Since blood pressure (BP) control is less often achieved by individuals with diabetes, the authors sought to determine whether receipt of and adherence to health behavior advice for hypertension control differs between people with and without diabetes, using data from the 2009 Survey on Living With Chronic Diseases in Canada. Individuals with coexisting diabetes were more likely to report receiving advice to control/lose weight (81% vs 66%), be physically active (79% vs 68%), limit alcohol consumption (78% vs 55%), and modify diet (70% vs 61%) but not limit dietary salt (65% vs 64%) compared with individuals with hypertension alone (n=4.965). People with and without diabetes were equally likely to report following the advice they received, with receipt of advice positively associated with engagement in healthy behaviors. Since receipt of advice appears to influence behavior, health professionals should be encouraged to further promote BP self-management strategies.
Elevated blood pressure (BP) is an important risk factor for cardiovascular disease among individuals with diabetes, accounting for up to 44% of deaths and 41% of cardiovascular disease events in this group. Although BP control among people with coexisting diabetes has improved in recent years,[2-4] individuals with diabetes are still less likely to have their BP controlled (<130/80 mm Hg) relative to individuals without diabetes (<140/90 mm Hg).[5-7] In Canada, BP control is similar between the two groups when control is defined based on a <140/90 mm Hg threshold. Furthermore, although individuals with diabetes take more antihypertensive medications on average, they are more likely to have other risk factors for poor BP control (eg, unhealthy body weight and physical inactivity).
In order to manage BP, clinical practice guidelines recommend that individuals with hypertension: (1) reduce dietary salt to ≤1500 mg/d depending on age, (2) eat a healthy diet, (3) limit alcohol consumption, (4) participate in aerobic exercise, (5) attain or maintain a healthy body weight, and (6) use stress management strategies where needed.[8, 9] Many of these behaviors are also indicated for management of diabetes. People who are advised by their health professional to make these changes may be more likely to do so.[11, 12] Understanding the extent to which healthy behaviors are recommended by clinicians and followed by individuals with both diabetes and hypertension may identify areas for intervention to reduce previously observed disparities in BP control.
Using a large population-based survey of people with hypertension, we sought to determine whether (1) receipt of clinical advice for nonpharmacologic management strategies and (2) engagement in these strategies differ between individuals with and without diabetes. Furthermore, we sought to determine whether likelihood of following advice differed by diabetes status, as well as by sex, level of education, and time since diagnosis.
The 2009 Survey on Living With Chronic Disease in Canada (SLCDC) Hypertension Component is a cross-sectional survey that collected information related to the experiences of Canadians with hypertension. Details surrounding the survey objectives, questionnaire development, sampling frame, and associated methodology are reported in detail elsewhere.[13, 14] In brief, the questionnaire was developed by a panel of hypertension and survey development experts assembled by the Public Health Agency of Canada. Survey questions were identified from publicly available population surveys and peer-reviewed instruments and scales, with some survey questions modified to reflect national hypertension guidelines.
Adults 20 years and older, living in the 10 Canadian provinces, and who reported having been diagnosed with high BP (n=7862) were selected from the 2008 Canadian Community Health Survey (CCHS). Of these, 6142 individuals (representing approximately 5 million Canadians) agreed to participate, for a response rate of 78.2%. Excluded from the CCHS, and subsequently from the 2009 SLCDC, were full-time members of the Canadian Forces, persons living on Indian reserves or Crown lands, and residents of institutions or of certain remote regions. Together these exclusions are estimated to represent <2% of the population. Residents of the 3 northern territories were excluded from the SLCDC due to insufficient sample sizes, which would lead to an inability to properly weight findings to represent all residents of the territories.
Computer-assisted telephone interviews were conducted between February and April 2009. Data from the SLCDC were linked to data from the 2008 Canadian Community Health Survey (collected between January and December 2008). The latter provided information on household income, education, marital status, race/ethnicity, body mass index (BMI; based on self-reported height and weight), and diabetes status. Individuals who reported having previous gestational diabetes only (n=2) were classified as not having diabetes. Individuals who answered “do not know” (n=7) were excluded, leaving 1170 and 4965 individuals with and without diabetes, respectively.
Clinical Recommendations for Self-Management
Participants were asked about advice ever received (yes/no) from health professionals for BP control including: limiting salt intake, eating certain foods (such as fruits and vegetables, fish or lean meats, foods high in fiber, or foods low in fat), engaging in physical activity, maintaining or losing weight, cutting down on smoking, limiting alcohol consumption, reducing stress levels, and correct use of a home BP monitor.
A composite measure of receipt of clinical advice for self-management was derived by summing positive responses for advice received for salt restriction, dietary changes, weight control/weight loss, and physical activity, based on the results of factor analysis of categorical data (ie, latent trait analysis), which showed that these factors comprise a single domain (Supplemental Table S1) and that advice for medication use and self-monitoring of BP did not fit well within this domain. These variables had reasonable internal consistency (Cronbach's α=.73).
Self-Management of High BP
Participants were asked whether, as a result of being diagnosed with high BP, they ever: (1) limited their daily salt intake, (2) changed the types of food they eat (choosing more fruits and vegetables, fish or lean meats, foods high in fiber, or foods low in fat), (3) exercised or participated in physical activities, and (4) tried to control or lose weight. The latter was described among individuals who were overweight or obese (ie, BMI ≥25 kg/m2 based on self-reported height and weight). Those who reported smoking at any time since their hypertension diagnosis were asked whether they ever quit or cut down on smoking to help control their BP. Respondents who reported regularly drinking more alcohol than recommended (ie, >14 drinks of alcohol for men or >9 drinks of alcohol for women per week) since their diagnosis were asked whether they ever limited their alcohol consumption to help control their BP.
For each behavior, individuals who answered “yes” to ever engaging in the activity were asked whether they continued to maintain the change “all of the time,” “most of the time,” “some of the time,” or “none of the time.” Patients who indicated having never engaged or no longer engage in a respective behavior were asked the open-ended question: “What are the reasons that you are not [engaging in the behavior] to help control your BP?” A substantial number of respondents reported not engaging in behaviors for BP control because they were already doing so “for other reasons” and were categorized as a separate group.
In order to correspond to the composite measure of clinical advice, an overall healthy behavior score was derived by summing ordinal responses for salt restriction, dietary changes, weight control/weight loss, and physical activity (with each variable given a score of 0=“none of the time,” 1=“for other reasons,” 2=“some of the time,” 3=“most of the time,” and 4=“all of the time”). Latent trait analysis revealed agreement between these 4 behaviors (Supplemental Table S1; Cronbach's α=.64). The possible range for the derived score was 0 (does not engage in any of the 4 behaviors) to 16 (engages in all 4 behaviors all of the time).
Participants were asked about monitoring their own BP outside of a health professional's office or medical clinic. Specifically, respondents were asked whether (yes/no), how often (daily, weekly, monthly, 3 or 4 times a year, once a year, less than once a year, never), and where (home, pharmacy, workplace, gym/fitness facility, other) they measure their BP. Regular use of home BP monitoring was defined as measuring one's BP at home at least weekly.
Data were analyzed using SAS Enterprise Guide 4.1 (Cary, NC). Point estimates were weighted to reflect the Canadian household population. The 95% confidence intervals (CIs) were calculated using exact standard errors obtained using bootstrap re-sampling methods. Latent trait analysis (ie, factor analysis of binary or ordinal data) was used to develop the clinical advice score and the healthy behavior score (http://support.sas.com/kb/22/558.html).
Sociodemographic characteristics were compared between individuals with and without diabetes using chi-square (χ2) tests of association based on a weighted bootstrap procedure. The association between diabetes status and (1) receipt of advice for self-management behaviors and (2) engagement in self-management behaviors all/most of the time were evaluated using crude and multivariate prevalence rate ratios (RRs), estimated using log-binomial regression. Potential confounders (sex, age group, ethnicity, education, total household income, marital status, region, time since diagnosis) were identified a priori based on previous descriptive analyses of the 2009 SLCDC.[16-18] Covariates were retained if their exclusion caused a ≥5% change in the primary effect estimate observed from fully adjusted models. Number of consultations with health professionals in the previous year was not considered as a potential confounder as it would lie on the causal pathway between diabetes status and receipt of advice, since diagnosis of diabetes may result in more referral care and thus more opportunity to receive advice.
In order to determine whether receipt of clinical advice for BP control is associated with engagement in self-management behaviors to the same or different extent among people with and without diabetes, we modeled the relationship between the clinical advice score and the healthy behavior score in a linear regression model that included an interaction term between diabetes status and clinical advice.
Sample Characteristics and Comparison of Individuals With and Without Diabetes
One in five (19%) people with hypertension reported having coexisting diabetes. The average age of the sample was 62 years, with an even split of men and women (Table 1). Reflecting this population in Canada, the majority were white (87%), relatively affluent (≥54% household income $50,000), married (68%), urban-dwelling (86%), educated (52% with post-secondary education) adults. The majority (63%) had been diagnosed with hypertension >5 years before the survey.
Table 1. Characteristics of Canadian Adults Aged 20 Years and Older Diagnosed With Hypertension Overall and by Diabetes Status: 2009 Survey on Living With Chronic Disease in Canada (N=6135)
No Diabetes (n=4965)
Abbreviations: CI, confidence interval; F, could not be reported due to high sampling variability (coefficient of variation >33.3%); SD, standard deviation.
Estimate should be interpreted with caution due to high sampling variability (coefficient of variation 16.6% to 33.3%).
As shown in Table 1, compared with individuals without diabetes, those with diabetes were older and had lower household income. Individuals with diabetes were also more likely to have been diagnosed with hypertension ≥10 years ago (42% vs 57%, P=.0003). The majority of individuals with and without coexisting diabetes considered their general practitioner as the health professional most responsible for treating their high BP (88% vs 92%, P=.11), had at least one visit to a health professional in the previous year (97% vs 90%, P<.0001), and had their BP measured in the previous year (93% vs 84%, P<.0001).
Clinical Advice for Self-Management of Hypertension
Compared with individuals without diabetes, individuals with diabetes were more likely to report having received clinical advice to control or lose weight if overweight or obese, to engage in physical activity, to limit alcohol consumption if consuming more than recommended, and to make dietary changes, after controlling for sex, age, ethnicity, income, and time since diagnosis (Table 2). Individuals with diabetes were not more likely to have received advice on limiting salt intake or reducing smoking.
Table 2. Associations Between Diabetes Status and Clinical Advice for Management of BP Among Canadian Adults Aged 20 Years and Older With Hypertension: 2009 Survey on Living With Chronic Disease in Canada
Adjusted for age group, ethnicity, income, and time since diagnosis based on 5% change in any of the effect estimates, with sex forced into the models.
Controlling or losing weight
1.4 (1.3, 1.5)
1.3 (1.2, 1.4)
Among overweight or obese, n=4214
1.2 (1.2, 1.3)
1.2 (1.1, 1.3)
Participating in physical activity
1.2 (1.1, 1.2)
1.1 (1.0, 1.2)
Limiting alcohol consumption (of those who drank, n=603)
1.4 (1.1, 1.8)
1.1 (1.0, 1.2)
Eating certain foods
1.2 (1.1, 1.2)
1.1 (1.0, 1.2)
Reducing levels of stress
1.1 (1.0, 1.3)
1.1 (1.0, 1.2)
Correct use of home BP monitor
1.1 (0.9, 1.2)
1.1 (0.9, 1.3)
Limiting daily salt intake
1.0 (0.9, 1.1)
1.0 (0.9, 1.1)
Quitting or cutting down smoking (of those who smoked, n=1484)
1.0 (0.9, 1.1)
1.0 (0.9, 1.0)
Engagement in Healthy Behaviors for BP Control
After controlling for sex, age, education, income, and time since diagnosis, individuals with diabetes were more likely than those without diabetes to report limiting alcohol consumption all or most of the time (among those who consumed more alcohol than recommended since diagnosis), measuring their BP at home on a weekly basis, trying to control or lose weight all or most of the time (among those who were overweight or obese), and making changes to their diet all or most of the time. Proportions reporting limiting their salt intake, reducing smoking, or engaging in physical activity did not differ significantly (Table 3).
Table 3. Crude and Adjusteda Association Between Diabetes Status and Management of BP Among Canadians Adults Aged 20 Years and Older With Hypertension: 2009 Survey on Living With Chronic Disease in Canada
Adjusted for education, income, and time since diagnosis based on 5% change in any of the effect estimate, with sex and age group forced into the models.
Limits alcohol consumption all/most of the time (of those who drank, n=603)
1.7 (1.3, 2.2)
1.4 (1.1, 1.7)
Measures own BP at home at least weekly
1.2 (1.0, 1.5)
1.2 (1.0, 1.5)
Tries to control or lose weight all/most of the time (among those who were overweight or obese, n=4214)
1.1 (1.0, 1.3)
1.2 (1.0, 1.3)
Changes to types of food eaten all/most of the time
1.2 (1.1, 1.2)
1.1 (1.1, 1.2)
Limits salt consumption all/most of the time
1.1 (1.0, 1.2)
1.0 (1.0, 1.1)
Quit/reduced smoking all/most of the time (of those who smoked, n=1484)
0.9 (0.8, 1.1)
1.0 (0.9, 1.2)
Engages in physical activity all/most of the time
1.0 (0.9, 1.1)
1.0 (0.9, 1.2)
Relationship Between Receipt of Clinical Advice and Engagement in Self-Management Strategies in Individuals With and Without Diabetes
Average composite health behavior scores were marginally higher in individuals with diabetes than in those without diabetes (10.4±0.2 vs 9.6±0.1, P=.0002). Individuals who reported receiving more clinical advice for self-management (in terms of number of recommended health behaviors) reported higher levels of engagement in health behaviors for self-management (Figure). Specifically, the average healthy behavior score for an individual who did not receive any advice for health behaviors change was 6.0. This increased, on average, by 1.1 points (P<.0001) for each additional recommendation received after adjustment for potential confounding factors (Table 4). The relationship did not differ significantly by presence of diabetes (pinteraction=.7) nor by sex (pinteraction=.4). Highest level of education modified the relationship. In people with high school education or lower, healthy behavior scores increased by 1.4 points for each additional recommendation received compared with an increase of 1.1 in people with higher levels of education (pinteraction=.03). Time since diagnosis appeared to modify the relationship to a similar extent, but did not reach statistical significance.
Table 4. Linear Model for the Association Between Behavior Scores and Clinical Advice Scores Among Canadians Aged 20 Years and Older With Hypertension: 2009 Survey on Living With Chronic Disease in Canada
Diabetes × advice score
Male × advice score
≤High school education
≤High school education × advice score
Time since diagnosis ≤5 y
Time since diagnosis ≤5 y × advice score
Total household income
Advice on the correct use of a BP monitor was associated with regular home BP monitoring. In individuals without diabetes, those who reported receiving advice on correct use of a BP monitor were 2.6 times (95% CI, 1.7–4.2) more likely to report regular home BP monitoring, after controlling for covariates (data not shown). In the separate stratum of individuals with diabetes, a stronger 3.1-fold increase (95% CI, 1.3–7.6) was observed (pinteraction=.004).
Adoption of healthy behaviors is a cornerstone of both hypertension[8, 9] and diabetes management. The current study suggests that among people with hypertension, individuals with coexisting diabetes are slightly more likely to receive advice for healthy behaviors (eg, exercise, dietary change, and weight control), but as likely to receive advice for dietary salt reduction and smoking cessation. This is consistent with findings from the 1998 National Health Interview Survey (NHIS), which, while not considering advice for salt reduction or smoking cessation, showed that people with diabetes were more likely to receive hypertension-specific advice for dietary change (odds ratio, 1.9) and exercise (odd ratio, 2.0). The relationships observed in the current study were not as strong, likely because the odds ratio overestimates the prevalence ratio when the outcome studied is not rare. People with diabetes were more likely to report making changes to diet, engaging in physical activity, trying to control or lose weight, and regularly monitoring their BP at home, but not limiting their salt intake. Receipt of clinical advice was similarly positively associated with engaging in healthy behaviors in people with and without diabetes (ie, these groups were equally likely to follow the advice they received). This is in contrast to the 1998 NHIS, which found that people with diabetes were more likely to adhere to advice. The current findings do not support our hypothesis that disparities in receipt of clinical advice and adherence to healthy behaviors relate to previously observed disparities in BP control in those with and without diabetes and further study is required to explain observed disparities.
In the current study, people who received advice for multiple behaviors reported higher levels of engagement in lifestyle self-management. The World Health Organization suggests that “moderately intense” primary care interventions, which include targeted information and follow-up, are effective in promoting adoption of healthy behaviors in those at risk for chronic disease. In the current survey, the method used to deliver advice was not measured (ie, what types of information were provided and whether this included follow-up). It is possible that individuals who received more intensive counseling may have been more likely to act on the advice. At the individual level, lifestyle counseling in primary care has been shown to confer small benefits in achieved BP, with 3 of 6 randomized controlled trials showing small (<4 mm Hg), significant decreases in BP. Small reductions in BP can reduce the burden of cardiovascular diseases at the population level; for example, in persons with and without diabetes, small reductions in BP (of −6/−4.6 mm Hg and −3.7/−3.3 mm Hg, respectively) have been associated with 36% and 11% reductions in total stroke events. Furthermore, face-to-face lifestyle counselling is associated with faster achievement of BP control in diabetes. Morrison and colleagues showed that monthly counselling was associated with a 3.7-week median time to BP control <130/85 mm Hg compared with 5.6 months in patients who received counselling less than once every 6 months. Individuals who receive advice from health professionals to modify their behavior may be more confident and motivated to attempt change and sustain changes over time.[24, 25]
Although potentially effective, provision of advice for multiple health behavior change may present a challenge for individual physicians in primary care, as suggested by the current finding that only 45% to 80% of people with hypertension reported ever having received advice for the various behaviors and by like findings in the United States.[12, 26] Clinicians may not have sufficient expertise to address the complexities of multiple health behavior changes nor have sufficient time to do so during brief clinical encounters. Thus, multidisciplinary teams (comprised of physicians, nurses, pharmacists, nutritionists, physical activity specialists, and others) may aid in providing behavioral counseling to patients and have been shown to improve BP control.[29-32] Specifically, in a meta-analysis of 27 cluster and randomized control trials of diabetes care with BP as an endpoint, Tricco and colleagues showed that changes to team structure (such as multidisciplinary teams, shared cared, or expansion/revision of professional roles) was associated with an average 4.3-mm Hg improvement in systolic BP. Establishment of multidisciplinary teams in rural populations may present challenges because of limited access and resources. Potential solutions may include the development of mobile multidisciplinary outreach services, which have been shown to improve BP control in rural populations, or telehealth/internet-based strategies.[35, 36] Because health behaviors are also influenced by the environments in which people live, health professional–led interventions could be further supported by workplace, community, and national initiatives targeting the environmental influences on general health.[20, 37]
Study Strengths and Limitations
Some strengths and weaknesses of the current study are noted. This research is strengthened by its use of a large, population-based survey of people diagnosed with hypertension, a survey that provides comprehensive data on factors associated with management of hypertension. Furthermore, we produced a novel composite score of clinical advice and of engagement in health behaviors for BP control that could be used in future research. The survey had a good response rate, and participants did not differ from the source population in terms of body weight, physical activity levels, or daily smoking; participants were slightly less likely than the source population to report taking a medication for high BP (83% vs 89%). The study is limited by the use of cross-sectional data, which precluded examination of how behaviors might change over time or how advice may have an impact over time. Furthermore, recall bias may have occurred if people who engage in the recommended behaviors are more likely to remember the advice that they received. Secondly, we could not establish the temporal relationship between advice and behavior; it is possible that people who engaged in healthy behaviors subsequently initiated conversation with their health care provider about lifestyle change. Canada's Aboriginal on-reserve communities and the 3 territories were excluded from the sampling frame. As a result, the findings may not generalize to these populations given their higher rates of hypertension and diabetes,[38, 39] as well as poorer access to primary care. Finally, the use of self-reported data may have led to an overestimate of engagement in behaviors and an underreporting of advice received. If individuals who reported receiving advice were more likely to over-report engaging in the behavior because of social desirability, this would have biased the observed associations.
People with hypertension and diabetes were slightly more likely to receive advice from health care professionals on behaviors for hypertension management, with the exception of salt reduction and home BP monitoring, and were as likely as people without diabetes to follow such advice. Health professionals may encourage health behavior changes in people both with and without diabetes by providing advice and recommendations for these behaviors, with the aim of further improving BP control.
M.E.G. analyzed the data and wrote the manuscript. W.P., I.J., J.A.J., N.R.C. reviewed, edited, and contributed to the development of the manuscript.
The 2009 SLCDC was conducted by Statistics Canada and sponsored by the Public Health Agency of Canada.
Sources of Funding
M.E.G. was supported by a Canadian Institutes of Health Research (CIHR) doctoral award and is supported by a Heart and Stroke Doctoral Award. I.J. holds a CIHR Canada Research Chair in Physical Activity and Obesity. J.A.J. holds a Canada Research Chair in Diabetes Health Outcomes and a Senior Health Scholar award from Alberta Innovates–Health Solutions. N.R.C. holds the Heart and Stroke Foundation–CIHR Research Chair in Hypertension Prevention and Control.