RESUMO
OBJECTIVE: To assess compliance with oral hormone replacement therapy (HRT) over time in postmenopausal women in clinical trials and to examine aspects of study conduct (e.g. randomized vs. nonrandomized design) that might be associated with observed compliance rates. METHODS: Aggregation of data from the available published studies using meta-analysis. Eligible studies had to be clinical trials reporting compliance with oral HRT among postmenopausal women and have a sample size of at least 20. MEDLINE and EMBASE databases were used through January 1998 to find studies that assessed compliance with oral HRT in postmenopausal women in clinical trials. Summary estimates of compliance, and of associations between compliance and features of study conduct, were computed allowing for clustering of data within studies. RESULTS: Thirty reports met the inclusion criteria. Overall compliance at about 1 year on average was 83.4%. In a multivariable model, one variable, whether compliance was the primary objective of the study or not, remained statistically significant, even after adjustment for clustering (odds ratio=0.47, 95% confidence interval: 0.22-1.00, P=0.05). CONCLUSIONS: This study supports the importance of using appropriate methods for measuring compliance, if valid estimates of compliance are a serious objective of an investigation. There has been much discussion about the need for increased medication compliance, particularly for those medications used for prevention, but the means and methods to improve compliance remain elusive.
