For the past two years, the investigators have been developing, implementing and studying a system for providing feedback to physicians regarding their inpatient practice patterns in order that they might more closely monitor and adjust their own utilization behavior. The proposed supplement to that study seeks to improve the validity of the practice pattern indices used in the physicians' feedback reports by introducing a severity adjustment into the calculation of these measures. The first step in the project will be to determine in which DRG's the severity measure (Disease Staging, version 3.0) consistently explains variability in length of stay and total charges. Only those DRG's will be severity-adjusted. The second step is to calculate and compare physicians' practice pattern indices (developed in the parent study) with and without the severity adjustment. If severity adjustment is effective in reducing the variability of index scores, it will be detected in 3 ways. First, visual inspection of a plot of individual physicians' adjusted vs. unadjusted index scores will reveal that there are physicians whose scores are changed substantially. Second, the degree of correlation between physicians' scores from four consecutive time periods will be greater using severity adjusted indices then using unadjusted indices (i.e., adjusted indices will be more stable over time). Third, a previously observed direct variation between physician's index scores and death and readmission rates will be weakened or eliminated. The physician feedback system offers a readily disseminated method of involving physicians in the regulation of their own clinical behavior. Before it can be used, however, the validity of its practice pattern indices must be maximized. Exploring the importance of severity adjustment, the aim of the proposed study, is a critical component of this task.
