Agentic Workflow Automation For Ghost Billing Detection In U.S. Home Health Agencies: An Autonomous Compliance Monitoring Framework Via GPS Correlation, EVV Integration, And Behavioral AI
DOI:
https://doi.org/10.29284/7g8zq523Keywords:
agentic artificial intelligence; healthcare fraud detection; ghost billing; Electronic Visit Verification (EVV); behavioral AI; GPS correlation; compliance monitoring; explainable AI; anomaly detection; home health agencies.Abstract
Background: Ghost billing remains one of the most persistent forms of healthcare fraud in U.S. home health agencies, contributing to the estimated $1.1 billion in annual Medicare home health improper payments. Although Electronic Visit Verification (EVV) systems mandated under the 21st Century Cures Act have strengthened visit documentation, existing implementations primarily collect data rather than intelligently detect fraud, allowing GPS spoofing, fabricated visits, duplicate billing, and falsified documentation to evade rule-based auditing.
Objective: This study proposes and empirically evaluates an Agentic Workflow Automation Framework for autonomous ghost billing detection that integrates EVV records, GPS correlation, behavioral artificial intelligence, and explainable decision-making in a coordinated multi-agent architecture.
Methods: Six specialized agents — EVV validation, GPS validation, behavioral intelligence, compliance rule evaluation, fraud risk assessment, and explainability — analyze caregiver activity and fuse evidence into dynamic risk scores. The framework was evaluated on a synthetic dataset of 77,624 home-health visits (120 caregivers, 180 patients, 180 days) containing 6,212 injected fraudulent visits (8.0%) across five fraud types, using a temporal train/test split and comparison against rule-based auditing, logistic regression, isolation forest, and random forest baselines.
Results: The proposed framework achieved the highest F1-score (0.863) and ROC-AUC (0.995) of all evaluated methods, with 89.3% precision, 83.5% recall, and a 0.9% false-positive rate — a 73% reduction in false positives relative to rule-based auditing — at 0.013 ms per visit. Detection recall reached 100% for ghost visits, GPS spoofing, and upcoded durations.
Conclusions: Coordinated agentic evidence fusion materially outperforms rule-based EVV auditing and matches or exceeds strong machine learning baselines while operating at real-time speed, supporting proactive pre-payment fraud screening in home healthcare. Validation on operational EVV data is the essential next step.
References
1. Adadi, A., & Berrada, M. (2018). Peeking inside the black-box: A survey on explainable artificial intelligence (XAI). IEEE Access, 6, 52138–52160. https://doi.org/10.1109/ACCESS.2018.2870052
2. Aggarwal, C. C. (2017). Outlier analysis (2nd ed.). Springer. https://doi.org/10.1007/978-3-319-47578-3
3. Centers for Medicare & Medicaid Services. (2023). Electronic Visit Verification (EVV). U.S. Department of Health and Human Services. https://www.medicaid.gov/medicaid/home-community-based-services/home-community-based-services-guidance-additional-resources/electronic-visit-verification
4. Centers for Medicare & Medicaid Services. (2024). Fiscal year 2024 improper payments fact sheet. U.S. Department of Health and Human Services. https://www.cms.gov/newsroom/fact-sheets/fiscal-year-2024-improper-payments-fact-sheet
5. Chatterjee, A., & Ahmed, B. S. (2022). IoT anomaly detection methods and applications: A survey. Internet of Things, 19, 100568. https://doi.org/10.1016/j.iot.2022.100568
6. Choi, E., Bahadori, M. T., Schuetz, A., Stewart, W. F., & Sun, J. (2016). Doctor AI: Predicting clinical events via recurrent neural networks. Proceedings of the 1st Machine Learning for Healthcare Conference, 301–318.
7. Doshi-Velez, F., & Kim, B. (2017). Towards a rigorous science of interpretable machine learning. arXiv. https://arxiv.org/abs/1702.08608
8. European Parliament & Council of the European Union. (2024). Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act). Official Journal of the European Union, L 2024/1689. http://data.europa.eu/eli/reg/2024/1689/oj
9. Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep learning. MIT Press. https://www.deeplearningbook.org
10. Gupta, D., Gupta, M., Bhatt, S., & Tosun, A. S. (2021). Detecting anomalous user behavior in remote patient monitoring. Proceedings of the 2021 IEEE International Conference on Smart Computing. https://doi.org/10.1109/SMARTCOMP52413.2021.00051
11. Kshetri, N. (2021). Artificial intelligence in healthcare: Applications, challenges, and future directions. IT Professional, 23(2), 12–16. https://doi.org/10.1109/MITP.2020.3035468
12. Lundberg, S. M., & Lee, S. I. (2017). A unified approach to interpreting model predictions. Advances in Neural Information Processing Systems, 30.
13. McCarthy, J., Powell, M., Stouffer, K., Tang, C., Zimmerman, T., Barker, W., Ogunyale, T., Wynne, D., & Wiltberger, J. (2020). Securing manufacturing industrial control systems: Behavioral anomaly detection (NISTIR 8219). National Institute of Standards and Technology. https://doi.org/10.6028/NIST.IR.8219
14. National Institute of Standards and Technology. (2023). Artificial intelligence risk management framework (AI RMF 1.0) (NIST AI 100-1). U.S. Department of Commerce. https://doi.org/10.6028/NIST.AI.100-1
15. OECD. (2022). OECD framework for the classification of AI systems (OECD Digital Economy Papers No. 323). OECD Publishing. https://doi.org/10.1787/cb6d9eca-en
16. Office of Inspector General. (2012). Personal care services: Trends, vulnerabilities, and recommendations for improvement (Report No. OIG-12-12-01). U.S. Department of Health and Human Services. https://oig.hhs.gov/reports/all/2012/personal-care-services-trends-vulnerabilities-and-recommendations-for-improvement/
17. Office of Inspector General. (2016). Nationwide analysis of common characteristics in OIG home health fraud cases (Report No. OEI-05-16-00031). U.S. Department of Health and Human Services. https://oig.hhs.gov/oei/reports/oei-05-16-00031.pdf
18. Office of Inspector General. (2024). Kansas's implemented Electronic Visit Verification system could be improved (Report No. A-07-23-03255). U.S. Department of Health and Human Services. https://oig.hhs.gov/reports/all/2024/kansass-implemented-electronic-visit-verification-system-could-be-improved/
19. Peffers, K., Tuunanen, T., Rothenberger, M. A., & Chatterjee, S. (2007). A design science research methodology for information systems research. Journal of Management Information Systems, 24(3), 45–77. https://doi.org/10.2753/MIS0742-1222240302
20. Ribeiro, M. T., Singh, S., & Guestrin, C. (2016). "Why should I trust you?": Explaining the predictions of any classifier. Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 1135–1144. https://doi.org/10.1145/2939672.2939778
21. Russell, S., & Norvig, P. (2021). Artificial intelligence: A modern approach (4th ed.). Pearson.
22. Sage, R., Mashinchi, G. M., Lissau, A., Standley, K., & Ender, J. (2022). "Less time committed to care": Beliefs about electronic visit verification among adults using home-based personal assistance services. Home Healthcare Now, 40(2), 82–91. https://doi.org/10.1097/NHH.0000000000001038
23. Shaik, T., Tao, X., Higgins, N., Li, L., Gururajan, R., Zhou, X., & Acharya, U. R. (2023). Remote patient monitoring using artificial intelligence: Current state, applications, and challenges. Wiley Interdisciplinary Reviews: Data Mining and Knowledge Discovery, 13(2), e1485. https://doi.org/10.1002/widm.1485
24. Topol, E. (2019). Deep medicine: How artificial intelligence can make healthcare human again. Basic Books.
25. U.S. Department of Justice. (2024). National health care fraud enforcement action results in 193 defendants charged and over $2.75 billion in false claims. https://www.justice.gov/archives/opa/blog/combating-health-care-fraud-2024-national-enforcement-action
26. World Health Organization. (2021). Ethics and governance of artificial intelligence for health. World Health Organization.
Downloads
Published
Issue
Section
License
This work is licensed under a Creative Commons Attribution 4.0 International License.
This work is licensed under a Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
