Fraud detection for healthcare

Fraud detection in healthcare is an important yet difficult problem. We present a fraud screening solution to identify suspicious pharmacies from a large dataset of pharmacy claims. Our solution has stemmed from collaboration with medical claim investigators and proven usefulness to investigators by discovering real fraud cases. We focus on a concrete problem of probabilistic outlier detection from a feature set designed for pharmacy claims. Although the reported results are specific to pharmacy claims, this work is not restricted to pharmacy claims. We are currently extending the solution to fraud screening of more general medical claims and fraud detection in other verticals.

Citation

Eldardiry, H.; Liu, J. J.; Zhang, Y.; Fromherz, M. P. J. Fraud detection for healthcare. KDD2013 Workshop on Data Mining for Healthcare; 2013 August 11; Chicago, IL.