A Multivariate Spatio-Temporal Model of Opioid Overdose Deaths in Ohio

• Activity Number: 83 - Applications in Surveys and Social Science
• Type: Contributed
• Date/Time: Sunday, July 28, 2019 : 4:00 PM to 5:50 PM
• Sponsor: Government Statistics Section
• Author(s): Staci Hepler* and David Kline and Lance Waller
• Companies: Wake Forest University and The Ohio State University and Emory University
• Keywords: spatio-temporal; opioids; multivariate; change point; auto-regressive; Bayesian

Abstract

In 2017, Ohio had the second highest age-adjusted fatal drug overdose rate. In earlier years it was believed prescription opiates were driving the opioid crisis in Ohio. However, following policy changes aimed at limiting access to prescription opiates, opioid overdose deaths due to fentanyl have drastically increased. In this work, we develop a Bayesian multivariate spatio-temporal model for Ohio county overdose death rates from 2007 to 2017 due to different types of opiates. The log-odds are assumed to have a spatially varying change point regression model. By assuming the regression coefficients are a multivariate conditional autoregressive process, we capture spatial dependence within a drug type and also dependence across drug types. A Polya-Gamma data augmentation scheme makes it computationally efficient to fit this model in a Bayesian framework. This model allows us to not only study spatio-temporal trends in overdose death rates, but also to detect county-level shifts in these trends over time for various types of opiates.