Search

Abstract Background The World Health Organization recommends a 10% total energy (TE%) limit for free sugars (i.e., added sugars and naturally occurring sugars in fruit juice, honey, and syrups) based on evidence linking higher intakes with overweight and dental caries. Evidence for cardiovascular disease (CVD) is limited. Impacts may differ by sex, age group, and solid vs. liquid sources; liquids may stimulate more adverse CVD profiles (due to their rapid absorption in the body along along with triggering less satiety). We examined associations of consuming total free sugars ≥ 10 TE% with CVD within four sex and age-defined groups. Given roughly equal free sugar intakes from solid and liquid sources, we also evaluated source-specific associations of free sugars ≥ 5 TE% thresholds. Methods In this retrospective cohort study, we estimated free sugars from 24-h dietary recall (Canadian Community Health Survey, 2004–2005) in relationship to nonfatal and fatal CVD (Discharge Abstract and Canadian Mortality Databases, 2004–2017; International Disease Classification-10 codes for ischemic heart disease and stroke) through multivariable Cox proportional hazards models adjusted for overweight/obesity, health behaviours, dietary factors, and food insecurity. We conducted analyses in separate models for men 55 to 75 years, women 55 to 75 years, men 35 to 55 years, and women 35 to 55 years. We dichotomized total free sugars at 10 TE% and source-specific free sugars at 5 TE%. Results Men 55 to 75 years of age had 34% higher CVD hazards with intakes of free sugars from solid sources ≥ 5 TE% vs. below (adjusted HR 1.34, 95% CI 1.05- 1.70). The other three age and sex-specific groups did not demonstrate conclusive associations with CVD. Conclusions Our findings suggest that from a CVD prevention standpoint in men 55 to 75 years of age, there may be benefits from consuming less than 5 TE% as free sugars from solid sources.

Abstract Background Modern health surveillance and planning requires an understanding of how preventable risk factors impact population health, and how these effects vary between populations. In this study, we compare how smoking, alcohol consumption, diet and physical activity are associated with all-cause mortality in Canada and the United States using comparable individual-level, linked population health survey data and identical model specifications. Methods The Canadian Community Health Survey (CCHS) (2003–2007) and the United States National Health Interview Survey (NHIS) (2000, 2005) linked to individual-level mortality outcomes with follow up to December 31, 2011 were used. Consistent variable definitions were used to estimate country-specific mortality hazard ratios with sex-specific Cox proportional hazard models, including smoking, alcohol, diet and physical activity, sociodemographic indicators and proximal factors including disease history. Results A total of 296,407 respondents and 1,813,884 million person-years of follow-up from the CCHS and 58,232 respondents and 497,909 person-years from the NHIS were included. Absolute mortality risk among those with a ‘healthy profile’ was higher in the United States compared to Canada, especially among women. Adjusted mortality hazard ratios associated with health behaviours were generally of similar magnitude and direction but often stronger in Canada. Conclusion Even when methodological and population differences are minimal, the association of health behaviours and mortality can vary across populations. It is therefore important to be cautious of between-study variation when aggregating relative effect estimates from differing populations, and when using external effect estimates for population health research and policy development.

Abstract Background Sugar-sweetened beverages have obesogenic and diabetogenic effects ascribed to free sugars. These include added sugars and naturally occurring sugars in juices. A meta-analysis indicates that some foods with added sugars are associated with lower type 2 diabetes rates. To expand the evidence relevant to free sugars from solid sources, we examined a young to middle-aged population with respect to overweight and gestational diabetes (GDM) outcomes. Methods We studied female participants (12–50 years old) from the 2004–2005 Canadian Community Health Survey 2.2 (CCHS) with data linked to the hospital Discharge Abstract Database (DAD) until 2017, providing 13 years of follow-up. We estimated free sugars by solid and liquid sources from 24-h dietary recalls as percent total energy intake (TE%), and computed body mass index (BMI). We applied ICD-10 diagnostic codes for deliveries and GDM to DAD. We conducted multivariable logistic regression analyses to evaluate associations between free sugars with overweight at baseline (cross-sectional component) and, in those who delivered, with GDM during follow-up (nested case control component). We compared those with consumption above versus below various thresholds of intake for free sugars, considering solid and liquid sources separately (2.TE%, 5TE%, 10TE% and 15TE% thresholds). Results Among 6305 participants, 2505 (40%) were overweight, defined as BMI ≥ 85th percentile below 18 years and BMI ≥ 25 kg/m2 for adults. Free sugars from solid sources were associated with lower odds of overweight above versus below the 2.5TE% (adjusted odds ratio [adjOR] 0.80, 95%CI 0.70–0.92), 5TE% (adjOR 0.89, 95%CI 0.79–0.99), and 10TE% (adjOR 0.86, 95%CI 0.75–0.97) thresholds. Free sugars from liquid sources were associated with greater odds of overweight across the 2.5TE% (adjOR 1.20, 95%CI 1.07–1.36), 10TE% (adjOR 1.17, 95%CI 1.02–1.34), and 15TE% (adjOR 1.43, 95%CI 1.23–1.67) thresholds. There were 113 cases of GDM among the 1842 women who delivered (6.1%). Free sugars from solid sources were associated with lower odds of GDM above versus below the 5TE% threshold (adjOR 0.56, 95%CI 0.36–0.85). Conclusions Our findings support limiting free sugars from liquid sources, given associations with overweight. We did not identify adverse associations of free sugars from solid sources across any of the thresholds examined.

Background: Hospital stays that are prolonged due to non-clinical factors are costly to health care systems and are likely suboptimal for patient well-being. We assessed the influence of psychosocial factors on hospital length of stay (LOS) for older Canadians in a retrospective cohort study. Data and Methods: Data from the Canadian Community Health Survey were linked with the Discharge Abstract Database. Analyses were stratified by age, 55–64 ( n = 1,060) and 65 and older ( n = 2,718). Main predictor variables of interest included four measures of social support, sense of belonging, and living alone. Multivariate models of LOS adjusted for age, sex, income, smoking, and frailty. Results: Among the younger respondents, low positive social interactions, low emotional/informational support, and living alone were associated with a longer LOS. Among respondents 65 and older, low affection, low positive social interactions, low emotional/informational support, and a weak sense of belonging were associated with a longer LOS. Discussion: Having low social support is associated with longer hospital stays in this Canadian cohort. Social support may influence LOS as risk factors for poor health and precarious care in the community. Mitigating these risk factors could reduce the economic burden that is played out through longer hospital stays.

Objectives Only a few cardiovascular risk prediction models have been developed based on modifiable exposures and even fewer utilize complex dietary factors. Our objective was to develop and validate the Cardiovascular Disease Population Risk Tool (CVDPoRT)-Nutrition as a tool for estimating 5-year risk of incident cardiovascular disease (CVD) using lifestyle factors and dietary pattern scores. Approach The CVDPoRT developed and validated using the Canadian Community Health Survey (CCHS) linked with health administrative databases was modified to remove limited measures of dietary intakes (i.e., fruit and vegetable, potato, and juice intake frequency) and instead incorporate 5 different dietary pattern scores (i.e., Dietary Guidelines for Americans Adherence Index, Dietary Approaches to Stop Hypertension, Healthy Eating Index, Alternative HEI, and Mediterranean Style Dietary Pattern Score). Outcome data (i.e., CVD events and CVD-related mortality) came from linkage with the Canadian Vital Statistics – Death Database and Discharge Abstract Database. CVDPoRT-Nutrition was tested in 61 policy-relevant subgroups. Results Performance after adding in each dietary pattern score was similar to the original CVDPoRT (Brier score=2.6%, Harrell’s c-stat=0.87(0.85-0.88) for female models; Brier score=1.6%, Harrell’s c-stat=0.82(0.81-0.84) for male models). The algorithm was calibrated in 53 (female models) and 57 (male models) of 61 policy relevant subgroups. The most important predictors of CVD and CVD-related mortality were age, sex, and smoking. Conclusion Altering the dietary measures included in the CVDPoRT algorithm did not greatly improve the predictive capacity. The original CVDPoRT can continue to be used for predicting CVD and CVD-related mortality, while CVDPoRT-Nutrition may be used for predicting CVD incidence associated with poor dietary patterns.

Introduction Neighbourhoods have the potential to influence population-wide modifiable risk factors such as physical inactivity and obesity. Built environments that encourage active living hold promise as a policy lever for reducing health care burden, particularly that related to cardiometabolic disease. Objectives and Approach We examined the role of active living environments on hospitalization risk, frequency, and cumulative length of stay for all-causes and cardiometabolic diseases. The linked dataset is a combination of survey data from Canadian respondents aged 45+, records from a national census of acute hospitalizations, and the Canadian Active Living Environment (Can-ALE) - a 5-class measure of how conducive one’s neighbourhood is to active living based on street connectivity, points of interest, and population density. We modelled the risk of all-cause and cardiometabolic hospitalizations for respondents living in more and less favourable environments using logistic regression. Frequency and cumulative length of stay were modelled using truncated negative binomial regression. Models were adjusted for individual-level factors and proximity to a hospital. An offset variable was included to account for different follow-up times. Results 232,000 respondents were included with a mean follow-up time of 5.37 years. Those living in progressively more favourable active living environments (classes 2, 3, 4, and 5) exhibited incrementally lower risk of hospitalization compared to those living in the least favourable (class 1). Relative to respondents living in the least favourable environments (class 1), odds ratios were 0.84 (95% CI 0.76-0.93) for all-cause hospitalization and 0.80 (95% CI 0.68-0.93) for cardiometabolic hospitalization for respondents living in the most favourable environments (class 5). There was little evidence of similar associations for hospitalization frequency and cumulative length of stay. Conclusion / Implications Living in neighbourhoods that are more conducive to active living are associated with lower risk of all-cause and cardiometabolic hospitalization.

Abstract Background In Canada, long waiting times for core specialized services have consistently been identified as a key barrier to access. Governments and organizations have responded with strategies for better access management, notably for total joint replacement (TJR) of the hip and knee. While wait time management strategies (WTMS) are promising, the factors which influence their sustainable implementation at the organizational level are understudied. Consequently, this study examined organizational and systemic factors that made it possible to sustain waiting times for TJR within federally established limits and for at least 18 months or more. Methods The research design is a multiple case study of WTMS implementation. Five cases were selected across five Canadian provinces. Three success levels were pre-defined: 1) the WTMS maintained compliance with requirements for more than 18 months; 2) the WTMS met requirements for 18 months but could not sustain the level thereafter; 3) the WTMS never met requirements. For each case, we collected documents and interviewed key informants. We analyzed systemic and organizational factors, with particular attention to governance and leadership, culture, resources, methods, and tools. Results We found that successful organizations had specific characteristics: 1) management of the whole care continuum, 2) strong clinical leadership; 3) dedicated committees to coordinate and sustain strategy; 4) a culture based on trust and innovation. All strategies led to relatively similar unintended consequences. The main negative consequence was an initial increase in waiting times for TJR and the main positive consequence was operational enhancement of other areas of specialization based on the TJR model. Conclusion This study highlights important differences in factors which help to achieve and sustain waiting times. To be sustainable, a WTMS needs to generate greater synergies between contextual-level strategy (provincial or regional) and organizational objectives and constraints. Managers at the organizational level should be vigilant with regard to unintended consequences that a WTMS in one area can have for other areas of care. A more systemic approach to sustainability can help avoid or mitigate undesirable unintended consequences.

Introduction The current opioid crisis is recognized by governments at many levels as an urgent priority. While there is basic demographic information on who experiences opioid related adverse events, there is little information to provide a more fulsome profile of those experiencing these events for targeted intervention and forward projection estimation. Objectives and Approach This study is the first to use a nationally-representative Census of Population linked with health administrative data to examine opioid-related hospitalization patterns across income and Aboriginal status. Results Preliminary analyses using 2006 Census-hospital linked database found cohort rates of opioid-related hospitalizations are up to 7 times higher among Aboriginal youth, and also for young adults (12-19/ 20-24) compared with non-Aboriginal youth or young adults. Rates among these youth living on reserves were 8.4 times higher; among those off reserve 8.7 times higher. Rate among all youth living in lower income households was 5 times higher compared with those living in highest income households, For Aboriginal persons in lowest income quintile, the rate was 3 times higher relative to non-Aboriginal persons in same quintille. Conclusion/Implications New linked health data reveal new information regarding the profile of those who experienced opioid-related adverse events. This information will serve to inform targeted intervention strategies, models for forward estimation of events.

Introduction Linking health and socio-demographic data is providing new opportunities for measuring health inequalities in Canada. Measuring inequalities across relevant population sub-groups is a key step for understanding progress towards achieving health equity and informing interventions to reduce inequities. Objectives and Approach This study uses two linked national datasets (excluding Quebec): 1) Statistics Canada’s linkage of the 2006 long-form Census to hospital data from the Canadian Institute for Health Information’s 2006-2008 Discharge Abstract Database (DAD); 2) hospital data linked to area-level income using Statistics Canada’s Postal Code Conversion File tool. Study objectives are to compare hospitalization rates between the linked census cohort and the general population and to inform the measurement of inequalities by income and education. Analysis was based on three hospitalization indicators: chronic obstructive pulmonary disease (COPD) among Results The linked cohort weighted to the population represented 67\% of DAD cases for asthma, 77\% for COPD and 80\% for heart attack. The lower coverage for asthma was because babies born after Census day were not linked. For all three indicators, rates were lower in the linked cohort than the DAD, likely because the linked cohort and DAD capture somewhat different populations (e.g., institutionalized population is in DAD but not Census). Large income and education-related inequalities were observed for all three indicators. Further analyses revealed inequalities of similar magnitude using before and after-tax income data, and larger inequalities by individual-level income compared to area-level (for COPD only). Inequalities were similar using individual and household education, with a clear gradient using multiple education categories. Conclusion/Implications Updating and expanding existing linkages to include recent data has the potential to expand and integrate health inequalities measurement into routine health system performance reporting. Challenges include measuring inequalities for smaller populations (e.g., health regions) and gaps in Census coverage (e.g., newborns/newcomers, urban indigenous populations, people experiencing homelessness, institutionalized population).

Introduction The built environment can shape modifiable risk factors such as obesity, poor diet, and physical inactivity, and could be a policy lever for the reduction of chronic disease. In Canada, the health care costs related to chronic disease continue to rise and there have been few policy options offered. Objectives and Approach We examine the role of the built environment in hospital burden and premature mortality, with an emphasis on one of the highest burden diseases, Type 2 Diabetes (T2D). Neighbourhood built environment measures for active living were derived using geographic information systems for respondents of the Canadian Community Health Survey, for whom we have linked hospitalization and mortality records. A combination of ICD codes, self-reported diabetes status, as well as a population-based algorithm identifying those at higher risk of developing diabetes were used to identify cases. Differences in hospitalization frequency, cumulative length of stay, and mortality are investigated. Results Over half a million hospitalization records were identified in our cohort of roughly 450,000 survey respondents. Key factors such as age, gender, race, and socioeconomic status are accounted for in modelling the association between neighborhood environment and hospitalization. Hospital burden and mortality in T2D patients are much higher than that of patients who do not report having the condition, and those at elevated risk of T2D display intermediate levels of hospitalization. Two-part hurdle models show evidence of an association between more walkable neighborhoods and lower hospitalization risk in non-T2D patients as well as those at elevated risk of developing T2D. The relationship between neighborhoods and the volume of chronic-disease related episodes as well as mortality is unclear, and under further investigation. Conclusion/Implications Elucidating the role of neighbourhood built environments on hospital burden and premature mortality for individuals with diabetes will provide insight as to the full range of clinical and non-clinical interventions that could feasibly address the needs of some the highest health care system users.

Introduction Dietary pattern modeling and linkage with health outcomes is essential for development of evidence-based dietary guidelines to support reduction of chronic diseases. National nutrition surveys are not routinely linked with health administrative databases, resulting in a lack of evidence on the health impact of unhealthy diets at the population level. Objectives and Approach This study is the first to use a nationally-representative nutrition survey (i.e., Canadian Community Health Survey-Nutrition-2004) linked with health administrative databases to examine the association of 5 key dietary quality indices with mortality risk. In total, 16,212 adults ≥20 years were followed for an average of 7.3 years. Two 24-hour dietary recalls were used to estimate the usual dietary intakes using the National Cancer Institute’s method. Weighted regression calibration was performed to obtain a true parameter relating diet (continuous) to mortality. Population Attributable Fractions were calculated to estimate the burden of all-cause mortality attributable to poor dietary patterns in Canada. Results The 5 dietary quality indices examined were Dietary Approaches to Stop Hypertension(DASH); Dietary Guidelines for Americans Adherence Index 2015(DGAI); Healthy Eating Index-2010(HEI); Alternative HEI-2010(AHEI); and Mediterranean Style Dietary Pattern Score(MSDPS). Having a better diet quality (90%ile vs. 10%ile of index score) was associated with a significant 31-51% reduction in all-cause mortality hazard ratio among adults 45 to 80 years and 10-35% reduction in those ≥20 years (in order of significance: DASH, DGAI, HEI, AHEI and MSDPS). Survival benefit was incrementally greater for higher diet quality scores; however, even the 90%ile scores (Reference) were notably lower than the recommended levels (45.99% of recommended score). On average, 26.42% of all mortality in Canada was attributable to poor dietary patterns (range: 19.01% for MSDPS to 31.39% for DGAI). Conclusion/Implications The diet-attributable burden of mortality was higher than those reported for other behavioural risks (e.g., smoking). This research informs future formulation of nutrition interventions and policies with a focus on dietary patterns. This project demonstrates the importance of leveraging linked data and analytical capacity to inform future evidence-based nutrition policies.

Introduction National health surveys, available in over 100 countries, are the most common data used for health behaviour surveillance and are increasingly being linked to individual-level health outcomes. We propose that improved health behaviour hazard estimates can be obtained from pooled international population health surveys linked to outcome data. Objectives and Approach The objective of this study was to compare smoking, alcohol, diet and physical activity all-cause mortality hazards in Canada and the United States using individual-level, linked population health survey data and common model specifications. The Canadian Community Health Survey (CCHS) (2003-2007) and the United States National Health Interview Survey (NHIS) (2000, 2005) linked to individual-level mortality outcomes with follow up to December 31, 2011 were used. Variable definitions consistent across the CCHS and NHIS were developed and used to estimate country-specific mortality hazards with sex-specific Cox proportional hazard models, including health behaviours, sociodemographic indicators and proximal factors including disease history. Results A total of 296,407 respondents and 1,813,884 million person-years of follow-up from the CCHS and 62,226 respondents and 497,909 person-years from the NHIS were included. Hazards of smoking, alcohol consumption, diet and physical activity in Canada and the United States are of similar magnitude and direction, with similar dose response relationships. The largest health behaviour mortality hazards were associated with female heavy smokers in both Canada (HR: 3.36, 95% CI: 2.86, 3.95) and the United States (Female HR: 2.63, 95% CI: 2.11, 3.27), compared to non-smokers. Conclusion/Implications Health behaviour mortality hazards are comparable in Canada and the United States, supporting the use of hazards obtained from pooled analyses for population heath. These hazards can replace those obtained from independent epidemiology studies that are often incompletely adjusted, rarely population-based and often not generalizable to the population of interest.

Introduction Hospital utilization varies across socioeconomic and demographic strata in Canada, which has a universal health care system that grants essential services to everyone. Rates of adverse birth outcomes are known to differ among high and low SES women, but less is known of the excess burden attached to those outcomes across Canadian provinces. Objective To examine length of stay for childbirth relative to women’s socio-demographic characteristics, in the context of the Canadian universal health care system. Methods A population-based record linkage between the Canadian Community Health Survey (CCHS) cycles 3.1 (2005) and 4.1 (2007/8), and the Discharge Abstract Database (DAD) allowed the tracking of hospital utilization for linked survey respondents between 2005 and 2009. Hourly length of stay for delivery was modeled by socio-demographic factors, controlling for other clinical and individual-level characteristics. Results There were 7,166 complete delivery records from 5,570 female CCHS respondents who agreed to link and share their information. Women with the lowest income had on average, four-hour longer stays for vaginal delivery as compared to high-income women (IRR 1.07, 95% CI 1.02-1.13, p=0.01), and eight-hour longer stays for Caesarian delivery (IRR 1.08, 95% CI 0.95-1.22, p=0.23). A greater proportion of teenage pregnancy was seen for Aboriginal girls. Aboriginal status and rural area of residence were co-determinants of elevated length of stay. Conclusion The absence of egregious socio-demographic differences regarding childbirth is reassuring for the Canadian health care system. However, the persistence of marginally longer, and in turn, costlier visits for low-income and rural Aboriginal women is suggestive that policies of cash transfers during the prenatal period might be highly cost-effective if they achieve population-wide reductions in length of stay.

Information is increasingly digital, creating opportunities to respond to pressing issues about human populations in near real time using linked datasets that are large, complex, and diverse. The potential social and individual benefits that can come from data-intensive science are large, but raise challenges of balancing individual privacy and the public good, building appropriate socio-technical systems to support data-intensive science, and determining whether defining a new field of inquiry might help move those collective interests and activities forward. A combination of expert engagement, literature review, and iterative conversations led to our conclusion that defining the field of Population Data Science (challenge 3) will help address the other two challenges as well. We define Population Data Science succinctly as the science of data about people and note that it is related to but distinct from the fields of data science and informatics. A broader definition names four characteristics of: data use for positive impact on citizens and society; bringing together and analyzing data from multiple sources; finding population-level insights; and developing safe, privacy-sensitive and ethical infrastructure to support research. One implication of these characteristics is that few people possess all of the requisite knowledge and skills of Population Data Science, so this is by nature a multi-disciplinary field. Other implications include the need to advance various aspects of science, such as data linkage technology, various forms of analytics, and methods of public engagement. These implications are the beginnings of a research agenda for Population Data Science, which if approached as a collective field, can catalyze significant advances in our understanding of trends in society, health, and human behavior.

ABSTRACT Objectives Birth and delivery in hospital is one of the most common medical procedures in Canadian hospitals, and can be used to assess equity in the delivery of health care. This study investigates the association between socioeconomic status and hospital burden for childbirth using linked Canadian survey and administrative databases, accounting for a wide array of other individual and health-care related characteristics. Approach A population-based record linkage between national health survey data and the Canadian Discharge Abstract Database (a census of all Canadian hospital separations) allowed the tracking of hospital utilization between 2005 and 2009 for which individual-level socioeconomic and demographic factors were also available. Length of stay for delivery, risk of pre-delivery hospitalization within 30 days of admission for delivery, and risk of maternal readmission within 30 days of discharge for delivery were the three measures of hospital utilization modeled. Results Complete information for 7,163 deliveries of 5,568 women was available and used in the models of length of stay and risk of maternal admission pre- and post-delivery. In fully adjusted models, predicted length of stay was graded by household income with longest stays for lowest income women (2.79 days, 95% CI 2.61-2.92), followed by middle income women (2.63 days, 95% CI 2.50-2.76) and high income women (2.56 days, 95% CI 2.49-2.63). Factors intrinsic to routine hospital care and protocol such as province and vaginal versus Cesarean section delivery were stronger predictors of length of stay than income. Additionally, Aboriginal status, compounded with residing in a rural setting was associated with higher predicted probability of maternal readmission. Conclusions Low income women have marginally longer stays in hospital following birth events than do middle and high-income women in Canada that persist after adjustment for strong drivers of length of stay (parity and birth mode). While this may suggest that more complicated cases are benefiting from longer stays in hospital, it provokes the question as to whether there may be additional reductions in length of stay that could be achieved through resource redistribution to the prenatal period. Overall, the results suggest that equity goals of the Canadian health care system are being achieved for birth-related hospitalizations, and the availability of these linked data will be key for further comprehensive evaluation of population health status.

ABSTRACT Objectives While existing data sources, such as the 2011 Census, provide an accurate count of who is currently institutionalized, there is a significant gap in terms of our understanding of who is at risk for institutionalization and what the future demand for care will be. The objective of this study is to use linked national health survey, the 2005 Canadian Community health Survey (CCHS), to the 2011 Census to identify factors associated with transitions from private households to alternative living arrangements, specifically long-term care (nursing homes), and retirement homes among Canadians 55 years of age and older. Approach Hierarchical deterministic methods were used to link the 2005 CCHS (n=114,000) to the 2011 Census (n=35 million) using identifying variables common to both data sets (i.e. name, birthdate, sex, postal code, social insurance number). Sex specific multivariate regression models with multiple outcomes were used to assess the impact of a comprehensive set of factors (i.e. demographic, socio-economic, health status, chronic conditions and marital status) available in the CCHS on the likelihood of residing in three possible home environments, relative to living in a private dwelling (PD): long-term care (LTC), retirement homes (RH) or private dwelling with support (PDS) as identified in the 2011 Census. Analyses were adjusted for mortality. Results Over 85% of CCHS records were linked to the 2011 Census (n=92,849). Among those 55 years of age and older (n=29,934), approximately 2.0% and 1.6% were living in LTC and RH respectively: an additional 7.2% were living in PDS. Results of the regression analyses, revealed that those with Alzheimer’s disease were at highest risk of transition to LTC (OR=11.7 females; 6.8 males). Losing a spouse was significantly associated with transitions to LTC, RH and PDS for both men and women. Being an immigrant was protective, with immigrant seniors less likely to transition to LTC and RH. Other factors significantly associated with transitions to LTC included low income, poor mental health (women only), and assistance with activities of daily living (meal preparation for women, finances for men). Regional variations were also noted. Conclusion Newly linked health survey and census data provide a unique opportunity to take a comprehensive look at those most at risk for institutionalization.

ABSTRACT Objectives In Canada, refugees represent between 10% and 15% of the total immigrant annual in-take of immigrants. Refugees are generally admitted with different selection criteria and have different settlement challenges which may lead to higher health risk, compared to other categories of immigrants. The purpose of this study is to use linked hospital and immigrant landing data to report hospitalization rates among refugees arriving to Canada between 1980 and 2006 with special focus on those from refugee-dominant source areas, namely Poland and Vietnam, and the Middle East. Approach Data from the 2006/07-2008/09 Discharge Abstract Database (DAD) were linked to the 1980-2008 Immigrant Landing File (ILF) to identify hospitalizations among immigrants by category of admission. Age-standardized hospitalization rates (ASHR) for overall causes (excluding pregnancy), and for selected causes are derived for refugees overall and by specific source country. Rates are compared with those for the Canadian-born population and with economic class immigrants from the same source country. Results Overall, refugees had substantially lower overall ASHR compared with the Canadian-born population (494 per 10,000, 95% CI, 487, 500 vs 891 per 10,000, 95% CI, 890, 892, retrospectively). Among refugees, those who landed in Canada (refugee claimants) had the highest ASHR (539, 95% CI 524,555), especially for circulatory disease. Among refugee-dominant areas, Government-assisted refugees from the Middle East tended to have higher ASHR at levels comparable with that of the Canadian born population (for circulatory disease). Refugees had higher rates of hospitalization compared with their economic class counterparts, except among those from Poland. Conclusion Refugees, like other immigrants, generally had lower hospitalization rates compared with the Canadian born population, but the heterogeneity within groups reveals specific vulnerability for certain types of health care use.

ABSTRACT Objectives This project links data on acute inpatient hospitalizations from the Canadian Discharge Abstract Database (DAD) with data on income and employment from various taxation- and employment-based administrative files. The goal was to create a linked database that will support research on the labour market and financial outcomes experienced by individuals and families following acute illness requiring hospitalization. Approach Data from the 1999/00 to 2014/15 Discharge Abstract Database (DAD) were linked to the 1981-2013/14 T1 Tax filer data and the Canadian Child Tax Benefit data. We sought to create a unique association between Health Insurance Numbers (HIN) available in the DAD and Social Insurance Numbers (SIN) available in the tax data by using variables common to both data sets – date of birth, postal code and sex. Both transactional data sets were “individualized” such that unique combinations of the linkage variables were identified and eligible for linkage. The linkage was conducted using deterministic methods. Results Approximately 97% of combinations involving date of birth, postal code and sex in the hospitalization data were uniquely related to a single valid HIN (n=18.8 million). Similarly, approximately 96% of the keys on the Tax data file were associated with a unique person. Approximately 86% of HINs were associated with a unique identifier in the tax file and these HINs account for approximately 83% of the hospital records. The linkage was consistent over time, with linkage rates between 85% and 88% of HINs for all years. Some variation in linkage rates were observed by jurisdiction and by age. (Error estimates to be reported) Conclusion This project has created a unique linked database that will support research on the economic consequences of ‘health shocks’ for individuals and their families, and the implications for income, labour and health policies. This database represents a new and unique resource that will fill an important national data gap, and enable a wide range of relevant research.

ABSTRACT Objectives Increasingly, national health surveys are being linked to vital statistics and health care information, providing a new and unique source of individual population health data. Given that nationally-representative health surveys are performed in over a hundred countries, these linkages create comprehensive data sets that are potentially larger than most existing cohort studies. To date, this resource has not been utilized. Approach The purpose, study base, content and methods of the Canadian Community Health Survey (CCHS) cycles 2.1 (2003-04) and 3.1 (2005-06) and the United States 2000 and 2005 National Health Interview Survey (NHIS) were examined for comparability and consistency. Smoking, alcohol, physical activity and diet questions were identified, question construct and response categorization were compared, and variable constructions possible for both national health surveys were created. All respondents 20+ years of age were identified and stratified by country and sex. Cox proportional hazard models were used to estimate 5-year hazards of mortality associated with the common smoking, alcohol, physical activity and diet variables. Results The CCHS and NHIS are highly consistent and comparable. Health behaviour questions are similar and permit the creation of smoking, alcohol, physical activity and diet variables that are comparable across surveys. A total of 284 475 survey respondents from Canada and the United States (CCHS, N= 226 731; NHIS, N= 57 744) were included. The largest mortality hazards were associated with female heavy smokers in both Canada (HR: 2.91; 95% CI: 2.52, 3.37) and the United States (Female HR: 2.96; 95% CI: 2.59, 3.38), compared to non-smokers. Moderate variation in the age adjusted all-cause mortality hazard ratios was observed; both smoking and physical activity hazard ratios were consistently higher in the United States than in Canada. Conclusion This study provides initial support for the methodological feasibility of pooling linked population health surveys however, challenges introduced by dissimilarities will require the use of innovative methodologies, and discussions regarding how to manage jurisdictional data restrictions and privacy issues are needed. Pooled population health data has the potential to improve national and international health surveillance and public health.

ABSTRACT Objectives To demonstrate the methodology and results for linking measures of neighbourhood walking-friendliness or "walkability" to Canadian health surveys and Canadian health surveys linked to administrative health care records. Approach We linked multiple measures of neighbourhood walkability to hundreds of thousands of 6-digit postal codes of respondents to three large Canadian surveys using geographic information systems and anonymized banks of postal codes. Results Long term exposure to walkable neighbourhoods was associated with increased reporting of walking for utilitarian purposes. Moving to a high walkable neighbourhood from a low walkable neighbourhood was associated with a full unit decrease in the body mass index of Canadian men over a 12-year follow-up. In a linkage of walkability measures to respondents who wore biosensors for a one-week period in several Canadian cities, we found that neighbourhood walkability was associated with increased reporting of utilitarian walking but not overall physical activity and step counts as measured by biosensors. Conclusion There is potential for walkable neighbourhoods to influence physical activity and body weight of Canadians which is more evident when individuals are followed for long periods of time.

BackgroundBehaviours such as smoking, poor diet, physical inactivity, and unhealthy alcohol consumption are leading risk factors for death. We assessed the Canadian burden attributable to these behaviours by developing, validating, and applying a multivariable predictive model for risk of all-cause death.MethodsA predictive algorithm for 5 y risk of death-the Mortality Population Risk Tool (MPoRT)-was developed and validated using the 2001 to 2008 Canadian Community Health Surveys. There were approximately 1 million person-years of follow-up and 9,900 deaths in the development and validation datasets. After validation, MPoRT was used to predict future mortality and estimate the burden of smoking, alcohol, physical inactivity, and poor diet in the presence of sociodemographic and other risk factors using the 2010 national survey (approximately 90,000 respondents). Canadian period life tables were generated using predicted risk of death from MPoRT. The burden of behavioural risk factors attributable to life expectancy was estimated using hazard ratios from the MPoRT risk model.FindingsThe MPoRT 5 y mortality risk algorithms were discriminating (C-statistic: males 0.874 [95% CI: 0.867-0.881]; females 0.875 [0.868-0.882]) and well calibrated in all 58 predefined subgroups. Discrimination was maintained or improved in the validation cohorts. For the 2010 Canadian population, unhealthy behaviour attributable life expectancy lost was 6.0 years for both men and women (for men 95% CI: 5.8 to 6.3 for women 5.8 to 6.2). The Canadian life expectancy associated with health behaviour recommendations was 17.9 years (95% CI: 17.7 to 18.1) greater for people with the most favourable risk profile compared to those with the least favourable risk profile (88.2 years versus 70.3 years). Smoking, by itself, was associated with 32% to 39% of the difference in life expectancy across social groups (by education achieved or neighbourhood deprivation).ConclusionsMultivariable predictive algorithms such as MPoRT can be used to assess health burdens for sociodemographic groups or for small changes in population exposure to risks, thereby addressing some limitations of more commonly used measurement approaches. Unhealthy behaviours have a substantial collective burden on the life expectancy of the Canadian population.

Health behaviours, important factors in cardiovascular disease, are increasingly a focus of prevention. We appraised whether stroke risk can be accurately assessed using self-reported information focused on health behaviours.Behavioural, sociodemographic and other risk factors were assessed in a population-based survey of 82,259 Ontarians who were followed for a median of 8.6 years (688,000 person-years follow-up) starting in 2001. Predictive algorithms for 5-year incident stroke resulting in hospitalization were created and then validated in a similar 2007 survey of 28,605 respondents (median 4.2 years follow-up).We observed 3236 incident stroke events (1551 resulting in hospitalization; 1685 in the community setting without hospital admission). The final algorithms were discriminating (C-stat: 0.85, men; 0.87, women) and well-calibrated (in 65 of 67 subgroups for men; 61 of 65 for women). An index was developed to summarize cumulative relative risk of incident stroke from health behaviours and stress. For men, each point on the index corresponded to a 12% relative risk increase (180% risk difference, lowest (0) to highest (9) scores). For women, each point corresponded to a 14% relative risk increase (340% difference, lowest (0) to highest (11) scores). Algorithms for secondary stroke outcomes (stroke resulting in death; classified as ischemic; excluding transient ischemic attack; and in the community setting) had similar health behaviour risk hazards.Incident stroke can be accurately predicted using self-reported information focused on health behaviours. Risk assessment can be performed with population health surveys to support population health planning or outside of clinical settings to support patient-focused prevention.

BACKGROUND: Management of chronic diseases requires patients to adhere to recommended health behavior change and complete tests for monitoring. While studies have shown an association between low income and lack of adherence, the reasons why people with low income may be less likely to adhere are unclear. We sought to determine the association between household income and receipt of health behavior change advice, adherence to advice, receipt of recommended monitoring tests, and self-reported reasons for non-adherence/non-receipt. METHODS: We conducted a population-weighted survey, with 1849 respondents with cardiovascular-related chronic diseases (heart disease, hypertension, diabetes, stroke) from Western Canada (n = 1849). We used log-binomial regression to examine the association between household income and the outcome variables of interest: receipt of advice for and adherence to health behavior change (sodium reduction, dietary improvement, increased physical activity, smoking cessation, weight loss), reasons for non-adherence, receipt of recommended monitoring tests (cholesterol, blood glucose, blood pressure), and reasons for non-receipt of tests. RESULTS: Behavior change advice was received equally by both low and high income respondents. Low income respondents were more likely than those with high income to not adhere to recommendations regarding smoking cessation (adjusted prevalence rate ratio (PRR): 1.55, 95%CI: 1.09-2.20), and more likely to not receive measurements of blood cholesterol (PRR: 1.72, 95%CI 1.24-2.40) or glucose (PRR: 1.80, 95%CI: 1.26-2.58). Those with low income were less likely to state that non-adherence/non-receipt was due to personal choice, and more likely to state that it was due to an extrinsic factor, such as cost or lack of accessibility. CONCLUSIONS: There are important income-related differences in the patterns of health behavior change and disease monitoring, as well as reasons for non-adherence or non-receipt. Among those with low income, adherence to health behavior change and monitoring may be improved by addressing modifiable barriers such as cost and access.