How-To Guides

Referral to specialty psychiatry was uncommon overall: 131 of 757 veterans in the antidepressant monitoring collaborative care program were referred to a higher level of mental health service, which was 17.31% of the cohort. That means more than 82% remained managed within the primary care collaborative care model rather than needing transfer to specialty psychiatric care.

In the multivariable logistic regression model, referral to specialty psychiatric services was associated with higher baseline depression severity, sleep problems, alcohol use disorder, and cannabis use disorder. Specifically, baseline PHQ-9 score was associated with referral at OR=1.067 per point (95% CI, 1.01671.121), sleep issues at enrollment had OR=1.752 (95% CI, 1.15072.671), current or past alcohol use disorder had OR=1.702 (95% CI, 1.02472.829), and current or past cannabis use disorder had OR=2.070 (95% CI, 1.11673.840).

Baseline anxiety severity differed between referred and nonreferred veterans in the initial group comparison, where baseline GAD-7 scores were significantly different (P=.001), but it did not remain associated with referral after adjustment in the multivariable model. In this study, depressive symptom severity and selected comorbidities were more informative independent predictors of referral than baseline GAD-7 score alone.

Referral was driven primarily by inadequate symptomatic improvement. Of the 131 veterans referred to specialty mental health services, 121 (92.37%) still had at least moderate depression and/or anxiety symptoms at the time of referral. The remaining referrals were for specific clinical reasons such as patient preference for traditional face-to-face psychiatric care, residual PTSD symptoms, worsening objective presentation despite measured improvement, persistent sleep problems, or ongoing substance use problems.

Among the 131 veterans referred to specialty mental health services, 64 (48.85%) were referred to the general outpatient mental health clinic, 27 (20.61%) to the women6s center psychiatry clinic, 17 (12.98%) to the PTSD clinic, 14 (10.69%) to geriatric psychiatry, 5 (3.82%) to the substance use disorders clinic, and 5 (3.82%) to other psychiatrist-run specialty clinics such as traumatic brain injury or polytrauma clinics.

This study included 757 veterans with depression and/or anxiety disorders who were receiving treatment exclusively in primary care clinics within a VA antidepressant monitoring collaborative care program. The cohort had a mean age of 50.36 years (SD=15.34), 77.01% were male, and 53.24% were white. Baseline symptom burden was moderate on average, with mean PHQ-9 score 13.57 (SD=5.65) and mean GAD-7 score 12.85 (SD=5.38), and co-occurring conditions were common, including PTSD in 31.44%, alcohol use disorder in 27.74%, suicidal ideation in 15.97%, and sleep disturbance in 87.22%.

The program was a pharmacologic collaborative care model in primary care in which registered nurse care managers contacted veterans by telephone every 3 to 4 weeks and collected symptom, adherence, and tolerability data. Depression and anxiety symptoms were tracked using the PHQ-9 and GAD-7, and consulting psychiatrists reviewed each new case and each patient contact to guide treatment decisions in collaboration with the patient6s primary care physician. Patients were followed for a goal minimum of 6 months, with longer follow-up possible if medication changes or other clinical needs arose.

No. The authors state that referral to a higher level of service should not be considered a failure of the collaborative care model, because the program operates within a stepped-care framework in which treatment begins at a lower level of service and is escalated when needed. In that context, referral represents movement to a more appropriate intensity of care rather than failure of care.

The main limitations were its retrospective design, reliance on the accuracy of the electronic health record, and the difficulty of controlling for bias and confounding in retrospective research. The authors also note the potential for missing data, that this was a secondary analysis, and that the cohort consisted entirely of veterans, which may limit generalizability. At the same time, they note strengths including the large sample size, longitudinal monitoring, and pragmatic clinical design.

The findings suggest that veterans entering collaborative care with higher baseline depressive symptoms, sleep problems, alcohol use disorder, or cannabis use disorder may be more likely to require later referral to specialty psychiatric services. The authors propose that such baseline factors could potentially inform future treatment flow algorithms to guide initial placement in collaborative care versus specialty psychiatry, but they also state that prospective studies are needed to examine these associations further.