In this study, we found social inequality in health behaviours, mental health and health literacy in married/cohabiting individuals with CVD using data from a representative population-based survey combined with register data. Thus, low educational attainment was associated with higher levels of CVD risk factors in married/cohabiting individuals with CVD compared to their high educational attainment counterparts.
We found an increased occurrence of daily smoking, physical inactivity, unhealthy diet and obesity in married/cohabiting individuals with CVD and low educational attainment. This constitutes an increased risk of recurrent CVD events (10, 11). Furthermore, we found an increased level of perceived stress along with reduced health literacy and MHRQOL, which also increase the risk of recurrent CVD events (6, 9). The accumulation of several risk factors in married/cohabiting individuals with CVD and low educational attainment further elevates the risk of recurrent CVD events (5). Our findings are in agreement with previous studies reporting social inequality in health in individuals with CVD in health behaviour (31, 32), mental health (33, 34) and health literacy (23). The findings are novel, as no previous study has addressed social inequality in health behaviour, mental health and health literacy among married/cohabiting individuals with CVD.
Despite the observed protective effects, in terms of reduced CVD risk, mortality and morbidity of marriage/cohabitation, the causal mechanisms remain to be elucidated (35). Furthermore, it remains unclear whether marriage or cohabitation best describes the protective effects of "living with someone" in terms of reduced CVD risk (3). Wong and colleagues suggest that marital status may actually be a surrogate marker for other underlying cardiovascular risks. This is supported by a recent study reporting that cohabitation provided similar health benefits as marriage in terms of self-rated health (36). Various explanations for the protective effect of marriage/cohabitation have been suggested including adherence to healthy lifestyle and treatment, increased social support and reduced stress (3, 4, 35, 37). Despite the protective effects of marriage/cohabitation, low educational attainment negatively affects health behaviour, mental health and health literacy in the present study. Rosengren and colleagues argued that low educational attainment may be a proxy for broader social disadvantages including low awareness of the need for seeking care, reduced access to information and reduced life opportunities due to poor economy, housing, healthcare, etc. (38). Somehow these or similar conditions negatively affected health in married/cohabiting individuals with CVD and low educational attainment and to some extent overruled the protective effects of marriage/cohabitation in the present study.
The major strength of the present study is its use of a large-scale representative population-based survey combined with register data. Furthermore, the study focus on age groups with a high prevalence of CVD (39). The survey has a high response rate, data are weighted and analyses are adjusted, which is expected to reduce non-response bias. Danish register data are generally of high validity and accuracy (40). Finally, the study includes ten highly relevant self-reported outcome measures in terms of describing the risk of recurrent CVD events (10, 11).
The study also has limitations. First, register information on marriage/cohabitation are collected at the time of the survey. Due to this marriage/cohabitation at the time of the initial diagnosis is unknown. Since the study focus on recurrent CVD events, we believe that simultaneous collection of data on health behaviours, mental health, health literacy and marriage/cohabitation is a sound approach. Second, non-response and mortality bias cannot be entirely excluded. However, we have a high response rate and use calibrated weights to reduce non-response bias. Since only survivors participate in the survey, we expect the most ill individuals with CVD to be underrepresented in the study, which could lead to survival bias towards the null. Furthermore, mortality bias may lead to an underestimation of the differences between the educational groups and thus an underestimation of the social inequality in health in married/cohabiting individuals with CVD. Third, time since diagnosis at the time of the survey varies from 0 to 23 years. This may influence self-reported data in the survey, treatment at the time of the initial CVD event and survival rates. In order to compensate for the large time span, we adjust the analyses for time since diagnosis. Furthermore, we conducted sensitivity analyses using a shorter timespan (0–15 years) and 1-year age groups (0–23 years), which only resulted in minor changes in the estimates. We conducted further sensitivity analyses leaving out multimorbidity, which affected some estimates, but no systematically. Thus, we decided to adjust for multimorbidity since it was of clinical relevance. Fourth, multiple simultaneous heart diagnoses are not handled in the study, since we focus solely on the initial CVD diagnosis. Fifth, in the present study, social inequality is addressed using educational attainment. This is a generic measure of social position which has been used as marker for social inequality in a wide range of studies. However, it may also be relevant to apply other measures of social inequality, such as labour market status or income to further elucidate social inequality in health among married/cohabiting individuals with CVD.
The results of the present study should be addressed in public health strategies, targeted secondary prevention, treatment and rehabilitation in relation to CVD. Married/cohabiting individuals with CVD are usually appraised as a group with a reduced risk of recurrent CVD events compared with non-married/cohabiting individuals with CVD (10, 11). However, due to the social inequality in health in married/cohabiting individuals with CVD found in this study, different approaches in targeted secondary prevention, treatment and rehabilitation may be required depending on educational attainment. It may also be useful to train or involve the partner of married/cohabitating individuals with low educational attainment and CVD in treatment and rehabilitation. Previous research has suggested that reduction of socioeconomic inequalities in CVD outcomes may require different approaches for men and women (41). This may also hold true for combinations of cohabitation status and educational attainment. Furthermore, the social inequality found in married/cohabiting individuals with CVD may also apply to married/cohabiting individuals with other chronic conditions.