Psilocybin with therapy for hard-to-treat depression: a small clinical trial

The study was an open-label, single-arm, pilot trial of two 25 mg doses of psilocybin with psychotherapeutic support, conducted by a co-therapist dyad. All on-site study activities were conducted at Swinburne University.

All data were collected in accordance with approvals obtained from The Swinburne University Human Research Ethics Committee (#20231367-15628). The trial was conducted under the Clinical Trial Notification scheme (CT-2020-CTN-02260-1) and registered on the Australian New Zealand Clinical Trials Registry (ACTRN12621001097831). The investigational product was provided by Usona Institute (WI, USA). This manuscript was prepared in accordance with the TREND (Transparent Reporting of Evaluations with Nonrandomized Designs) statement. A completed TREND checklist is provided in the Supplemental Materials.

This study aimed to recruit 15 participants through study advertisements, clinical trial registries, and direct referrals from health practitioners.

Inclusion criteria included age between 18 and 65, and moderate to severe Major Depressive Disorder (MDD) – Montgomery–Åsberg Depression Rating Scale (MADRS) score ⩾ 30, deemed treatment-resistant (defined as having failed to respond to two or more antidepressant medications in the current episode). Medical exclusion criteria included conditions and use of medications contraindicated for psilocybin administration, pregnancy or breastfeeding, and past 12-month use of psychedelic substances or 3,4-Methylenedioxymethamphetamine (MDMA) (or past 1-month for microdosing). Psychiatric exclusion criteria included very severe anxiety, depression or suicidality; current or past history of schizophrenia, other Psychotic Disorders or Bipolar I or II Disorder, or a first-degree relative with these conditions; current or past 5 year history with alcohol or drug dependence; current Dissociative Disorder, Anorexia Nervosa, Bulimia Nervosa, or any psychiatric condition judged by screening clinicians to be incompatible with the establishment of rapport or safe exposure to psilocybin; situational or personal factors that may interfere with involvement; or a 'complex case' of depression (e.g. extensive trauma history or complex psychiatric or medical comorbidities, or comorbidities where depression appears to be secondary).

Eligibility criteria were assessed in a four-phase screening process: (1) online pre-screening survey, (2) screening interview video-call, (3) medical exam, (4) consultation with treating healthcare professional. Following successful completion of all screening procedures, participants tapered off contraindicated medications under the supervision of their usual care team and were enrolled only after cessation of these agents.

The intervention had three distinct phases:

The second dosing session was optional but encouraged unless contraindicated. All sessions were conducted by a co-therapy dyad. Therapists were qualified mental healthcare workers who underwent training designed specifically for this trial. The treatment protocol was not manualised, instead basic treatment guides were provided to therapists. The intervention was informed by contemporary practices in psychedelic therapy rather than a specific psychological framework. See Supplemental Methods for comprehensive information on trial screening and intervention.

The primary outcome was change in depression scores from baseline to 3 weeks post-dose 2, measured using the self-rated Quick Inventory of Depressive Symptomatology (QIDS-SR)⁴³; Secondary health and well-being outcomes included the World Health Organization Quality of Life Questionnaire – Brief Version (WHOQoL-Bref)⁴⁴; Generalized Anxiety Disorder 7-item scale (GAD-7)⁴⁵; Brief Experiential Avoidance Questionnaire (BEAQ)⁴⁶; two subscales from the Revised Self-Efficacy Scale (RSES)⁴⁷; Watts Connectedness Scale (WCS)⁴⁸ and the Persisting Effects Questionnaire (PEQ).⁴⁹

Participants' treatment expectations were assessed with the Patients' Therapy Expectation and Evaluation questionnaire (PATHEV),⁵⁰ containing three subscales (hope of improvement, fear of change, and suitability), administered at baseline and 1 day prior to dose 1. The Mindset questionnaire⁴⁰ was used to assess participants' preparedness, encompassing comfort in the treatment context, rapport with therapists, and intentions, administered on the morning of each dosing session. The 11 Dimension Altered States of Consciousness scale (11D-ASC),⁵¹ the Psychological Insight Questionnaire (PIQ)⁵² and Emotional Breakthrough Inventory (EBI)³³ were administered the day following each dose.

Participants were asked about their experience of any adverse events (AEs) at all study visits. Blood pressure and heart rate were measured at minimum set points throughout each dosing session. A blood test was conducted at baseline and 1-day post-dose 1 to check participant health. At long-term follow-up, the Hallucinogen Persisting Perceptual Disorder (HPPD) Questionnaire⁵³ was administered. Consent to, use of and participants' experience of therapeutic touch were documented.

One-hour qualitative interviews were conducted via video call at baseline, 3 weeks post-dose 2 and long-term follow-up. For a full list of trial assessments, refer to the Supplemental Methods.

Statistical analyses were performed using R Studio version 2024.4.1.748.⁵⁴ Due to the small sample size, exact paired permutation tests were used to assess the effect of treatment on QIDS-SR scores and secondary outcome measures at the primary endpoint (3-week post-dose 2). Permutation tests offer robust results by relying on data-driven resampling to assess significance without making assumptions about the distribution of data.⁵⁵ Effect sizes were calculated with Hedge's g formula with 95% confidence intervals.

For comparison with previous research, the number of participants showing clinically-meaningful treatment response, defined as a reduction of 50% or more from baseline on QIDS-SR score,⁵⁶ and remission, defined as a QIDS-SR score of 5 or less^57,58 was calculated.

Upon data inspection, participants were grouped descriptively into different treatment trajectory categories based on their clinical outcomes throughout the study. Multinomial logistic regressions were run to test possible predictors of treatment trajectory: PATHEV subscales, 11D-ASC subscales, PIQ, EBI, and Mindset. Reference category was rotated to capture a fuller picture of the relationships among all outcome categories.

Due to the small sample size, analyses of secondary and exploratory outcomes were conducted using a descriptive approach, focusing on effect size estimation and confidence intervals rather than significance testing. Results with confidence intervals that did not cross zero (for Hedges' g) or 1 (for odds ratios) were prioritised, and p-values are reported in the supplementary results for completeness. Quotes from qualitative interviews were selectively extracted to contextualise and complement quantitative findings.

Recruitment commenced in February 2022, with enrolment spanning August 2022 to August 2023. From an initial 209 complete pre-screening survey responses, eight participants were enrolled in the study (Figure 2). The most common reasons for exclusion at pre-screening were having taken a psychedelic drug in the previous 12 months (28.41%) and/or not meeting criteria for TRD (26.14%); and at interview, MADRS score < 30 (52.63%), and/or 'complex case' depression (34.21%). Of the eight enrolled; one was withdrawn prior to starting the treatment after discovery of factors leading to a 'complex case' of depression, determined to pose a safety concern (Table S1↗). The remaining seven participants completed the full treatment protocol and all outcome assessments reported here, and were included in all analyses. Due to delays related to the COVID-19 pandemic, and time and funding constraints, recruitment was closed prior to reaching the planned sample size. Final follow-up assessments were completed in January 2024.

The majority of potential and enrolled participants found out about the study from australianclinicaltrials.gov.au (49.03% and 75%, respectively) or a healthcare professional (18.93% and 25%, respectively; Table S2↗). Full demographic information for completing participants is displayed in Table 1.

All participants displayed a reduction in QIDS-SR scores from baseline to the primary endpoint (3 weeks following dose two; Figure 3), with a significant mean reduction of 7.14 points (SD = 4.41; p = 0.02; Hedges' g = –1.27, 95% CI [–2.40, –0.37]), and average decrease from baseline of 47.96%. Treatment response and remission was observed in three of seven participants at the primary endpoint (ID 3, 4 and 6). At the long-term follow-up (20-week post-dose 2), aggregated reduction in QIDS-SR scores was statistically significant (mean reduction of 7.14 points; SD = 5.84; p = 0.02; Hedge's g = –1.16; 95% CI (–2.77, –0.46)). However, three participants resumed or started a new antidepressant medication prior to follow-up (ID 1, 2 and 5).

In terms of individual response trajectories, we determined three categories:

Mean improvements at the primary endpoint compared to baseline were observed in all secondary mental health and well-being outcomes (Table S3 and Figure S1↗), with all confidence intervals excluding zero, except for anxiety scores. The Persisting Effects Questionnaire scores indicated largely positive lasting changes to attitudes, mood and behaviour. Both QIDS scores and secondary health and well-being outcomes appeared to differ by treatment trajectory (Tables S4 and S5↗).

Multinomial logistic regression modelling showed that Mindset and some elements of the 11D-ASC reliably predicted treatment trajectory. Higher Spiritual Experience scores decreased the odds of relapsing and non-response compared to sustained response. Higher Changed Meaning of Percepts scores increased the odds of sustained and relapsing responses compared to non-response. Higher Mindset scores decreased the odds of non-response compared to both sustained and relapsing responses and increased the odds of sustained and relapsing responses compared to non-response. These models displayed good fit metrics (Table 2 and Figure 4). A combined multinomial logistic regression model was not run due to small sample size and multicollinearity between predictors. All other predictors displayed less certain effects (Table S6↗).

Inspection of qualitative data validated the categorisation of participants into treatment trajectory groups and revealed variations in acute subjective dosing experiences (Table 3).

All participants received both doses of psilocybin. No serious adverse events occurred at any point during the trial (Table 4). See Tables S7 and S8↗ for full lists of adverse events.

One participant (ID 2) reported anxiety, insomnia and suicidal ideation following a very emotionally challenging first dose. Daily phone calls for 1 week from therapists were used to monitor well-being and provide support. This participant also received support from external mental healthcare providers and commenced psychiatric medications. These adverse events resolved by long-term follow-up, except for anxiety, which persisted and required ongoing treatment. A second participant (ID 5) received an additional integration session conducted via video call between doses, due to well-being concerns.

All participants consented to both participant- and therapist-initiated therapeutic touch, and five of the seven participants received touch. No changes to therapeutic touch consent were made between dose 1 and dose 2. All participants receiving touch reported it to be helpful.

While response and remission rates at the primary endpoint facilitate comparison with prior studies, categorising participants by overall treatment trajectory provided a fuller picture of response patterns. Two participants showed a sustained response, three experienced relapse after initial improvement and two had no substantial symptom reduction. These distinct patterns were linked to lasting changes in attitude, mood and behaviour, highlighting the importance of understanding patterns of individual responses.

Mindset prior to dosing and aspects of the acute psychedelic experience predicted treatment trajectory, with the latter supported by qualitative data. Our findings suggest that psychological readiness, comfort, rapport and clear intention before dosing (mindset), combined with emotionally intensified, novel or meaningful perceptual changes during dosing, are associated with an initial positive response to psilocybin therapy for TRD. In contrast, a sustained response appears linked to profound spiritual experiences during dosing.

This link between spiritual experience and benefit aligns with previous research,⁶³ although other dimensions related to the mystical experience, such as unity of experience or insightfulness, did not reliably predict treatment outcomes in this study. Changed Meaning of Percepts emerged as a novel predictor, supporting the view that therapeutic effects of psychedelics may stem from emotional insights or perspective shifts that come from processing experiences in new ways. The link between Mindset and clinical benefit may occur through these or other changes to acute subjective experience.^39,40 Subject to confirmation in future research, this finding supports screening of participants prior to dosing to ascertain their level of preparedness and providing additional preparation when indicated.

Treatment expectations did not emerge as a consistent predictor of treatment trajectory. However, positive expectations have been associated with improved outcomes in psychotherapy⁶⁴ and with conventional antidepressant treatments.⁶⁵ In this study, the lack of association may reflect limited variability in expectancy scores, with some subscales showing consistently high ratings across participants. While a prior study⁶⁶ has also reported expectations lacked predictive value, a non-standardised, single-question measure was used. Further research is needed to clarify the relationship between expectations and treatment outcomes and to identify or develop tools capable of capturing meaningful variability in expectancy within this context. All interpretations from the current study regarding exploratory predictors should be considered tentative due to the small sample size.

Documented adverse events were largely congruent with previous research. Additional support outside of pre-scheduled trial treatment sessions was required for two participants, highlighting the need for flexibility in trial design and the ability to respond to individual participant needs. Transient suicidal ideation recorded here adds to the emerging understanding of suicidality as a risk in psychedelic trials.^16,19 Furthermore, persisting anxiety reported by one participant underscores the importance of addressing the potential for long-term adverse psychological reactions following psilocybin therapy in future studies and the need for more consistent reporting on the duration and resolution of adverse events.

This trial monitored AEs in a similar manner to previous trials – using open-ended questions rather than structured questionnaires, and not documenting expected acute effects of psilocybin (e.g. changes in visual perception) as AEs; these methods likely result in underreporting of AEs.⁶⁷ Furthermore, it has been argued that standard monitoring methods may not adequately capture risks specific to psychedelic therapy, such as sociocultural and interpersonal events.^27,68,69 While this trial used additional questionnaires to assess interpersonal risks related to therapeutic touch, finding no adverse events, future research would benefit from the use of standardised tools such as those recently developed to assess psychedelic-specific harms.^68,70

These results highlight challenges with antidepressant discontinuation, including withdrawal-related AEs and increased QIDS-SR scores from screening to baseline, likely due to symptom exacerbation during withdrawal, which may inflate baseline scores.⁷¹ This is important to consider when interpreting symptom change across the trial, as some apparent improvement may reflect the natural resolution of withdrawal effects over time rather than true treatment-related change. While a longer washout period might mitigate this, it is often not feasible due to individuals' reluctance to remain without treatment. Discontinuation is recommended due to safety concerns and potential dampening of psychedelic effects,^72,73 however, preliminary findings suggest acute subjective effects,^23,74 safety and therapeutic outcomes²³ may remain intact when antidepressants are continued during psilocybin therapy, challenging the necessity of this protocol.

Beyond the small sample size, reliance on a single self-report depression measure and self-referral recruitment, the study's conclusions are limited by the open-label design. Double-blind randomised controlled trials are the gold standard for efficacy, though psychedelic therapy complicates this design due to functional unblinding. Active placebos and incomplete disclosure have been suggested to address this, though questions remain regarding their effectiveness.⁷⁵ Combining and modifying these methods in novel trial designs may help reduce participant certainty about treatment allocation.^38,76

A further important limitation of this study is the extensive, highly selective exclusion criteria. Individuals presenting with a variety of comorbid psychiatric conditions, complex clinical profiles or traits considered likely to interfere with the development of therapeutic rapport were excluded from participation. Although this approach is consistent with prior psychedelic research protocols,¹⁴ it limits the extent to which the findings can be generalised to broader clinical populations, where such complexities are frequently encountered.