| Previous research suggests that cardiac rehabilitation programs produce favorable effects on patients' physical and psychological health; however, the therapeutic processes by which these treatment gains are made have not been delineated. Both specific therapeutic processes, such as changes in self-efficacy and exercise capacity, and nonspecific therapeutic processes, such as mood changes and the quality of the relationship with the staff, may represent therapeutic mechanisms (i.e., processes) by which treatment gains are made. The present study used a cross-lagged panel design to examine whether early-treatment changes in exercise capacity, diet self-efficacy, and exercise self-efficacy predicted and preceded late-treatment changes in outcome factors, over and above the influence of the general therapeutic processes. To support these hypotheses, it was necessary to show that: (a) changes in exercise capacity, diet self-efficacy and exercise self-efficacy correlated with changes in outcome; (b) changes in these specific process factors from pre- to midtreatment predicted changes in outcome variables from mid- to posttreatment, but not vice versa; (c) the correlation coefficient for the first cross-lagged regression was statistically larger than the correlation coefficient for the second cross-lagged regression; (d) changes in specific process variables from pre- to midtreatment predicted changes in outcome variables from mid- to posttreatment, controlling far general process factors during the pre- to midtreatment; and (e) changes in exercise capacity, diet self-efficacy, and exercise self-efficacy preceded improvements in outcome. The results did not support the hypotheses. Increases in exercise capacity and confidence in patients' abilities to perform diet and exercise behaviors did not lead to later treatment gains in general. Further analyses revealed no clear causal agent operating in cardiac rehabilitation, suggesting that the associations among these variables were characterized by reciprocity, not by causality. Possible explanations for these findings are discussed. |
