Keywords: spatial, Bayesian, opioid, dynamic, conditional autoregressive
Opioid misuse is a national epidemic, and the problem is even more stark in Ohio which led the nation in fatal overdoses in 2015. In previous years, it was believed that prescription opiates were driving this crisis, and in 2011 legislation was implemented to shut down Ohio's "pill mills." However, opioid overdose deaths due to heroin and then fentanyl have drastically increased in Ohio in recent years. To combat the epidemic, it is imperative to account for the different types of opiates when studying changing trends of overdose deaths. In this talk, we develop a Bayesian multivariate spatio-temporal model for Ohio county overdose death rates due to different types of opioids. This model allows us to not only study how socio-environmental factors relate to opioid overdose deaths, but also spatio-temporal trends in the types of opioids contributing to death.