IDEAS home Printed from https://ideas.repec.org/p/osf/socarx/c7rvx_v1.html

Predicting Opioid Cost Burden through Integrated PBM and SDOH Modeling: An Explainable AI Framework

Author

Listed:
  • Singhal, Mohit

Abstract

Background: The U.S. opioid epidemic imposes a persistent and disproportionate economic burden on payers and communities, driven by complex interactions among clinical utilization, pharmacy benefit management (PBM) policies, and social determinants of health (SDOH). Despite extensive research on overdose and prescribing risk, few studies have quantified the cost burden of opioid use or examined how policy and community-level factors jointly shape it. Objective: This study develops an interpretable, machine-learning framework integrating PBM cost levers and SDOH indicators to predict and explain county-level variation in opioid-related spending from 2013–2023. The goal is to identify structural drivers of cost, simulate potential savings under policy and social interventions, and support data-driven resource allocation. Methods: Using CMS Medicare Part D data linked with County Health Rankings and U.S. Census indicators, a Random Forest regression model was trained on ten years of county-level data (≈3,000 counties, 2013–2023). Key predictors included unemployment, income ratio, obesity, smoking, provider density, and PBM variables such as cost per claim and opioid prescribing rate. Model interpretability was achieved through SHAP (SHapley Additive exPlanations) analysis and policy simulations testing both PBM and SDOH interventions. Results: The model achieved high predictive accuracy (R² ≈ 0.97), explaining nearly all observed variation in opioid cost per capita. SHAP analysis revealed unemployment, mental-health-provider density, and income inequality as dominant drivers, while provider access and preventive-care variables exerted cost-mitigating effects. Simulated PBM levers (e.g., formulary tightening, utilization management) reduced predicted costs by 4–6%, while integrated SDOH improvements (e.g., +20% behavioral-health access, +10% primary care) achieved up to 17% savings—equivalent to approximately $23 billion nationally. The combined model demonstrated both statistical robustness and policy relevance. Conclusion: This study reframes the opioid crisis through an economic and structural lens, demonstrating that predictive modeling can translate public-health and PBM data into actionable fiscal insights. The proposed PBM–SDOH integration provides a scalable, transparent framework for targeting interventions in high-burden counties, optimizing healthcare spending, and informing evidence-based opioid policy. Keywords: Opioid costs, pharmacy benefit management, social determinants of health, Medicare Part D, machine learning, SHAP, predictive modeling, public health economics, health policy

Suggested Citation

  • Singhal, Mohit, 2025. "Predicting Opioid Cost Burden through Integrated PBM and SDOH Modeling: An Explainable AI Framework," SocArXiv c7rvx_v1, Center for Open Science.
  • Handle: RePEc:osf:socarx:c7rvx_v1
    DOI: 10.31219/osf.io/c7rvx_v1
    as

    Download full text from publisher

    File URL: https://osf.io/download/68ff1d5681d15973f8c3ff98/
    Download Restriction: no

    File URL: https://libkey.io/10.31219/osf.io/c7rvx_v1?utm_source=ideas
    LibKey link: if access is restricted and if your library uses this service, LibKey will redirect you to where you can use your library subscription to access this item
    ---><---

    References listed on IDEAS

    as
    1. Hui Zou & Trevor Hastie, 2005. "Addendum: Regularization and variable selection via the elastic net," Journal of the Royal Statistical Society Series B, Royal Statistical Society, vol. 67(5), pages 768-768, November.
    2. Hui Zou & Trevor Hastie, 2005. "Regularization and variable selection via the elastic net," Journal of the Royal Statistical Society Series B, Royal Statistical Society, vol. 67(2), pages 301-320, April.
    Full references (including those not matched with items on IDEAS)

    Most related items

    These are the items that most often cite the same works as this one and are cited by the same works as this one.
    1. Tutz, Gerhard & Pößnecker, Wolfgang & Uhlmann, Lorenz, 2015. "Variable selection in general multinomial logit models," Computational Statistics & Data Analysis, Elsevier, vol. 82(C), pages 207-222.
    2. Hauzenberger, Niko & Huber, Florian & Klieber, Karin & Marcellino, Massimiliano, 2025. "Bayesian neural networks for macroeconomic analysis," Journal of Econometrics, Elsevier, vol. 249(PC).
    3. Oxana Babecka Kucharcukova & Jan Bruha, 2016. "Nowcasting the Czech Trade Balance," Working Papers 2016/11, Czech National Bank, Research and Statistics Department.
    4. Carstensen, Kai & Heinrich, Markus & Reif, Magnus & Wolters, Maik H., 2020. "Predicting ordinary and severe recessions with a three-state Markov-switching dynamic factor model," International Journal of Forecasting, Elsevier, vol. 36(3), pages 829-850.
    5. Hou-Tai Chang & Ping-Huai Wang & Wei-Fang Chen & Chen-Ju Lin, 2022. "Risk Assessment of Early Lung Cancer with LDCT and Health Examinations," IJERPH, MDPI, vol. 19(8), pages 1-12, April.
    6. Margherita Giuzio, 2017. "Genetic algorithm versus classical methods in sparse index tracking," Decisions in Economics and Finance, Springer;Associazione per la Matematica, vol. 40(1), pages 243-256, November.
    7. Nicolaj N. Mühlbach, 2020. "Tree-based Synthetic Control Methods: Consequences of moving the US Embassy," CREATES Research Papers 2020-04, Department of Economics and Business Economics, Aarhus University.
    8. Wang, Qiao & Zhou, Wei & Cheng, Yonggang & Ma, Gang & Chang, Xiaolin & Miao, Yu & Chen, E, 2018. "Regularized moving least-square method and regularized improved interpolating moving least-square method with nonsingular moment matrices," Applied Mathematics and Computation, Elsevier, vol. 325(C), pages 120-145.
    9. Dmitriy Drusvyatskiy & Adrian S. Lewis, 2018. "Error Bounds, Quadratic Growth, and Linear Convergence of Proximal Methods," Mathematics of Operations Research, INFORMS, vol. 43(3), pages 919-948, August.
    10. Mkhadri, Abdallah & Ouhourane, Mohamed, 2013. "An extended variable inclusion and shrinkage algorithm for correlated variables," Computational Statistics & Data Analysis, Elsevier, vol. 57(1), pages 631-644.
    11. Lucian Belascu & Alexandra Horobet & Georgiana Vrinceanu & Consuela Popescu, 2021. "Performance Dissimilarities in European Union Manufacturing: The Effect of Ownership and Technological Intensity," Sustainability, MDPI, vol. 13(18), pages 1-19, September.
    12. Andrea Carriero & Todd E. Clark & Massimiliano Marcellino, 2025. "Specification Choices in Quantile Regression for Empirical Macroeconomics," Journal of Applied Econometrics, John Wiley & Sons, Ltd., vol. 40(1), pages 57-73, January.
    13. Kim, Hyun Hak & Swanson, Norman R., 2018. "Mining big data using parsimonious factor, machine learning, variable selection and shrinkage methods," International Journal of Forecasting, Elsevier, vol. 34(2), pages 339-354.
    14. Jaime Oliver Huidobro & Alberto Antonioni & Francesca Lipari & Ignacio Tamarit, 2022. "Social capital as a network measure provides new insights on economic growth," PLOS ONE, Public Library of Science, vol. 17(8), pages 1-18, August.
    15. Shuichi Kawano, 2014. "Selection of tuning parameters in bridge regression models via Bayesian information criterion," Statistical Papers, Springer, vol. 55(4), pages 1207-1223, November.
    16. Yize Zhao & Matthias Chung & Brent A. Johnson & Carlos S. Moreno & Qi Long, 2016. "Hierarchical Feature Selection Incorporating Known and Novel Biological Information: Identifying Genomic Features Related to Prostate Cancer Recurrence," Journal of the American Statistical Association, Taylor & Francis Journals, vol. 111(516), pages 1427-1439, October.
    17. Chuliá, Helena & Garrón, Ignacio & Uribe, Jorge M., 2024. "Daily growth at risk: Financial or real drivers? The answer is not always the same," International Journal of Forecasting, Elsevier, vol. 40(2), pages 762-776.
    18. Enrico Bergamini & Georg Zachmann, 2020. "Exploring EU’s Regional Potential in Low-Carbon Technologies," Sustainability, MDPI, vol. 13(1), pages 1-28, December.
    19. Jie Jian & Peijun Sang & Mu Zhu, 2024. "Two Gaussian Regularization Methods for Time-Varying Networks," Journal of Agricultural, Biological and Environmental Statistics, Springer;The International Biometric Society;American Statistical Association, vol. 29(4), pages 853-873, December.
    20. Qianyun Li & Runmin Shi & Faming Liang, 2019. "Drug sensitivity prediction with high-dimensional mixture regression," PLOS ONE, Public Library of Science, vol. 14(2), pages 1-18, February.

    More about this item

    Keywords

    ;
    ;
    ;
    ;
    ;
    ;
    ;
    ;
    ;

    NEP fields

    This paper has been announced in the following NEP Reports:

    Statistics

    Access and download statistics

    Corrections

    All material on this site has been provided by the respective publishers and authors. You can help correct errors and omissions. When requesting a correction, please mention this item's handle: RePEc:osf:socarx:c7rvx_v1. See general information about how to correct material in RePEc.

    If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. This allows to link your profile to this item. It also allows you to accept potential citations to this item that we are uncertain about.

    If CitEc recognized a bibliographic reference but did not link an item in RePEc to it, you can help with this form .

    If you know of missing items citing this one, you can help us creating those links by adding the relevant references in the same way as above, for each refering item. If you are a registered author of this item, you may also want to check the "citations" tab in your RePEc Author Service profile, as there may be some citations waiting for confirmation.

    For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: OSF (email available below). General contact details of provider: https://arabixiv.org .

    Please note that corrections may take a couple of weeks to filter through the various RePEc services.

    IDEAS is a RePEc service. RePEc uses bibliographic data supplied by the respective publishers.