|Home | About | Journals | Submit | Contact Us | Français|
In 2002, I was part of a group that published a report of a clinical trial designed to test whether arthroscopic surgery for osteoarthritis of the knee contributed to pain relief and to improvement in functioning.1 The procedure was widely used, but it had never been validated in a clinical trial, and there was no obvious physiologic basis for its working. Because the endpoint (pain relief and associated improved functioning) was very subjective, a placebo control group was required to ensure scientific validity. This meant performing sham surgery (incisions were made but no instruments entered the portal) on the control group. The trial demonstrated that the actual procedure was no more effective than the sham procedure. This trial raised two very different types of ethical questions. The first set of questions concerns the conduct of the trial, while the second set of questions concerns the diffusion of this surgery before it had been validated in a clinical trial.
Crucial ethical issues pertaining to the conduct of the trial include
Crucial ethical issues pertaining to the diffusion of the surgery include
A group headed by Drs. Carol Ashton and Nelda Wray from the Department of Surgery at The Methodist Hospital and by this author, from Baylor College of Medicine, is now systematically examining these issues.
The 1st of our studies, involving the ethical issues in clinical trials of surgical innovations, has been funded by the National Cancer Institute (R01CA134995) and already has been described in the Journal of Medical Ethics.2 It involves a) developing a comprehensive and integrated set of ethical and scientific standards related to clinical trials, organized into a rational taxonomy; b) applying those standards to more than 300 influential surgical trials, to see whether the standards were followed; and c) determining whether any deviations from the standards can be justified as necessary, in light of the special features of surgical trials. That analysis is in progress, but an example can be presented here. A crucial standard for clinical trials is that they not be underpowered to detect real differences so that the null hypothesis is not accepted because of inadequate data. Our preliminary analysis suggests that many of the trials with negative results were underpowered to discover real differences, so valuable innovations might have been inappropriately rejected. Can this lack of statistical power be explained or justified because of difficulty in obtaining support to run non-device surgical trials? Or is it simply a failure on the part of those who conducted the trial to appreciate the importance of adequate power?
A proposal to conduct the 2nd part of our research is under consideration for funding. Preliminary data from a systematic PubMed search for evaluative studies of surgical procedures reveal how few of these studies meet the minimum requirement that there be treatment and control groups together with a systematic plan for assigning subjects to these groups. The percentage of studies in compliance ranges from 1.2% of studies on prostate cancer to 15.1% of studies on urinary incontinence. We propose to examine both the barriers to running controlled trials of surgical procedures and the factors that, in some instances, lead to their being conducted. We also hope to advance suggestions for improving this situation.
Address for reprints: Baruch A. Brody, PhD, Center for Medical Ethics & Health Policy, Baylor College of Medicine, 1 Baylor Plaza, Room 310D, Houston, TX 77030
Presented at the Joint Session of the Denton A. Cooley Cardiovascular Surgical Society and the Michael E. DeBakey International Surgical Society; Austin, Texas, 10–13 June 2010