We will study the relationship between small area variations in Medicare hospitalization rates and rates of inappropriate (i.e. medically unnecessary) hospitalization. The main question is whether higher rates of inappropriateness are found in areas with higher hospitalization rates, especially for clinical conditions which display much between variability. The findings will be important for cost containment programs, outcomes research, and the development of clinical practice protocols. This application is for a 30-month continuation of one-year study. In the first study we developed and applied improved methods for estimating systematic differences in hospital admission rates between small geographic areas, for estimating area-specific rates, and for forming hospital market areas, using six year of data on Massachusetts Medicare admissions by condition. Based on that work, clinical conditions have been ranked by their estimated """"""""systematic"""""""" or underlying between-area variability. In the continuation, for each condition geographic areas will be ranked by their estimated underlying hospitalization rates. Seven diagnostic conditions will be selected, four with high and three with low variability, plus a sample of medical cases as an eighth """"""""condition"""""""". For each condition, some discharges will be selected from areas with high hospitalization rates, some from areas with intermediate rates, and some from areas with low rates. To improve the efficiency of data collection, all discharges will be sampled from a subset of approximately 30 hospitals. Six thousand Medicare records will be reviewed. Modified, condition-specific versions of the Appropriateness Evaluation Protocol (AEP) will be used to screen sampled cases for inappropriateness. The AEP is a widely used criteria-based instrument of documented reliability and validity; two AEP modifications have already been developed and tested. PRO nurse reviewers will perform initial medical record reviews. Cases they identify as inappropriate will be evaluated further by physicians. The data will be analyzed 1) to determine whether rates of inappropriateness are related to admission rates and 2) to assess the influence of differences in population and supply characteristics on the relationship between hospitalization rates and inappropriate use of the hospital, using data bases on need, supply and demand characteristics of each area developed during the first study.
