Article Title

Adolescent SBIRT in Pediatric Primary Care: Patient Outcomes From a Randomized Trial in an Integrated Health Care System

Publication Date



child and adolescent health, behavioral and mental health, substance abuse, addiction


Background: Many adolescents in need of specialty treatment for substance use and mental health problems never access care. Screening, brief intervention and referral to treatment (SBIRT) is a promising approach to early identification and intervention for adolescent substance use. We describe patient outcomes from a trial of different modalities of SBIRT for adolescents in primary care.

Methods: We randomized pediatricians (N = 52) in a pediatrics clinic to three study arms: 1) pediatrician-only, in which pediatricians were trained to deliver SBIRT; 2) embedded behavioral health clinician (BHC), in which pediatricians referred adolescents who endorsed risk factors (substance use or mood symptoms) to a BHC; and 3) usual care (UC). We used electronic health record (EHR) data to examine risk factors, CRAFFT scores, treatment referral, initiation and engagement.

Results: 1,871 patients were eligible for assessments, brief interventions and referrals. Differences in outcomes between the index and the next well visit within 2 years were examined across the three arms. The odds of risk factor endorsement decreased over time for all patients (adjusted odds ratio [aOR]: 0.07; 95% confidence interval [CI]: 0.05–0.11); the embedded-BHC arm had significantly lower odds of risk factor endorsement compared to UC (aOR: 0.65; 95% CI: 0.43–0.97); there were no differences between UC and the physician-only arm. There were no differences in CRAFFT scores over time or between intervention arms. The embedded-BHC arm had fewer referrals to specialty treatment compared to the other study arms; referral rates between the physician-only and UC arms did not differ. Among those referred, the physician-only arm had significantly lower odds of treatment initiation compared with UC (aOR: 0.53; 95% CI: 0.28–0.99) and the embedded-BHC arm (aOR: 0.25; 95% CI: 0.12–0.49); no differences were found between UC and embedded-BHC. There were no differences in treatment engagement (at least 2 visits within 30 days) across the arms.

Conclusion: Patients in the embedded-BHC arm reported lower behavioral health risk at subsequent screenings, and the embedded-BHC intervention arm was more effective at facilitating treatment initiation for those adolescents needing specialty behavioral health services.




June 29th, 2017


August 10th, 2017