Patient Predictors of Substance Use Disorder Treatment Initiation in Primary Care

Published in: Journal of Substance Abuse Treatment, Volume 90 (July 2018), Pages 64-72. doi:10.1016/j.jsat.2018.04.004

Posted on on June 28, 2018

by Allison J. Ober, Katherine E. Watkins, Colleen M. McCullough, Claude Messan Setodji, Karen Chan Osilla, Sarah B. Hunter


Primary care clinics are opportune settings in which to deliver substance use disorder (SUD) treatment, but little is known about which patients initiate treatment in these settings.


Using secondary data from a RCT that aimed to integrate SUD treatment into a federally qualified health center (FQHC) using an organizational readiness and collaborative care (CC) intervention, we examined patient-level predictors of initiation of evidence-based practices for opioid and/or alcohol use disorders (OAUDs): a brief behavioral treatment (BT) based on motivational interviewing and cognitive behavioral therapy and medication-assisted treatment (MAT) (extended-release injectable naltrexone (XR-NTX) for patients with an alcohol use disorder or opioid use disorder and buprenorphine/naloxone (BUP/NX) for patients with an opioid use disorder). Using the Andersen model of health care access, we tested bivariate and multivariate logistic regression models to assess associations between patient factors and initiation of BT and MAT.


Twenty-three percent of all participants (N = 392) received BT and 13% received MAT. In the multivariate model examining factors associated with initiation of BT, being of "other" or "multiple" races compared with being White (OR = 0.45, CI = 0.22, 0.92), being homeless (OR = 0.45, CI = 0.21, 0.97) and having been arrested within 90 days of baseline (OR = 0.21 CI = 0.63, 0.69) were associated with significantly lower odds of initiating BT. Greater self-stigma (OR = 1.60, CI = 1.06, 2.42), receiving MAT (OR = 5.52, CI = 2.34, 12.98), and having received the CC study intervention (OR = 12.95, CI = 5.91, 28.37) were associated with higher odds of initiating BT. In the multivariate model examining patient factors associated with initiating MAT, older age (OR = 1.07, CI = 1.03, 1.11), female gender (OR = 3.05, CI = 1.25, 7.46), having a diagnosis of heroin abuse or dependence (with or without alcohol abuse or dependence compared with have a diagnosis of alcohol dependence only (OR = 3.03, CI = 1.17, 7.86), and having received at least one session of BT (OR = 6.42, CI = 2.59, 15.94), were associated with higher odds of initiating MAT.


Individuals who initiate BT for OAUDs in a FQHC are less likely to be homeless and more likely to have greater self-stigma. Those who receive MAT are more likely to be of older age, female, and to have a diagnosis of heroin abuse or dependence, with or without concomitant alcohol abuse or dependence, rather than alcohol abuse or dependence alone. Receiving collaborative care (e.g., a warm handoff, and follow-up by a care coordinator) may be critical to initiating BT. Receiving at least one session of BT is associated with higher odds of receiving MAT, and receiving MAT is associated with higher odds of receiving BT. The Andersen model of health care access provides some insight into who initiates BT and MAT for OAUD treatment in FQHC-based primary care; further research is needed to explore system-level factors that may also influence treatment initiation.

Research conducted by

This report is part of the RAND Corporation External publication series. Many RAND studies are published in peer-reviewed scholarly journals, as chapters in commercial books, or as documents published by other organizations.

The RAND Corporation is a nonprofit institution that helps improve policy and decisionmaking through research and analysis. RAND's publications do not necessarily reflect the opinions of its research clients and sponsors.