century Europe? Diabetes Care.

151

2001;24:1377–1383.

219 Balarajan R, Raleigh VS. The ethnic populations of England and Wales: the 1991 Census. BMJ. 1992:24:113-116.

220 Balarajan R, Raleigh VS. Ethnicity and health in England. London, U.K.: HMSO; 1995.

221 McKeigue P, Sevak L. Coronary heart disease in South Asian communities: a manual for health promotion. London, U.K.: Health Education Authority; 1994.

222 Nazroo JY. The health of Britain’s ethnic minorities: findings from a national survey. London, U.K.: Policy Studies Institute; 1997.

223 Simmons D, Williams DRR, Powell MJ. Prevalence of diabetes in a predominantly Asian community: preliminary findings of the Coventry diabetes study. BMJ. 1989;298:18-21.

224 UK Prospective Diabetes Study Group. UK Prospective Diabetes Study XII: differences between Asian, Afro-Caribbean and white Caucasian type 2 diabetic patients at diagnosis of diabetes. Diabetic Medicine. 1994;11:670-677.

225 Hayes L, White M, Unwin N, et al. Patterns of physical activity and relationship with risk markers for cardiovascular disease and diabetes in Indian, Pakistani, Bangladeshi and European adults in a UK population. J Public Health Med. 2002;24:170–178.

226 Oldroyd J., Banerjee M, Herald A, Cruickshank K. Diabetes and ethnic minorities. Postgraduate Medical Journal. 2005;81(958):486-490.

227 Herald AH, Cruickshank JK, Riste LK, et al. Close relation of fasting insulin-like growth factor binding protein-1 (IGFBP-1) with glucose tolerance and cardiovascular risk in two populations. Diabetologia. 2001;44:333–339.

228 Cruickshank JK, Heald AH, Anderson S, et al. Epidemiology of the insulin-like growth factor system in three ethnic groups. Am J Epidemiol. 2001;154:504–513.

229 Heald AH, Cade JE, Cruickshank JK, et al. The influence of dietary intake on the insulin-like growth factor (IGF) system across three ethnic groups: a population-based study. Public Health Nutr. 2003;6:175–180.

230 Roderick PJ, Raleigh VS, Hallam L, Mallick NP. The need and demand for renal replacement therapy amongst ethnic minorities in England. J Epidemiol Comm Health. 1996;50:334-339.

231 Raleigh VS, Kiri VA, Balarajan R. Variations in mortality from diabetes, hypertension and renal disease in England and Wales by country of birth: why the NHS should act now. BMJ. 1997;314(7075);213- 215.

232 Leggetter S, Chaturvedi N, Fuller JH, et al. Ethnicity and risk of diabetes related lower extremity amputation: a population-based, case-control study of African Caribbeans and Europeans in the United kingdom. Arch Intern Med. 2002;162:73–78.

233 Chaturvedi N, Fuller JH. Ethnic differences in mortality from cardiovascular disease in the UK: do they

152

persist in people with diabetes? J Epidemiol Community Health 1996;50:137–139.

234 Perez-Escamilla R, Putnik P. The role of acculturation in nutrition, lifestyle, and incidence of T2DM among Latinos. The Journal of Nutrition. 2007;137(4);860-870.

235 Marshall MC. Diabetes in African Americans. Postgraduate Medical Journal. 2005;81(962);734-740.

236 United States Census Bureau Web site. http://www.census.gov/population/www/pop- profile/hisppop.html. Accessed February 14, 2013.

237 Centers for Disease Control and Prevention. National Center for Chronic Disease Prevention and Health Promotion National Diabetes Fact Sheet. Atlanta, G.A.: Centers for Disease Control and Prevention; 2007.

238 Centers for Disease Control and Prevention. Prevalence of diabetes and impaired fasting glucose in adults, United States, 1999–2000. Morb Mortal Wkly Rep . 2003;52:833–837.

239 Lipton R, Losey L, Giachello AL, Corral M, Girotti MH, Mendez JJ. Factors affecting diabetes treatment and patient education among Latinos: results of a preliminary study in Chicago. JMed Syst . 1996;20:267–276.

240 Centers for Disease Control and Prevention. Results from the Diabetes surveillance system. Atlanta, G.A.: Centers for Disease Control and Prevention; 2005.

241 American Diabetes Association. Diabetes 2001 vital statistics (p.87-109). Alexandria, V.A.: American Diabetes Association; 2001.

242 Young BA, Maynard C, Reiber G, et al. Effects of ethnicity and nephropathy on lower extremity amputation risk among diabetic veterans. Diabetes Care. 2003;26:495–501.

243 Arfken CL, Reno PL, Santiago JV, et al. Development of proliferative diabetic retinopathy in African- Americans and whites with type 1 diabetes. Diabetes Care. 1998;21:792–795.

244 Young BA, Maynard C, Boyko EJ. Racial differences in diabetic nephropathy, cardiovascular disease, and mortality in a national population of veterans. Diabetes Care. 2003;26:2392–2393.

245 De Rekeniere N, Rooks RN, Simonson EM, et al. Racial differences in glycemic control in a well- functioning older diabetic population. Findings from the health, aging and body composition study. Diabetes Care. 2003;26:1986–1992.

246 El-Kebbi IM, Ziemer DC, Musey VC, et al. Diabetes in urban African- Americans: Provider adherence to management protocols. Diabetes Care. 1997;20:698–703.

247 Gaskin R. Diet, diabetes, hypertension and blacks. Ethn Dis. 1999;9:272–277.

248 Robbins JM, Vaccarino V, Zhang H, et al. Socioeconomic status and type 2 diabetes in African American and non-Hispanic white women and men: evidence from the third National Health and Nutrition Examination Survey. Am J Public Health. 2001;91:76-83.

249 Neel JV, Weber AB, Julius S. Type II diabetes, essential hypertension, and obesity as syndromes of

153

impaired genetic homeostasis: the "thrifty genotype" hypothesis enters the century. Perspect Bioi Med. 1998;42:44-74.

250 Dagogo-Jack, Samuel. Ethnic Disparities in Type-2 Diabetes: Pathophysiology and Implications for Prevention and Management. Journal of the National Medical Association. 2003;95:9.

251 Curtsinger JW, Service PM, Prout T. Antagonistic pleiotropy, reversal of dominance, genetic polymorphism. Am Naturalist . 1994;144:210-228.

252 Lev-Ran A. Thrifty genotype: how applicable is it to obesity and T2DM? Diabetes Rev . 1999;7:1-22.

253 Turner RC, Levy JC, Clark A. Complex genetics or type 2 diabetes: thrifty genes and previously neutral polymorphisms. Q I Med .1993;86:413-417.

254 Serjeantson S, Owerbach D, Zimmet P, Nerup J, Thoma K. Genetics of diabetes in Nauru: effects of foreign admixture, HLA antigens and the insulin-gene-linked polymorphism. Diabetologica .1983;25:13-17.

255 Gushulak B. Healthier on arrival? Further insight into the "healthy immigrant effect”. CMAJ. 2007;176(10):1439-1440.

256 Chui T, Tran K, Maheux H, et al. Immigration in Canada: A Portrait of the Foreign-born Population, 2006 Census. Ottawa, ON: Statistics Canada; 2008.

257 Migration Policy Institute Web site. Size of the foreign-born population and foreign born as a percentage of the total population, for the United States: 1850 to 2006. http://www.migrationinformation.org/DataHub/charts/final.fbs.html. Accessed March 1, 2013.

258 Newbold KB, Danforth J. Health status and Canada’s immigrant population. Social Science and Medicine 2003;57:981-995.

259 Newbold KB. Self-rated health within the Canadian immigrant population: risk and the healthy immigrant effect. Social Science and Medicine. 2005;60:1359-1370.

260 Rotermann M. The impact of considering birthplace in analyses of immigrant health. Health Reports. 2011;22(4):37-43.

261 Betancourt MT, Roberts KC. Chronic Disease Patterns for Immigrants to Canada—A Recent Data Analysis. Ottawa, ON: Statistics Canada; 2010.

262 Newbold B. Health status and healthcare of immigrants in Canada: A longitudinal analysis. Journal of Health Services Research and Policy. 2005;10(2):77-83a.

263 Kobayashi KM, Prus S, Lin Z. Ethnic differences in self-rated and functional health: Does immigrant status matter? Ethnicity and Health . 2008;13(2):129-147.

264 Chen J, Wilkins R, Ng E. The health of Canada’s immigrants in 1994-95. Health Reports. 1996;7(4):29- 38.

154

265 Gushulak BD, Pottie K, Hatcher Roberts J, Torres S, DesMeules M, Canadian Collaboration for Immigrant and Refugee Health. Migration and health in Canada: Health in the global village. CMAJ. 2011;183(12):e952-958.

266 Ng E, Wilkins R, Gendron F, et al. Dynamics of immigrants’ health in Canada: evidence from the National Population Health Survey. Ottawa, ON: Statistics Canada; 2005.

267 Perez CE. Health status and health behaviour among immigrants. Health Rep . 2002;13(Suppl):s1-13.

268 McDonald JT, Kennedy S. Insights into the ‘healthy immigrant effect’: health status and health service use of immigrants to Canada. Soc Sci Med . 2004;59:1613-1627.

269 Newbold KB. Self-rated health within the Canadian immigrant population: risk and the healthy immigrant effect. Soc Sci Med. 2005;60:1359-1370.

270 Newbold KB. The short-term health of Canada’s new immigrant arrivals: evidence from LSIC. Ethn Health. 2009;14:315-336.

271 Newbold KB. Healthcare use and the Canadian immigrant population. Int J Health Serv. 2009;39:545- 565.

272 Pottie K, Ng E, Spitzer D, et al. Language proficiency, gender and self-reported health: an analysis of the first two waves of the Longitudinal Survey of Immigrants to Canada. Can J Public Health. 2008;99:505-510.

273 Oxman-Martinez J, Abdool S, Loiselle-Léonard M. Immigration, women and health in Canada. Can J Public Health. 2000;91:394-395.

274 Newbold KB, Danforth J. Health status and Canada’s immigrant population. Soc Sci Med. 2003;57(10):1981-95.

275 McDonald JT, Kennedy S. Insights into the ‘healthy immigrant effect’: health status and health service use of immigrants to Canada. Soc Sci Med. 2004;59:1613-1627.

276 Hyman I. Immigration and health: reviewing evidence of the healthy immigrant effect in Canada. CERIS working paper no. 55. Toronto, ON: Joint Centre of Excellence for Research on Immigration and Settlement; 2007.

277 Keane VP, Gushulak BD. The medical assessment of migrants: current limitations and future potential. Int Migr. 2001;39:29-42.

278 Ng E. The healthy immigrant effect and mortality rates. Health Reports. 2011;22(4):25-29.

279 Newbold B. Health status and healthcare of immigrants in Canada: a longitudinal analysis. J Health Serv Res Policy. 2005;10(2):78 –83.

280 Mazur RE, Marquis GS, Jensen HH. Diet and food insufficiency among Hispanic youths: acculturation and socioeconomic factors in the third National Health and Nutrition Examination Survey. Am J Clin Nutr. 2003;78:1120-1127.

155

281 Sundquist J, Winkleby MA. Cardiovascular risk factors in Mexican American adults: a transcultural analysis of NHANES III, 1988 –1994. Am J Public Health. 1999;89(5):723 –730.

282 Koya DL, Egede LE. Association between length of residence and cardiovascular disease risk factors among an ethnically diverse group of United States immigrants. J Gen Intern Med. 2007;22(6):841 – 846.

283 Kandula NR, Diez-Roux AV, Chan C, Daviglus ML, Jackson SA, Ni H, Schreiner PJ. Association of acculturation levels and prevalence of diabetes in the multi-ethnic study of atherosclerosis (MESA). Diabetes Care. 2008;31(8):1621 –1628.

284 Pan YL, Dixon Z, Himburg S, Huffman F. Asian students change their eating patterns after living in the United States. J Am Diet Assoc. 1999;99(1):54 –57.

285 Kaplan MS, Chang C, Newsom JT, McFarland BH. Acculturation status and hypertension among Asian immigrants in Canada. J Epidemiol Community Health. 2002;56(6):455 –456.

286 Gordon-Larsen P, Harris KM, Ward DS, Popkin BM. Acculturation and overweight-related behaviors among Hispanic immigrants to the US: the National Longitudinal Study of Adolescent Health. Soc Sci Med 2003;57(11):2023 –2034.

287 Jaber LA, Brown MB, Hammad A, Zhu Q, Herman WH. Lack of acculturation is a risk factor for diabetes in Arab immigrants in the US. Diabetes Care. 2003;26(7):2010 –2014.

288 Vega WA, Amaro H. Latino outlook: good health, uncertain prognosis. Annu Rev Public Health . 1994;15:39 –67.

289 Abraido-Lanza AF, Chao MT, Florez KR. Do healthy behaviors decline with greater acculturation? Implications for the Latino mortality paradox. Soc Sci Med. 2005;61(6):1243 –1255.

290 Alkerwi A, Sauvageot N, Pagny S, Beissel J, Delagardelle C, Lair ML. Acculturation, immigration status and cardiovascular risk factors among Portuguese immigrants to Luxembourg: Findings from ORISCAV-LUX study. BMC Public Health. 2012;12(864):2458.

291 Chun KM, Chesla CA, Kwan CM. So we adapt step by step: Acculturation experiences affecting diabetes management and perceived health for Chinese American immigrants. Social Science & Medicine. 2011;(2):256-264.

292 Chun KM, Akutsu PD. Assessing Asian American family acculturation in clinical settings: guidelines and recommendations for mental health professionals. In: N.-H. Trinh, Y. C. Rho, F. G. Lu, & K. M. Sanders (Eds.), Handbook of mental health and acculturation in Asian American families (pp. 99 -122). New York, N.Y.: Humana Press; 2009.

293 Chun KM, Balls Organista P, Marin G. Acculturation: Advances in theory, measurement, and applied research. Washington, D.C.: American Psychological Association; 2003.

294 Alegria M. The challenge of acculturation measures: what are we missing? A commentary on Thomson & Hoffman-Goetz. Soc Sci Med . 2009;69(7):996 –998.

295 Newbold KB. Self-rated health within the Canadian immigrant population: risk and the healthy

156

immigrant effect. Soc Sci Med . 2005;60:1359-1370.

296 Dunn JR, Dyck I. Social determinants of health in Canada’s immigrant population: Results from the national population health survey. Soc Sci Med. 2000;51:1573-1593.

297 Lara M, Gamboa C, Kahramanian MI, Morales LS, Bautista DE. Acculturation and Latino health in the United States: a review of the literature and its sociopolitical context. Annu Rev Public Health . 2005; 26:367–397.

298 Roshania R, Venkat Narayan KM, Oza-Frank R. Age at arrival and risk of obesity among US immigrants. Obesity (Silver Spring) . 2008;16(12):2669–2675.

299 Huang B, Rodriguez BL, Burchfiel CM, Chyou PH, Curb JD, Yano K. Acculturation and prevalence of diabetes among Japanese-American men in Hawaii. Am J Epidemiol . 1996;144:674–681.

300 Hosler AS, Melnik TA. Prevalence of diagnosed diabetes and related risk factors: Japanese adults in Westchester County, New York. Am J Public Health . 2003;93:1279–1281.

301 Lee MM, Wu-Williams A, Whittemore AS, Zheng S, Gallagher R, Teh C-Z, et al. Comparison of dietary habits, physical activity and body size among Chinese in North America and China. Int J Epidemiol . 1994;23:984–990.

302 Australian Bureau of Statistics. A picture of the nation: the statisticians report on the 2006 census. Canberra, AUS: Australian Bureau of Statistics; 2006.

303 Cunningham J. Socio-economic gradients in self-reported diabetes for Indigenous and non-Indigenous Australians aged 18–64. Aust NZ J Public Health. 2010;34(Supp1):s18–24.

304 Australian Bureau of Statistics. National Health Survey: Summary of Results, 2007–2008 (Reissue), Publication 4364.0. Canberra, AUS: Australian Bureau of Statistics; 2009.

305 Ahonen EQ, Benavides FG, Benach J. Immigrant populations, work and health- a systematic literature review. Scandinavian Journal of Work, Environment & Health. 2007;33(2):96-104.

306 2005 Immigration Overview. In: The Monitor. 2006; issue 2: p 2-17. Ottawa, ON: Citizenship and Immigration Canada; 2006.

307 Green AR, Ngo-Metzger Q, Legedza ATR, Massagli MP, Phillips RS, Iezzoni LI. Interpreter services, language concordance, and healthcare quality: experiences of Asian Americans with limited English proficiency. Journal of General Internal Medicine . 2005;20:1050 -1056

308 Ngo-Metzger Q, Massagli MP, Clarridge BR, Manocchia M, Davis RB, Iezzoni LI, et al. Linguistic and cultural barriers to care: perspectives of Chinese and Vietnamese immigrants. Journal of General Internal Medicine . 2003;18:44-52.

309 Newbold KB. Healthcare use and the Canadian immigrant population. Int J Health Serv. 2009;39:545- 565.

310 Mendoza FS. Health disparities and children in immigrant families: A research agenda. Pediatrics. 2009;124(Suppl 3);s187-195.

157

311 Miedema B, Hamilton R, Easley J. Climbing the walls: structural barriers to accessing primary care for refugee newcomers in Canada. CMAJ . 2008;54(3):335 –336.

312 Muggah E, Dahrouge S, Hogg W. Access to primary healthcare for immigrants: Results of a patient survey conducted in 137 primary care practices in Ontario, Canada. BMC Family Practice. 2012;13:128.

313 Setia M. Access to health-care in Canadian Immigrants: a longitudinal study of the National Population Health Survey. Health and Social Care. 2011;19(1):70 –79.

314 Glasgow RE, Hampson SE, Strycker LA, Ruggiero L. Personal model beliefs and social-environmental barriers related to diabetes self-management. Diabetes Care. 1997;20:556–561.

315 Frosch DL, Kaplan RM. Shared decision making in clinical medicine: past research and future directions. Am J Prev Med. 1999; 17:285–294.

316 Surmond J, Seeleman C. Shared decision-making in an intercultural context: Barriers in the interaction between physicians and immigrant patients. Patient Educ Couns. 2006;60:253–259.

317 Hjelm K, Bard K, Nyberg P, Apelqvist J. Religious and cultural distance in beliefs about health and illness in women with diabetes mellitus of different origin living in Sweden. Int J Nurs Stud. 2003;40:627–643.

318 Thabit H, Shah S, Nash M, Brema I, Nolan JJ, Martin G. Globalization, immigration and diabetes self- management: An empirical study amongst immigrants with T2DM in Ireland. QJM. 2009;102(10):713-720.

319 Donnelly TT, McKellin W. Keeping healthy! Whose responsibility is it anyway? Vietnamese Canadian women and their healthcare providers’ perspectives. Nurs Inq. 2007;14:2-12.

320 DeBellonia RR, Marcus S, Shih R, et al. Curanderismo: consequences of folk medicine. Pediatr Emerg Care. 2008;24:228-229.

321 Creatore MI, Moineddin R, Booth G, Manuel DH, DesMeules M, McDermott S, Glazier RH. Age- and sex-related prevalence of diabetes mellitus among immigrants to Ontario, Canada. CMAJ. 2010;182(8):781-789.

322 Booth GL, Creatore MI, Gozdyra P, Glazier RH. Ethnicity, Immigration and Diabetes. In: Glazier RH, Booth GL, Gozdyra P, Creatore MI, Tynan A-M, eds. Neighbourhood Environments and Resources for Healthy Living - A Focus on Diabetes in Toronto: ICES Atlas. Toronto, ON: Institute for Clinical Evaluative Sciences; 2007:57-86.

323 Caulford P, Vali Y. Providing healthcare to medically uninsured immigrants and refugees. CMAJ. 2006;174:1253-1254.

324 Green AG, Green DA. The Economic Goals of Canada’s Immigration Policy, Past and Present. Vancouver, BC: Department of Economics, University of British Columbia; 1996.

325 Thompson EN. Immigrant Occupational Skill Outcomes and the Role of Region-Of- Origin-Specific

158

Human Capital. Ottawa, ON: Human Resources Development Canada; 2000.

326 Di Biase S, Bauder H. Immigrants in Ontario: Linking Spatial Settlement Patterns and Labour Force Characteristics. Guelph, ON: Department of Geography, University of Guelph; 2004.

327 Schellenberg G. Trends and Conditions in Census Metropolitan Areas: Immigrants in Canada’s Census Metropolitan Areas. Ottawa, ON: Statistics Canada; 2004.

328 Statistics Canada. Aboriginal Peoples in Canada in 2006. Ottawa, ON: Statistics Canada; 2009.

329 Statistics Canada. Aboriginal Peoples Technical Report, 2006 Census, 2nd edition. Ottawa, ON: Statistics Canada; 2010.

330 Royal Commission on Aboriginal Peoples. Residential schools: Report of the Royal Commission on Aboriginal Peoples. Ottawa, ON: Canada Communications Group; 1996.

331 King M, Smith A, Gracey M. Indigenous health part 2: the underlying causes of the health gap. Lancet . 2009;374(9683):76 –85.

332 Reading C, Wien F. Health inequalities and social determinants of Aboriginal peoples’ health. Prince George, BC: National Collaborating Centre for Aboriginal Health; 2009.

333 Chandler MJ, Lalonde C. Cultural continuity as a hedge against suicide in Canada ’s First Nations. Transcult. Psychiatry . 1998;35(2):191 –219.

334 Gionet L. Inuit in Canada: Selected findings of the 2006 Census. Ottawa, ON: Statistics Canada; 2008.

335 Bhattacharyya OK, Rasooly LR., Naqshbandi M, Estey EA, Esler J, Toth E, et al. Challenges to the provision of diabetes care in first nations communities: Results from a national survey of healthcare providers in Canada. BMC Health Services Research. 2011;11:283

336 Public Health Agency of Canada. Building a National Diabetes Strategy: Synthesis of Research and Collaborations (Consultation Findings). Ottawa, ON: Statistics Canada; 2011.

337 Statistics Canada Web site. Aboriginal Peoples of Canada. http://www23.statcan.gc.ca/imdb/p2SV.pl?Function=getSurvey&SDDS=3226&Item_Id=50653&lang =en. Accessed on February 26 th , 2013.

338 Barton SS, Anderson N, Thommasen HV. The diabetes experiences of Aboriginal people living in a rural Canadian community. Aust J Rural Health . 2005;13(4):242-246.

339 Elliott CT, De Leeuw SN. Our aboriginal relations: When family doctors and aboriginal patients meet. Can Fam Phys. 2009;55(4):443.

340 Garner R, Carrière G, Sanmartin C, Longitudinal Health and Administrative Data Research Team. The health of First Nations living off-reserve, Inuit, and Métis adults in Canada: The impact of socioeconomic status on inequalities in health. Ottawa, ON: Statistics Canada; 2010.

341 Macaulay AC. Improving aboriginal health: How can healthcare professionals contribute? Can Fam Phys. 2009;55(4):334.

159

342 Towle A, Godolphin W, Alexander T. Doctor-patient communications in the Aboriginal community: Towards the development of educational programs. Patient Educ Couns. 2006;62(3):340-346.

343 Bruce S. Prevalence and determinants of diabetes mellitus among the metis of western Canada. American Journal of Human. 2000;12(4):542-551.

344 Young TK, Reading J, Elias B, O'Neil JD. T2DM in Canada's First Nations: Status of an epidemic in progress. CMAJ . 2000;163(5):561-566.

345 Dyck R, Osgood N, Lin TH, Gao A, Stang MR. Epidemiology of diabetes mellitus among First Nations and non-First Nations adults. CMAJ . 2010;182(3):249-256.

346 Adelson N. The embodiment of inequity: Health disparities in Aboriginal Canada. Can J Public Health. 2005;96(Suppl 2):s45-61.

347 Ho LS, Gittelsohn J, Rimal R, Treuth MS, Sharma S, Rosecrans A, Harris SB. An integrated multi- institutional diabetes prevention program improves knowledge and healthy food acquisition in Northwestern Ontario First Nations. Health Educ Behav. 2008;35:561 –573.

348 Anand SS, Davis AD, Ahmed R, Jacobs R, Xie C, Hill A, et al. A family-based intervention to promote healthy lifestyles in an aboriginal community in Canada. Can J Public Health. 2007;98(6):447 –452.

349 Public Health Agency of Canada. Unpublished analysis using 2009-2010 data from the Canadian Community Health Survey. Ottawa, ON: Statistics Canada; 2011.

350 Blanchet C, Rochette L. Nutrition and food consumption among the Inuit of Nunavik: Nunavik Inuit Health Survey 2004, Qanuippitaa? How are we? Quebec City, QC: Institut national de santé publique du Québec, Nunavik Regional Board of Health and Social Services; 2008.

351 Gittelsohn J, Sharma S. Physical, consumer, and social aspects of measuring the food environment among diverse low-income populations. Am J Prev Med. 2009;36(Suppl 4):s161 –165.

352 Joseph P, Davis AD, Miller R, Hill K, McCarthy H, Banerjee A, et al. Contextual determinants of health behaviours in an aboriginal community in Canada: Pilot project. BMC Public Health. 2012;12:952

353 Statistics Canada. Spending patterns in Canada 2009. In: 2009 SpiC, ed. Ottawa, ON: Statistics Canada; 2010.

354 Wolever TMS, Hamad S, Gittelsohn J, Gao J, Hanley AJG, Harris SB, Zinman B. Low dietary fiber and high protein intakes associated with newly diagnosed diabetes in a remote aboriginal community. Am J Clin. Nutr . 1997;66(b):1470-1474.

355 Gittelsohn J, Wolever TM, Harris SB, Harris-Giraldo R, Hanley AJ, Zinman B. Specific patterns of food consumption and preparation are associated with diabetes and obesity in a native Canadian community. The Journal of Nutrition. 1998;128(3):541-547.

356 Harris SB, Gittelsohn J, Hanley A, Barnie A, Wolever TMS, Gao J, et al. The prevalence of NIDDM and associated risk factors in native Canadians. Diabetes Care. 1997;20:185-187.

160

357 Kuhnlein VH, Receveur O, Soueida R, Berti PR. Unique patterns of dietary adequacy in three cultures of Canadian Arctic indigenous peoples. Public Health Nutr. 2011;11:349-360.

358 Noakes M, Keogh JB, Foster PR, Clifton PM. Effect of an energy-restricted, high-protein, low-fat diet relative to a conventional high-carbohydrate, low-fat diet on weight loss, body composition, nutritional status, and markers of cardiovascular health in obese women. Am J Clin Nutr . 2005;81(6):1298–1306.

359 Public Health Agency of Canada. Unpublished analysis using 2006 data from the Aboriginal Peoples Survey. Ottawa, ON: Statistics Canada; 2011.

360 Young TK, Dean HJ, Flett B, Wood-Steiman P. Childhood obesity in a population at high risk for T2DM. J Pediatr. 2000;136(3):365-369.

361 Shaw J. Epidemiology of childhood T2DM and obesity. Pediatr Diabetes. 2007;8(Suppl 9):s7-15.

362 Anand SS, Yusuf S, Jacobs R, Davis AD, Yi Q, Gerstein H, Montague PA, Lonn E. Risk factors, atherosclerosis, and cardiovascular disease among aboriginal people in Canada: The study of health assessment and risk evaluation in aboriginal peoples (SHARE-AP). Lancet. 2001;358(9288):1147 – 1153.

363 First Nations Regional Health Survey. National Report on the Adult, Youth and Children Living in First Nations Communities. First Nations Regional Health Survey (RHS) 2008/2010. Ottawa, ON: First Nations Information Governance Centre; 2011.

364 Ralph-Campbell K, Oster RT, Connor T, et al. Increasing rates of diabetes and cardiovascular risk in Métis settlements in Northern Alberta. Int J Circumpolar Health . 2009;68(5):433-442.

365 Aljohani N, Rempel BM, Ludwig S, et al. Gestational diabetes in Manitoba during a twenty-year period. Clin Invest Med . 2008;31(3):e131-137.

366 Dyck RF, Tan L, Hoeppner VH. Body mass index, gestational diabetes and diabetes mellitus in three northern Saskatchewan Aboriginal Communities. CDIC . 1995;16: 24-26.

367 Harris SB, Caulfield LE, Sugamori ME, Whalen EA, Henning B. The epidemiology of diabetes in pregnant Native Canadians. a risk profile. Diabetes Care. 1997;20(9):1422-1425.

368 Mohamed N, Dooley J. Gestational diabetes and subsequent development of NIDDMin Aboriginal women of northwestern Ontario. Int J Circumpolar Health. 1998;57(Suppl 1):s355-358.

369 Rodrigues S, Robinson E, Gray-Donald K. Prevalence of gestational diabetes mellitus among James Bay Cree women in Northern Quebec. CMAJ . 1999;160(9):1293-1297

370 Chamberlain C, Yore D, Li H, Williams E, Oldenburg B, Oats J, et al. Diabetes in pregnancy among indigenous women in Australia, Canada, New Zealand, and the united states: A method for systematic review of studies with different designs. BMC Pregnancy and Childbirth. 2011;11:104.

371 Liu SL, Shah BR, Naqshbandi M, Tran V, Harris SB. Increased rates of adverse outcomes for gestational diabetes and pre-pregnancy diabetes in on-reserve first nations women in Ontario, Canada. Diabetic Medicine. 2012;29(8):e180-183.

161

372 Wenman WM, Joffres MR, Tataryn IV, Edmonton Perinatal Infections Group. A prospective cohort study of pregnancy risk factors and birth outcomes in aboriginal women. CMAJ. 2004;171(6):585-589.

373 Tait H. Aboriginal Peoples Survey, 2006: Inuit Health and Social Conditions. Ottawa, ON: Statistics Canada; 2008.

374 Reading J. The crisis of chronic disease among Aboriginal Peoples: A challenge for public health, population health and social policy. Victoria, BC: University of Victoria, Centre for Aboriginal Health Research; 2009.

375 Harris SB, Perkins BA, Whalen-Brough E. Non-insulin-dependent diabetes mellitus among First Nations children: New entity among First Nations people of Northwestern Ontario. Can Fam Physician. 1996;42:869-876.

376 Amed S, Dean HJ, Panagiotopoulos C, et al. T2DM, medication-induced diabetes, and monogenic diabetes in Canadian children: A prospective national surveillance study. Diabetes Care. 2010;33(4):786-791.

377 Southam L, Soranzo N, Montgomery S, et al. Is the thrifty genotype hypothesis supported by evidence based on confirmed T2DM- and obesity-susceptibility variants? Diabetologia. 2009;52:1846-1851.

378 Hegele RA, Cao H, Harris SB, Hanley AJ, Zinman B. Hepatocyte nuclear factor-1 alpha G319S. A private mutation in Oji-Cree associated with T2DM. Diabetes Care. 1999;22:524.

379 Sellers EA, Triggs-Raine B, Rockman-Greenberg C, Dean HJ. The prevalence of the HNF-1alpha G319S mutation in Canadian aboriginal youth with type 2 diabetes. Diabetes Care . 2002;25:2202– 2206.

380 Millar K, Dean HJ. Developmental Origins of Type 2 Diabetes in Aboriginal Youth in Canada: It Is More Than Diet and Exercise. Journal of Nutrition and Metabolism. 2012;2012.

381 Statistics Canada. Aboriginal Peoples in Canada in 2006: Inuit, Métis and First Nations. Ottawa, ON: Statistics Canada; 2008.

382 Public Health Agency of Canada. Report from the National Diabetes Surveillance System: Diabetes in Canada, 2009. Ottawa, ON: Statistics Canada; 2009.

383 Willows ND, Hanley AJ, Delormier T. A socioecological framework to understand weight-related issues in aboriginal children in Canada. Applied Physiology, Nutrition, and Metabolism. 2012;37(1):1-13.

384 Egeland GM, Cao Z, Young TK. Hypertriglyceridemic- waist phenotype and glucose intolerance among Canadian Inuit: the International Polar Year Inuit Health Survey for Adults 2007 –2008. CMAJ. 2011;183(9):e553 –558.

385 Harris SB, Naqshbandi M, Hanley AJG, Bhattacharyya OK, Esler JG, Zinman B. Comparison of the Clinical Management of T2DM in Canada ’s First Nations Peoples to National Guidelines: The CIRCLE Study. Canadian Journal of Diabetes. 2009;33(3):202.

386 Harris SB, Naqshbandi M, Bhattacharyya OK, Hanley AJG, Esler JG, Zinman B. Burden of T2DM-

162

Associated Complications in Canada ’s First Nations Peoples in 2007: The CIRCLE Study. Canadian Journal of Diabetes. 2008;33(3):196.

387 Hanley AJG, Harris SB, Mamakeesick M, Goodwin K, Fiddler E, Hegele RA, et al. Complications of T2DM among Aboriginal Canadians. Diabetes Care. 2005;28(8):2054-2057.

388 Anderson JF. Diabetes in aboriginal populations. CMAJ. 2000;162(1):11-12.

389 Dyck R, Naqshbandi Hayward M, Harris S, CIRCLE Study Group. Prevalence, determinants and comorbidities of chronic kidney disease among First Nations adults with diabetes: results from the CIRCLE study. BMC Nephrology. 2012;13(1):57.

390 Wilson R, Krefting LH, Sutcliffe P, Van-Bussel L. Incidence and prevalence of end-stage renal disease among Ontario’s James Bay Cree. Can J Public Health. 1992;83:143–146.

391 Pavkov ME, Bennett PH, Knowler WC, Krakov J, Seivers ML, Nelson RG. Effect of youth-onset T2DM on end-stage renal disease and mortality in young and middle-aged Pima Indians. JAMA. 2010;296(4):421 –426.

392 Dyck RF, Sidhu N, Klomp H, Cascagnette PJ, Teare GF. Differences in Glycemic Control and Survival Predict Higher ESRD Rates in Diabetic First Nations Adults. Clin Invest Med . 2010;33(6):e390 –397.

393 Hanley AJ, Harris SB, Mamakeesick M, Goodwin K, Fiddler E, Hegele RA, et al. Complications of T2DM among Aboriginal Canadians: Prevalence and associated risk factors. Diabetes Care. 2005;28(8):2054-2057.

394 Anand SS, Yusef S, Jacobs R, Davis AD, Yi Q, Gerstein H, et al. Risk factors, atherosclerosis, and cardiovascular disease among Aboriginal people in Canada: the Study of Health Assessment and Risk Evaluation in Aboriginal Peoples. Lancet. 2001;358(9288):1147-1153.

395 Harris SB, Naqshbandi M, Hanley AJG, Bhattacharyya OK, Esler JG, Zinman B. Comparison of the Clinical Management of T2DM in Canada ’s First Nations Peoples to National Guidelines: The CIRCLE Study. Canadian Journal of Diabetes. 2009;33(3):202.

396 Harris SB, Naqshbandi M, Bhattacharyya OK, Hanley AJG, Esler JG, Zinman B. Burden of T2DM- Associated Complications in Canada ’s First Nations Peoples in 2007: The CIRCLE Study. Canadian Journal of Diabetes. 2008;33(3):196.

397 Hanley AJG, Harris SB, Mamakeesick M, Goodwin K, Fiddler E, Hegele RA, et al. Complications of T2DM among Aboriginal Canadians. Diabetes Care. 2005;28(8):2054-2057.

398 Anderson JF. Diabetes in aboriginal populations. CMAJ. 2000;162(1):11-12.

399 Canadian Community Health Survey. User Guide for the Public use Microdata File 2001. Ottawa, ON: Statistics Canada; 2001.

400 Canadian Community Health Survey. User Guide for the Public use Microdata File 2003. Ottawa, ON: Statistics Canada; 2003.

163

401 Canadian Community Health Survey. User Guide for the Public use Microdata File 2005. Ottawa, ON: Statistics Canada; 2005.

402 Canadian Community Health Survey. User Guide for the Public use Microdata File 2007. Ottawa, ON, ON: Statistics Canada; 2007.

403 Canadian Community Health Survey. User Guide for the Public use Microdata File 2008. Ottawa, ON: Statistics Canada; 2008.

404 Canadian Community Health Survey. User Guide for the Public use Microdata File 2009. Ottawa, ON: Statistics Canada; 2009.

405 Canadian Community Health Survey. User Guide for the Public use Microdata File 2010. Ottawa, ON: Statistics Canada; 2010.

406 Statistics Canada Web site. Summary of changes over time – Canadian Community Health Survey - Annual Component (CCHS). http://www23.statcan.gc.ca/imdb/p2SV.pl?Function=getMainChange&SurvId=3226&SurvVer=1&Ins taId=15282&SDDS=3226&lang=en&db=imdb&adm=8&dis=2. Accessed February 18th, 2013.

407 Statistics Canada Web site . Health Regions. http://www.statcan.gc.ca/concepts/health-sante/maps- cartes/ontario_south-eng.jpg. Accessed February 18th, 2013.

408 Tremblay MS, Wolfson M, Connor Gorber S. Canadian Health Measures Survey: Rationale, background and overview. Health Reports. 2007;18:7-20.

409 Shields M, Connor Gorber S, Tremblay MS. Estimates of obesity based on self-report versus direct measures. Health Reports. 2008;19(2):61-76.

410 Chiolero A, Peytremann-Bridevaux I, Paccaud F. Associations between obesity and health conditions may be overestimated if self-reported body mass index is used. Obesity Reviews. 2007;8(4):373-374.

411 Béland Y, St-Pierre M. Mode Effects in the Canadian Community Health Survey: A Comparison of CATI and CAPI. Advances in Telephone Survey Methodology: John Wiley & Sons, Inc.; 2007:297- 314.

412 Martin LM, Leff M, Calonge N, Garrett C, Nelson DE. Validation of self-reported chronic conditions and health services in a managed care population. American Journal of Preventive Medicine . 2000;18(3):215-218.

413 Statistics Canada. Methodology of the Canadian labour force survey. Statistics Canada Catalogue: Ministry of Industry. 2008;71:526.

414 Statistics Canada Web site. Health Reports. http://www.statcan.gc.ca/pub/82-003- x/2009001/article/10795/findings-resultats-eng.html. Accessed February 26, 2013.

415 Kish L. Cumulating/combining population surveys. Statistics Canada Catalogue: Survey Methodology .1999;25(2):129-38.

164

416 Statistics Canada Web site. Canadian Community Health Survey, Annual Content Questionnaire (2010 Master File Documentation). http://www23.statcan.gc.ca/imdb/p2SV.pl?Function=getSurvey&SDDS=3226&Item_Id=50653&lang =en. Accessed February 16, 2013.

417 Mendoza FS. Health disparities and children in immigrant families: A research agenda. Pediatrics. 2009;124(3):s187-95. 418 Statistics Canada Web site. 2006 Census: Aboriginal Peoples in Canada in 2006: Inuit, Métis and First Nations, 2006 Census: First Nations people. http://www12.statcan.ca/census-recensement/2006/as- sa/97-558/p16-eng.cfm. Accessed June 1 2013.

419 Liu R, So L, Mohan S, Khan N, King K, Quan H. Cardiovascular risk factors in ethnic populations within Canada: Results from national cross-sectional surveys. Open Medicine. 2010;4(3):e143-53.

420 Public Health Agency of Canada Web site. Report from the National Diabetes Surveillance System: Diabetes in Canada, 2009. http://www.phac-aspc.gc.ca/publicat/2009/ndssdic-snsddac-09/1-eng.php. Accessed June 3 2013.

421 Dannenbaum D, Kuzmina E, Lejeune P, Torrie J, Gangbe M. Prevalence of diabetes and diabetes-related complications in First Nations communities in Northern Quebec (Eeyou Istchee), Canada. Can J Diabetes. 2008;32:46-52. 422 Statistics Canada. Aboriginal Peoples Survey, 2006: An overview of the health of the Métis population. Ottawa, ON: Statistics Canada; 2009.

423 Statistics Canada. Aboriginal Peoples Survey 2001: Initial Release - Supporting Tables. Ottawa, ON: Statistics Canada; 2003.

424 Hsin-Chieh Y, Duncan B, Schmidt M, Wang NY, Frederick L. Smoking, Smoking Cessation, and Risk for Type 2 Diabetes Mellitus, A Cohort Study. Annals of Internal Medicine . 2010;152(1):10-17.

425 Hofstetter A, Schutz Y, Jequier E, Wahren J. Increased 24-hour energy expenditure in cigarette smokers. N Engl J Med. 1986;314(2):79–82.

426 Williamson DF, Madans J, Anda RF, Kleinman JC, Giovino GA, Byers T. Smoking cessation and severity of weight gain in a national cohort. N Engl J Med. 1991;324:739–45.

427 Chiolero A, Faeh D, Paccaud F, Cornuz J. Consequences of smoking for body weight, body fat distribution, and insulin resistance. Am J Clin Nutr . 2008;87(4);801-809.

428 Braveman P, Cubbin C, Egerter S, et al. Socioeconomic Status in Health Research, One Size Does Not Fit All. JAMA . 2005;294(22):2879-2888.

429 Chalabian J, Dunnington G. Impact of language barrier on quality of patient care, resident stress and teaching. Teaching and Learning in Medicine. 1997;9:84-90.

430 Chugh U, Dillmann E, Kurtz SM, Lockyer J, Parboosingh J. Multicultural issues in medical curriculum: implications for Canadian physicians. Med. Teach. 1993;15:83-91.

165

431 Pasick RJ. Socioeconomic and cultural factors in the development and use of theory. In: Glanz K, Lewis FM, Rimer BK, editors. Health Behavior and Health Education—Theory, Research, and Practice. 2. San Francisco, CA: Jossey-Bass Inc; 2002.

432 U.S. Department of Health and Human Services Office of Minority Health. Assuring Cultural Competence in Health Care: Recommendations for National Standards and Outcomes-Focused Research Agenda. Washington, DC: U.S. Government Printing Office; 2000.

433 Trivedi AN, Ayanian JZ. Perceived discrimination and use of preventive health services. J Gen Intern Med. 2006;21(6):553-558. 434 Krieger N. Shades of difference: theoretical underpinnings of the medical controversy on black/white differences in the United States, 1830-1870. Int J Health Serv. 1987;17(2):259-278.

435 Bellon JA, Lardelli P, Luna JD, Delgado A. Validity of self reported utilisation of primary health care services in an urban population in Spain . J Epidemiol Community Health . 2000;54:544–551.

436 Dunlop S, Coyte PC, McIsaac W. Socio-economic status and the utilisation of physicians' services: results from the Canadian National Population Health Survey. Soc Sci Med . 2000;51:123–133.

437 Hyman I, Patychuk D, Zaidi Q, et al. Self-management, health service use and information seeking for diabetes care among recent immigrants in Toronto. Chronic diseases and injuries in Canada . 2012;33:12.

166

VITA

Name: Michael James Taylor

Education: University of Western Ontario ( Candidate ) 2011 – Present Master of Science, Epidemiology & Biostatistics

University of Waterloo 2006 – 2011 Bachelor of Arts, Psychology & Business, Co-operative Program (Hons.)

Honours 3 Minute Thesis Competition – People’s Choice Award (Ontario) 2013 & Awards: 3 Minute Thesis Competition – First Place Award (Western) 2013 Western Research Forum – Second Place Award 2013 Ontario Graduate Scholarship 2011-2012 Schulich Graduate Scholarship 2011-2013 University of Waterloo Dean’s Honours List 2007-2011 University of Waterloo President’s Scholarship 2006 Millennium Excellence Scholarship for Innovation 2006

Presentations: Risk factors for Diabetes Mellitus - a Comparative Analysis of Racial, Cultural, and Geographic Differences in a Large Canadian Sample [Oral Presentation ]. Centre for Migration Studies Graduate Research Symposium, London, Ontario, 04/04/2013. Risk factors for Diabetes Mellitus - a Comparative Analysis of Racial, Cultural, and Geographic Differences in a Large Canadian Sample [Oral Presentation ]. Centre for Migration Studies Graduate Research Symposium, London, Ontario, 04/04/2013. Risk factors for Diabetes Mellitus - a Comparative Analysis of Racial, Cultural, and Geographic Differences in a Large Canadian Sample [Poster Presentation ]. London Health Research Day, London, Ontario, 19/03/2013. Risk factors for Diabetes Mellitus - a Comparative Analysis of Racial, Cultural, and Geographic Differences in a Large Canadian Sample [Oral Presentation ]. Western Research Forum, London, Ontario, 15/03/2013 Immigrant Depression & Diabetes: a Comparative Analysis of Differences in a Canadian-based Sample [ Poster Presentation ]. Diabetes Research Day, London, Ontario, 13/11/2013

Work Vice-President, Western Diabetes Association, London, ON 2011-2013 Experience: Research Student, St. Michael’s Hospital, Toronto, ON 2011 Project Coordinator, eHealth Ontario, Toronto, ON 2010 Analyst, Cancer Care Ontario (CCO), Toronto, ON 2009-2010