This study, conducted in four general hospitals, evaluates the effectiveness of feedback from utilization review coordinators in reducing the number of inappropriate hospital days. Experimental control was exerted over the type of feedback provided in the utilization review process to produce four treatment groups (including a control group) that vary according to 1) the channel used to inform an attending physician of a probable inappropriately located patient and 2) the amount of discretion permitted the nurse coordinator in deciding whether, when and to whom this information is provided. The results of the study indicate that physician and hospital performance in utilization review, measured both by inappropriate patient days and length of stay, are affected by the feedback strategy employed. For patients inappropriately located for a portion of their hospital stay, provision of concurrent feedback resulted in reduction of approximately two thirds of an inappropriate day and two and one half total days, compared with patients for whom no feedback was provided or for whom physician advisor involvement was required. In addition, the effect of reasons causing patients to be inappropriately located (barriers) was assessed. Barriers outside the realm of influence of the hospital or the physician were found to impede the effectiveness of the utilization review systems.