AI Medicare Advantage Upcoding
DOI:
https://doi.org/10.63282/3050-9262.IJAIDSML-V6I3P123Keywords:
AI in Healthcare, Medicare Advantage, HCC Coding, Algorithmic Upcoding, Risk Adjustment, CMS Policy, Healthcare Economics, Explainable AIAbstract
The steep rise of artificial intelligence (AI) use in healthcare finance has brought about a major change in risk adjustment and reimbursement mechanisms under Medicare Advantage (MA) programs. AI, based predictive systems are now very commonly used to extract patient records for the assignment of Hierarchical Condition Category (HCC) codes, which in turn results in systematic inflation of risk scores and thus higher government reimbursements. The practice of algorithmic upcoding, as it is often called, constitutes a major risk for the financial viability of public healthcare systems and the Medicare trust fund's ability to continue operating over the long term. In this paper, we suggest an integrated analytical framework that links predictive modelling, healthcare economic analysis, and policy modelling in order to formally define the trade, off between AI predictive accuracy and payment distortion. Through the development of mathematical models, we connect AI, based HCC prediction, risk score calculation, and reimbursement mechanisms to measure the extent of the fiscal distortion effects resulting from algorithmic optimization. The findings reveal the existence of a structural dilemma between the accuracy of the model and the interpretability of the payment; they also show how AI performance, driven optimization can unintentionally increase the economic incentives for systematic upcoding. In addition, we propose a policy, aware simulation framework that helps to assess regulatory measures taken by the Centers for Medicare and Medicaid Services (CMS) under different enforcement and transparency conditions. The study provides a formal foundation for designing regulation-aware AI systems in healthcare finance and proposes a CMS reimbursement reform framework that integrates explainability, auditability, and economic accountability as core design principles. This work represents the first unified AI–economics–policy model for understanding and mitigating algorithmic upcoding in Medicare Advantage systems.
References
[1] Kronick, R., & Welch, W. P. (2016). Measuring coding intensity in Medicare Advantage. Health Services Research, 51(2), 693–709.
[2] Pope, G. C., Kautter, J., Ellis, R. P., Ash, A. S., Ayanian, J. Z., Ingber, M. J., Levy, J. M., & Robst, J. (2016). Risk adjustment of Medicare capitation payments using the CMS-HCC model. Health Care Financing Review, 25(4), 119–141.
[3] Newhouse, J. P., Price, M., McWilliams, J. M., Hsu, J., & Souza, J. (2017). How much favorable selection is left in Medicare Advantage? American Journal of Health Economics, 3(1), 1–26.
[4] Layton, T. J., McGuire, T. G., & Sinaiko, A. D. (2017). Risk adjustment, adverse selection, and health insurance payment policy. Journal of Health Economics, 56, 15–30.
[5] Geruso, M., & Layton, T. (2018). Selection in health insurance markets and its policy remedies. Journal of Economic Perspectives, 32(1), 23–50.
[6] Jacobs, P. D., Kronick, R., & Sommers, B. D. (2018). The risk adjustment of Medicare Advantage: evidence and implications. Health Affairs, 37(1), 103–110.
[7] Meyers, D. J., Trivedi, A. N., & Mor, V. (2019). The Medicare Advantage coding intensity adjustment. Health Affairs, 38(7), 1157–1165.
[8] Obermeyer, Z., Powers, B., Vogeli, C., & Mullainathan, S. (2019). Dissecting racial bias in an algorithm used to manage population health. Science, 366(6464), 447–453.
[9] Geruso, M., & Layton, T. (2020). Upcoding: Evidence from Medicare on squishy risk adjustment. Journal of Political Economy, 128(3), 984–1026.
[10] McWilliams, J. M., Gilstrap, L. G., Stevenson, D. G., Chernew, M. E., Huskamp, H. A., & Grabowski, D. C. (2020). Changes in Medicare Advantage payments and coding practices. Health Affairs, 39(1), 55–63.
[11] McGuire, T. G., Zink, A. L., & Rose, S. (2021). Improving the performance of risk adjustment systems: constrained regressions, reinsurance, and variable selection. American Journal of Health Economics, 7(4), 497–521.
[12] Chen, I. Y., Pierson, E., Rose, S., Joshi, S., Ferryman, K., & Ghassemi, M. (2021). Ethical machine learning in healthcare. Annual Review of Biomedical Data Science, 4, 123–144.
[13] Parikh, R. B., Kakad, M., & Bates, D. W. (2022). Integrating machine learning into health policy payment models. JAMA, 327(15), 1467–1468.
[14] MedPAC (Medicare Payment Advisory Commission). (2022). Report to the Congress: Medicare Payment Policy—Medicare Advantage coding intensity and risk adjustment.
[15] Weissman, G. E., & Joynt Maddox, K. E. (2023). Guiding risk adjustment models toward machine learning methods. JAMA, 330(9), 807–808.
[16] Meyers, D. J., Mor, V., & Trivedi, A. N. (2023). Medicare Advantage risk score growth and plan incentives. Health Services Research, 58(3), 621–630.
[17] Palanisamy, P., Urooj, S., Arunachalam, R., & Lay-Ekuakille, A. (2023). A novel prognostic model using chaotic CNN with hybridized spoofing for enhancing diagnostic accuracy in epileptic seizure prediction. Diagnostics, 13(21), 3382.
[18] Preethi, P., Vasudevan, I., Saravanan, S., Prakash, R. K., & Devendhiran, A. (2023, December). Leveraging network vulnerability detection using improved import vector machine and Cuckoo search based Grey Wolf Optimizer. In 2023 1st International Conference on Optimization Techniques for Learning (ICOTL) (pp. 1-7). IEEE.
[19] Andriola, C., Ellis, R. P., Siracuse, J. J., Hoagland, A., Kuo, T. C., Hsu, H. E., ... & Ash, A. S. (2024). A novel machine learning algorithm for creating risk-adjusted payment formulas. JAMA Health Forum, 5(4), e240625.
[20] Carlin, C. S., Feldman, R., & Jung, J. (2024). The mechanics of risk adjustment and incentives for coding intensity in Medicare. Health Services Research, 59(3), e14272.
