An examination of the impact of treatment on later risk for a substance use disorder in young people with major depressive disorder and bipolar disorder.

Objective: We examined the impact of treatment of young people with major depressive disorder (MDD) and bipolar disorder (BPD) on risk for subsequent substance use disorder (SUD). Methods: Patients aged 16-30 years seen at Mass General Brigham between 02/16/15 and 07/14/23 with MDD or BPD and no history of SUD were included. Diagnoses were determined using electronic health record data (billing codes, problem lists, patient-reported outcome measures, and other proxies of diagnosis(es)). Patients were treated or untreated based on pharmacological and/or psychosocial treatment following the onset of MDD/BPD and prior to SUD onset. Groups were matched on propensity scores (1:1) and compared on the development of SUD using a Cox regression model. Results: A total of 3601 and 796 patients (mean age 22.6 ± 4.3 years) were identified with no SUD and MDD or BPD, respectively. Final analysis of matched patients with and without treatment included 1666 with MDD and 314 with BPD. The most common treatment was pharmacologic only for MDD (89.3 %) and BPD (94.3 %). Overall, there was no significant difference in the development of SUD between those treated and not treated for MDD (HR=1.06 [0.84, 1.32], p = 0.63) nor those treated and not treated for BPD (HR=0.80 [0.49, 1.30], p = 0.37). However, longer duration of treatment for both MDD and BPD was associated with a significant reduction in risk for SUD. There was a 2.1 % risk reduction for each additional cumulative month of treatment for MDD (HR=0. 979 [0.976, 0.981], p < 0.001) and a 2.6 % risk reduction for each additional cumulative month of treatment for BPD (HR=0.974 [0.968, 0.979], p < 0.001). Conclusion: Longer treatment duration for young people with MDD and BPD is associated with decreased SUD risk.