PROVIDENT: Development and Validation of a Machine Learning Model to Predict Neighborhood-level Overdose Risk in Rhode Island.

BACKGROUND: Drug overdose persists as a leading cause of death in the United States, but resources to address it remain limited. As a result, health authorities must consider where to allocate scarce resources within their jurisdictions. Machine learning offers a strategy to identify areas with increased future overdose risk to proactively allocate overdose prevention resources. This modeling study is embedded in a randomized trial to measure the effect of proactive resource allocation on statewide overdose rates in Rhode Island (RI). METHODS: We used statewide data from RI from 2016 to 2020 to develop an ensemble machine learning model predicting neighborhood-level fatal overdose risk. Our ensemble model integrated gradient boosting machine and super learner base models in a moving window framework to make predictions in 6-month intervals. Our performance target, developed a priori with the RI Department of Health, was to identify the 20% of RI neighborhoods containing at least 40% of statewide overdose deaths, including at least one neighborhood per municipality. The model was validated after trial launch. RESULTS: Our model selected priority neighborhoods capturing 40.2% of statewide overdose deaths during the test periods and 44.1% of statewide overdose deaths during validation periods. Our ensemble outperformed the base models during the test periods and performed comparably to the best-performing base model during the validation periods. CONCLUSIONS: We demonstrated the capacity for machine learning models to predict neighborhood-level fatal overdose risk to a degree of accuracy suitable for practitioners. Jurisdictions may consider predictive modeling as a tool to guide allocation of scarce resources.

Translating Predictive Analytics for Public Health Practice: A Case Study of Overdose Prevention in Rhode Island.

Prior applications of machine learning to population health have relied on conventional model assessment criteria, limiting the utility of models as decision support tools for public health practitioners. To facilitate practitioners' use of machine learning as a decision support tool for area-level intervention, we developed and applied 4 practice-based predictive model evaluation criteria (implementation capacity, preventive potential, health equity, and jurisdictional practicalities). We used a case study of overdose prevention in Rhode Island to illustrate how these criteria could inform public health practice and health equity promotion. We used Rhode Island overdose mortality records from January 2016-June 2020 (n = 1,408) and neighborhood-level US Census data. We employed 2 disparate machine learning models, Gaussian process and random forest, to illustrate the comparative utility of our criteria to guide interventions. Our models predicted 7.5%-36.4% of overdose deaths during the test period, illustrating the preventive potential of overdose interventions assuming 5%-20% statewide implementation capacities for neighborhood-level resource deployment. We describe the health equity implications of use of predictive modeling to guide interventions along the lines of urbanicity, racial/ethnic composition, and poverty. We then discuss considerations to complement predictive model evaluation criteria and inform the prevention and mitigation of spatially dynamic public health problems across the breadth of practice. This article is part of a Special Collection on Mental Health.

Cycles of Chronic Opioid Therapy Following Mandatory Prescription Drug Monitoring Program Legislation: A Retrospective Cohort Study.

BACKGROUND: Mandates for prescriber use of prescription drug monitoring programs (PDMPs), databases tracking controlled substance prescriptions, are associated with reduced opioid analgesic (OA) prescribing but may contribute to care discontinuity and chronic opioid therapy (COT) cycling, or multiple initiations and terminations. OBJECTIVE: To estimate risks of COT cycling in New York City (NYC) due to the New York State (NYS) PDMP mandate, compared to risks in neighboring New Jersey (NJ) counties. DESIGN: We estimated cycling risk using Prentice, Williams, and Peterson gap-time models adjusted for age, sex, OA dose, payment type, and county population density, using a life-table difference-in-differences design. Failure time was duration between cycles. In a subgroup analysis, we estimated risk among patients receiving high-dose prescriptions. Sensitivity analyses tested robustness to cycle volume considering only first cycles using Cox proportional hazard models. PARTICIPANTS: The cohort included 7604 patients dispensed 12,695 prescriptions. INTERVENTIONS: The exposure was the August 2013 enactment of the NYS PDMP prescriber use mandate. MAIN MEASURES: We used monthly, patient-level data on OA prescriptions dispensed in NYC and NJ between August 2011 and July 2015. We defined COT as three sequential months of prescriptions, permitting 1-month gaps. We defined recurrence as re-initiation of COT after at least 2 months without prescriptions. The exposure was enactment of the PDMP mandate in NYC; NJ was unexposed. KEY RESULTS: Enactment of the NYS PDMP mandate was associated with an adjusted hazard ratio (HR) for cycling of 1.01 (95% CI, 0.94-1.08) in NYC. For high-dose prescriptions, the risk was 1.16 (95% CI, 1.01-1.34). Sensitivity analyses estimated an overall risk of 1.01 (95% CI, 0.94-1.11) and high-dose risk of 1.09 (95% CI, 0.91-1.31). CONCLUSIONS: The PDMP mandate had no overall effect on COT cycling in NYC but increased cycling risk among patients receiving high-dose opioid prescriptions by 16%, highlighting care discontinuity.

Neighborhood-level association between release from incarceration and fatal overdose, Rhode Island, 2016-2020.

The association between recent release from incarceration and dramatically increased risk of fatal overdose is well-established at the individual level. Fatal overdose and. arrest/release are spatially clustered, suggesting that this association may persist at the neighborhood level. We analyzed multicomponent data from Rhode Island, 2016-2020, and observed a modest association at the census tract level between rates of release per 1000 population and fatal overdose per 100,000 person-years, adjusting for spatial autocorrelation in both the exposure and outcome. Our results suggest that for each additional person released to a given census tract per 1000 population, there is a corresponding increase in the rate of fatal overdose by 2 per 100,000 person years. This association is more pronounced in suburban tracts, where each additional release awaiting trial is associated with an increase in the rate of fatal overdose of 4 per 100,000 person-years and 6 per 100,000 person-years for each additional release following sentence expiration. This association is not modified by the presence or absence of a licensed medication for opioid use disorder (MOUD) treatment provider in the same or surrounding tracts. Our results suggest that neighborhood-level release rates are moderately informative as to tract-level rates of fatal overdose and underscore the importance of expanding pre-release MOUD access in correctional settings. Future research should explore risk and resource environments particularly in suburban and rural areas and their impacts on overdose risk among individuals returning to the community.

Opioid Use Disorder Treatment Facilities With Programs for Special Populations.

INTRODUCTION: Special populations, including veterans, pregnant and postpartum women, and adolescents, benefit from opioid use disorder treatment tailored to their specific needs, but access to such services is poorly described. This study identifies the availability of opioid use disorder treatment facilities that use medications and have special programming and contextualizes facilities amid counties' opioid-related overdose mortality. METHODS: Data were compiled on 15,945 U.S. treatment facilities using medications for opioid use disorder listed in the Behavioral Health Services Treatment Locator in 2018. Facilities with programs tailored to special populations (veterans, pregnant and postpartum women, and adolescents) were identified and geocoded. Counties with such facilities were characterized. Cold spots (county clusters with poor treatment availability) were identified using Getis-Ord Gi* statistics. Data were extracted in October 2018 and analyzed from October 2018 to May 2019. RESULTS: Of all 3,142 U.S. counties, 1,889 (60.1%) had opioid use disorder treatment facilities. Facilities with tailored programs for veterans, pregnant and postpartum women, and adolescents were located in 701 (22.3%), 918 (29.2%), and 1,062 (33.8%) of the counties, respectively. Specific medications provided for opioid use disorder varied, with only a minority of facilities offering methadone (among facilities with tailored programs for veterans, 6.0%; pregnant and postpartum women, 13.2%; adolescents, 1.3%). Many counties reporting opioid-related overdose deaths lacked programs for special populations (veterans, 72.6%; pregnant and postpartum women, 54.8%; adolescents, 30.6%). Cold spots were located throughout the Midwest, U.S. Southeast, and portions of Texas. CONCLUSIONS: Facilities using medications for opioid use disorder with tailored programs for veterans, pregnant and postpartum women, and adolescents are limited. There is a need for improved access to evidence-based programs that address the unique treatment needs of special populations.