Estimating Racial and Ethnic Disparities in Substance Use Disorder Treatment and Harm Reduction Services: Findings From a Survey of People Who Use Drugs.

OBJECTIVE

Quantify racial and ethnic disparities in harm reduction and substance use disorder (SUD) treatment use among people who use drugs and compare estimates using a healthcare disparities measurement method aligned with the Institute of Medicine (IOM) definition of healthcare disparities against other regression approaches.

STUDY SETTING AND

DESIGN

45-minute telephone survey of people who use drugs administered from January 2023 to August 2024 across four US locations (i.e., Milwaukee County, Wisconsin; Flint/Detroit, Michigan; statewide in New Jersey; and Bernalillo County, New Mexico). Service use disparities were estimated using propensity score models with rank-replace methods and compared against standard covariate-adjusted logistic regression models.

DATA SOURCES AND ANALYTIC SAMPLE: Survey response data from 1651 respondents who identified as White non-Hispanic (N = 572), Black non-Hispanic (N = 479), Hispanic (N = 453), or American Indian/Alaska Native (N = 147) and had used drugs in the past 30 days. Exposures included respondent demographics, health status, social determinants of health (SDOH), and race/ethnicity. Outcomes included use of any harm reduction services, fentanyl test strip use, naloxone possession, any SUD treatment use, and receipt of buprenorphine, methadone, or naloxone in the past 30 days.

PRINCIPAL FINDINGS: Compared to White non-Hispanic (NH) respondents, Black NH respondents were 17.8 (95% CI: -0.24, -0.12) percentage points less likely to use harm reduction services and 8.8 (95% CI: -0.15, -0.03) points less likely to use SUD treatment. Hispanic respondents were 12.8 percentage points less likely to use any SUD treatment (95% CI: -0.20, -0.06), while American Indian/Alaska Native respondents were 11 percentage points less likely (95% CI: 0.20, -0.02). Standard adjustment models tended to mask or overestimate healthcare disparities relative to rank and replace with propensity scores.

CONCLUSION

Racial and ethnic disparities in harm reduction and treatment necessitate policy reform. Social determinants adjustment should be performed carefully to prevent inaccurately estimating health disparities.