Abstract:
|
The improper payment rate of government healthcare insurance programs such as Medicare is estimated to add up to billions of dollars. Hence, understanding the underlying factors for provider billing behavior and the attendant identification of abnormal activities are crucial. However, this is complicated because the data distribution of billing patterns is asymmetric and fat-tailed. Therefore, the relevant covariates can vary depending on the segment of interest in the distribution of billing aggressiveness. This paper proposes a Bayesian Information Criterion within a quantile regression framework to isolate key factors that best explain billing aggressiveness patterns. Using public data sources related to the CMS Medicare Part B program, we offer insights on the common characteristics for both conservative and aggressive billing by providers and identification of upcoding providers to be audited. This can help health care professionals better understand and manage billing patterns with respect to geography, provider and procedure type.
|