Speaker
Description
Medical claim expenses are inherently compositional, as fraud-relevant patterns often emerge from the relative allocation of costs across categories rather than from total expenditure alone. We propose a claim-level fraud screening framework based on compositional profiling, using the Aitchison distance to compare new claims with a historical reference distribution. Statistical significance is assessed via bootstrap resampling. Simulation results under a multivariate normal setting demonstrate effective false positive control, increased sensitivity to meaningful profile deviations, and robustness to scale-only changes. The framework offers an efficient, interpretable, and practically relevant approach to anomaly detection in healthcare expenditure data.
| Classification | Both methodology and application |
|---|---|
| Keywords | Fraud detection; Healthcare claims; Medical expenses; Claim-level profiling; Compositional data analysis; Aitchison distance; Anomaly detection; Bootstrap; |