Three years of experience with prospective randomized effectiveness studies

We developed methodology for prospective randomized effectiveness studies using a demonstration project at a multispecialty practice, health maintenance organization, and hospital in academic medical center. An operational unit called the effectiveness registry was developed to design and support comparisons of potential practice improvements with standard care. The studies differ from observational effectiveness studies in that they provide long-term follow-up of randomized comparison groups. Physician involvement in data collection is limited. No tests or observations are made other than those required for clinical care. Follow-up and data collection are modeled after tumor registry procedures. Patients who refuse randomization enter the study in whichever treatment arm they choose. The protocol for each study is approved by the institutional review board (IRB) before recruitment begins, and all patients, randomized and nonrandomized, sign an informed consent document. Between its beginning on October 7, 1993 and April 7, 1997, the IRB approved 14 trials. Four were terminated after entering at most a few patients. Recruitment is complete in four trials and continues in six. Randomization was accepted by 74% (596/804) of the patients. Over 800 patients in 10 studies are being followed at least annually. Major peer-reviewed journals have accepted reports of initial findings for two studies. Prospective randomized effectiveness studies are feasible in the multipractice setting and have potential to provide useful and reliable assessment of treatment outcomes. Collaborative arrangements between several institutions are needed to provide larger sample sizes.