Prevalence and patterns of chronic disease multimorbidity and associated determinants in Canada - HPCDP: Volume 35-6, August 2015

Prevalence and patterns of chronic disease multimorbidity and associated determinants in Canada

K. C. Roberts, MSc; D. P. Rao, MSc; T. L. Bennett, MSc; L. Loukine, MSc; G. C. Jayaraman, PhD

https://doi.org/10.24095/hpcdp.35.6.01

This article has been peer reviewed.

Author references: Surveillance and Epidemiology Division, Centre for Chronic Disease Prevention, Health Promotion and Chronic Disease Prevention Branch, Public Health Agency of Canada, Ottawa, Ontario, Canada

Correspondence: Deepa P. Rao, 785 Carling Avenue, Surveillance and Epidemiology Division, Centre for Chronic Disease Prevention, Public Health Agency of Canada, Ottawa, ON K1A 0K9; Tel: 613-946-5397; Email: deepa.rao@phac-aspc.gc.ca

Introduction

About one in three Canadians live with at least one major chronic disease,^{Footnote 1} and this proportion is expected to rise as the population ages and the risk factors for these diseases continue to climb.^{Footnote 2-4} As the leading cause of death worldwide, chronic diseases were responsible for 38 million (68%) deaths in 2012. Moreover, the World Health Organization projects that the total annual number of deaths due to chronic disease will increase to 52 million by 2030.^{Footnote 5} In Canada, heart disease, stroke, cancer, chronic respiratory diseases and diabetes accounted for 62.7% of all deaths in 2011.^{Footnote 6} Having two or more chronic diseases concurrently, referred to as multimorbidity, adds another layer of complexity to their prevention and management, in part because people with multiple diseases are at a greater risk of adverse health outcomes, more frequent hospitalizations and greater health care needs.^{Footnote 7}

Multimorbidity is increasingly recognized as a key public health issue. Despite that, its epidemiology is not well understood. Estimates of the prevalence of multimorbidity vary from 13.1%^{Footnote 8} to 90%^{Footnote 9}, largely due to differences in case definitions, data sources and study populations.^{Footnote 10} The most common way to measure multimorbidity is to count whether a person has two or more of a predefined list of chronic conditions.^{Footnote 16} However, the choice of conditions can influence prevalence estimates. For example, Agborsangaya et al.^{Footnote 11} found that 36% of Albertans had two or more chronic conditions, but their definition included obesity, which itself had a prevalence of 28.1%. Limiting a study sample to certain sub-populations, such as institutionalized populations or older populations, would also affect estimates.^{Footnote 9,Footnote 12-16}.

The main objective of this study was to examine the prevalence of multimorbidity in the adult Canadian population by selected sociodemographic variables and risk behaviours. Our secondary objectives were to examine the prevalence of disease dyads and triads.

Methods

Data source

This study is based on an analysis of the Canadian population aged 20 and older using data derived from the 2011/12 Canadian Community Health Survey (CCHS). The CCHS is a cross-sectional population health survey that includes Canadians aged 12 years and older but not those living in nursing homes or long-term care facilities or on reserves, full-time members of the Canadian Armed Forces, or civilian residents of military bases. The survey was designed to derive estimates at the national and provincial levels and for 110 health regions in Canada. Details about survey methodology are described elsewhere.^{Footnote 17} A multi-stage sample allocation strategy ensures relatively equal weighting to the health regions and the provinces.

Study variables

To identify the prevalence of specific chronic diseases, CCHS respondents were asked about those of their long-term conditions that were "expected to last or have already lasted 6 months or more and that have been diagnosed by a health professional." Each respondent was asked about their demographic and socioeconomic characteristics including sex, age, household characteristics (i.e. number of people in the household, highest level of education, household income) and Aboriginal identity. For our analysis, we assigned a relative measure of socioeconomic status to each respondent based on their reported total household income and household characteristics, according to Statistics Canada's household income adequacy quintiles. We used these quintiles to examine differences between the lowest income level and the highest. The Survey captures behaviours such as smoking (daily or occasional vs. not at all), dietary habits (5 or more servings of fruits and vegetables per day^{Footnote 18} vs. less than 5), leisure-time physical activity (active or moderately active vs. inactive) and obesity (body mass index ≥ 30) based on self-reported information, as well as the risk conditions of high blood pressure and stress (not at all, not very or a bit stressful vs. quite a bit or extremely stressful).^{Footnote 18} These behavioural risk factors for chronic disease were identified through a consultative process between the Public Health Agency of Canada and relevant stakeholders.^{Footnote 19}

Multimorbidity case definition

We defined multimorbidity as having two or more (2+) and three or more chronic diseases (3+) from a list of nine. This definition was established by an expert working group convened by the Public Health Agency of Canada in 2012.^{Footnote 20} The diseases were selected based on their chronic duration, high population prevalence in Canada, significant societal and/or economic impact, and amenability to primary prevention.^{Footnote 20} Based on these criteria, and data availability in the CCHS, the nine diseases we included in this study were arthritis, mental disorder (mood disorder and/or anxiety), asthma, diabetes mellitus, heart disease, chronic obstructive pulmonary disease (COPD, only available for ages 35 years and older), cancer, stroke, and Alzheimer's disease and related dementias. We analysed data using the 2+ and 3+ chronic disease case definitions.

Statistical analyses

The prevalence of multimorbidity for the general population, and stratified by sex, were estimated. Common dyads and triads of diseases were identified within each of the four age groups (20-34, 35-49, 50-64, 65+ years).^{Footnote 21} Data are reported for the top five combinations that accounted for at least 5% of Canadians with multimorbidity within the respective pairing group (dyad or triad) and age group.

Univariate logistic models were used to examine the relationship between multimorbidity and various sociodemographic, health and behavioural risk factors, including current age, sex, household income, household education level, Aboriginal status, area of residence (urban vs. rural), physical activity, fruit and vegetable consumption, smoking status, stress level, high blood pressure and obesity. Odds ratios were used to measure associations and the Wald ℵ² test for statistical analyses. A multivariate model using backward stepwise logistic regression analysis identified factors significantly (p <.05) associated with multimorbidity (defined using the more conservative 3+ conditions vs. 2, 1 or 0 conditions). Multivariate analyses were stratified by age to account for effect modification. Chi-square analyses were used to test for trend of physical activity across age groups.

CCHS data were weighted to be representative of the Canadian population, and 95% confidence intervals (CIs) of prevalence estimates and odds ratios (ORs) were calculated using a bootstrap resampling technique (500 replications) for all analyses to account for the complexity of survey design. Statistical analyses were performed using SAS procedure statements (SAS EG, version 5.1, SAS Institute Inc., Cary, North Carolina, US).

Results

Prevalence and distribution of multimorbidity by demographics and risk factors

In 2011/12, 12.9% (95% CI: 12.6-13.2; data not shown) of Canadians aged 20 years and over (n = 105 416) reported having two or more (2+) chronic diseases and 3.9% (95% CI: 3.7-4.1; data not shown) reported having three or more (3+) chronic diseases. Whatever the case definition used (2+ or 3+ diseases), the prevalence of multimorbidity showed similar trends across sociodemographic characteristics and risk factors (Table 1). Of those reporting 3+ diseases, the prevalence of multimorbidity was significantly higher (p <.05) in women (4.5% vs. 3.3% for men; OR = 1.3, 95% CI: 1.2-1.4), older Canadians (11.3% of the 65+ year age group vs. 0.4% of the 20-34 year age group; OR = 35.0, 95% CI: 26.2-46.8), those living in the lowest income quintile (7.4% vs. 1.5% of the highest income quintile; OR = 3.7, 95% CI: 3.1-4.5), those living in households where the highest education attained was less than high school (11.6% vs. 2.9% of households with post-secondary school education; OR = 1.8, 95% CI: 1.6-2.1), those self-identifying as Aboriginal (6.6% vs. 4.0% of non-Aboriginal populations; OR = 2.7, 95% CI: 2.2-3.4) and those born in Canada (4.2% vs. 3.4% of immigrants living in Canada for ≥ 5 years; OR = 0.7, 95% CI: 0.6-0.8).

Mood disorder and/or anxiety accounted for 51.6% (95% CI: 49.2-53.9; data available on request) of reported multimorbidity in younger Canadians (aged 20-34 years). In those aged 35 years and older, the combinations of diseases begin to show more heterogeneity, with the prevalence of particular multimorbidity combinations decreasing with age. Mood disorder and/or anxiety explained less of the overall multimorbidity burden (25.3%, 95% CI: 24.5-26.1; data available on request) within the older age group. Not surprisingly, the more common or prevalent an individual disease is within the population (see Table 2), the more likely that it contributed to rates of multimorbidity. As a consequence, diseases such as diabetes or COPD contributed more to the multimorbidity burden in the older age groups. It is worth noting that the top five multimorbidity combinations follow the same general trends across sexes, except that the fifth most common disease combination was diabetes and heart disease in men and arthritis and COPD in women (data not shown).

Determinants of multimorbidity (3+ chronic diseases)

After adjusting for all variables in the model, we found multimorbidity to be significantly associated with sex, age, household income, Aboriginal status, physical activity, smoking status, stress level and obesity. Stratification by age suggests that physical inactivity progressively, and significantly as of age 50 years and over, contributes to increases in the prevalence of multimorbidity (Table 4; p <.001).

In the younger age group (20-34 years), low income quintile was the only demographic risk factor significantly associated with multimorbidity (lowest quintile adjusted OR = 4.5, 95% CI: 1.3-15.7). Health conditions such as obesity (adjusted OR = 2.7, 95% CI: 1.4-5.2) and high stress (adjusted OR = 2.9, 95% CI: 1.6-5.2) were also significantly associated with multimorbidity.

In the 35- to 49-year age group, low income was once again a significant contributor to multimorbidity (lowest quintile adjusted OR = 7.5, 95% CI: 4.0-13.7) as were obesity (adjusted OR = 2.5, 95% CI: 1.7-3.8); high stress (adjusted OR = 1.7, 95% CI: 1.2-2.4); and smoking (adjusted OR = 2.6, 95% CI: 1.8-3.8).

Among 50- to 64-year-olds, women (adjusted OR = 1.3, 95% CI: 1.1-1.6), low income (lowest quintile adjusted OR = 5.3, 95% CI: 4.4-7.9), and Aboriginal status (adjusted OR = 2.0, 95% CI: 1.5-2.8) were significantly associated with multimorbidity (Table 4). Being inactive emerged as a significant risk factor (adjusted OR = 1.4, 95% CI: 1.1-1.7). Smoking (adjusted OR = 1.5, 95% CI: 1.2-1.8), high stress (adjusted OR = 2.1, 95% CI: 1.7-2.5), and obesity (adjusted OR = 2.24, 95% CI: 1.8-2.7) remained significantly associated with multimorbidity.

In those aged 65 years and over, low income (lowest quintile adjusted OR = 2.5, 95% CI: 1.8-3.5) and Aboriginal status (adjusted OR = 2.4, 95% CI: 1.7-3.3) were significantly associated with multimorbidity as were risk factors such as inactivity (adjusted OR = 1.7, 95% CI: 1.5-2.0), high stress (adjusted OR = 2.1, 95% CI: 1.7-2.5), and obesity (adjusted OR = 1.9, 95% CI: 1.6-2.2)

Discussion

The analyses we present in this report align with identified calls for improvements in measuring multimorbidity to respond to program and policy needs.^{Footnote 22} Unlike comorbidity, which has a primary condition, multimorbidity is the concurrent existence of multiple conditions without an index condition. As mentioned previously, published prevalence estimates of multimorbidity vary widely,^{Footnote 10} likely due to differences in how multimorbidity is defined and the methodology used, including population characteristics, sample size and recruitment methods.^{Footnote 10} Using data representative of the general adult population in Canada, our analysis suggests that, in 2011/12, the overall prevalence of multimorbidity was 12.9% (95% CI: 12.6-13.2) in those who reported having 2+ chronic diseases and 3.9% (95% CI: 3.7-4.1) in those who reported having 3+ chronic diseases.

The strong association between age and multimorbidity is well recognized,^{Footnote 12-14} but some key findings from our study merit attention. First, although multimorbidity was higher in the older population (65+ years), 2.2% of Canadians aged 20 to 65 years reported 3 or more coexisting chronic conditions, and 8.8% reported 2 or more coexisting chronic conditions, in keeping with previous Canadian studies.^{Footnote 11}

Second, the association of risk factors and multimorbidity increased with age. However, demographic characteristics such as low income were also significantly associated with multimorbidity irrespective of age. In the overall population, social deprivation resulted in a 3.7 odds of multimorbidity, but when examined across age groups, the odds of multimorbidity were notably higher in middle age, with odds of 7.5 for those aged 35 to 49 years and 5.9 for those aged 50 to 64 years. This differentiation of multimorbidity prevalence in middle age based on social deprivation supports the findings of Barnett et al.^{Footnote 23} who noted that the onset of multimorbidity occurred 10 to 15 years earlier in people living in the lowest socioeconomic conditions. Groups living in lower socioeconomic conditions may therefore have greater clinical management needs earlier in their lives and may live with multimorbidity for longer. Whether the progressive increase in prevalence of multimorbidity with age is due to a concentration of common risk factors, such as high stress or smoking, or due to ecological reasons, such as unemployment and gender inequality, is beyond the scope of our analyses. However, it warrants further investigation, particularly as unemployment, income inequality, physical activity level and stress are preventable risk factors that benefit from multisectoral action.

Third, we noticed patterns in the prevalence of chronic diseases. Arthritis, mood disorder and/or anxiety and asthma, which were the most common chronic conditions overall, were prominent in disease groupings, and were each more significantly prevalent in women. Anxiety and/or mood disorders accounted for over half of reported multimorbidity in younger Canadians (20-34 years). In addition, younger women had higher rates of mental disorders than their male counterparts.

Finally, our finding that arthritis and mental disorder are the most common pairing of multimorbidity supports the findings from a 2012 Canadian study that identified arthritis as a top condition in multimorbidity dyads.^{Footnote 24} Our analysis into common disease dyads and triads underscores the need to further investigate clusters of diseases to inform programs focused on integrated and holistic care.^{Footnote 25}

Strengths and limitations

Although the prevalence of multimorbidity and associated determinants we found are generalizable to the non-institutionalized Canadian population, our study has its limitations. Our results are likely to be subject to recall bias since the data are based on self-reports of clinical diagnosis. The results may also be explained, in part, by the operational case definition used for multimorbidity, namely, the number and type of chronic diseases considered. Fortin et al.^{Footnote 10} suggested that the more diseases considered, the higher the prevalence estimates. Rapoport et al.^{Footnote 26} looked at 22 diseases and reported a higher prevalence of multimorbidity in Canada than did Cazale et al.,^{Footnote 27} who considered only 7 diseases. However, Fortin et al.^{Footnote 10} also noted a threshold effect, with prevalence not further affected when 15 or more diseases were considered. We considered 9 chronic diseases, based on a set of public health criteria and limited to those captured via the CCHS. Finally, given the exploratory nature of our analyses and the multiple comparisons, a type 1 error is likely.

Despite these limitations, our results contributed to understanding the epidemiology of multimorbidity to inform prevention efforts, reduce disease burden and align health care services with holistic patient needs. Traditional, single disease-centric approaches may be increasingly inappropriate given that multimorbidity is becoming the norm for people living with chronic conditions. Rather, an integrated approach that addresses multiple diseases and upstream factors, such as socioeconomic determinants and behavioural risk factors, could have a broad-reaching, systematic effect on a range of outcomes including quality of life, health care costs and mortality.^{Footnote 16,Footnote 28} Further studies to understand the societal and economic impact of multimorbidity are warranted, although any findings in this area will certainly underline the need for prevention, particularly in tackling the common risk factors and determinants of common chronic diseases.

References