Discussion

In this large sample of PLWH seen for routine HIV primary care, 35% of patients had an indication for depression treatment at their most recent PHQ-9 assessment, which is consistent with previous estimates of depression in patients receiving HIV primary care in the US [1, 2, 4, 43]. Large gaps in treatment were present along the depression treatment cascade, confirming results from previous studies conducted in similar settings [13, 22, 43, 44]. Among patients with an indication for depression treatment in our total sample, just over half were receiving treatment and of those, one-third had achieved remission. Of patients receiving antidepressant treatment with an indication for treatment adjustment, only 9% received a dose escalation or medication change within 30 days.

The present study suggests that patients with depression and psychiatric comorbidities relating to illicit drug use or panic disorder are even less likely to receive and respond to antidepressant treatment relative to depressed PLWH who do not have these comorbidities.

Abstract

Little is known about disparities in depression prevalence, treatment, and remission by psychiatric comorbidities and substance use among persons living with HIV (PLWH). We conducted a cross-sectional analysis in a large cohort of PLWH in routine care and analyzed conditional probabilities of having an indication for depression treatment, receiving treatment, receiving indicated treatment adjustments, and achieving remission, stratified by alcohol use, illicit drug use, and panic symptoms. Overall, 34.7% (95% CI 33.9–35.5%) of participants had an indication for depression treatment and of these, 55.3% (53.8–56.8%) were receiving antidepressants. Among patients receiving antidepressants, 33.0% (31.1–34.9%) had evidence of remitted depression. In a subsample of sites with antidepressant dosage data, only 8.8% (6.7–11.5%) of patients received an indicated treatment adjustment. Current drug users (45.8%, 95% CI 43.6–48.1%) and patients reporting full symptoms of panic disorder (75.0%, 95% CI 72.9–77.1%) were most likely to have an indication for antidepressant treatment, least likely to receive treatment given an indication (current drug use: 47.6%, 95% CI 44.3–51.0%; full panic symptoms: 50.8%, 95% CI 48.0–53.6%), or have evidence of remitted depression when treated (22.3%, 95% CI 18.5–26.6%; and 7.3%, 95% CI 5.5–9.6%, respectively). In a multivariable model, drug use and panic symptoms were independently associated with poorer outcomes along the depression treatment cascade. Few differences were evident by alcohol use. Current drug users were most likely to have an indication for depression treatment, but were least likely to be receiving treatment or to have remitted depression. These same disparities were even more starkly evident among patients with co-occurring symptoms of panic disorder compared to those without. Achieving improvements in the depression treatment cascade will likely require attention to substance use and psychiatric comorbidities.

Results

Our primary analysis sample included 12,776 patients who completed at least one PRO. The sample was largely male (84%) and identified as white non-Hispanic (49%) (Table 1). The majority of patients were currently taking ART (89%), had a CD4 count > 500 (55%), and had an undetectable viral load (72%) within 6 months prior to their most recent PRO completion. Twenty-three percent of the sample had PHQ-9 scores ≥ 10 indicating probable major depression, and 19% were taking antidepressants at the time of PRO completion. Full panic symptoms were evident in 13% of patients, 17% were classified as having high-risk alcohol use, and 16% percent of patients reported current illicit drug use. Among patients with a PHQ-9 score ≥ 10 and/or taking antidepressants, 29% had evidence of full panic symptoms, 18% had high-risk alcohol use, and 21% reported current drug use (data not shown).

Panic disorder: 13.2%. - 13.6% “some panic symptoms”. Table 1

Across all sites, 34.7% (95% CI 33.9–35.5%) of patients had an indication for depression treatment, only 55.3% (53.8–56.8%) of whom were receiving antidepressant treatment. Of those receiving antidepressant treatment, only 33.0% (31.1–34.9%) had evidence of remitted depression based on PHQ-9 score (Table 2a). When restricted to the subsample of three sites with dosing data (n = 5484), a higher proportion of patients had an indication for depression treatment (40.8%, 95% CI 39.6–42.2%) and received antidepressant treatment (66.3%, 95% CI 64.3–68.2%), but the proportion of those receiving antidepressant treatment who had evidence of remitted depression was similar to the overall sample (34.1%, 95% CI 31.7–36.6%) (Table 2b). Of those receiving antidepressant treatment who had an indication for a treatment adjustment based on PHQ-9 score (37.5%, 95% CI 35.1–40.0%), only 8.8% (95% CI 6.7–11.5%) received a treatment adjustment within 30 days.