Same Presentation, Different Referrals: Race, Beliefs, and Other Drivers of Lcsw's Antidepressant Referral Decisions

Background and Purpose: Social workers provide a large share of mental health services in the United States and are well-positioned to respond to the growing demand for care. Depression is among the most commonly diagnosed and treated conditions in mental health practice, and social workers frequently initiate client referrals to psychiatrists for medication consultation. While referral decisions constitute a routine professional role, the explicit and implicit factors that shape social workers' referral decisions remain poorly understood. Aligned with the grand challenges' emphasis on closing health gaps and ensuring equitable access to treatment, this study employed a rigorous hybrid experimental and survey methodology to isolate factors that influence social workers’ antidepressant referral decisions.

Methods: The factorial survey method was used to present participants with three randomly generated clinical vignettes depicting a depressed adult client. Participants were asked if they would refer the fictive client for antidepressant consultation using a binary outcome variable. Twelve vignette case characteristics that each had multiple levels were randomized to create the vignette universe of ~1,68 million unique scenarios. Licensed clinical social workers (LCSWs) were recruited across eleven states, with the analytic sample comprising ~1,390 vignette judgments nested within 466 practitioners who completed a clinician questionnaire (81% female, 80% White non-Hispanic, 66% licensed 16+ years, 59% private practice). Binary multilevel logistic regression with random intercepts was estimated using cluster-robust standard errors, with a parsimonious final model encompassing client case features, practitioner demographics, beliefs about depression, and personal mental health experiences.

Results: Overall, the antidepressant referral rate was 65.3%. Three pathways to referral emerged. First, a case-driven pathway: the five strongest predictors of increased odds of antidepressant referral were prior positive antidepressant response (OR=3.68), severe symptoms (OR=3.37), chronic/recurrent history (OR=3.29), suicidal ideation (OR=3.18), and moderate symptoms (OR=1.93), all p ≤ .001. Prior poor response to antidepressants reduced referral odds (OR=0.63, p =.001). Second, a practitioner belief-driven pathway: beliefs that antidepressants produce rapid improvement (OR=1.49), depression has a biomedical cause (OR=1.41), and that antidepressants are a necessary treatment (OR=1.23) each increased referral odds (all p≤.001). Practitioner belief in psychosocial causes reduced odds (OR=0.83, p =.003). Third, an experience-driven pathway: LCSWs with personal psychotherapy history were less likely to refer (OR=0.46, p<.001) and those with a personal mental health history were more likely to refer (OR=1.50, p=.027). Notably, across all models, vignettes depicting a Black client were associated with lower referral odds compared to White client vignettes (OR=0.763–0.769), all else the same, reaching significance in the model with only clinician-level variables (OR=0.764, p=.039).

Conclusions and Implications: While LCSWs demonstrated evidence-informed and case-driven judgments, their referral decisions were also shaped by judgments prone to biases such as beliefs and personal experiences. Crucially, the consistent directional evidence of lower referral rates for Black clients in the vignettes was concerning and requires further evaluation with adequate power to consistently detect this effect. Overall, these findings suggest that enhancing social workers’ reflective practices through de-biasing interventions, training, and education may bolster evidence-based practices for the treatment of depression and antidepressant referral decisions.