Publications
Methodology preprints
Eight working papers documenting the methods behind the Medicare program integrity portfolio. Each is posted to SSRN and paired with a public code repository that reproduces every numerical claim from public CMS data. Each paper presents anomaly detection for investigative prioritization, not fraud determination.
Working papers, 2026
Anomaly Detection and Investigative Prioritization at National Scale: A Reproducible Framework for Medicare Part D Prescriber Risk Analytics
Statistical peer benchmarking combined with isolation forest scoring across 1.38 million Part D prescribers, validated against an honest 2023 exclusions holdout.
Anomaly Detection in Medicare Part D Opioid Prescribing: A Reproducible Dual-Method Risk Stratification Framework with Temporal Holdout Validation
Six-feature opioid prescribing risk scoring with state and specialty peer benchmarking across more than two thousand peer cohorts.
An Unsupervised Composite Risk Scoring Framework for Detecting Anomalous Medicare Part B Injectable Drug Waste
Pareto concentration analysis of injectable drug discard amounts, isolating the small share of drugs that drive most national Medicare waste.
Anomaly Detection and Risk Stratification for Medicare Durable Medical Equipment Billing Patterns: A Reproducible Multi-Year Analysis of 198,621 Supplier-Year Records
Dual-method supplier risk scoring across 2021 to 2023 with bootstrap confidence intervals on national excess projections.
Medicare Telehealth Fraud and Utilization Analytics: An Unsupervised Anomaly Detection Framework on Post-COVID HCPCS-Level Provider Service Data
Provider-level risk scoring of telehealth utilization following the COVID-19 expansion, aligned with OIG telehealth fraud typologies.
Hospital Inpatient DRG Billing Anomaly Detection and Risk Stratification: A Reproducible Composite Risk Score Methodology Applied to the CMS Medicare Inpatient Public Use Files, Fiscal Years 2019 through 2023
Five-feature anomaly detection across 27.3 million inpatient discharges over five fiscal years, combining statistical and machine learning flags.
Unsupervised Anomaly Detection for Medicare Home Health and Hospice Program Integrity: A Composite Risk Score Framework Validated on Real Medicare Payment Data, 2020 to 2023
Dual-track risk scoring across home health agencies and hospice providers with state-level baselines and threshold sensitivity analysis.
A County-Level Composite Equity Index for Medicare Access Disparity Stratification: Methodology, Validation, and Reproducible Federal Data Integration Across 3,198 U.S. Counties, 2019 to 2022
A county-level composite equity index for Medicare access disparities, integrating CMS geographic variation data with HRSA resource files.