Accepted for/Published in: JMIR Public Health and Surveillance
Date Submitted: Jul 26, 2022
Date Accepted: Dec 26, 2022
Small Area Forecasting of Opioid-related Mortality: A Bayesian Spatiotemporal Dynamic Modeling Approach
ABSTRACT
Background:
Opioid overdose mortality continued to be at crisis levels across the United States, worsened during the COVID pandemic. The ability to provide prediction of opioid overdose mortality may guide pre-emptive public health responses. Current forecasting models focus on prediction at a large geographical scale such as states or counties, lacking the spatial granularity that local public health officials desire.
Objective:
We aimed to develop Bayesian spatiotemporal dynamic models to predict opioid overdose morality by geographically granular scales for Massachusetts.
Methods:
We developed Bayesian spatiotemporal dynamic models using retrospective fatal opioid OD data for 2005 through 2019 in Massachusetts by 537 ZIP Code Tabulations Areas (ZCTA). The prediction performances were evaluated using 1-year ahead approach. We investigated the potential improvement of prediction accuracy by incorporating ZCTA-level demographic and socioeconomic determinants. We identified ZCTAs with the highest predicted opioid OD levels in terms of rates or counts, and stratified by rural and urban areas.
Results:
The prediction assessment showed that the Bayesian dynamic models with the full spatial and temporal dependency performed the best. Including the ZCTA-level demographic and socioeconomic variables as predictors improved the prediction accuracy, but only in the model without accounting for the spatial and temporal dependency. Predictions were better for urban areas than for rural areas. Using the best performing model and the MA fatal opioid OD data from 2005 through 2019, our models suggested a stabilizing pattern in opioid overdose mortality in 2020 and 2021 if no disruptive changes to the trends observed for 2005-2019.
Conclusions:
Sparse data from small locales pose special challenges in small area predictions. Bayesian spatiotemporal dynamic models, which maximize information borrowing across geographic areas and time points, may be used to provide more accurate predictions for small areas. We encourage the formation of a modeling consortium in the field for fatal opioid OD predictions, where different modeling techniques could be ensembled. Clinical Trial: NA
Citation
Request queued. Please wait while the file is being generated. It may take some time.
Copyright
© The authors. All rights reserved. This is a privileged document currently under peer-review/community review (or an accepted/rejected manuscript). Authors have provided JMIR Publications with an exclusive license to publish this preprint on it's website for review and ahead-of-print citation purposes only. While the final peer-reviewed paper may be licensed under a cc-by license on publication, at this stage authors and publisher expressively prohibit redistribution of this draft paper other than for review purposes.