Automatic Detection of Providers with Excess Healthcare Spending

Abstract

This thesis aims to develop techniques to help large hospital systems to detect providers with excess spending. Identifying fraud, waste, and abuse resulting in super uous expendi- tures associated with care delivery is central to the success of these large hospital systems and for making the cost of healthcare sustainable. In theory, such expenditures should be easily identiable with large amounts of historical data. However, to the best of our knowledge there is no data mining framework that systematically addresses the problem of identifying unwarranted variation in expenditures on high dimensional claims data using unsupervised machine learning techniques. In this thesis, we propose methods to uncover unwarranted variation in healthcare spending by automatically extracting reference groups of peer-providers from the data and then detecting high cost outliers within these groups. Besides we also implement existing graph based techniques and compare the results with our methods. We demonstrate the utility of our proposed framework on datasets from a large ACO (Accountable Care Organization) in the Pacic Northwest of the United States to successfully identify unwarranted variation in the provision of therapeutic procedures that had previously gone unnoticed.