Background
Multimorbidity, the concurrent occurrence of two or more chronic conditions [
1], is increasingly common, probably due to aging populations, lowered threshold of diagnosis, inclusion of traditional risk factors such as obesity into its definition, longevity achieved through advances in medical care or possibly a true increase in the prevalence of some chronic diseases [
2].
As in other industrialized countries, Canadian healthcare delivery is typically guided by clinical practice guidelines that are oriented towards single-diseases [
3]. This poses a challenge for primary care professionals who try to implement evidence from these guidelines in caring for patients with multimorbidity. Individuals with multimorbidity are therefore at increased risk of receiving less than best practice care [
4,
5], more frequent and longer hospitalizations, higher health care costs and increased use of polypharmacy with the potential for adverse drug effects [
6]. The challenges have prompted calls for patient care guidelines and health programs that are multiple disease-centered [
2,
7,
8].
Furthermore, from a public health perspective, surveillance systems for chronic diseases tend to focus on single conditions. In Canada, for example, the National Diabetes Surveillance System (NDSS) was developed to track diabetes incidence, prevalence and mortality in all provinces and territories [
9]. In the province of Alberta, this system has been embellished to report more extensively on a variety of comorbidities in people with diabetes [
10], but it remains focused on a single, albeit common, condition in the population. The Public Health Agency of Canada has recently expanded the model from the NDSS to provide surveillance data on other conditions under the umbrella of the Canadian Chronic Disease Surveillance System (CCDSS), but this approach still retains the single disease focused model, with some attention to relevant comorbidities. Given that several common chronic conditions may cluster as multimorbidity in the general population, it would seem appropriate to take a multimorbidity approach to population health surveillance. Moreover, given a common set of shared risk factors (e.g., smoking, obesity, physical activity), multimorbidity surveillance may be more appropriate to evaluate the efficiency of more general or broader public health interventions.
Estimates of the prevalence of multimorbidity vary from 17% to over 90% [
1,
11,
12]. The wide variation is due to dissimilar study populations or data sources, usually entailing differences in demographic characteristics and disease types or classification [
11‐
13]. Most studies have been limited to patients in the primary care setting [
1,
14‐
17], having a specific index disease [
18‐
21] or to just the elderly [
13,
22‐
25]. Few studies have evaluated the prevalence of multimorbidity across age groups of the general population, including younger adults [
11,
26].
In a recent study on the prevalence of multimorbidity in a population-based cohort in South Australia [
11], the authors concluded that multimorbidity is not just a condition of the elderly. However, the definition of multimorbidity in their study was based on participants having two or more of a limited number of chronic conditions; asthma, cardiovascular disease, chronic obstructive pulmonary disease, diabetes, a mental health condition, arthritis and osteoporosis. Validity of the prevalence estimates potentially increases when a broader list of common chronic conditions is included in the study.
In Canada, for example, a study by Fortin and colleagues [
26] observed an overall prevalence of 11.6% in the general population and 32.5% in practice-based population, using data obtained from adults (25+ years) in the province of Quebec. This study highlighted the higher prevalence of multimorbidity across different age groups in the primary care sample compared to the general population. However, Fortin and colleagues did not elaborate the particular clusters of chronic conditions that comprise the patterns of multimorbidity. Indeed, there are currently no published Canadian data on the specific patterns of multimorbidity combinations in the general population. Therefore, the aim of this study was to estimate the prevalence and patterns of multimorbidity in different adult age groups, as well as determine the association of multimorbidity with socio-demographic factors.
Methods
The study is based on data from the Health Quality Council of Alberta (HQCA) 2010 Patient Experience Survey [
27]. The survey evaluated a sample of adult Albertans, representative of the general adult population, on their experience and satisfaction with the quality of health services they receive. The survey instrument, a telephone-based questionnaire, was administered by Random-Digit Dialing (RDD) approach to ensure that households in each of the five health zones had an equal chance to be contacted. Data were collected from 5010 adult (≥ 18 years of age) respondents during the fall of 2010. Sampling weights were derived to account for the stratified sampling approach, with a provincially representative sample of the population across the five health zones of Alberta.
To determine the occurrence of chronic conditions, respondents were asked the question "Do you have any of the following chronic conditions or diseases?"; diabetes, chronic obstructive pulmonary disorder, asthma, hypertension, high cholesterol, sleep apnea, congestive heart failure, obesity, depression or anxiety, chronic pain, arthritis, heart disease, stroke (or related conditions) and cancer. Apart from these 14 chronic conditions, respondents were also asked to report any other chronic conditions not listed. Based on the most frequent responses to the open-ended "other conditions" query, two additional chronic conditions were identified: gastro-intestinal tract (GIT disease) and kidney diseases. This study, therefore, considered a total of 16 chronic conditions.
Disease status was based on self-reports, and multimorbidity was defined as the presence of two or more chronic conditions [
1,
15]. Demographic data were also based on self-report, and included respondents' age, sex, educational level, annual household income level, number of adults (≥ 18 years of age) and number of children (≤ 16 years of age) living in the same household.
We undertook primarily a descriptive statistics evaluation of multimorbidity in this analysis. The prevalence of multimorbidity was estimated in relation to age, sex, household income and educational level. Educational level was collapsed into three categories; high school (at most high school education), college (more than high school education, including completion of college education), and university (at least university degree). All prevalence measures were direct- standardized to the 2006 age and sex distribution of the Alberta population (2006 Canadian Census). We were also interested to study the most common clusters of chronic conditions. We therefore determined the most common pairs, triads, quartets and quintets of chronic conditions. These were assessed by clustering the disease types per individual, then reporting the most common (frequencies) within pairs, triads, etc. Data are reported only for combinations of chronic conditions that occurred six or more times in the sample [
11].
Univariate logistic regression models were then used to evaluate the association between socio-demographic variables and multimorbidity within age strata (i.e. 18-24, 25-44, 45-64, 65 + years). Variables that were statistically significant (p = 0.05) within at least one age group were entered into a multivariate regression analysis, to examine the factors independently associated with multimorbidity. Analyses were adjusted for survey sampling weights. Statistical analyses and data management were performed using STATA V11 package. The Health Research Ethics Board (HREB) at the University of Alberta approved the data collection protocols and instruments.
Discussion
This study, based on a selection of community dwelling Albertans, describes the epidemiology of multimorbidity. The overall prevalence of multimorbidity in the study population was 19% in the general adult population. Age, household income and family structure were the most important measured predictors of the multimorbidity status. Multimorbidity tended to be more common in females than in males, an observation made in previous studies [
23,
26].
Our estimates of the overall prevalence of multimorbidity is comparable to studies in Quebec, Canada [
26] and Australia [
12], and lower than reports from hospital-based practice [
12,
28]. Patients consulting at a hospital for a chronic condition are more likely to have another chronic condition. In the Canadian study, Fortin and colleagues [
26] compared the prevalence of multimorbidity in practice-based and general population samples in Quebec. They observed that the overall prevalence was significantly higher for the primary care sample (32.3%) than in the general population (11.6%), highlighting the importance of the study population characteristics in the interpretation of findings on the prevalence of multimorbidity. Their study, however, defined multimorbidity based on 7 chronic conditions in adults aged 25 years and over. The lower prevalence of multimorbidity for the general population observed in their study compared to the present study may be due to the limited number of chronic conditions included [
26].
Studies examining the prevalence of multimorbidity have largely been limited to the elderly [
13,
22,
25,
29,
30], indicating that multimorbidity is a condition of old age. We observed, however, that 70% of persons with multimorbidity were less than 65 years of age, consistent with previous observations that multimorbidity affects not just older people [
11]. Mercer and colleagues [
15] argued that future studies "must begin to investigate multimorbidity across a life-course". Our findings provide further evidence on the importance of multimorbidity in young adults.
There have recently been calls for a more holistic definition of the term, with the inclusion of not just chronic disease "labels" [
15], but also morbidities suggesting emotional and psychological distress. The present study included anxiety and depression as a morbidity. The inclusion of another important chronic condition, obesity, remains controversial and has been considered elsewhere as a risk factor of multimorbidity, rather than a disease on its own right [
11]. Nagel and colleagues [
31] in a prospective study noted that obesity rates increase with the number of chronic conditions. While the direction of the relationship between obesity and multimorbidity is yet to be ascertained, there is need for public health policy to emphasize the importance of a healthy weight in reducing the burden of multimorbidity.
Age, household income and family structure (Not living with children) were independently association with multimorbidity. Although there is ample evidence for the inverse association between increasing age and decreasing income with multimorbidity [
11,
16], the importance of family structure has received little attention in the past. Taylor and colleagues [
11] showed that independent of age, multimorbidity was more common among adults living alone or with partner, compared to those living with children. The reasons underlying these findings are not clearly understood. There is evidence that family support, also known as family-centered care [
32], may be vital in the management and control of chronic diseases [
33,
34]. The importance of family support, through chronic disease management, may be an important component in reducing the likelihood of developing other chronic conditions. However, this hypothesis remains to be tested.
A major strength of this study is the population representativeness of the study sample, that allows for generalization of the findings. Thus, findings represent prevalence estimates in the general adult population. Population-based prevalence estimates of multimorbidity are important for reporting about the health status of the population. Our study entails a modest number of chronic conditions, including the core chronic conditions recommended for inclusion in measures of multimorbidity [
35]. Also, important chronic conditions such as obesity, anxiety and depression were included in this study.
Our study also has some limitations. The cross-sectional nature of the data prevents the examination of the temporality of the associations between socio-demographic factors and multimorbidity. The study included a limited number of morbidities, which are based on self-reports. Self-reported chronic disease status is subject to self-declaration bias due to under-reporting of diagnosis or forgetfulness [
36,
37]. Surveyed patients with only one or none of the listed morbidities, who were counted as having no multimorbidity in this study, may have other unlisted chronic conditions. In the interpretation of these findings, it is therefore important to note that the reported prevalence of multimorbidity is only based on the set of chronic conditions in the HQCA survey. Moreover, some individuals who report having multimorbidity may essentially be reporting a single chronic condition and its symptom, e.g. arthritis and chronic pain. This, potentially, may lead to over-estimation of the true prevalence of multimorbidity. It is also possible that some important groups, such as immigrants (e.g. due to language barriers), were under-sampled. A further limitation of this study is the absence of an indicator of disease severity, as provided in the Kaplan Index, the Index of Coexisting Diseases, Charlson Index or the Cumulative Illness Rating Scale [
38]. Some studies have characterized conditions such as hypertension, high cholesterol and obesity as risk factors, rather than as chronic conditions [
11]. A further step may be to incorporate such differences in the analysis, while weighting conditions by severity.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
ACB conception and design, statistical analysis and interpretation of data, drafting manuscript, revision of manuscript. DL conception and design, interpretation of data, critical revision of manuscript. ML: data acquisition, survey instrument and design, critical revision of manuscript. TC data acquisition, survey instrument and design, critical revision of manuscript. JAJ conception and design, data acquisition and interpretation of data, critical revision of manuscript. All authors read and approved the final manuscript.