Relationship between multimorbidity and physical activity: Secondary analysis from the Quebec health survey

Finding strategies to deal with chronic medical conditions is a priority for improving the health of the world population [1]. Studies [2–4] in various countries demonstrate that many people have multimorbidity, defined as the co-occurrence of two or more chronic diseases [5]. Persons with multimorbidity require more frequent appointments and hospitalizations [2, 6–9] and are at greater risk for drug interactions, acute deteriorations, disability, and mortality [10–14]. They also have higher psychological distress [15] and their quality of life is affected [16, 17]. Among the strategies for improving the health of people with chronic diseases, physical activity is extensively supported in the published literature [18–20]. Physical activity improves psychological well-being, and decreases stress, anxiety, and the feelings associated with depression [21, 22]. Physical activity also decreases pain during the treatment of painful conditions [23], favors resistance and vigor of people with chronic diseases [18], and decreases their risk of developing functional incapacities [19].

However, sixty percent of the world population and more than 50% of American adults do not attain the minimum of physical activity levels recommended by the American College of Sports Medicine and the Centers for Disease Control and Prevention, namely, 30 minutes of moderate physical activity at least five days a week [24]. Levels of physical activity decrease with age, and women are less physically active than men [2]. Physical activity levels are lower among populations with low income or low education [25, 26], and among people with functional incapacities or limited ability to be active [25, 27]. Studies also suggest a relationship between employment [28, 29], perceived health status or psychological distress, and physical activity levels [25, 30–32].

Considering the numerous preventive and curative effects of physical activity, health care professionals should get involved in promoting physical activity among their patients with chronic diseases and multimorbidity [18]. We already know that patients with multimorbidity are more likely to have limited activity levels [33], psychological distress [15], and perception of a lower health status [34]. Kaplan et al. examined predictors of frequent and infrequent self-reported physical activity lasting more than 15 min in community-dwelling people ≥ 65 yr of age. Using logistic regression analyses, they found that the absence of chronic conditions was associated with more frequent physical activity in late life [25]. The relationship between multimorbidity and physical activity levels remains however poorly documented among non geriatric people. A better understanding of this relationship would facilitate development of more appropriate counseling strategies among this clientele.

The purpose of the present exploratory study was to examine the relationship between a) multimorbidity and physical activity levels; and b) long-term limitations on activity, self-rated general health, psychological distress, and physical activity levels for both sexes in adults, after age, education, income, and employment factors have been controlled for. Our hypothesis, suggested by Kaplan's study, was that multimorbidity may be associated with a decreased physical activity level in adults.

To our knowledge, this is the first study to explore the relationship between multimorbidity and physical activity in this age group. We found that none of the cut-offs of multimorbidity (number of chronic conditions) that we used was associated with physical activity levels when long-term limitations on activity, psychological distress, perceived health status, age, sex, education, income, and employment were controlled for.

Our results differ from those of Kaplan's study [25]. Kaplan et al. used data for 12,611 community-dwelling people ≥ 65 yr of age from the 1996–1997 Canadian National Population Health Survey to examine predictors of frequent and infrequent self-reported physical activity lasting more than 15 min, including many independent and confounding variables (geographic location, psychological distress measured with the Generalized Distress Scale, age, gender, educational level, marital status, perceived social support, chronic medical conditions, physical limitations due to injury, functional limitations, smoking behavior and body mass index). Using logistic regression analyses, they found that the absence of chronic conditions was associated with more frequent physical activity in late life. Our two studies shared many similarities: 1) Population Health Surveys; 2) Larges samples; 3) Comparable lists of chronic conditions; 4) Comparable self-reported questionnaires for physical activity; 5) Logistic regression analysis; 6) Separated analysis for males and females. However, Kaplan et al documented a relationship between the absence of chronic conditions and an increased level of physical activity, a relationship that we could not ascertain in our sample. A few hypotheses may be raised to explain different results. First, the ages of our study populations were different. Kaplan et al. [25] recruited people ≥ 65 yr of age, whereas we deliberately excluded people > 69 yr so that we could evaluate a non-geriatric population. Relationship between multimorbidity and physical activity may be different in these two groups. Second, we don't know if the independent "chronic conditions" was considered continuous or categorical in Kaplan's study. This choice may have impact on results. Finally, they include different variables in their model but did not include self-rated general health that is significant in our models. Addition of this variable may have changed their results.

Neither study used a comorbidity or multimorbidity index that accounted for the severity of any single disease or the involvement of a system. One of our previous studies [15] showed that the count of chronic diseases was not associated with an outcome such as psychological distress, whereas chronic disease was associated with psychological distress when an index that takes severity into account was used.

Even if multimorbidity does not seem to be related to physical activity levels, it is related to other variables such as long-term limitations on activity [33], psychological distress [15], and perception of a lower health status [34] that can influence physical activity. Our findings for these variables are in accordance with those of previous studies that demonstrated a relationship between functional incapacities or long-term limitations on activity [27, 37–39], perceived health status, and physical activity level [30, 31]. We found a significant relationship between psychological distress and physical activity levels only among women. Other studies [25, 26, 32] demonstrated a relationship between psychological distress and physical activity, but did not distinguish between the sexes.

The following limitations of this study should be kept in mind. In our analyses we used the variables available from the QHS. First, despite the good quality of the data from the QHS 1998 survey (it had a large sample, four data collections to take seasonal variations into account, and a good response rate), it is possible that recall bias and social desirability may have influenced the respondents' answers. Second, our measure of multimorbidity was limited to a number of chronic diseases. All of these chronic diseases had equal weight in the analysis without any assessment of their severity. Third, although self-efficacy and physical activity levels are strongly related [40, 41], the original QHS survey did not measure the relationship between self-efficacy and physical activity prospectively, so we could not include this variable in our model. We were not able to include other variables such as cholesterol, blood glucose, medication and renal function as they were not available in the population survey. Also, other characteristics such as diet, body mass index and smoking may have been included. However, we wanted this exploratory study to be as focused as possible and decided not to include these as covariables in the analysis. Finally, since the data from this study are cross-sectional, we cannot assign any causality to the relationships that were identified. Despite these limitations, our study remains one of the first to explore the topic.

Multimorbidity was not associated with physical activity levels for either sex in adults when age, education, income, and employment factors were controlled for. Long-term limitations on activity and poor-to-average self-rated general health status seem to be related to lower physical activity levels for both sexes, whereas psychological distress was associated with lower physical activity levels only among women.