Mediators of Social Support in Coronary Disease

BACKGROUND:

Besides extending our understanding of the mechanisms of social support effects on health, the findings of this project helped in the design and development of more effective and efficient approaches to secondary prevention in coronary artery disease.

DESIGN NARRATIVE:

Social support was assessed in a large consecutive cohort of coronary disease patients referred for diagnostic catheterization (Group A) and in a subgroup of medically treated patients (Group B) with severe coronary artery disease and/or poor left ventricular function with an expected two year 'hard' cardiac event rate (death or nonfatal myocardial infarction) of 25 percent or more. A brief baseline questionnaire assessment of structural and functional aspects of social support as well as other aspects of quality of life was obtained on all coronary artery disease patients without prior revascularization who were referred to the Duke University Cardiac Catheterization Laboratory over a three year period (Group A). Detailed questionnaire and interview assessment of perceived and received social support and psychological traits, that is hostility, was obtained on a high risk subgroup (Group B) and a randomly selected 10 percent subgroup of other Group A patients. Potential behavioral mediators (including smoking behavior, physical activity, medical care utilization) and biological mediators,(including vagal tone, ambulatory ischemic burden) of the social support effects on outcomes were measured in Group B patients and the random subset of Group A. Group A patients were followed by mailed questionnaire at three months and one year and then annually. Group B and the random subset of Group A returned for a one month clinic visit. At that time, repeat social support interviews were administered and patients were sent home with a 48 hour ambulatory ECG monitor to allow measurement of total ischemic burden and heart rate variability (vagal tone).

Group B patients and the random subset of Group A were then followed by telephone interview at one year and then annually. In addition, these patients had brief bimonthly telephone contacts to assess interval changes in social support as well as levels of environmental stress and mood states including depression and anger. All patients were followed for up to three years. Outcome events, including death and myocardial infarction, were ascertained at each point in follow-up. Multivariable analyses using the spline proportional hazards regression model tested the prognostic importance of the social support and psychological measures on outcome and evaluated the role of biological and behavioral variables as mediators, controlling for baseline disease severity.