Examining NSDUH's Assessment of Fentanyl Use: A Comparison of Trends in Fentanyl Use and Fentanyl Overdose Deaths From 2015 to 2020.

Objective: The National Survey on Drug Use and Health (NSDUH), as the primary source of epidemiological substance use data in the United States, could illuminate trends in fentanyl use behaviors contributing to the opioid overdose crisis. We hypothesized that the trend in NSDUH prevalence of lifetime fentanyl injection would match the direction and magnitude of the trend in synthetic opioid overdose deaths.

Method: Using logistic regression, we modeled the 2015-2020 trend in synthetic opioid overdose deaths as a proportion of all deaths. We modeled contemporary trends from cross-sectional NSDUH data for (a) lifetime fentanyl injection, (b) past-year prescription fentanyl misuse, (c) prescription tramadol misuse (the other synthetic opioid counted alongside fentanyl in the overdose deaths category), and (d) combined prescription fentanyl or tramadol misuse. The average annual NSDUH weighted sample size was 272,519,038 (51.5% female, 48.5% male).

Results: Synthetic opioid overdose deaths increased from 2015 to 2020 (odds ratio = 3.39, meaning that the odds of a death being from synthetic opioid overdose in 2020 were 3.39 times the odds of death from that cause in 2015, 95% CI [3.34, 3.44]). None of the substance use trends significantly increased.

Conclusions: Per NSDUH data, the prevalence of fentanyl misuse did not significantly increase in tandem with synthetic opioid overdose deaths from 2015 to 2020. Scrutiny of NSDUH's approach to assessing fentanyl misuse casts doubt on the utility of NSDUH fentanyl data collection. We acknowledge recent changes to the survey and recommend two further changes to optimize a vital source of data on behaviors related to the opioid overdose crisis.