Understanding the impact of co-occurring substance use disorders on receipt of medications for opioid use disorder in the Veterans Health Administration to inform care improvement

Abstract

Background: A growing majority of people with opioid use disorder (OUD) have co-occurring substance use and/or substance use disorders (SUDs), which are associated with reduced likelihood of receiving life-saving OUD treatment medications (MOUD). Efforts to increase and improve MOUD care should explicitly address the role of co-occurring substance use/SUDs. This work is particularly important in the context of the Veterans Health Administration (VA) healthcare system – the VA is the nation’s largest OUD treatment provider, expanding MOUD access is an important VA priority, and over half of VA patients with OUD have co-occurring SUDs. However, more information is needed to effectively address this issue. Objectives: This dissertation involved three aims which sought to: 1) examine associations between distinct types of co-occurring SUDs and either initiation or continuation of MOUD among outpatients with OUD in the national VA healthcare system; 2) assess whether the VA Stepped Care for Opioid Use Disorder Train the Trainer (SCOUTT) initiative’s effectiveness in increasing MOUD receipt was modified by the presence of co-occurring SUDs; and 3) qualitatively assess the perspectives of VA clinicians providing buprenorphine care in primary care, mental health, and pain settings to understand their approach to addressing OUD in patients with co-occurring substance use/SUDs, barriers and facilitators to MOUD receipt for this patient population, and support needed to increase MOUD receipt in this patient population. Methods: In Aim 1, among national VA outpatients with OUD who received care 8/1/2016-7/31/2017, we used adjusted regression models to estimate the likelihood of following-year MOUD initiation (among patients without prior-year MOUD receipt) and continuation (among patients with prior-year MOUD receipt) for patients with each co-occurring SUD relative to those without. In Aim 2, we used a controlled interrupted time series design to examine the monthly proportion of patients who received MOUD in SCOUTT and/or SUD specialty clinics during pre- and post-implementation years (9/1/2017-8/31/2018 and 9/1/2018-8/31/2019, respectively) among patients with OUD who received care in SCOUTT intervention or comparison clinics. We fit segmented logistic regression models to examine pre-post changes in outcomes (immediate level change, change in trend/slope) in intervention vs. comparison clinics, adjusting for patient characteristics and pre-implementation trends, with interaction terms to assess effect modification by the presence of co-occurring SUDs. In Aim 3, we interviewed 27 clinicians in the VA northwest regional network. Interviews were transcribed and qualitatively analyzed using inductive content analysis. Results: In Aim 1, we found that among 23,990 patients without prior-year MOUD receipt, 12% initiated MOUD in the following year. Alcohol use disorder (adjusted incidence rate ratio [aIRR] 0.80, 95% confidence interval [CI] 0.72-0.90) and cannabis use disorder (aIRR 0.78, 95% CI 0.70-0.87) were negatively associated with initiation. Among 11,854 patients with prior-year MOUD receipt, 83% continued MOUD in the following year. Alcohol use disorder (aIRR 0.94, 95% CI 0.91-0.97), amphetamine/other stimulant use disorder (aIRR 0.94, 95% CI 0.90-0.99), and cannabis use disorder (aIRR 0.95, 95% CI 0.93-0.98) were negatively associated with continuation. In Aim 2, we found that the impact of the SCOUTT initiative on MOUD receipt was not significantly modified by the presence of co-occurring SUDs. However, among patients without co-occurring SUDs, the trend change in MOUD receipt in SCOUTT clinics was greater in intervention vs. comparison clinics (adjusted odds ratio [aOR]: 1.06, 95% confidence interval [CI]: 1.02-1.10), and the immediate level change in MOUD receipt in SUD clinics was greater in intervention vs. comparison clinics (aOR: 1.12, 95% CI: 1.02-1.22), while these changes were not significantly greater in intervention vs. comparison clinics among patients with co-occurring SUDs. In Aim 3, we found that participants reported varied approaches to identifying co-occurring substance use/SUDs and to treating OUD in this patient population. Although they reported that this topic was not clearly addressed in clinical guidelines or training, participants generally felt that patients with co-occurring substance use should receive MOUD. Some viewed their primary role as providing this care, others as facilitating linkage to OUD care in SUD specialty settings. Participants reported multiple barriers and facilitators to providing buprenorphine care to patients with co-occurring substance use/SUDs and linking them to SUD specialty care. They discussed their perceptions of how provider factors, patient factors, organizational factors, and external factors impact MOUD receipt for this population. Conclusions: Current VA efforts to implement MOUD outside of SUD specialty settings may be primarily increasing MOUD receipt for patients without co-occurring SUDs. Providers in these settings report multiple barriers to as well as facilitators of providing MOUD to patients with co-occurring substance use/SUDs and linking them to SUD specialty care. Ongoing and future MOUD implementation efforts can work to address these barriers and leverage these facilitators to improve MOUD care for patients with co-occurring substance use/SUDs, keeping in mind that barriers specifically related to co-occurring alcohol, cannabis and amphetamine use disorders may need particular focus.