Hostility, Depression, Social Environment and CHD Risk

BACKGROUND:

Hostility, depression/depressive personality, and socioeconomic status (SES) have all been shown to influence the risk of coronary heart disease (CHD). While the three factors have been studied separately in previous work, there is evidence that they are interrelated and there is reason to hypothesize that they would interact in a multiplicative fashion to dramatically increase risk when they are present in combination.

DESIGN NARRATIVE:

Three studies were conducted designed to describe the interrelationships of hostility, depression/depressive personality, and socioeconomic status and document their joint impact on several health outcomes. Study 1 took advantage of two existing databases to study over 3000 coronary artery disease patients who had been followed for over a decade. Both self-report and behavioral indicators of hostility and depression were studied in combination with SES to predict survival while controlling for initial disease severity. These measures were also used in analyses designed to discriminate the psychosocial profile of patients who died early in follow-up from those who died after five years. Data on health behavior profiles and medical care utilization were examined to evaluate the possibility that those factors could account for the relationships between psychosocial characteristics and survival.

Study 2 examined the ability of depressive personality, hostility, and SES to jointly predict acute myocardial infarction (AMI) and mortality in a sample of 730 initially healthy people who had been followed for 30 years. It also investigated the associations of these psychosocial measures with changes in risk factor patterns and health indicators over the follow-up period.

Study 3 administered a large battery of measures to community volunteers in order to document the interrelationships between hostility, depressive personality, and SES in detail. These measures were also used to predict the magnitude and duration of cardiovascular reactivity to mood induction tasks.

The study was renewed in FY 2000 to continue data analysis.

The study has been extended through June, 2009 to conduct four studies that address several questions. All will target significant clinical events such as coronary events and mortality as the primary dependent variables. Two of these are designed to test aspects of a model of mechanisms that account for the connections between stress and health outcomes. One of these is the notion that tangible, social, and psychological resources can moderate stress reactions. This leads to the hypothesis that good social relationships can be especially protective for those of low SES. This will be tested in a large representative population sample (Study One) and a large sample of military veterans. A corollary of this hypothesis is that the interaction of social support and SES should be most apparent in those under high levels of job strain, which will also be tested in the population sample. Both physical health and depressive symptoms will be the outcomes. Another resource is effective coping strategies for stress producing conditions. One of these may be tension reduction through moderate alcohol consumption, a proposition will be examined as an interaction between moderate alcohol usage and job strain in the same sample. Alcohol consumption is also hypothesized to affect health through other pathways in the conceptual model. The investigators have found that some psychosocial variables (hostility and depression) are associated with patterns of binge drinking and alcohol-related behavior problems, although this effect may be moderated by SES. This could account for some of the associations between hostility, depression, and health, a proposition that will be tested in the sample of veterans. The other studies will concentrate on the task of better illuminating the health effects due to particular components of two complex psychosocial variables that are often treated as more global constructs. One study of cardiac patients will rely on an extensive battery of depressive symptom measures collected during hospitalization to predict subsequent cardiac events and mortality. The other is a study of community volunteers who were assessed with a comprehensive set of hostility measures in the early 1990's. These will serve as predictors of subsequent cardiac events and mortality. All studies are expansions of past and current work, building on those findings to test new hypotheses, extend notions based previous findings to the prediction of significant clinical outcomes, and improve our conceptualizations of widely studied psychosocial risk factors.

• Study One will use the data of the Danish National Cohort Study (DANCOS; Helwig-Larsen, et al., 2003)
• Study Two Part A will employ the same DANCOS sample that was used in Study One.
• Study Two Part B will use data from the Vietnam Experience Study (Center for Disease Control, 1988; Tegan, 2004)
• Study Three will be based on data from the patient sample previously collected under the investigator's former protocol (IRB #0608-04-5R5ER)
• Study Four will involve the follow-up of approximately 250 community volunteers who were intensively evaluated for hostility in laboratory sessions during the period of 1991-1993