A number of studies have found marked small area variation in hospital discharge rates for many common medical diagnoses and surgical procedures. A large fraction of these hospitalizations have been shown to be clinically unnecessary. If small area variation in inappropriate care accounts for the variation in discharges, then reducing high rates could lead to substantial cost reductions with an improvement in the quality of care. In our study we will test the effectiveness of an educational intervention in use rates and appropriateness of care for two types of procedures: the use of cholecystectomies, and the procedures used in the screening and treatment for cancer of the colon. These procedures were chosen because we expect there to be differences in the problems/questions/interventions/solutions between the treatment of gallbladder disease with its largely inpatient focus, and the screening and treatment of cancer of the colon with its largely outpatient focus. In addition, we have identified two conditions that are: done frequently; have high a cumulative cost; have a significant amount of variation; have a significant amount of physician discretion; for which limited small area data on the reasons for the variation exist; have a high potential intervention; and have measurable outcomes. In order to address the issues raised by the application of these technologies we will make use of existing expertise at The University of Michigan in the area of clinical practice, analysis of small area variation, and community-based intervention. The overall design of the project follows a six step process for each disease. 1) Construct a disease-specific meta analysis based on the clinical literature. 2) A clinician panel will add important clinical details about the care of the patients in question, further refine the meta analysis and derive a set of accepted approaches to the patient. Thus, the meta analysis/clinical panel process identifies normative and aberrant practice patterns. 3) Concurrently, we will examine small area variation in Michigan communities for cholecystectomies and procedures used in the diagnosis, treatment, and screening of colon cancer. Communities with significantly higher and lower utilization rates will serve as intervention sites. 4) An assessment of the appropriateness of the technology usage and patient outcomes will be conducted. 5) An educational intervention will be designed and conducted to improve care and patient outcomes. 6) The effect of the educational intervention will be assessed between the control and intervention communities. We expect that analysis of these important clinical problems, in the context of small areas, will generate new insights as to the reasons and for potential solutions to address small area variation in discharge rates and clinically inappropriate care with their associated costs.
