Psilocybin Therapy for Clinicians With Symptoms of Depression From Frontline Care During the COVID-19 Pandemic

Frontline physicians, advanced practice practitioners, and nurses experienced substantial psychological morbidity during the COVID-19 pandemic, including burnout, depression, and posttraumatic stress disorder (PTSD). At pandemic peaks these clinicians faced prolonged exposure to critically ill patients, high mortality, extended work hours, fear for personal and family safety, isolation, and politically motivated hostility; these stressors contributed to a pandemic-related syndrome with features overlapping first-responder trauma, burnout, and depression. Earlier clinical studies reported that psilocybin administered with psychological support produced reductions in depression and anxiety in populations such as people with major depressive disorder, treatment-resistant depression, or cancer-related distress, but it remained unclear whether psilocybin therapy could help clinicians whose symptoms emerged specifically from pandemic frontline work. Back and colleagues therefore designed a randomized clinical trial to test whether a single episode of psilocybin therapy (25 mg, given with preparation and integration sessions) would improve symptoms of depression, burnout, and PTSD in US clinicians who developed moderate to severe symptoms after providing frontline care during the pandemic. The trial compared psilocybin plus psychological support with an active placebo (niacin, 100 mg) plus the same therapeutic milieu, with change in Montgomery–Åsberg Depression Rating Scale (MADRS) scores from baseline to day 28 specified as the primary outcome.

This was a double-blind, randomized clinical trial conducted at the University of Washington between February and December 2022, approved by the institutional review board and conducted according to CONSORT guidelines. Eligible participants were physicians, APPs, or nurses who had provided more than 1 month of direct frontline COVID-19 care, endorsed at least 2 of 4 high-exposure items on a COVID-19 occupational exposure index, reported no prepandemic mental health diagnosis (self-reported), and had persistent moderate to severe depressive symptoms (participant-scored MADRS ≥ 21) for at least 6 months despite at least one prior trial of medication and/or therapy. Key exclusions included a first-degree family history of schizophrenia, bipolar or paranoid disorder, current substance use disorder, recent psychedelic use (within 12 months), pregnancy, and unstable medical conditions. Recruitment used a study website and newsletters; screening combined web surveys, telephone screening, and in-person medical and psychiatric assessment including ECG and labs. Randomisation was performed by a faculty member unaffiliated with the study using block sizes of 6; an investigational pharmacist dispensed numbered envelopes and both participants and study staff (including facilitators and raters) were blinded. The intervention episode comprised two 60–90 minute preparation sessions (day −8 and day −1), a medication session (day 0) of approximately 7 hours, and three integration sessions (days 1, 8, and 15 reported in some places, with follow-up at days 28, 56, 84, and 180). Facilitators were interprofessional and used a facilitator manual with specified elements for preparation, medication, and integration. Sessions combined secular ceremonial elements, a curated music programme matched to the pharmacologic trajectory, encouragement of eyeshades, and explicit guidance on touch and support. The experimental medication was a fixed oral psilocybin dose of 25 mg; the control was oral niacin 100 mg, both provided in identical capsules by Usona Institute. Participants were intentionally not given a diagnostic label in records; the study obtained a certificate of confidentiality. The prespecified primary outcome was the change in MADRS score (blinded clinician-administered) from baseline (preparation 1) to day 28. Secondary outcomes, in a hierarchical order, were change in burnout measured by the Stanford Professional Fulfillment Index (SPFI) burnout subscale and change in PTSD symptoms measured by the PCL-5. The SPFI burnout subscale assesses work exhaustion and interpersonal disengagement; the PCL-5 is a 0–80 symptom checklist for PTSD. The Mystical Experience Questionnaire (MEQ-30) and adverse events including suicidal ideation (Columbia Suicide Severity Rating Scale) were collected at the medication session. The trial was powered a priori for the primary outcome assuming an effect size of 1.06 SD; with 15 participants per arm the study had 80% power at a two-sided α = .05. Analyses followed the intention-to-treat principle, used two-sample t tests for mean change comparisons after testing assumptions, applied the prespecified hierarchical testing sequence to control type I error, and evaluated the correlation between MEQ-30 and MADRS change with linear regression (Pearson coefficient). Participants were unblinded after completing day 28 questionnaires; those assigned to niacin were offered open-label psilocybin with the same protocol and follow-up.

Thirty US clinicians were enrolled and randomised, 15 to psilocybin and 15 to niacin; the sample was 50% female and 50% male with a mean age of 38 years (range 29–60). All participants completed the intervention sessions. For the primary outcome, the mean (SD) change in MADRS score from preparation 1 to day 28 was −21.33 (7.84) in the psilocybin arm and −9.33 (7.32) in the niacin arm, yielding a mean difference in change scores of −12.00 (95% CI, −17.67 to −6.33; P < .001), favouring psilocybin. The reduction in depressive symptoms in the psilocybin arm was sustained through month 6 (mean decrease −24.00; 95% CI, −26.87 to −21.13). For context, the authors note that a 6–9 point MADRS change is typically considered clinically meaningful and that the minimum clinically important difference is around 1.6–1.9 points. For the first secondary outcome, burnout measured by the SPFI burnout subscale decreased more in the psilocybin arm than the niacin arm (mean [SD] change −6.40 [5.00] vs −2.33 [5.97]; reported P = .05), though the authors state this did not meet their prespecified significance threshold for continuing hierarchical testing. Because that secondary outcome did not reach the prespecified significance, the PCL-5 PTSD outcome was reported descriptively: mean change in PCL-5 was −16.67 (15.04) in the psilocybin arm versus −6.73 (10.69) in the niacin arm, a numerically larger decrease in PTSD symptoms that was not formally hypothesis-tested under the hierarchical plan. Immediate post-session MEQ-30 scores indicated a much deeper subjective medication experience after psilocybin than niacin (mean 129.40 [range 88–119] vs 15.07 [0–52]), and higher MEQ-30 scores correlated modestly with greater MADRS improvement (R = 0.701). Safety outcomes showed no serious adverse events. Common acute adverse events on the day of psilocybin included mild nausea (4 participants, 27%), mild headache (reported as 27% in the text), mild tachycardia (2 participants, 13%), and transient hypertension (mild in 6 [40%], moderate in 8 [53%], severe in 1 [7%]); these cardiovascular changes resolved within 20 minutes without medical intervention. There were no psychosis episodes or attempts to leave the treatment room. One participant experienced transient passive death wishes on day 28 after open-label psilocybin but reported no intent or plan and symptoms did not recur. Functional unblinding was evident: immediately after ingesting the capsule 100% of participants reported not knowing which treatment they had received, but by the end of the medication session 100% reported knowing and were correct. Illustrative participant comments from preparatory and integration sessions described feelings of disposability and guilt and, after the intervention, increased openness to support, perspective, and realignment of professional meaning; examples included “I feel more disposable than a used COVID-19 swab” and “I'm reaching out and leaning on others…and trusting that the people around me can support me.” Of the 15 participants initially randomised to niacin, 12 (80%) later received open-label psilocybin; in those open-label episodes the mean baseline MADRS was 20.17 (10.11) and the mean MADRS change from preparation 1 to day 28 was −12.83 (95% CI, −18.29 to −7.38). During follow-up 21 participants (70%) reported substantial changes in work role or full-time equivalent status, 8 (27%) remained in the same position, and 1 (3%) remained unemployed; no participants left health care.

Back and colleagues interpret their findings as evidence that a single episode of psilocybin therapy administered with structured preparation and integration can produce large, rapid, and sustained reductions in depressive symptoms among clinicians who developed moderate to severe depression while providing frontline COVID-19 care. The authors propose a biological rationale—psilocybin is a partial 5-HT2A receptor agonist that facilitates neuroplasticity—together with a therapeutic rationale that the structured psychotherapeutic context enabled participants to process complex pandemic-related emotions. They emphasise that both arms received the same counselling and support, suggesting that the pharmacologic component substantially augmented the therapeutic effect. Compared with prior trials of psilocybin in major depressive disorder and treatment-resistant depression, the present sample differed in that participants generally had no prepandemic mental health diagnoses; nonetheless, most had previously tried counselling and more than half had used antidepressants. The authors highlight that the observed 21-point mean MADRS decrease exceeds typical thresholds for clinical meaningfulness and is larger than many prior reported effects. Although the SPFI burnout outcome did not reach the prespecified significance threshold, the authors note that the trial was small and possibly underpowered for burnout; similarly, the 16-point PCL-5 decrease in PTSD symptoms exceeded commonly used thresholds for clinical meaningfulness (10 points) but was not formally tested under the hierarchical analysis. The investigators acknowledge several limitations. Chief among them is the small sample size and consequent limits on generalisability. Recruitment interest greatly exceeded enrolment (2,247 expressing interest vs 30 enrolled), raising the possibility of selection biases despite randomisation at each recruitment step. Functional unblinding—participants and facilitators could accurately distinguish psilocybin from niacin by the end of the medication session—was a limitation inherent to psychedelic trials and complicates interpretation of expectancy effects. The trial also did not include a psilocybin arm without therapy, so the independent contribution of the psychological support cannot be separated from the pharmacologic effect. Finally, the authors report that participants were unblinded after day 28 and many in the control group received open-label psilocybin, which affects longer-term between-group comparisons. Despite these caveats, the study team concludes that psilocybin therapy was associated with substantial improvements in depressive symptoms in this specific clinician population and that the intervention was generally well tolerated.