Methodological issues of randomized controlled trials for the evaluation of reproductive health interventions

Randomized controlled trials provide the most valid basis for comparing alternative treatment modalities. Unfortunately, the quality of many such trials has not been as high as our patients deserve. We offer here methodological and practical insights to hey issues related to the design, implementation, and analysis of randomized controlled trials in obstetrics and gynecology, based on our experience in exploring research hypotheses in this field. Specifically, we present randomization as a method for the prevention of confounding; we discuss the process of randomization and the bias that can be introduced to randomized controlled trials by some popular methods of randomization. The strategies for presenting baseline comparisons and evaluating the impact of a treatment or intervention are also reviewed. We hope to contribute to a more rigorous selection of treatments for our clinical and public health armamentarium. (C) 1996 Academic Press, Inc.