Online Program

A multivariate spatio-temporal model of opioid overdose deaths in Ohio (306598)

Keywords: opioid, change point, spatial, Bayesian, policy, fentanyl

In 2017, Ohio had the second highest age-adjusted fatal drug overdose rate. In earlier years it was believed prescription opioids were driving the opioid crisis in Ohio, and in 2011 legislation was implemented to shut down Ohio's "pill mills." However, opioid overdose deaths due to fentanyl have drastically increased in Ohio in recent years. To better understand the changing waves of the epidemic, it is imperative to account for the different types of opioids when studying trends of overdose deaths. In this talk, we develop a Bayesian multivariate spatio-temporal model for Ohio county overdose death rates from 2007 to 2017 due to different types of opioids. By assuming a change point regression model for the log-odds, we can estimate when a shift in the trend occurred for each type of opioid considered. This model allows us to study spatio-temporal trends in the types of opioids contributing to death and also to estimate whether a change in the trend occurred following policy changes.