Literature Collection

In 2007, the Veterans Health Administration (VHA) began national implementation of Primary Care-Mental Health Integration (PCMHI) services. A major goal was to enhance mental health access and address unmet treatment needs. Integrated care services have been shown to improve depression identification1 and enhance engagement in specialty mental health treatment.2 Szymanski et al. examined a sample of VHA users in fiscal year (FY) 2010 and documented positive associations between receipt of PCMHI services on the day of a positive depression screen and initiation of depression treatment within 12 weeks.3 The authors noted as a study limitation that individuals who had received same-day PCMHI services may have had unmeasured differences in symptom severity or willingness to initiate depression treatment, as compared with other study patients.3 Also, in the years since 2010, VHA implementation of PCMHI services has expanded substantially. The present analysis re-evaluates the influence of same-day PCMHI services on initiation of depression treatment, for a more recent period and adjusting for patient propensity to have received PCMHI and/or Specialty Mental Health (SMH) clinic services on the day of the initial positive depression screen in primary care (PC). METHODS For FY2017 data, we applied methods from the FY2010 study,3 adding use of propensity score weights to adjust for patient likelihood of receiving same-day PCMHI and/or SMH services.4, 5 Generalized boosted models4, 5 were used to create the weights, with location of same-day services as the outcome. Separate propensity score–weighted logistic regressions assessed initiation within 12 weeks of antidepressant pharmacotherapy, psychotherapy, and either treatment.4 Analyses used SAS (version 9.3) and R software (version 3.4.2). The study was conducted as part of ongoing operations in the VHA Office of Mental Health and Suicide Prevention. RESULTS Table 1 presents sample characteristics, by type of services received. Compared with patients receiving same-day PCMHI, the “PC only” patients were older (38.4% vs. 23.1% were 65+ years old); less likely to be Hispanic (7.7% vs. 8.9%); more likely to be male (87.8% vs. 85.7%), married (49.2% vs. 45.8%), with a prior mental health diagnosis (16.7% vs. 13.3%) and prior VHA outpatient use (67.1% vs. 64.6%); and had lower baseline PHQ-2 (Patient Health Questionnaire) scores (4.64 vs. 4.83).

BACKGROUND: People living with HIV and opioid use disorder (OUD) are disproportionally affected by adverse socio-structural exposures negatively affecting health, which have shown inconsistent associations with uptake of medications for OUD (MOUD). This study aimed to determine whether social determinants of health (SDOH) were associated with MOUD uptake and trajectories of substance use in a clinical trial of people seeking treatment. METHODS: Data are from a 2018 to 2019 randomized trial comparing the effectiveness of different MOUD to achieve viral suppression among people living with HIV and OUD. SDOH were defined by variables mapping to Healthy People 2030 domains: education (Education Access and Quality), income (Economic Stability), homelessness (Neighborhood and Built Environment), criminal justice involvement (Social and Community Context), and recent SUD care (Health Care Access and Quality). Associations between SDOH and MOUD initiation were assessed with Cox proportional hazards models, and SDOH and substance use over time with generalized estimating equation models. RESULTS: Participants (N = 114) averaged 47 years old, 63% were male, 56% were Black, and 12% Hispanic. Participants reported an average of 2.3 out of 5 positive SDOH indicators (SD = 1.2). Stable housing was the most commonly reported SDOH (61%), followed by no recent criminal justice involvement (59%), having a high-school level education or greater (56%), income stability (45%), and recent SUD care (13%). Each additional favorable SDOH was associated with a 25% increase in the likelihood of MOUD initiation during the study period [adjusted HR = 1.25, 95% CI = (1.01, 1.55), P = .044]. Positive SDOH were also associated with a decrease in the odds of baseline opioid use and a greater reduction in opioid use during subsequent weeks of the study (P < .001 for a joint test of baseline and slope differences). CONCLUSIONS: Positive social determinants of health, in aggregate, may increase the likelihood of MOUD treatment initiation among people living with HIV and OUD.

IMPORTANCE: There are high rates of maternal and newborn morbidity and mortality associated with opioid use disorder (OUD). Integrating OUD treatment in obstetric practices for pregnant and postpartum women via telemedicine can increase access to care and reduce the consequences of OUD. Evaluation of this care delivery model, however, is needed before widespread adoption. OBJECTIVE: To compare maternal and newborn outcomes among pregnant women with OUD receiving care via telemedicine vs in person. DESIGN, SETTING, AND PARTICIPANTS: A nonrandomized controlled trial including 98 women receiving perinatal OUD treatment in 4 outpatient obstetric practices by telemedicine or in person and followed up until 6 to 8 weeks post partum was conducted from September 4, 2017, to December 31, 2018. Logistic regression with propensity score adjustment was applied to reduce group selection bias and control for potentially confounding variables. INTERVENTIONS: Participants were seen weekly for 4 weeks, every 2 weeks for 4 weeks, and monthly thereafter and provided relapse prevention therapy and buprenorphine. MAIN OUTCOMES AND MEASURES: The outcomes were retention in treatment, defined as uninterrupted addiction treatment during pregnancy through 6 to 8 weeks post partum; urine drug screen results at delivery and 6 to 8 weeks post partum; and a neonatal abstinence syndrome (NAS) diagnosis collected via electronic health records. RESULTS: The mean (SD) age of the 98 pregnant women was 30.23 (5.12) years. Of these, 41 of 44 women (93.2%) in the telemedicine group and 48 of 54 women (88.9%) in the in-person group chose to continue treatment in the program after an initial evaluation. After propensity score weighting and doubly robust estimation, no significant differences were found between groups in retention in treatment at 6 to 8 weeks post partum (telemedicine: 80.4% vs in person: 92.7%; treatment effect, -12.2%; 95% CI, -32.3% to -4.4%). Similarly, after propensity score weighting and doubly robust estimation, there were no significant group differences in rates of NAS (telemedicine: 45.4% vs in person: 63.2%; treatment effect, -17.8%; 95% CI, -41.0% to 8.9%). CONCLUSIONS AND RELEVANCE: In this nonrandomized controlled trial, virtually integrated OUD care in obstetric practices produced similar maternal and newborn outcomes compared with in-person care. These findings may have important public health implications for combatting the opioid crisis and its consequences on pregnant women and their families. Future large randomized clinical trials are needed. CLINICAL TRIAL REGISTRATION: ClinicalTrials.gov identifier: NCT04049032.

BACKGROUND: Persistence with medication-assisted therapy among patients with opioid use disorder has been associated with reduced likelihood of illicit opioid use. OBJECTIVE: We aimed to describe treatment persistence and identify factors associated with 1-year persistence among insured patients newly initiating buprenorphine-containing pharmacotherapy. METHODS: The retrospective observational cohort included employer-sponsored and managed Medicaid patients newly started on buprenorphine-containing therapy between June 30, 2010, and January 1, 2015. Persistence was measured as both a continuous and dichotomous variable (proportion of patients persistent for 1 year). Multivariable logistic regression analysis was used to identify factors associated with 1-year persistence. RESULTS: A total of 302 patients met inclusion criteria. The median [range] number of treatment episodes was 1 [1-4]. Mean number of days on therapy during the first episode was 206 (SD = 152) days, with 40.4% (n = 122) of patients persisting for 1 year. Presence of concomitant fills of prescription opioid analgesics (odds ratio [OR] = 0.25; 95% CI = 0.12-0.51), being in care of an addiction specialist (OR = 0.40; 95% CI = 0.21-0.76), and Medicaid insurance coverage (OR = 0.33; 95% CI = 0.13-0.84) were significantly and negatively associated with 1-year persistence. There was also a strong inverse relationship between persistence and inpatient hospitalization (OR = 0.30; 95% CI = 0.12-0.76). CONCLUSIONS: Several health care delivery and use variables were significantly associated with nonpersistence. Concomitant use of prescription opioids is the most easily modifiable risk factor that health care providers and policy makers may act on to improve treatment continuation.