Article Text
PDF
Abstract
Introduction Patients with diabetes have a higher risk of mortality compared with the general population. Large population-based studies that quantify variations in mortality risk for patients with diabetes among subgroups in the population are lacking. This study aimed to examine the sociodemographic differences in the risk of all-cause mortality, premature mortality, and cause-specific mortality in persons diagnosed with diabetes.
Research design and methods We conducted a population-based cohort study of 1 741 098 adults diagnosed with diabetes between 1994 and 2017 in Ontario, Canada using linked population files, Canadian census, health administrative and death registry databases. We analyzed the association between sociodemographics and other covariates on all-cause mortality and premature mortality using Cox proportional hazards models. A competing risk analysis using Fine-Gray subdistribution hazards models was used to analyze cardiovascular and circular mortality, cancer mortality, respiratory mortality, and mortality from external causes of injury and poisoning.
Results After full adjustment, individuals with diabetes who lived in the lowest income neighborhoods had a 26% (HR 1.26, 95% CI 1.25 to 1.27) increased hazard of all-cause mortality and 44% (HR 1.44, 95% CI 1.42 to 1.46) increased risk of premature mortality, compared with individuals with diabetes living in the highest income neighborhoods. In fully adjusted models, immigrants with diabetes had reduced risk of all-cause mortality (HR 0.46, 95% CI 0.46 to 0.47) and premature mortality (HR 0.40, 95% CI 0.40 to 0.41), compared with long-term residents with diabetes. Similar HRs associated with income and immigrant status were observed for cause-specific mortality, except for cancer mortality, where we observed attenuation in the income gradient among persons with diabetes.
Conclusions The observed mortality variations suggest a need to address inequality gaps in diabetes care for persons with diabetes living in the lowest income areas.
- mortality
- socioeconomic status
- income
- immigrant status
- type 2
Data availability statement
Data may be obtained from a third party and are not publicly available. The data set from this study is held securely in coded form at ICES. While legal data sharing agreements between ICES and data providers (eg, healthcare organizations and government) prohibit ICES from making the data set publicly available, access may be granted to those who meet prespecified criteria for confidential access, available at www.ices.on.ca/DAS (email: das@ices.on.ca). The full data set creation plan and underlying analytic code are available from the authors upon request, understanding that the computer programs may rely upon coding templates or macros that are unique to ICES and are therefore either inaccessible or may require modification.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Individuals with diabetes are at higher risk of all-cause mortality and cardiovascular disease mortality compared with the general population.
It remains unclear how mortality risk varies among subgroups in the diabetes population.
WHAT THIS STUDY ADDS
This population-based study included 1 741 098 individuals diagnosed with diabetes and found that those who lived in the lowest income neighborhoods had a higher risk of all-cause mortality, premature mortality, and cause-specific mortality, relative to individuals with diabetes living in the highest income areas.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
The findings of this study suggest a need to address inequities in diabetes care among persons with diabetes who live in low-income areas.
Introduction
Diabetes contributes to health resource burdens in healthcare systems. Most people with type 2 diabetes have more than one chronic condition,1 known as multimorbidity, which increases the complexity of care and resource coordination.2 3 The global estimate of the health expenditure due to diabetes was US$673 billion in 2015,4 and the estimated direct healthcare costs of people with diabetes have been shown to be more than double than those without diabetes.5
Individuals with diabetes are at a higher risk of all-cause mortality and cardiovascular disease mortality than the general population.6 7 Multimorbidity in those with type 2 diabetes is associated with higher all-cause mortality.8 Other well-known risk factors that increase mortality risk in the diabetes population include age at diabetes diagnosis, poor lipid control, and history of hypertension and cardiovascular diseases.6 There is a lack of studies that quantify if variation in mortality risk exists among vulnerable groups in the diabetes population, such as those with low socioeconomic status (SES) and immigrants, due to an inability to identify these groups in health databases since this information is not collected clinically.
Understanding variations in mortality risk can reveal subgroups among the diabetes population who are disproportionately affected and health service needs for persons with diabetes since effective interventions exist.9 10 Immigrants in Canada have a health advantage, known as the healthy immigrant effect, in which they have been shown to have generally better health than their Canadian-born counterparts, including lower all-cause mortality and cause-specific mortality rates.11–13 Income disparities associated with all-cause and premature mortality in the general population are well documented,14 15 although the magnitude of these disparities is shown to be less pronounced among immigrants compared with long-term residents.16 Despite an immigrant health advantage, immigrant populations experience a large burden of diabetes and certain ethnic groups are at higher risk of type 2 diabetes, including South Asian, African, and Indigenous populations.17–19 In addition, the immigrant health advantage tends to decline as residency in the host country lengthens.20 In Canada, age-standardized mortality rates were shown to be lower among recent and medium-term immigrants compared with long-term immigrants.13 21
In this population-based study of individuals diagnosed with diabetes, we examine the sociodemographic differences in the risk of all-cause mortality, premature mortality, and cause-specific mortality. We also quantify the health system impact in persons with diabetes by measuring chronic conditions at the time of diagnosis and healthcare costs in the last 2 years of life.
Research design and methods
Study population and design
We conducted a population-based cohort study of patients with diabetes with diagnosis dates between 1 January 1994 and 31 December 2017. Patients were residents of the province of Ontario, Canada and aged between 18 and 120 years. We excluded those patients with invalid health card, missing age or sex, non-Ontario residence or when a death date was recorded before the diabetes diagnosis date. Data sets were linked using unique encoded identifiers and analyzed atICES in Toronto, Ontario. ICES is an independent, non-profit research institute whose legal status under Ontario’s health information privacy law allows it to collect and analyze healthcare and demographic data, without consent, for health system evaluation and improvement.
Data sources
The Ontario Diabetes Database (ODD) was used to identify patients with a diabetes diagnosis.22 Briefly, the ODD uses a validated algorithm applied to inpatient hospitalization (Canadian Institute for Health Information Discharge Abstract Database, DAD) records, same-day surgery (SDS) records, and physician billing claims (Ontario Health Insurance Plan, OHIP) data to determine the diagnosis of incident cases of diabetes in Ontario. The ODD has demonstrated 90% sensitivity and 97% specificity. The definition for diabetes is two physician billing claims with a diagnosis for diabetes (OHIP diagnosis code: 250) or one inpatient hospitalization or SDS record with a diagnosis for diabetes (International Classification of Diseases 9th Revision (ICD-9) diagnosis code: 250; ICD-10 diagnosis codes: E10, E11, E13, E14; in any diagnostic code space) within a 2-year period. Physician claims and hospitalizations with a diagnosis of diabetes occurring within 120 days prior to and 180 days after a gestational hospitalization record are excluded.
Information on deaths recorded in the study period, including death dates and causes of death, was captured from the Office of the Registrar General-Deaths (ORG-D) file. Cause of death was coded using the ICD-9 and ICD-10 with Canadian Enhancements (ICD-9-CA/ICD-10-CA). The ORG-D was linked probabilistically to the Registered Persons Database (RPDB) with an overall linkage rate of 96.5%.23 The RPDB is a central population registry file that permits linkage with data holdings held at ICES and contains basic demographic information, such as birth date, sex, and postal code, for those who have ever received an Ontario health card number for the province’s universal healthcare system. We used the RPDB to identify death dates that were not captured in the ORG-D and patient’s demographic information.
Other administrative databases linked to identify history of chronic conditions and healthcare utilization costs included: hospital admissions and day surgery (DAD/SDS); the National Ambulatory Care Reporting System (NACRS) for all records of emergency room visits; the OHIP claims database for outpatient physician visits; Continuing Care Reporting System; National Rehabilitation Reporting System; Ontario Mental Health Reporting System; Home Care Cost Database; and Ontario Drug Benefit claims database, which captures prescription drugs for patients 65 years and older. We used several disease-specific registries that contain regularly updated population cohorts and were created by applying validated data algorithms to the DAD and OHIP, specifically: Ontario Asthma Database,24 Ontario Cancer Registry,25 Ontario Congestive Heart Failure Database,26 Ontario Chronic Obstructive Pulmonary Disease Database,27 Ontario Rheumatoid Arthritis Database,28 Ontario Hypertension Database,29 and Ontario Crohn’s and Colitis Cohort Database.30
The Ontario portion of the Immigration, Refugees and Citizenship Canada (IRCC) Permanent Resident Database was used to identify immigrants in our study population. The IRCC contains records for over 3 million individuals at the time of landing in Ontario from January 1985 to May 2017, and was linked to the RPDB with an 85.8% overall linkage rate.23 Neighborhood-level measures of income were calculated using data from the 1996, 2001, 2006, 2011 and 2016 Canadian census.
Exposure
Comorbidity burden was captured in the form of total Aggregated Diagnosis Group (ADG) scores which were derived from diagnosis codes patients received in both ambulatory (NACRS and OHIP) and inpatient care (DAD/SDS) in a period of 2 years prior to the index diagnosis date. Individual diseases or conditions are placed into a single ADG based on five clinical dimensions: duration of the condition (acute, recurrent, or chronic); severity of the condition (eg, minor and stable vs major and unstable); diagnostic certainty (symptoms focusing on diagnostic evaluation vs documented disease focusing on treatment services); etiology of the condition (infectious, injury or other); and specialty care involvement (medical, surgery, obstetric, hematology, etc). Individuals are assigned up to 32 ADGs, with higher ADG scores reflecting a greater burden of illness. ADG scores were calculated using The Johns Hopkins ACG System V.10.0.1.31
Area-level income information was obtained from the census and applied to individual cases according to the dissemination area (DA), which represents the smallest geographic census area in which the individual resided. Individuals were assigned to a DA based on their postal code at time of diabetes diagnosis. We applied 1996 census data for the years 1994 and 1998; 2001 census data for the years 1999 and 2003; 2006 census data for the years 2004 and 2008; 2011 census data for the years 2009 and 2013; and 2016 census data for diabetes diagnoses in 2014 and onward. Individuals were then categorized into income quintiles ranging from 1 (20% lowest income) to 5 (20% highest income). Immigrant status was defined as landed immigrants or refugees who were not Canadian citizens by birth and arrived in Ontario between 1985 and 2017.
Outcomes
The outcomes of interest included all-cause mortality, premature mortality, and cause-specific mortality. Mortality was determined by death date, from diabetes diagnosis until the end of follow-up (31 December 2018). Premature mortality was defined as a death before 75 years of age, from any cause. Cause of death was assigned using the ICES-derived cause of death variable in the ORG-D, which is based on Medical Certificate of Death coding, enhanced via linkages with other provincial data holdings, and converted to ICD-9 codes. Cause of death groupings were based on chapters of ICD-9: diseases of the cardiovascular and circulatory systems (ICD-9 codes 390–459), cancers (ICD-9 codes 140–239), external causes of injury and poisoning (ICD-9 codes 800–999), and diseases of the respiratory system (ICD-9 codes 460–519). All other deaths were assigned the cause of death category of ‘other’.
Health system impact was quantified according to number of chronic conditions and end of life healthcare spending, as calculated over the last 2 years of life. Chronic conditions included: acute myocardial infarction, asthma, cancer, cardiac arrhythmia, chronic coronary syndrome, chronic obstructive pulmonary disorder, congestive heart failure, Crohn’s or colitis disease, dementia, diabetes, hypertension, mood disorder, other mental health disorders, osteoarthritis and other arthritis, osteoporosis, rheumatoid arthritis, renal failure, and stroke. Multimorbidity was defined as the co-occurrence of two or more of these chronic conditions. The lookback period included all available data before diabetes diagnosis date. Diagnostic codes and details are found in online supplemental table S1. Healthcare costs in the last 2 years of life for those patients who died between 2004 and 2017 were computed by applying a person-centered costing approach to the linked administrative databases.32 Costs covered healthcare encounters accrued through Ontario’s single payer government insurer, including inpatient hospitalizations, emergency department visits, physician services, SDS, prescriptions, rehabilitation, complex continuing care, mental health inpatient stays, long-term care, and home care services. Cost values were adjusted to the year 2017. The costing methodology is described elsewhere.32
Supplemental material
Statistical analysis
We summarized baseline characteristics with stratification of sex and comparisons between female and male patients were carried out by Pearson’s χ2 test for categorical variables and analysis of variance or Kruskal-Wallis test for continuous variables. Kaplan-Meier survival curves were plotted and crude survival probabilities between these two groups were compared by the log-rank test.
The HRs for all-cause mortality and premature mortality were obtained by Cox proportional hazards models. Cause-specific mortality was modeled using competing risk models. Survival time was calculated from index diagnosis date to death date or the end of follow-up, whichever occurred first. Premature mortality models were restricted to those patients who were less than 75 years old at their diabetes diagnosis date, and survival over the follow-up period was modeled up to age 75, after which individuals were censored. Analyses were adjusted for sex, age at diabetes diagnosis grouping (18–34, 35–64, 65–79, 80+ years old), immigrant status (immigrant, refugee, long-term resident), area income quintile, and ADG score. The proportionality assumption was assessed by the log-negative-log of the Kaplan-Meier estimates of the survival function versus the log of time (log-log plot) and Schoenfeld residuals. Visual inspection of the Kaplan-Meier survival graph showed the proportional hazards assumption was not violated as the curves remain parallel (see online supplemental figure S2). The competing risk analysis of each cause of death was conducted by Fine-Gray subdistribution hazards model while taking all other causes of death as competing risks. Due to data availability, cause-specific mortality was followed up to 31 December 2017.
All analyses were performed using SAS V.9.4 (SAS Institute) in a UNIX environment. P value ≤0.05 was considered significant.
Results
The cohort was composed of those diagnosed with diabetes from 1 January 1994 to 31 December 2017 who were followed for a maximum of 25 years to assess for mortality outcomes, as indicated in the ORG-D and RPDB from 1 January 1994 to 31 December 2018. Those aged <18 and ≥120 years (n=24 128), non-Ontario residents (n=2911), and those whose death date occurred before the diabetes diagnosis date (n=41) were excluded, leaving a final cohort of 1 741 098 individuals. The cohort flow chart can be found in online supplemental figure S1.
Cohort characteristics
Table 1 displays the baseline characteristics of the cohort by sex at time of diabetes diagnosis. On average, females were 1 year older and more likely to be an immigrant compared with males. A greater proportion of the cohort resided in the lowest area income quintiles compared with the highest for both sexes. Overall morbidity was high in the cohort and more pronounced among females who had a median ADG score of 7 (IQR 5–10) compared with 6 (IQR 4–9) among males. Within the cohort, 197 596 (24.0%) females and 236 023 (25.7%) males died during the study follow-up period. There were 61 383 (7.5%) premature deaths in females and 103 148 (11.2%) premature deaths in males.
Baseline characteristics of all adults diagnosed with diabetes at diagnosis date, by sex, Ontario, Canada, 1994–2017
Presence of chronic conditions
Figure 1 displays the distribution of chronic conditions at time of diabetes diagnosis by sex. Conditions ranged from being prevalent in only 0.6% of the cohort (Crohn’s colitis disease) to 50.1% (hypertension). The top three most prevalent conditions across both sexes were hypertension, osteoarthritis and other arthritis, and mood disorder. The presence of multimorbidity was highly prevalent, with a median of three (IQR 2–5) conditions among females and three (IQR 2–4) conditions among males. In the overall cohort, 41.3% of individuals had four or more conditions at the time of diabetes diagnosis.
Presence of chronic conditions at time of diabetes diagnosis by sex, 1994–2017. COPD, chronic obstructive pulmonary disease.
Associations of sociodemographics on all-cause and premature mortality
Table 2 presents the Cox proportional hazards associated with immigrant status, area-level income and other covariates on all-cause mortality and premature mortality, respectively. After full adjustment, immigrants with a diabetes diagnosis experienced a 54% (HR 0.46, 95% CI 0.46 to 0.47) reduced hazard of all-cause mortality and 60% (HR 0.40, 95% CI 0.40 to 0.41) reduced hazard of premature mortality, compared with long-term residents with diabetes. In the fully adjusted models, individuals with diabetes who lived in the lowest income neighborhoods had a 26% (HR 1.26, 95% CI 1.25 to 1.27) increased hazard of all-cause mortality and 44% (HR 1.44, 95% CI 1.42 to 1.46) increased hazard of premature mortality, compared with individuals with diabetes living in the highest income neighborhoods. Differential risk was also observed by sex where, after full adjustment, males with diabetes had an increased hazard of all-cause mortality by 1.37-fold (95% CI 1.35 to 1.37), and 1.61-fold increased hazard for premature mortality (95% CI 1.59 to 1.61), compared with females with diabetes.
Unadjusted and adjusted HRs for all-cause and premature mortality among adults with a diabetes diagnosis, 1994–2017*
Associations of sociodemographics on cause-specific mortality
The results of the adjusted subdistribution competing risk analyses for cause-specific mortality are presented in table 3. The adjusted competing risk models reiterated a trend toward decreased subdistribution HRs for cause-specific mortality in immigrants with diabetes, relative to long-term residents with diabetes. As well, similar findings with those observed for all-cause mortality were found for the association of area-level income and cause-specific mortality; however, the income gradient was less pronounced for cancer mortality.
Adjusted subdistribution HRs (SDHR) for cause-specific mortality among adults with a diabetes diagnosis, 1994–2017*
Healthcare expenditures
The total accumulated healthcare costs in the last 2 years of life for those who died between 2004 and 2017 are displayed in figure 2A. Of the 323 694 individuals included in the analysis, total costs were $29.83 billion overall. The middle usage group (top 6%–50%) accounted for 60% of the total accumulated costs ($17.90 billion), while expenditures for the top 5% usage group accounted for 20% of the total costs ($6.05 billion). The average per-person spending in the last 2 years of life is depicted in figure 2B. Average per-person spending during this time period ranged from $36 356 in the bottom 50% usage group to $373 570 in the top 5% usage group.
Healthcare costs in the last 2 years of life for those who died with diabetes from 2004 to 2007. A) Total health care costs; B) Average per person health care costs
Discussion
This population-based cohort included all persons living with diabetes in Ontario, Canada between 1994 and 2017 and all deaths that occurred in this population up until 2018. In our study, individuals with diabetes had a large multimorbidity burden and high per-person healthcare costs in the last 2 years of life, underscoring the impact of diabetes on the healthcare system. Our study found reduced mortality risk among immigrants with diabetes, compared with long-term residents with diabetes, which is consistent with the immigrant health advantage pertaining to mortality observed in previous studies.16 21
We found socioeconomic variations in mortality risk among persons living with diabetes, in that those who lived in the lowest income neighborhoods had a higher risk of all-cause, premature, and cause-specific mortality, compared with those who lived in the highest income neighborhoods. The results are consistent with socioeconomic gradients in diabetes-related mortality that were reported in the USA in a study linking population survey data with vital statistics.33 Other studies that examined the effect of SES on cardiovascular disease mortality have shown increasing cardiovascular mortality with increased material deprivation in people with type 2 diabetes.34 35 The observed mortality variations suggest a need to address inequality gaps in diabetes care and management for persons with diabetes living in the lowest income areas. In contrast, although there is an increased risk of cancer mortality in type 2 diabetes,7 36 we observed a narrow income gradient for cancer mortality, which suggests gains in cancer prevention and treatment across SES groups in persons with diabetes.
A considerable strength of our study is that we identified all incident diabetes cases in Ontario from 1994 to 2017 using a validated population-based administrative data registry.22 The analysis was strengthened by the linkage of 24 years of vital statistics death data with administrative health data, several disease registries, census data on SES and a population registry of immigrants. With these data linkages, we improve on existing studies that rely on death certificate data by accurately capturing multimorbidity and information on vulnerable population groups (ie, immigrants and low area-level SES). The study also has some limitations. Our study relied on neighborhood-level income information based on postal code information at the time of death which is not a proxy for individual-level income37; however, it is a useful socioeconomic indicator that is informative for program and policy decision-making. As well, not all immigrants were identified in this analysis because the IRCC linkage starts in 1985 and does not include immigrants who initially landed in a province outside of Ontario. In addition, we were unable to differentiate between type 1 and type 2 diabetes in the administrative data; however, only a small proportion of total cases are expected to represent type 1 diabetes. Due to data limitations, our analysis did not adjust for obesity and hypertension status in our sample, which are important comorbidities related to diabetes. However, our analysis does adjust for comorbidity burden in the 2 years prior to diabetes diagnosis, as measured by ADG scores. Finally, our study did not examine explanatory factors, such as diabetes duration or the presence of diabetes complications, that could elucidate why differences in mortality exist among the diabetes population. For example, diabetes duration is associated with long-term complications, such as cardiovascular disease,38 and complications of diabetes could be indicative of access to treatment and services.39 This line of inquiry was beyond the scope of the current study but is a direction for future research.
The study illustrated an income gradient in all-cause, premature, and cause-specific mortality among people with diabetes. The socioeconomic variations in mortality risk among persons with diabetes suggest a need to better understand the factors that drive these inequalities to inform improvements in diabetes care and management.
Data availability statement
Data may be obtained from a third party and are not publicly available. The data set from this study is held securely in coded form at ICES. While legal data sharing agreements between ICES and data providers (eg, healthcare organizations and government) prohibit ICES from making the data set publicly available, access may be granted to those who meet prespecified criteria for confidential access, available at www.ices.on.ca/DAS (email: das@ices.on.ca). The full data set creation plan and underlying analytic code are available from the authors upon request, understanding that the computer programs may rely upon coding templates or macros that are unique to ICES and are therefore either inaccessible or may require modification.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and was approved by the Research Ethics Boards at the University of Toronto and Sunnybrook Health Sciences Centre (reference number: 32405). ICES is a prescribed entity under Ontario’s Personal Health Information Protection Act (PHIPA). Section 45 of PHIPA authorizes ICES to collect personal health information, without consent, for the purpose of analysis or compiling statistical information with respect to the management, evaluation or monitoring of the allocation of resources to or planning for all or part of the health system. The use of the data in this project is authorized under Section 45 and approved by ICES Privacy and Legal Office.
Acknowledgments
This study was supported by ICES, which is funded by an annual grant from the Ontario Ministry of Health (MOH) and the Ministry of Long-Term Care (MLTC). Parts of this material are based on data and information compiled and provided by Immigration, Refugees and Citizenship Canada, Cancer Care Ontario, the Canadian Institute for Health Information, and Ontario Registrar General (ORG) information on deaths, of which the original source is ServiceOntario. We thank IQVIA Solutions Canada Inc. for use of their Drug Information File.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Contributors Conceptualization: LCR. Methodology: LCR. Formal analysis: LZ. Writing—original draft preparation: KK. Writing—review and editing: LCR, EN and LZ. Project administration: KK. Funding acquisition: LCR. All authors have read and agreed to the published version of the manuscript. LCR is the guarantor of this work and as such, had full access to the data and takes responsibility for the work and conduct of the study.
Funding This study was funded by the Canadian Institutes for Health Research Foundation Grant (FRN72051628) and ICES. LCR is supported by a Canada Research Chair in Population Health Analytics.
Disclaimer The analyses, conclusions, opinions, and statements expressed herein are solely those of the authors and do not reflect those of the funding or data sources; no endorsement is intended or should be inferred.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.