Correlates of depression outcomes in collaborative care for HIV

Collaborative care can treat depression in HIV but existing studies have been limited by excluding patients with acute or severe depression. The purpose of this analysis is to determine if real-world implementation of collaborative care in HIV is associated with improvement in depression, and to identify correlates of depression outcomes.

Collaborative care was implemented as part of a large practice transformation initiative. Change in depression, measured by PHQ-9 score, at baseline compared to 12 months post-enrollment was the outcome, which was operationalized as remission, response, and neither response nor remission. Bivariate and multivariate associations were assessed between several variables at baseline and the outcome.

Out of 416, 99 (23.79%) patients remitted and 89 (21.39%) responded (without remission). In the bivariate analysis having a documented psychiatric comorbidity was associated with low remission [31 (16.58%)]; p = 0.008. Having generalized anxiety disorder was associated with low remission [18 (15.00%)] and response rates [26 (21.67%)]; p = 0.022. Having a substance use disorder (alcohol, cocaine, or amphetamine) - was associated with poor remission [29 (16.67%)] and response [33 (18.97%)]; p = 0.004. Social isolation was correlated with lower response and remission rates (p = 0.0022). In the multivariate analysis older age was associated with higher remission rates (OR: 1.10; 95% CI: 1.005-1.194) whereas being a Medicaid beneficiary (OR: 0.652; 95% CI: 1.123-2.797), having comorbid generalized anxiety disorder (OR: 0.267; 95% CI: 0.122-0.584) or a stimulant use disorder (cocaine [OR: 0.413; 95% CI: 0.222-0.926] or amphetamines [OR: 0.185; 95% CI: 0.037-0.766]), was associated with lower remission rates.

We found that depression improved in our study subjects. We identified several modifiable correlates of depression outcomes.