Two articles in this issue of THE JOURNAL pertain to the National Institutes of Health Consensus Development Program. One is the Consensus Conference statement on the management of clinically localized prostate cancer.1 The other is an investigation by Kosecoff and colleagues2 into the effect of Consensus Conference recommendations on how physicians practice. The former summarizes clinical and histological grading of, and surgical and radiation therapy for, a common malignancy. The latter suggests that the former will make little difference.
Kosecoff et al2 reviewed charts from ten Washington State hospitals before and after four Consensus Conferences, two on various aspects of breast cancer and one each on coronary artery bypass surgery and cesarean childbirth. Although there was a trend across time toward greater "compliance" with consensus-panel recommendations, that trend peaked prior to the conferences. And compliance with six of 11 recommendations remained at less than 50% two years