Evaluations of patient education programs: a meta-analysis.

A literature search located 23 evaluations of patient education programs that utilized a randomly selected control group (i.e., an experimental design) or a nonequivalent control group (i.e., a quasi-experimental design), Meta-analysis procedures were used to combine the findings of these studies. The effect size for each of 102 dependent variables was derived from the mean of an experimental group and the mean and standard deviation of an untaught control group. The effect size is a =-score that can be converted into the mean percentile rank of taught patients relative to the untaught controls. The mean effect size for all studies was .74, which is equivalent to improving the average taught patient from the fiftieth to the seventy-seventh percentile relative to the untaught patient group. Suggestions for improving the usefulness of patient education research design and report content are discussed.