Modeling the effect of drug courts in North Carolina counties

Background. Illicit opioid overdose death rates drastically increased across North Carolina through 2023. We sought to study the impact of drug courts and what conditions can make their presence more or less effective. Methods. We analyzed counts of illicit opioid overdose deaths for each county in North Carolina from 2017 - 2023. Bayesian Poisson autoregressive models are used to model the change in illicit opioid overdose death rates. We included an indicator of drug court presence in the model and used interaction terms to quantify effect heterogeneity. We used our model to estimate counterfactual outcomes and quantify the effect of drug courts. Results. We found drug courts were associated with decreases in illicit opioid overdose death rates, and that the effect was heterogeneous across North Carolina. We estimated a large protective effect in counties with fewer drug possession arrests, after controlling for other covariates. Conclusion. Diversion into drug courts can provide an effective pathway to reducing the overdose crisis, but the effectiveness of drug courts depends on other features of the county.