Predicting metabolic control in diabetes: a pilot study using meta-analysis to estimate a linear model.

The purpose of this pilot study was to determine the feasibility of using data from replicated descriptive studies to test a four-variable model designed to explain metabolic control in diabetes. Predictors of metabolic control selected for this analysis were knowledge; health beliefs (and the subscales of barriers, commitment, cues, expectancies, impact on lifestyle, support, and susceptibility); and compliance/adherence. A total of 17 studies, published between 1982 and 1991, were located that met inclusion criteria. Findings indicated that health beliefs have direct and indirect effects on diabetes metabolic control, depending on the individual health belief subscale analyzed. For example, commitment to the benefits of therapy was found to have a statistically significant direct effect on metabolic control; barriers had a statistically significant indirect effect through compliance. The effects of knowledge were consistent throughout the five path models explored. An inverse direct effect was noted on metabolic control and a positive indirect effect was noted on metabolic control through compliance.