Assessing Spatial and Temporal Variation in Opioid-Related Incidents and Risk Factors in Lowell, Massachusetts, from 2011 to 2022: A Bayesian Spatial-Temporal Approach.

Objective: Lowell, Massachusetts, has been severely impacted by the opioid-related overdose crisis. Utilizing emergency medical services (EMS) data can inform local interventions by identifying opioid-related incidents (ORIs) with shorter lags in reporting. Our objective was to identify spatial and temporal variation in ORI and investigate its association with underlying socioeconomic indicators by coupling EMS data with Bayesian spatial-temporal analyses.

Methods: We obtained data on ORI occurrences within the City of Lowell from January 2011 to June 2022 from Pridestar Trinity EMS. The ORI occurrences were aggregated by month and census tracts. We gathered American Community Survey 5-year estimates (2011-2022) for census-tract percentages of white, black, Hispanic, poverty, unemployed, bachelor's degree, and rent-burdened populations. Using these data, we constructed a Bayesian spatial-temporal Poisson model to identify associations between quarterly ORI rates and these tract-level measures, along with seasonal effects.

Results: ORI rates in Lowell rose from 20 per 10,000 people in 2011 to 93 per 10,000 people in 2018, stabilizing around 60 per 10,000 people from 2019 to 2021, with annual peaks between July through September. Downtown Lowell had consistently higher ORI rates, which extended north-south after 2016. Census tracts with higher percentage of black (relative risk = 1.008; 95% credible interval [1.002, 1.014]) and Hispanic populations (1.014 [1.009, 1.018]) were associated with higher ORI rates. Higher rent burden (1.103 [1.095, 1.11]) and poverty rates (1.02 [1.015, 1.025]) were positively associated with ORI rates, while unemployment rates were inversely associated.

Conclusions: ORI rates in Lowell were associated with community-level sociodemographic factors and exhibited clear seasonal patterns. These findings could inform local prevention and response planning strategies for near-real-time ORI spike detection in communities to mitigate the impact of opioid overdose.