Process variables as predictors of risk factor changes in a family health behavior change program.

The purpose of the present study was to determine the extent to which intervention process measures are useful in predicting changes in cardiovascular disease (CVD) risk factors among subjects exposed to interventions. Subjects were 99 adult and 105 children who participated in an 18 session, family-based diet and exercise change program. During the intervention, each participant self-monitored diet and exercise for 12 weeks. Additionally, attendance, session evaluation, confidence to achieve goals and goal achievement data were collected each week. The intervention was successful in changing diet, blood pressure and cholesterol levels, but did not produce significant changes in exercise or body mass index. Stepwise multiple regression analyses were conducted for each outcome variable at the 1 and 2 year follow-ups. Ethnicity and sex were first forced into each regression. None of the process measures consistently predicted multiple outcomes in adults and children. The pattern of results provides limited support for the hypothesis that intervention process variables such as attendance, adherence to self-monitoring, achievement of goals and attitude toward sessions partially mediate intervention effects. It is concluded that process measures should be collected in health behavior change programs so that process-outcome relationships can be further explored.