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.