Outlier Identification in Utilization Data

摘要

Government regulators, community leaders, corporate benefits managers, the press and many others are scrutinizing case mix data in an attempt to develop controls on the excessively rapid rise in health care costs. Since many of these people are not health care professionals, there is always a risk of misinterpretation of such data. To minimize this, those responsible for providing health care utilization data and statistics should ensure that the information is as accurate as possible. Identification of 'outliers,' or atypical cases is one very important aspect of this responsibility. Exploiting certain characteristics of health care data, the process of identifying outliers involves three stages -- statistical analysis, clinical evaluation, and consideration of other aspects of the health care delivery system.