Leveraging Machine Learning to Predict High-Risk Opioid Overdose Cases in Massachusetts: A Jurisdiction-Level Analysis
Downloads
The opioid crisis remains a significant public health challenge in the United States, characterized by evolving drug supply dynamics and significant geographic and demographic disparities. This study leverages machine learning to analyze jurisdiction-level data from Massachusetts to identify the primary drivers of opioid overdose deaths. Using a descriptive quantitative design, this research employs a supervised machine learning framework, comparing Random Forest, Gradient Boosting, and Ridge Regression models to predict state-level opioid overdose rates. The analysis focuses on model interpretability to determine the relative importance of various substance use rates and demographic factors. The Gradient Boosting Regressor demonstrated exceptional predictive accuracy, achieving a Mean Absolute Error (MAE) of 0.81 and a coefficient of determination (R²) of 0.9930. The feature importance analysis revealed a singular, dominant predictor: the stimulant rate. This finding indicates that the co-use of stimulants is the most critical factor driving opioid-related fatalities, overshadowing the predictive power of other variables, including the fentanyl dominance ratio, heroin rate, and demographic characteristics. The results strongly suggest that the opioid epidemic has transitioned into a polysubstance crisis. Consequently, effective public health interventions must shift from an opioid-centric focus to integrated strategies that address the concurrent use of opioids and stimulants. This study underscores the power of machine learning to provide clear, actionable insights for tackling complex public health crises and recommends the reallocation of resources toward programs that target polysubstance use.
Downloads
1. Borquez, A., & Martin, N. K. (2022). Fatal overdose: predicting to prevent. The International journal on drug policy.
2. Bozorgi, P., Porter, D. E., Eberth, J. M., Eidson, J. P., & Karami, A. (2021). The leading neighborhood-level predictors of drug overdose: a mixed machine learning and spatial approach. . Drug and alcohol dependence, 109-143.
3. Cesare, N. L., Chandler, R., Gibson, E. B., Vickers-Smith, R., Jackson, R., & Oga, E. (2024). Development and validation of a community-level social determinants of health index for drug overdose deaths in the Healing Communities Study. Journal of substance use and addiction treatment.
4. Chatterjee, A., Weitz, M., Savinkina, A., Macmadu, A., Madushani, R. W., Potee, R. A., & Linas, B. P. (2023). Estimated costs and outcomes associated with use and nonuse of medications for opioid use disorder during incarceration and at release in Massachusetts. JAMA network open, e237036-e237036.
5. Davis, J. P., Rao, P., Dilkina, B., Prindle, J., Eddie, D., Christie, N. C., & Dennis, M. (2022). Identifying individual and environmental predictors of opioid and psychostimulant use among adolescents and young adults following outpatient treatment. . Drug and alcohol dependence.
6. Feldmeyer, B., Cullen, F. T., Sun, D., Kulig, T. C., Chouhy, C., & Zidar, M. (2022). The community determinants of death: comparing the macro-level predictors of overdose, homicide, and suicide deaths, 2000 to 2015. Socius.
7. Friedman, J. R., Nguemeni Tiako, M. J., & Hansen, H. (2024). Understanding and addressing widening racial inequalities in drug overdose. American journal of psychiatry, 181(5), 381-390.
8. Garnett, M. F., & Miniño, A. M. (2024). Drug overdose deaths in the United States, 2003-2023.
9. Garnett, M. F., & Miniño, A. M. (2025). Changes in Drug Overdose Mortality and Selected Drug Type by State: United States, 2022 to 2023.
10. Hébert, A. H., & Hill, A. L. (2023). Increasing burden of opioid overdose mortality in the United States: Years of life lost by age, race, and state from 2019–2021. . medRxiv.
11. Kariisa, M., Seth, P., & Jones, C. M. (2022). Increases in disparities in U.S. drug overdose deaths by race and ethnicity: opportunities for clinicians and health systems. JAMA, 328(5), 421-422.
12. Kimmel, S. D., Walley, A. Y., White, L. F., Yan, S., Grella, C., Majeski, A., & Larochelle, M. R. (2024). Medication for opioid use disorder after serious injection-related infections in Massachusetts. JAMA Network Open, e2421740-e2421740.
13. Kirtley, O. J., van Mens, K., Hoogendoorn, M., Kapur, N., & De Beurs, D. (2022). Translating promise into practice: a review of machine learning in suicide research and prevention. The Lancet Psychiatry, 243-252.
14. Liao, C. Y., Garcia, G. G., DiGennaro, C., & Jalali, M. S. (2022). Racial disparities in opioid overdose deaths in Massachusetts. JAMA Network Open.
15. Liu, J., Wu, J., Wang, J., Chen, S., Yin, X., & Gong, Y. (2024). Prevalence and associated factors for depressive symptoms among the general population from 31 provinces in China: The utility of social determinants of health theory. . Journal of Affective Disorder.
16. Liu, Y. S., Kiyang, L., Hayward, J., Zhang, Y., Metes, D., Wang, M., & Cao, B. (2023). Individualized prospective prediction of opioid use disorder. The Canadian Journal of Psychiatry, 54-63.
17. Liu, Y. S., Pierce, D. V., Metes, D., Song, Y., Kiyang, L., Wang, M., & Cao, B. (2025). Population-level individualized prospective prediction of opioid overdose using machine learning. Molecular Psychiatry, 1-6.
18. Luo, S. X., Feaster, D. J., Liu, Y., Balise, R. R., Hu, M. C., Bouzoubaa, L., & Nunes, E. V. (2024). Individual-level risk prediction of return to use during opioid use disorder treatment. . JAMA psychiatry, 45-56.
19. Maclean, J. C., Mallatt, J., Ruhm, C. J., & Simon, K. (2020). Economic studies on the opioid crisis: A review.
20. Maclean, J. C., Mallatt, J., Ruhm, C. J., & Simon, K. (2021). Economic studies on the opioid crisis: costs, causes, and policy responses. In Oxford Research Encyclopedia of Economics and Finance.
21. McCradden, M. D., Anderson, J. A., Stephenson, E., Drysdale, E., Erdman, L., Goldenberg, A., & Zlotnik Shaul, R. (2022). A research ethics framework for the clinical translation of healthcare machine learning. The American Journal of Bioethics, 8-22.
22. Pustz, J., Srinivasan, S., Larochelle, M. R., Walley, A. Y., & Stopka, T. J. (2022). Relationships between places of residence, injury, and death: Spatial and statistical analysis of fatal opioid overdoses across Massachusetts. Spatial and spatio-temporal. Spatial and spatio-temporal epidemiology.
23. Sabo, A., Kuan, G., Abdullah, S., Kuay, H. S., Goni, M. D., & Kueh, Y. C. (2024). Psychometric properties of the social determinants of health questionnaire (SDH-Q): development and validation. . BMC Public Health, 2507.
24. Schell, R. C., Allen, B., G. W., Hallowell, B. D., Scagos, R., Li, Y., & Ahern, J. (2022). Identifying predictors of opioid overdose death at a neighborhood level with machine learning. American Journal of Epidemiology, 526-533.
25. Schell, R. C., Allen, B., Goedel, W. C., Hallowell, B. D., Scagos, R., Li, Y., & Ahern, J. (2022). Identifying predictors of opioid overdose death at a neighborhood level with machine learning. American Journal of Epidemiology, 191(3), 526-533.
26. Shojaati, N. (2023). Systems Science Approaches to the Opioid Crisis: Exploring its Multifaceted Nature through Agent-Based Model Simulations (Doctoral dissertation).
27. Sistani, F., de Bittner, M. R., & Shaya, F. T. (2023). Social determinants of health, substance use, and drug overdose prevention. Journal of the American Pharmacists Association, 628-632.
28. Smith, M. K., Planalp, C., Bennis, S. L., Stately, A., Nelson, I., Martin, J., & Evans, P. (2025). Widening racial disparities in the U.S. overdose epidemic. American journal of preventive medicine, 68(4), 745-753.
29. Tai, M. Y. (2024). A machine learning approach to overdose risk assessment (Doctoral dissertation, University of British Columbia).
30. Williams, L. D., Lee, E., Kristensen, K., Mackesy-Amiti, M. E., & Boodram, B. (2023). Community-, network-, and individual-level predictors of uptake of medication for opioid use disorder among young people who inject drugs and their networks: A multilevel analysis. Drug and alcohol dependence.
31. Wind, K. S. (2021). What Causes Health? Revisiting the Social Determinants of Health (SDH) Through a Salutogenic Lens and Self-Reported Health (SRH) as the Main Outcome: A Realist Evaluation (Doctoral dissertation, University of Toronto (Canada)).
Copyright (c) 2025 Awele Okolie, Dumebi Okolie , Callistus Obunadike, Darlington Ekweli , Bello Abdul-Waliyyu, Paschal Alumona
This work is licensed under a Creative Commons Attribution 4.0 International License.
