Using drug checking data to model the relationship between the prevalence of fentanyl and xylazine to drug overdose deaths in New York City

Drug checking is an increasingly common harm reduction service available across the United States. However, evidence of whether drug-checking data is associated with population-level health measures is limited. Beginning in November 2021, the New York City (NYC) Health Department partnered with syringe service programs (SSP) to initiate a drug checking pilot study. SSP participants submitted drug samples for point-of-care drug checking and secondary lab testing. Time series generalized additive modeling (GAM) techniques were used to examine whether there was an association between the prevalence of specific compounds and drug-involved overdose death rates. Separate models were fit to evaluate: (1) xylazine prevalence and the xylazine-involved overdose death rate and (2) fentanyl prevalence and the overall drug-involved overdose death rate. Models were adjusted for calendar month A total of N = 26 months of study data were analyzed (November 2021 through December 2023), during which N = 5549 drug involved overdose deaths occurred and N = 594 drug checking samples contained an opioid. A significant nonlinear association was observed between the prevalence of xylazine in drug checking samples and the rate of overdose deaths involving xylazine. In addition, a significant nonlinear association was observed between fentanyl prevalence in drug checking samples and the overall rate of overdose deaths. These findings provide evidence that point-of-care drug checking can produce data that correlate with the drug-involved overdose death rate. This predictive capacity, as well as the timeliness of point-of-care drug checking data, lend support for the role of drug checking as both a harm reduction service and public health monitoring tool.