Faculty Bibliography

BACKGROUND AND AIMS/OBJECTIVE:The opioid crisis is still a public health emergency in the United States, despite recent declines in opioid overdose deaths (OODs) and increased availability of evidence-based practices (EBPs) for opioid use disorder (OUD). The geographic variability in OODs drives the need for localized decision-making, where interventions are tailored to the unique needs of communities. This study aimed to develop and calibrate a simulation model that evaluates the impact of enhanced implementation of EBP on OODs at the community-level. DESIGN/METHODS:We developed OPSiM (Opioid Policy Simulation Model), a community-level microsimulation model that simulates the course of opioid use, OUD, treatment, recovery and overdose-related events. The model was parameterized with data from the HEALing Communities Study and looked at six scenarios of EBPs implemented in 2025 with sustainment through 2030: (1) maintain 2024 EBP levels (status quo); (2) increase initiation of medications for opioid use disorder (MOUD); (3) increase MOUD retention; (4) increase MOUD initiation and retention; (5) increase distribution of naloxone; and (6) both scenarios 4 and 5. SETTING/METHODS:Twenty-nine communities in Massachusetts, New York, and Ohio, USA. PARTICIPANTS/METHODS:Simulated community residents with non-prescribed opioid use or OUD. MEASUREMENTS/METHODS:Estimated number of OODs per 100 000 individuals between 2025 and 2030 in each community, averaged across the 26 communities. FINDINGS/RESULTS:Under the status quo, the model projected 158 OODs (range across communities: 39-468) per 100 000 individuals between 2025 and 2030. Increasing medications for the treatment of OUD (MOUD) retention alone reduced OODs by 6% (range: 3-15%), while increasing MOUD initiation alone reduced OODs by 9% (range: 8-12%). Increasing both MOUD initiation and retention had a synergistic effect, reducing OODs by 21% (range: 15-31%). Reduction in OODs in response to increased MOUD initiation and/or retention was similar across urban and rural communities. The effect of increasing naloxone distribution varied substantially across communities due to differing saturation levels; in some communities, additional naloxone kits provided only marginal benefits. Rural communities were further from saturation whereas most urban communities were at or close to saturation. CONCLUSIONS:A tailored, multi-pronged approach that scales up medications for opioid use disorder alongside widespread naloxone distribution, and that addresses community-specific needs and capacities, will be most effective at reducing opioid overdose deaths in the United States.

IMPORTANCE/UNASSIGNED:Overdose prevention centers (OPCs) are interventions to reduce overdose mortality and support health care engagement. In the US, concerns have been raised that OPCs may be associated with reduced economic activity in their surrounding neighborhoods. OBJECTIVE/UNASSIGNED:To evaluate changes in the local economic activity in New York City (NYC), measured by neighborhood-level foot traffic and consumer spending, following the opening of the first 2 publicly recognized OPCs in the US. DESIGN, SETTING, AND PARTICIPANTS/UNASSIGNED:This cohort study used anonymized mobility and spending data from June 1, 2021, to June 13, 2022, for the areas surrounding the East Harlem and Washington Heights OPCs in NYC. These neighborhoods were defined using 5-minute and 10-minute walking buffers and Business Improvement Districts (BIDs). Synthetic control donors included walking buffers and BIDs around syringe service programs without OPCs and opioid treatment programs that were operational as of OPCs' opening. Analyses were conducted from February to July 2025. EXPOSURES/UNASSIGNED:Opening of the 2 NYC OPCs on November 30, 2021. MAIN OUTCOMES AND MEASURES/UNASSIGNED:Primary outcomes were foot traffic and in-person consumer spending within 10-minute walking buffers. Secondary analyses considered 5-minute walking buffers and BIDs. Augmented synthetic control models were adjusted for neighborhood-level demographic and socioeconomic features, with fit assessed using root mean squared error before OPC opening. Permutation tests and conformal inference were used to assess significance. RESULTS/UNASSIGNED:A total of 27 biweekly observations (13 in pre-OPC and 14 in post-OPC periods) were analyzed. The 10-minute walking buffer analyses captured 1259 consumer spending sites and 7816 foot traffic sites across 2 treated buffers and 56 donor buffers. In East Harlem, the average treatment effect on the treated (ATT) estimate (SE) was -$21.96 ($40.53) for consumer spending (P = .16) and 1.28 (5.40) visits for foot traffic (P = .19). In Washington Heights, ATT (SE) estimates were $14.94 ($37.38) for consumer spending (P = .13) and 0.44 (3.54) visits for foot traffic (P = .97). Secondary analyses produced consistent results. No statistically significant results were observed at any post-OPC time point. CONCLUSIONS AND RELEVANCE/UNASSIGNED:This cohort study found that OPC opening was not associated with significant changes in local economic activity. Given the absence of observed economic harms, policy debates should instead focus on the public health implications of OPCs.

Given substantial reporting delays in overdose deaths, state health departments increasingly use non-fatal overdose data to inform geographically targeted rapid overdose response efforts. We sought to evaluate the extent to which non-fatal overdose events were associated with concurrent and future overdose deaths in Rhode Island. We aggregated non-fatal overdose data from emergency medical services records (2019-2023) and fatal overdose data from the State Unintentional Drug Overdose Reporting System (2020-2023) in 1-, 3-, and 6-month intervals at census block group and census tract levels. Rates of fatal overdose were estimated, relative to non-fatal overdose lagged by 0-12 months, using negative binomial regression, and relative to monthly spikes in non-fatal overdose burden, using zero-inflated Poisson regression. Estimation was implemented using integrated nested Laplace approximation. Each additional non-fatal overdose event per census block group was associated with fatal overdose rates that were 48% higher (95% credible interval: 1.37-1.59) than expected in concurrent months, with smaller associations at the census tract level, in wider time intervals, and when non-fatal overdose data were lagged. Spikes in non-fatal overdose activity were associated with elevated overdose mortality in concurrent periods with fine temporal and geographic granularity, but not in larger time frames and geographic areas.

OBJECTIVE:The CDC recommends initiating opioids for pain treatment at the lowest effective dose and duration. We examine how interactions between dose, duration, and other medication factors (e.g., drug type) influence opioid use disorder (OUD) risk-a gap not considered by CDC guidelines. SUBJECTS/METHODS:Using Medicaid claims data (2016-2019) from 25 states, we analyzed opioid-naïve adults, newly diagnosed with musculoskeletal pain who initiated opioids within three months of diagnosis. A 6-month washout confirmed no prior opioid exposure or musculoskeletal diagnosis. METHODS:Initial opioids were categorized by "dose-days supplied" (low [>0-20 mg MME] to very high [>90 mg MME] dose, and short [1-7 days] to moderate [>7-30 days] supply), and by opioid type; physical therapy (PT) sessions were also recorded. Using Poisson regression models, we estimated the OUD risk associated with dose-days categories, adjusting for baseline demographics, clinical characteristics, and medications. We separately examined opioid dose-days and PT, and assessed PT's moderating effect on dose-days' impact. RESULTS:Among 30,536 patients, half initiated opioids at 20-50 MME for 1-7 days, and 20% received PT. OUD risk was 2-3 times higher for opioids initiated for >7-30 days compared to 1-7 days across doses, and 5.5 times higher for opioids initiated for >7-30 days at > 90 MME versus 1-7 days at < 20 MME. PT alone, neither affected OUD risk nor mitigated the increased risk from longer or higher-dose opioids. CONCLUSIONS:Our findings support the need for careful opioid prescribing and alternative pain management strategies, as the observed associations between initial prescription characteristics and OUD were not mitigated by adjunctive PT. PERSPECTIVE/CONCLUSIONS:This study demonstrated that initial opioid prescriptions of 7-30 days, especially above 90 MME/day, increased OUD risk in opioid-naïve patients with musculoskeletal pain; physical therapy did not mitigate the risk. Different opioids posed varied risks, even at the same dose and duration. Careful prescribing and alternative pain management are essential.

IMPORTANCE/UNASSIGNED:Since the start of the 21st century, more than 800 000 firearm deaths and more than 2 million firearm injuries have occurred in the US. All categories of firearm violence-homicide, suicide, unintentional-result in reverberating harms to individuals, families, communities, and society. The collective responsibility of society is to safeguard the health and safety of its members, including from firearm harms. The JAMA Summit on Firearm Violence convened 60 thought leaders from a wide array of disciplines to chart an innovations roadmap that will lead to substantial reductions in firearm harms by 2040. OBSERVATIONS/UNASSIGNED:The vision for 2040 is a country where firearm violence is substantially reduced and where all people and communities report feeling safe from firearm harms. The vision centers on practical solutions with an understanding of the country's constitutional protections for firearm ownership. Achieving the 2040 vision will require expansion of proven evidence-based strategies and the development of new, innovative approaches rooted in equity, accountability, and collective responsibility. Discussions centered on projecting a safer world, community violence interventions, technologic innovations, federal and state-level oversight of firearms, ethical considerations, and primordial prevention of firearm violence. The Summit charted a roadmap of 5 essential actions in the next 5 years to achieve this vision: (1) focus on communities and change fundamental structures that lead to firearm harms, (2) harness technological strengths responsibly, (3) change the narrative around firearm harms, (4) take a whole-government and whole-society approach, and (5) spark a research revolution on preventing firearm harms. CONCLUSIONS AND RELEVANCE/UNASSIGNED:A safer world will require investing in the discovery, implementation, and scaling of solutions that reduce firearm harms and center on the people and communities most affected by firearm violence.

OBJECTIVE:To examine nationwide trends in the prescribing of controlled medications to early adolescents, adolescents, and young adults enrolled in public insurance (Medicaid) from 2001 to 2019. METHODS:The study utilized US Medicaid data covering publicly insured enrollees from 43 states (2001-2019). Early adolescents (10-12y), adolescents (13-17y), and young adults (18-24y, 25-29y) with ≥ 10 months enrollment in each calendar year were included. Filled prescription for opioids, stimulants, benzodiazepines, Z-hypnotics, barbiturates, and gabapentin were identified. In each calendar year, annual proportions with 1 +  controlled medication, 2 +  classes of controlled medications, and each controlled medication were estimated. RESULTS:In 2019, the sample included 17.9 million enrollees (53 % female). The annual proportion prescribed any controlled medication peaked at 17.5 % in early adolescents (2003), 20.6 % in adolescents (2009), and 34.1 % (18-24y) and 47.0 % (25-29y) in young adults (2010). By 2019, the proportions declined to 11.7 % (early adolescents), 12.6 % (adolescents), 16.2 % (18-24y), and 23.9 % (25-29y). Trends varied by medication and age. The largest absolute decline was in the proportion with an opioid filled (2010 =29.8 %, 2019 =11.2 %, young adults 18-24y; 2003 =14.3 %, 2019 =4.4 %, adolescents). In contrast, the proportion with a stimulant fill increased, with eight-fold increases in young adults 25-29y (2001 =0.3 %, 2019 =2.6 %). Benzodiazepine and Z-hypnotic use peaked in 2010 and declined through 2019. CONCLUSIONS:In the past two decades, there were increases in stimulant prescriptions among young Medicaid enrollees. The declines in opioid, benzodiazepines, barbiturate and Z-hypnotic prescribing are encouraging and may indicate more cautious prescribing related to greater awareness of harms such as misuse and overdose, along with policy initiatives.

Policy Points We can leverage data science and artificial intelligence to inform state and local resource allocation for overdose prevention. Data science and artificial intelligence can help us answer four questions: (1) What is the impact of laws on access to interventions and overdose risk? (2) Where should interventions be targeted? (3) Which types of demographic subgroups benefit the most and the least from interventions? and (4) Which types of interventions should they invest in for each setting and population? Advances in data science and artificial intelligence can accelerate the pace at which we can answer these critical questions and help inform an effective overdose prevention response.

CONTEXT/BACKGROUND:Predictive modeling can identify neighborhoods at elevated risk of future overdose death and may assist community organizations' decisions about harm reduction resource allocation. In Rhode Island, PROVIDENT is a research initiative and randomized community intervention trial that developed and validated a machine learning model that predicts future overdose at a census block group (CBG) level. The PROVIDENT model prioritizes the top 20th percentile of CBGs at highest risk of future overdose death over the subsequent 6-month period. In CBGs assigned to the trial intervention arm, these predictions are then displayed for partnering community organizations via an interactive mapping dashboard. OBJECTIVE:To evaluate whether CBGs prioritized by the PROVIDENT model were associated with increased user engagement via an online dashboard for fatal overdose forecasting and resource planning. DESIGN/METHODS:We estimated prevalence ratios using modified Poisson regression models, adjusted for CBG-level characteristics that may confound the relationship between model predictions and dashboard engagement. SETTING/METHODS:We used CBG-level data in Rhode Island (N = 809) from November 2021 to July 2024. INTERVENTION/METHODS:Our exposure of interest was whether each CBG was prioritized by the PROVIDENT model and shown as prioritized on the interactive mapping dashboard. MAIN OUTCOME MEASURE/METHODS:Our primary outcome was whether a dashboard user from any partnering community organization engaged (eg, clicked, interacted with dashboard elements, or completed assessment or planning surveys) with each CBG on the interactive mapping dashboard. RESULTS:After adjusting for previous model predictions and dashboard engagement, nonfatal overdose counts, and distribution of race and ethnicity, poverty, unemployment, and rent burden, dashboard users were 1.0 to 2.4 times as likely to engage with CBGs prioritized by the PROVIDENT model that were shown as prioritized on the dashboard as compared to CBGs that were prioritized by the PROVIDENT model that were blinded on the dashboard. CONCLUSIONS:Interactive mapping tools with predictive modeling may be useful to support community-based harm reduction organizations in the allocation of resources to neighborhoods predicted to be at high risk of future overdose death.

UNLABELLED:Policy Points State "must-query" prescription drug monitoring programs (PDMPs) were associated with increased overdose deaths, suggesting these policies may have unintended consequences. Black and Hispanic populations experienced disproportionately higher increases in overdose deaths following must-query PDMP adoption, highlighting that these policies may contribute to health disparities. Addressing systemic inequities in health care access and substance use treatment may help supplement the effective components of PDMPs, ensuring that these policies reduce rather than exacerbate overdose deaths. CONTEXT/BACKGROUND:Despite recent declines in national overdose deaths, these reductions have not been equitably experienced. Black and Hispanic communities continue to face rising rates of opioid-related mortality, even as overdose death rates among White individuals have begun to decline. One of the most widely implemented policy responses to the overdose crisis has been the adoption of prescription drug monitoring programs (PDMPs), particularly "must-query" mandates requiring prescribers to consult the PDMP before issuing controlled substances. However, limited research has examined whether the impact of these mandates varies by race and ethnicity. METHODS:We used restricted-use National Vital Statistics System data from 2013 to 2020 to estimate county-level overdose mortality stratified by drug type and race and ethnicity. We categorized deaths as follows: (1) all drug overdoses, (2) all opioid overdoses, and (3) natural/semisynthetic opioid overdoses. Exposure to must-query mandates was modeled as the proportion of the prior year during which mandates were in effect. Using Bayesian spatiotemporal models with county random effects and spatial autocorrelation, we estimated relative rates (RRs) for each outcome overall and by race and ethnicity, adjusting for state policies and sociodemographic characteristics. FINDINGS/RESULTS:Must-query mandates were associated with increases in overdose deaths across all groups, with the largest relative increases among Hispanic (RR = 1.32, 95% credible interval [CrI]: 1.21-1.44) and Black individuals (RR = 1.23, 95% CrI: 1.14-1.33) compared with White individuals (RR = 1.14, 95% CrI: 1.10-1.19). These increases were also observed among Black and Hispanic individuals for natural/semisynthetic opioid overdoses. CONCLUSIONS:PDMP must-query mandates are not uniformly protective across racial and ethnic groups. Increases in overdose mortality following adoption, particularly among Black and Hispanic populations, underscore the need to evaluate drug policies through an equity lens and consider broader structural determinants of health that shape their effectiveness.

BACKGROUND:Between 2022 and 2023, overdose mortality decreased among non-Hispanic (NH) white people but stayed the same or increased among people of colour in the USA. County racialised economic segregation may contribute to overdose mortality. METHODS:measures, one for higher-income NH white and lower-income black residents and another for higher-income NH white and lower-income Hispanic residents. Models included random effects for county, year and county-year interaction, and fixed effects for proportion male, proportion aged 25-44, land area, state and year. We estimated relative risk (RR) by quintile (least vs most privileged) and the difference in overdose mortality per 100 000 (RD) had all counties shifted to the risk of the most advantaged counties (Q5). RESULTS:Counties with the highest proportion of lower-income racially minoritised residents (Q1) had an increased RR of overdose deaths compared with Q5 counties, both overall (aRRs 1.64 (1.51-1.78); 1.40 (1.29-1.52)), and among subgroups. Had all counties experienced the risk of Q5 counties, we estimated an average reduction in overdose deaths overall (RDs per 100 000: -7.20 (-8.25 to -6.10); -6.37 (-7.38 to -5.25)) and among subgroups. CONCLUSION/CONCLUSIONS:County racialised economic segregation was associated with overdose mortality risk in 2018-2022. Investment in evidence-based strategies to reduce overdose risk in places experiencing harms related to racialised economic segregation is critical.