Integrating Natural Language Processing and Machine Learning for Documentation-Driven Billing Anomaly Detection in Medicare and Medicaid: A Quantitative Framework for Long-Term Care Facilities
- Publication History
- Published online: September 19, 2026
- DOI
- https://doi.org/10.35877/454RI.qems4806
- Copyright
- Copyright (c) 2026 Michael Agyapong
- User License
- https://creativecommons.org/licenses/by-nc-sa/4.0
Abstract
Improper payments remain one of the most consequential operational and fiscal risks in U.S. public health insurance. For fiscal year 2025, the Centers for Medicare & Medicaid Services (CMS) reported improper payments of $28.83 billion in Medicare fee-for-service, $23.67 billion in Medicare Part C, $4.23 billion in Medicare Part D, $37.39 billion in Medicaid, and $1.37 billion in the Children’s Health Insurance Program. A large share of these losses is documentation-related rather than purely utilization-related, yet most operational detection systems still analyze structured claims without reading the clinical narratives that should substantiate those claims. This paper rebuilds the documentation-driven anomaly detection problem as a joint information problem: what was clinically documented, what was coded, and what was ultimately billed. Using current CMS, HHS, AHRQ, and OIG data, combined with the peer-reviewed literature on clinical natural language processing (NLP), automated coding, and healthcare fraud analytics, the paper develops a hybrid framework for long-term care (LTC) settings that links (1) NLP extraction of diagnoses, services, and severity indicators from unstructured notes, (2) machine-learning detection of anomalous claims behavior, and (3) a cross-referencing engine that scores documentation-billing consistency. Because linked, public LTC note-claim corpora are not available for true patient-level validation, the quantitative contribution of the paper is national and programmatic rather than encounter-level: it documents the financial exposure in nursing care, identifies documentation-dominant error structures, and models conservative savings scenarios from better documentation integrity and automated pre-bill review. The results show that directly quantified documentation-related exposure in FY2025 Medicare fee-for-service and Medicaid alone was approximately $47.65 billion. If integrated documentation-aware anomaly detection reduced that exposure by 10% to 30%, annual savings could plausibly range from $4.77 billion to $14.30 billion before considering secondary spillovers into Medicare Advantage, Part D, or downstream audit efficiencies. The article concludes that documentation-grounded anomaly detection is not merely a technical enhancement; it is a payment-integrity strategy with measurable fiscal and compliance relevance for LTC operators and public payers.
Keywords
References (29)
- References
- Agency for Healthcare Research and Quality. (2024). Challenges and opportunities for improvement. https://www.ahrq.gov/diagnostic-safety/resources/issue-briefs/dxsafety-ehr-impact4.html
- Ahmed, M., Mahmood, A. N., & Islam, M. R. (2016). A survey of anomaly detection techniques in financial domain. Future Generation Computer Systems, 55, 278-288.
- Alsentzer, E., Murphy, J. R., Boag, W., Weng, W.-H., Jin, D., Naumann, T., & McDermott, M. B. A. (2019). Publicly available clinical BERT embeddings. Proceedings of the 2nd Clinical Natural Language Processing Workshop, 72-78.
- Bauder, R. A., Khoshgoftaar, T. M., & Seliya, N. (2017). A survey on the state of healthcare upcoding fraud analysis and detection. Health Services and Outcomes Research Methodology, 17(1), 31-55.
- Centers for Medicare & Medicaid Services. (2026a). Fiscal year 2025 improper payments fact sheet. https://www.cms.gov/newsroom/fact-sheets/fiscal-year-2025-improper-payments-fact-sheet
- Centers for Medicare & Medicaid Services. (2026b). Medicare fee-for-service supplemental improper payment data, 2025. https://www.cms.gov/files/document/nov-2025-medicare-ffs-supplemental-improper-payment-data-2025922.pdf
- Centers for Medicare & Medicaid Services. (2026c). Medicare Part C improper payment measurement. https://www.cms.gov/data-research/monitoring-programs/improper-payment-measurement-programs/medicare-part-c-ipm
- Centers for Medicare & Medicaid Services. (2026d). Medicare Part D improper payment measurement. https://www.cms.gov/data-research/monitoring-programs/improper-payment-measurement-programs/medicare-part-d-ipm
- Centers for Medicare & Medicaid Services. (2026e). 2025 Medicaid and CHIP supplemental improper payment data. https://www.cms.gov/files/document/2025-medicaid-chip-supplemental-improper-payment-data.pdf
- Centers for Medicare & Medicaid Services. (2026f). National health expenditure accounts: 2024 highlights. https://www.cms.gov/files/document/highlights.pdf
- Centers for Medicare & Medicaid Services. (2026g). NHE fact sheet. https://www.cms.gov/data-research/statistics-trends-and-reports/national-health-expenditure-data/nhe-fact-sheet
- Centers for Medicare & Medicaid Services. (2026h). Brief summaries of Medicare & Medicaid. https://www.cms.gov/files/document/brief-summaries-medicare-medicaid-november-19-2025.pdf
- Davis, J., & Shepheard, J. (2024). Clinical documentation integrity: Its role in health data integrity, patient safety and quality outcomes. Health Information Management Journal, 53(1), 3-14.
- Dong, H., Suárez-Paniagua, V., Tikk, D., & Lashari, S. A. (2022). Automated clinical coding: What, why, and where we are? NPJ Digital Medicine, 5, 159.
- du Preez, A., de Vries, J., & de Wet, J. (2025). Fraud detection in healthcare claims using machine learning: A systematic literature review. Health Information Science and Systems, 13, 1-23.
- Herland, M., Khoshgoftaar, T. M., & Wald, R. (2020). A review of data mining using big data in health informatics. Journal of Big Data, 7, Article 113.
- Hossain, E., Rana, R., Higashi, N., Kobashi, S., & Hoque, M. M. (2023). Natural language processing in electronic health records in relation to healthcare decision-making: A systematic review. Computers in Biology and Medicine, 155, 106649.
- Johnson, J. M., & Khoshgoftaar, T. M. (2019). Medicare fraud detection using neural networks. Journal of Big Data, 6, Article 63.
- Li, J., Huang, K.-Y., Jin, J., & Shi, J. (2008). A survey on statistical methods for health care fraud detection. Health Care Management Science, 11(3), 275-287.
- Office of Inspector General. (2024). Nursing facility industry segment-specific compliance program guidance. https://oig.hhs.gov/compliance/nursing-facility-icpg/
- Office of Inspector General. (2026). Nursing homes. https://oig.hhs.gov/reports/featured/nursing-homes/
- Sanderson, A. L., Elkbuli, A., McKenney, M., Boneva, D., & Hai, S. (2025). The impact of clinical documentation integrity programs on case mix index and severity capture. Journal of Trauma and Acute Care Surgery, 98(1), 73-80.
- U.S. Department of Health and Human Services. (2025a). HHS unveils AI strategy to transform agency operations. https://www.hhs.gov/press-room/hhs-unveils-ai-strategy-to-transform-agency-operations.html
- U.S. Department of Health and Human Services. (2025b). Request for information: Accelerating the adoption and use of artificial intelligence as part of clinical care. https://www.federalregister.gov/documents/2025/12/23/2025-23641/request-for-information-accelerating-the-adoption-and-use-of-artificial-intelligence-as-part-of
- U.S. Department of Health and Human Services. (2026). TEFCA, America’s national interoperability network, reaches nearly 500 million health records exchanged. https://www.hhs.gov/press-room/tefca-americas-national-interoperability-network-reaches-nearly-500-million-health-records-exchanged.html
- Woo, B. F. Y., Cato, K., Cho, H., You, S. B., & Song, J. (2025). The use of large language models in clinical documentation: A scoping review. International Journal of Nursing Studies, 176, 105322.
- Yu, H., Beam, A. L., & Kohane, I. S. (2024). Large language models in biomedical and health informatics: A bibliometric review. Journal of the American Medical Informatics Association, 31(10), 2354-2365.
- Zhang, Y., Liu, K., & He, Y. (2025). A systematic review of automated International Classification of Diseases coding models using MIMIC databases. JAMIA Open, 8(1), ooae123.
How to Cite
Copyright & license
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
