Meta-analysis of diabetes patient education research: variations in intervention effects across studies.

Data from a previously reported meta-analysis of diabetes patient education literature were reanalyzed to determine the influence of study/subject characteristics, such as study quality and age of subjects, on patient outcomes. Patient knowledge and self-management skills, weight loss, glycosylated hemoglobin levels, and psychological outcomes were analyzed as outcome variables. Seventy-three relevant published and unpublished studies were located. Patient education appeared to be more effective in younger patients, particularly for the knowledge outcome. For all patients, glycosylated hemoglobin levels improved between 1 and 6 months postintervention, but decreased to 1-month levels after 6 months. Length of the educational intervention did not appear to influence outcomes. More rigorous, experimental research designs tended to produce more conservative effect size estimates. Implications for diabetes patient education are discussed.