Abstract
Background: This study was designed to describe the distribution of risk factors for cardiovascular disease by socioeconomic status in adult men and women across Canada using the Canadian Heart Health Surveys Database.
Methods: The data were derived from provincial cross-sectional surveys done between 1986 and 1992. Data were obtained through a home interview and a clinic visit using a probability sample of 29 855 men and women aged 18-74 years of whom 23 129 (77%) agreed to participate. The following risk factors for cardiovascular disease were considered: elevated total plasma cholesterol (greater than 5.2 mmol/L), regular current cigarette smoking (one or more daily), elevated diastolic or systolic blood pressure (140/90 mm Hg), overweight (body mass index and lack of leisure-time physical activity [less than once a week in the last month]). Education and income adequacy were used as measures of socioeconomic status and mother tongue as a measure of cultural affiliation.
Results: For most of the risk factors examined, the prevalence of the risk factors was inversely related to socioeconomic status, but the relationship was stronger and more consistent for education than for income. The inverse relationship between socioeconomic status and the prevalence of the risk factors was particularly strong for smoking and overweight, where a gradient was observed: 46% (standard error [SE] 1.4) of men and 42% (SE 4.3) of women who had not completed secondary school were regular smokers, but only 12% (SE 1.0) of men and 13% (SE 0.9) of women with a university degree were regular smokers. Thirty-nine percent (SE 1.4) of men and 19% (SE 3.8) of women who had not completed secondary school were overweight, compared with 26% (SE 2.6) of male and 19% of female university graduates. The prevalence of leisure-time physical inactivity and elevated cholesterol was highest in both men and women in the lowest socioeconomic category, particularly by level of education.
Interpretation: The differences in the prevalence of risk factors for cardiovascular disease between socioeconomic groups are still important in Canada and should be considered in planning programs to reduce the morbidity and mortality from cardiovascular disease.
Behaviour-related risk factors such as being a regular smoker, having high blood pressure, having an elevated cholesterol level, being overweight, being physically inactive and consuming excessive amounts of alcohol can increase the risk of cardiovascular disease (CVD).[1–18] About 70% of premature deaths related to CVD can be prevented by controlling these risk factors.19
The death rate from CVD is inversely related to socioeconomic status (SES).[1–3, 20–24] In 1986 in urban Canada, the number of CVD-related deaths of men in the poorest income quintile was 35% higher than it was for men in the richest quintile. For women, the number was 11% higher. Between 1971 and 1986 there was little change in the differences in the CVD death rate among income quintiles.20
The prevalence of CVD risk factors is also inversely related to SES.[1, 2, 4–9, 25–35] Studies in Canada and the United States have illustrated the relationship between SES and regular smoking,[1, 2, 4–7, 9, 28, 29, 31] between SES and high blood pressure,[2, 5, 7, 8, 27, 32] between SES and obesity[2, 4, 5, 29–31, 33] and between SES and physical inactivity.[2, 4, 6, 9, 28, 29, 30, 34] The findings of an association between SES and elevated cholesterol levels have been inconsistent. Some studies in the US have found no relationship, others have found a weak positive relationship and still others have found an inverse relationship;[4, 5, 7, 29–31, 35] previous surveys in Canada indicate no clear pattern.[1, 9]
The objective of this study was to determine the relationship between CVD risk factors and SES in Canada using the Canadian heart health database. This survey has 4 interesting features: the data are population-based; the main risk factors for CVD are all analysed at the same time; it is one of the few surveys presenting clinical measures for Canada, particularly by SES; and 2 SES measures - education and income adequacy - are used.
The relationship between each of 5 risk factors - having an elevated total plasma cholesterol level, being a regular cigarette smoker, having elevated diastolic (DBP) or systolic blood pressure (SBP), being overweight and being physically inactive - and SES for men and women are analysed by age group, region of residence and mother tongue. The prevalence rates for having diabetes mellitus and consuming excessive amounts of alcohol were not analysed because they were too low. Education and income adequacy were used as measures of SES, and mother tongue was used as a measure of cultural affiliation.
Methods
The prevalence of each of the 5 CVD risk factors was derived from 10 provincial heart health surveys conducted between 1986 and 1992. Each provincial survey targeted people between the ages of 18 and 74 years who did not live in an institution or on a military base and (except in Manitoba) were not aboriginal and living on a reserve. Data from each of the 10 provinces were combined to form the Canadian heart health database. Details of the survey methods are described elsewhere by MacLean and colleagues36 In brief, a probability sample of 29 855 people was selected using health insurance registration files from each province, and these people were invited to participate; 23 129 (77%) agreed. Specially trained nurses administered a standard questionnaire and recorded 2 blood pressure measurements during a home visit. Within 2 weeks, participants visited a survey clinic where 2 more blood pressure measurements were recorded, anthropometric measurements were taken and a fasting blood specimen was taken for plasma lipid determination.
We defined most risk factors described in this paper according to criteria used in a preliminary profile of the Canadian heart health surveys[18, 36–40] but used a modified definition of high blood pressure. Elevated cholesterol was defined as total plasma cholesterol of 5.2 mmol/L or more[18, 37] in people who had fasted 8 hours or more.37 Those who either did not provide a blood specimen or who had fasted for fewer than 8 hours were excluded from the analysis of cholesterol levels. People were considered to be "regular smokers" if they reported smoking at least one cigarette every day.[18, 38] People were considered to have high blood pressure if they had a DBP ≥ 90 mm Hg (using the 5th Korotkoff sound, or the 4th if the sound did not disappear when the pressure reached 0.0 mm Hg), an SBP ≥ 140 mm Hg, or they were being treated (either pharmacologically or non-pharmacologically [salt restriction or weight reduction]) for high blood pressure.[18, 39] Blood pressure values were the mean of the 2 measurements taken during the home visit and the 2 measurements taken during the clinic visit. The mean of the 2 measurements taken during the home interview39 was used for those who did not come to the clinic. The body mass index (BMI, kg/m2) was used; people with a BMI of 27 or more were considered to be overweight.40 People were considered to be physically inactive if they had not engaged in leisure-time physical activity at least once a week during the previous month.38
Three categories of income adequacy were defined: high, middle and low income. These are based on rough approximations of the ratio of household income to the 1990 Statistics Canada low-income cut-offs for each family size, using the 4 income categories available for all surveys. (High income is considered to be one person with an income of $25 000 or more, or 2 or more people with an income of $50 000 or more; middle income is considered to be one person with an income between $12 000 and $24 999, 2 people with an income between $12 000 and $49 999, or 3 or more people with an income between $25 000 and $49 999; low income is considered to be 1 or 2 people with an income of less than less than $12 000 or 3 or more people with an income of less than $25 000.)
The provincial surveys were carried out between 1986 (in Nova Scotia) and 1992 (in Ontario). Since only 4 income categories were available for all surveys it was not possible to adjust the cut-offs for inflation during those years. As a result, the extent of low income is overreported in the earliest surveys (Atlantic provinces) and underreported in the latest surveys (Quebec and Ontario). Caution should be used therefore when comparing regions, although the income adequacy measures are meaningful for Canada as a whole.
Education was also used as a measure of SES. Level of education was determined by asking, "What is the highest grade or year of school you have completed?"41
Mother tongue (English, French, other) refers to the first language learned in childhood and was used as a measure of cultural affiliation.
All prevalence rates (except those broken down by age) were age-standardized using the 1986 Canadian population (including both sexes) as the reference population. Because of the complex sampling design of the survey, standard statistical packages could not be used to provide estimates of standard errors or to calculate test statistics. The formulae used to calculate these quantities in such packages are based on the assumption that samples are selected by simple random sampling with replacement. To obtain standard error (SE) estimates of quantities of interest that take into account the design of the survey, a module written in SAS42 was developed to produce variance estimates specifically for this survey, using the jackknife approach.[41, 43, 44] Weighted standard deviation was used to describe the distribution of variables.
The small number of primary sampling units selected in most provinces necessitated that some of the categories be regrouped to obtain efficient jackknife estimates. Three categories of education were retained (secondary school not completed, secondary school completed and university degree obtained), as well as 3 income categories (low, middle and high). Three age groups were also retained (18 to 34, 35 to 64 and 65 to 74 years). The sample size of each provincial survey was not large enough to support an analysis of SES by province so the provinces were grouped into 3 regions: Atlantic (Newfoundland, Prince Edward Island, Nova Scotia and New Brunswick), Central (Quebec and Ontario) and Western (Manitoba, Saskatchewan, Alberta and British Columbia).
Results
Details of the sample sizes for each risk factor for men and women are provided in Appendix 1. Details about the exact sample sizes for both sexes by age group, region, mother tongue, education and income adequacy are presented in Appendices 2A and 3A. For each of the 5 CVD risk factors, men were generally more likely to have the risk factor than women, no matter what their age, region or mother tongue.
Education level
The prevalence of the 5 CVD risk factors by education level (Table 1) is consistently highest among men and women who have not completed secondary school, with one exception. Men who have completed secondary school are most likely to have high blood pressure. Men and women who have university degrees are least likely to have each of the risk factors, with one exception. Men who have a university degree and men who have completed secondary school are equally unlikely to be physically inactive. The gradient in prevalence from the lowest to the highest level of education varies for each risk factor. For example, 46% (SE 1.4) of men and 42% (SE 4.3) of women who have not completed secondary school are regular smokers, whereas only 12% (SE 1.0) of men and 13% (SE 0.9) of women with university degrees are regular smokers. But 46% (SE 1.5) of men and 43% (SE 1.5) of women who have not completed secondary school have elevated cholesterol levels, as opposed to 38% (SE 1.2) of men and 35% (SE 2.6) of women with university degrees. The prevalence of regular smokers varied inversely with age and education level and was highest in the Atlantic region (34%, SE 0.8; Appendix 2B) and in those whose mother tongue was French (32%, SE 1.5; Appendix 2B). The inverse association between smoking and education was most pronounced in the 18 to 34-year age group (Appendix 2B); in this age group, the prevalence of men who were regular smokers and had not completed secondary school was 62% (SE 2.4) and of women was 58% (SE 5.7).
There was no clear association between level of education and high blood pressure (Appendix 2C). However, more men had high blood pressure than women, and high blood pressure became more prevalent with age. Geographically, men in the Atlantic region were most likely to have high blood pressure (30%, SE 1.2). Men whose mother tongue was French were less likely to have high blood pressure than men whose mother tongue was not French. These differences were observed in all education levels, except that men whose mother tongue was French or anything other than English were equally unlikely to have high blood pressure.
Men were more likely to be overweight than women, and men and women in the Atlantic region were more likely to be overweight than people in the Central or Western regions (Appendix 2D). Differences between mother tongues were not very large. Men in the 35 to 64-year age group who had not completed secondary school were most likely to be overweight. For both men and women, the likelihood of being overweight decreased as level of education increased. In the 18 to 34- and 35 to 64-year age groups, there was a clear relationship between weight and level of education. For men and women in the 65 to 74-year age group, people who had completed secondary school were the least likely to be overweight. In all regions, women with higher levels of education were less likely to be overweight. Differences between linguistic groups were small. For each mother tongue, the likelihood of being overweight decreased as education level increased, except for men whose mother tongue was neither English nor French. In this group, men who had completed secondary school were least likely to be overweight.
More men than women reported being physically inactive for each education level. The highest rates are found in men in the middle age group (44%, SE 1.7; Appendix 2E) and in women in the middle (39%, SE 1.3) and upper (39%, SE 2.4) age groups, in the Atlantic region (44%, SE 1.3 for both men and women) and among those having neither English nor French as their mother tongue (47%, SE 6.1 for men, 42%, SE 2.3 for women). Men in each income (Appendix 3E) and education level (Appendix 2E) were more likely to be physically inactive than women. Physical inactivity was highest in the group with the least education whatever age, region or mother tongue, with 2 exceptions. For women in the 65 to 74-year age group, women with a university degree were most likely to be physically inactive, and for men whose mother tongue was neither English nor French, those who had completed secondary school were most likely to be physically inactive.
Men and women with a university degree were less likely to have elevated cholesterol levels than those with no university degree. The prevalence of elevated total blood cholesterol increased with age (Appendix 2F). People whose mother tongue was French were most likely to have elevated cholesterol levels. For men, the highest rates of elevated cholesterol were in the Atlantic region, whereas for women the highest rates were in Central Canada.
Income adequacy
When the prevalence of risk factors for CVD was examined by level of income adequacy (Table 2) there was a marked inverse relationship between being a cigarette smoker and income level in both men and women, particularly in the 18 to 34-year age group (Appendix 3B). There was also an inverse relationship between smoking and income level by region and mother tongue with one exception, although not as dramatic as that found when analysed by education level. The exception was that in the group of women whose mother tongue was neither English nor French, those in the middle-income group were more likely to be regular smokers. There was no clear association between income level and high blood pressure in either sex. Predictably, high blood pressure was most common in the older age group in both men and women (Appendix 3C) and in both men and women in the Atlantic region. In men, those whose mother tongue was French were least likely to have high blood pressure (20%, SE 1.2).
Although men and women in the low-income group were most likely to be overweight, the gradient between low and high income was smaller than that demonstrated between lower and upper levels of education (Tables 1 and 2). Men and women in the Atlantic region were more likely to be overweight (Appendix 3D), but income level seemed to have little effect on the likelihood of being over-weight. Men and women in the low-income group were most likely to be physically inactive (43%, SE 5.1 for men; 37%, SE 5.0, for women; Appendix 3E). Both men and women in the Atlantic region, and men and women who did not have English or French as their mother tongue were most likely to be physically inactive. The gradient in the rates of physical inactivity between low and high income is very small. Men and women in the oldest age group were the most likely to have high cholesterol levels (Appendix 3F), but when income level was analysed, no discernible pattern was evident between income level and cholesterol level.
Discussion
The likelihood of having any of the 5 CVD risk factors would be expected to be lower in the high income and education groups. The results obtained from the Canadian heart health database show that for most of the risk factors, the prevalence did vary as expected. Moreover, as was observed from the Stanford 5-year project in the US,7 the pattern of distribution was more consistent for education than for income.7 These results are similar to those found in previous surveys conducted in Canada and in the US.[1, 2, 4–9, 27–34]
The likelihood of being physically inactive or having an elevated cholesterol level varied by SES, particularly according to education level, but the gradient was less obvious from low to high levels of education or income. The 1991 General Social Survey also indicated a greater likelihood of physical inactivity among those who had not completed secondary school.6 When income adequacy was considered, this contrast was less evident.
Men and women with a university degree were less likely to have an elevated cholesterol level than those with no university degree, but no difference was found among income levels. Previous surveys conducted in Canada indicate no clear relationship between cholesterol level and education level.6 The relationship between elevated cholesterol levels and SES varies in the literature.[4, 5, 7, 29–31, 35]
Finally, the results of our survey showed no clear pattern between SES and the prevalence of high blood pressure. Previous Canadian surveys also indicated no clear relationship between high blood pressure and education level.[6, 45] A study conducted among British men showed that the prevalence of high blood pressure was lower in the highest social classes than it was in the rest of the population.2
The relationship between education and CVD risk factors was stronger for women than for men, except between education and the likelihood of being a regular smoker. Also, for each of the 5 risk factors, men had higher prevalence rates than women. For both men and women, level of education had the greatest effect on whether a person was a regular smoker or overweight.
In general for men and women the highest prevalence rates for each of the 5 CVD risk factors were found in the Atlantic region. CVD death rates showed the same pattern; the Atlantic provinces have consistently higher rates than other provinces.46 Within each region, there was an inverse relationship between SES and being a regular smoker, being overweight and being physically inactive. Differences among regions did not vary much from one education level to another.
People whose mother tongue was French were most likely to have an elevated cholesterol level and be a regular smoker, whereas people whose mother tongue was English were most likely to be physically inactive. This pattern was still present among the linguistic groups when analysed by level of income and education. Thus, differences among linguistic groups cannot be explained solely by SES. Canada's Health Promotion Survey of 1985 showed that French-speaking people were most likely of the 3 linguistic groups to be regular smokers and physically inactive.28
Although the prevalence of CVD risk factors and death rates have declined sharply during the last decade, the differences between SES are still important. Over time, reducing the differences in the prevalence of CVD risk factors between the low and high SES groups would diminish SES disparities in CVD mortality and reduce CVD mortality for the population as a whole. Smoking is the single most preventable cause of premature death.[47–49] These findings show that smoking prevalence rates vary greatly with SES. Moreover, 50% of all premature deaths each year would be preventable through control of smoking, blood pressure and cholesterol levels, diabetes and alcohol abuse.50 Studies show that increases in life expectancy that arise from changes in prevalence rates of CVD risk factors are more substantial to individuals at risk than for the population as a whole.48 These facts should be considered when planning health promotion programs.
This study has several limitations. The data were obtained through surveys conducted between 1986 and 1992 and any changes in prevalence of the risk factors that occurred during that period will not be evident. Furthermore, the effects of inflation may bias the data on income level, depending on when each provincial survey was done. Finally, using only 3 levels of education and income provides a relatively crude estimate of SES. Nevertheless, the differences in the prevalence of CVD risk factors between high and low levels of SES are still obvious and point to the continuing need to focus on programs designed to lower morbidity and mortality from CVD.
The authors thank the members of the Canadian Heart Health Surveys Research Group for access to the data of the Canadian heart health surveys: C. Balram, P. Connelly, A. Edwards, D. Gelskey, K. Hogan, M. Joffres, R. Lessard, S. MacDonald, D. MacLean, E. MacLeod, M. Nargundkar, B. O'Connor, G. Paradis, A. Petrasovits, B. Reeder, R. Schabas, S. Stachenko and T. Young. They also acknowledge the helpful comments made on previous drafts by the members of this monograph group, especially Doreen Neville, the monograph coordinator.
Funding has been provided in part by the National Health Research and Development Program, Health Canada; provincial ministries of health and the Heart and Stroke Foundation of Canada.
Footnotes
-
This article has been peer reviewed.
Reprint requests to: Dr. Robert Choinière, Direction de la santé publique, Régie régionale de Montréal-Centre, 3725 St-Denis, Montreal QC H2X 3L9