| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
| ||||||||||||||||||||||||
Correspondence: Address correspondence to Zhenmei Zhang, Department of Sociology, Bowling Green State University, Bowling Green, OH 43403. E-mail: zzhang{at}bgnet.bgsu.edu
 |
Abstract |
|---|
 TOP Abstract MethodsResultsDiscussionReferences |
|---|
Key Words: Cardiovascular disease • Marital history • Divorce • Cumulative disadvantage • Marital status
Although many researchers have examined marital status and health, the focus has been on the effects of current marital status on health. Drawing on the 1992 Health and Retirement Study (HRS), this study expands on previous research by examining the health consequences of marital history—that is, the number of marriages, the number and type of marital losses, and cohabitation. Previous research has shown that a marital loss is usually harmful to health (Waite & Gallagher, 2000). What remains unclear is (a) whether the negative effects of marital loss can be modified or erased when a person enters a new relationship, and (b) the health implications of experiencing multiple marital breakups.
Enlightened by the life-course perspective, recent research suggests that the timing and sequence of life events—like marriage and marital dissolution—may lead to different life outcomes in later life through cumulative advantages and disadvantages (Dannefer, 1987; O'Rand, 1996; Wilmoth & Koso, 2002). Specifically, negative events can accumulate over the life course through episodes of illness, adverse socioeconomic conditions, and unhealthy behaviors, resulting in differential lifetime exposure to underlying causal factors of diseases (Kuh & Ben-Shlomo, 1997). This so-called cumulative-effects model suggests that different life trajectories can be associated with different disease risks (Kuh & Ben-Shlomo).
A growing body of research supports the cumulative-effects model. Holden and Kuo (1996) found that people who experienced multiple marital transitions had significantly lower incomes and assets than couples in first marriages. Lower socioeconomic status might be one of the pathways linking multiple marital transitions and health. In terms of mortality, research in the United States as well as in Europe has shown that remarried persons have significantly higher mortality than the continuously married (Hemstrom, 1996; Tucker, Friedman, Wingard, & Schwartz, 1996).
An alternative perspective used in the literature on marital status and health—that is, the selection hypothesis—states that health is associated with marital status through the process of marital selection: Individuals suffering from health problems are less likely to marry and stay married than those who are healthy (Fu & Goldman, 1996; Joung, van de Mheen, Stronks, van Poppel, & Mackenbach, 1997; Lillard & Panis, 1996). In addition, previous research has shown that people with higher education and good economic prospects are more likely to get married (Smock, Manning, & Porter, 2005). This positive socioeconomic selection into marriage suggests that the observed advantage in health enjoyed by married couples cannot be entirely contributed to marriage itself. Nonetheless, there is little evidence that the selection process plays a major role in generating the health advantages of married people (Johnson & Wu, 2002; Waite & Gallagher, 2000).
Previous work in this area often focused on mortality and mental health as the primary health outcomes. The present study examines cardiovascular disease, the main cause of the burden of disability and the leading cause of death in midlife (National Center for Health Statistics, 2004). Two basic questions guide this research: (a) Are various characteristics of the marital history (e.g., current marital status, number and type of marital losses) associated with the likelihood of cardiovascular disease in midlife? and (b) Can socioeconomic status, health behaviors, and social integration explain the links between marital history and cardiovascular morbidity?
|
Methods |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
Measures
Dependent Variables
This study focuses on three cardiovascular outcomes: heart disease, heart attack, and stroke. Respondents reported whether a doctor ever told them that they had a particular cardiovascular problem. A dichotomous indicator of heart disease was created, where 1 = respondent reported that he or she had had a heart attack, coronary heart disease, angina, congestive heart failure, or other heart problems, and 0 = otherwise. The same method was used to create the indicators of heart attack and stroke. Table 1 shows that approximately 12.7% of the sample reported having heart disease, 5.6% reported having had a heart attack, and 2.6% reported having had a stroke.
|
Three mechanisms may potentially mediate the association between marital history and cardiovascular-disease morbidity: socioeconomic status, health behavior, and social integration.
Four indicators of socioeconomic status were created: education, household income, wealth, and health insurance coverage. Education measured the number of years of schooling completed. Household income measured the household income during 1991. Wealth represented the market value of respondents' assets minus debts. Household income and wealth were adjusted by adding constants to all households to eliminate zero income and negative wealth, respectively, and then logging the values. Respondents were considered uninsured when they were not covered by any health insurance programs.
Four types of health behaviors were examined: smoking status, alcohol consumption, exercise, and body mass index (BMI). Smoking status included current smokers and past smokers, with people who had never smoked as the reference group. Alcohol consumption included 1–2 drinks per day and 3 or more drinks per day, with 0 drinks per day as the reference group. Exercise was coded 1 if the respondent exercised three times a week or more. BMI was measured as a categorical variable including obesity (BMI 
30), overweight (25.0 
BMI 29.9), and underweight (BMI <18.5), with normal weight (18.5 BMI 24.9) as the reference group.
Social integration was measured by three variables: childlessness, parental survival status, and church attendance. Childlessness was coded 1 if the respondent had no children. Parental survival status was a categorical variable including one living parent and both parents living, with no living parents as the reference category. Church attendance was a categorical variable including attendance one or more times a week, monthly, or yearly, with non-churchgoers as the reference category.
Age, gender, race, and nativity were controlled in this study because previous research has found that they are associated with cardiovascular disease. The risk of such disease increases with age, and men are more likely than women to develop cardiovascular disease (Black, 1992). Blacks are more likely than Whites to have hypertension and stroke, and foreign-born individuals are less likely than their American-born counterparts to have cardiovascular disease (Hayward, Crimmins, Miles, & Yang, 2000; Jasso, Massey, Rosenzweig, & Smith, 2004). Age was measured as a continuous variable ranging from 51 to 61 years old. Gender (1 = female) and nativity (1 = foreign born) were dummy variables. Race was a categorical variable including Black and Other, with White as the reference category. The weighted descriptive statistics for the independent variables are available from the author upon request.
Analytic Strategy
A series of nested logistic regression models was used to examine differences in the odds of cardiovascular disease across marital history groups and to determine whether the effects of marital history were reduced after introducing the hypothesized mechanisms into the model. The main effects of marital history on the likelihood of cardiovascular disease were examined in Model 1, controlling for age, gender, race, and nativity. In Model 2, the four indicators of socioeconomic status were added to Model 1. And in Model 3, health behaviors and social integration were added to Model 2. Finally, the significance of the interaction between gender and marital history was examined by introducing a set of interaction terms for gender and marital history to Model 3. Although prior work indicated that gender differences existed in the association between marital status and a few health outcomes such as mental health and self-assessed health (e.g., Brown, Bulanda, & Lee, 2005; Williams & Umberson, 2004), only one gender difference reached statistical significance in this study. Therefore, the result is not presented in the table but is described later in the article.
|
Results |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
|
The addition of socioeconomic status in Model 2 of Table 2 reduced but did not explain away the health advantages enjoyed by the continuously married. For example, with the inclusion of socioeconomic status, the odds of heart disease, heart attack, or stroke for the remarried who had experienced multiple losses dropped about 5%, 9%, and 7%, respectively, but were still statistically significant. Socioeconomic status played a much larger role in explaining the higher likelihood of cardiovascular disease among the divorced who had experienced multiple marital losses. Their odds of heart disease, heart attack, and stroke dropped about 16%, 24%, and 24%, respectively, when socioeconomic status was controlled. The results suggest that socioeconomic status is an important mechanism linking marital history and cardiovascular problems. Model 3 introduced health behaviors and social integration to Model 2. Overall, the disparities of the odds of cardiovascular disease across marital-history groups persisted. Health behaviors and social integration did not explain much of the remaining health advantage of the continuously married once socioeconomic status was controlled.
In models not shown here, multiplicative interaction terms for gender and each marital history group were added in Model 3 of Table 2 to examine whether the effects of marital history on the likelihood of cardiovascular disease differed significantly between men and women. Results suggested that the effects of marital history on the risk of cardiovascular disease were similar for men and women with one exception: After controlling for demographic characteristics, socioeconomic status, health behaviors, and social integration, first-time widows were significantly more likely than their male counterparts to report having had a heart attack.
|
Discussion |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
There are several limitations to this study. First, because the 1992 HRS did not contain information about the relative timing of marital transitions and the onset of cardiovascular disease (HRS only asked about the year of the most recent heart attack and stroke), no causal inferences can be drawn between marital history and disease. Both health selectivity into stable marriages and cumulative negative effects of marital losses can produce the cardiovascular-disease morbidity patterns observed in this study. It is highly possible that both processes are operating. The relative good cardiovascular health of the never married, however, casts some doubt on the selection hypothesis. In order to disentangle causal processes, longitudinal data with detailed life-course information about respondents' childhood socioeconomic status, personality traits, family history of cardiovascular disease, and health status before and after marital losses are needed. Second, recent studies suggest that different dimensions of marital quality are related to well-being. The incorporation of quality of current and/or previous marriages may shed light on issues such as whether the health disadvantage of the divorced is due more to poor marital quality before the divorce than to events and conditions after the divorce. Nonetheless, the present results suggest that substantial health disparities exist across different marital trajectories in the risk of cardiovascular disease.
As the trend of marital instability shows no sign of abating, and more and more people live longer than ever before, the number of people with multiple marital losses will certainly increase in the future. This group of people, regardless of current marital status, has a higher risk of cardiovascular disease than the continuously married. Considering the debilitating nature of cardiovascular disease, they will need significant formal and informal care as they advance into old age. This study demonstrated the importance of going beyond current marital status in the study of marital status and health in old age as the marital history of baby boomers becomes increasingly complex.
|
Footnotes |
|---|
1 Center for Family and Demographic Research and the Department of Sociology, Bowling Green State University, OH.
Decision Editor: Linda S. Noelker, PhD
Received for publication December 13, 2004. Accepted for publication December 5, 2005.
|
References |
|---|
|
TOPAbstractMethodsResultsDiscussionReferences |
|---|
| ||||||||||||||||||||||||
| HOME | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|---|
