Psychedelic-Assisted Group Therapy: A Systematic Review

Psychedelic substances have a long history of group use in ceremonial and clinical contexts, and some landmark early studies of psilocybin were conducted in group settings. Trope and colleagues note that, despite this precedent and recent renewed scientific interest in classic psychedelics, contemporary clinical trials have largely used individual psychotherapy models and no published 21st-century clinical trial has tested a group therapy approach. Prior qualitative and quantitative findings from modern trials have suggested that social connectedness and interpersonal processes can be important mechanisms of change following psychedelic sessions, and many participants in individual trials request contact with other participants for support and integration. To address the gap in the literature, the investigators carried out a systematic review of English- and Spanish-language publications from 1900 to 2018 that described a group element during preparation, drug administration, or integration in psychedelic-assisted psychotherapy. The review excluded nonclinical ritual uses (for example, ayahuasca ceremonies) but included MDMA only when it was used alongside a classic psychedelic. The stated aim was to characterise published studies by clinical approach, experimental method, and outcomes, and to identify hypotheses and design considerations for future research on psychedelic-assisted group therapy and interpersonal functioning.

The review followed PRISMA guidelines. Searches were conducted in PsycINFO and MEDLINE in November 2018 using terms for classic psychedelics (for example LSD, psilocybin, mescaline) combined with terms for group therapy or group administration. Supplementary searches included the Erowid Reference Vaults and the MAPS Psychedelic Bibliography, plus a specialist bibliography of psycholytic and psychedelic therapy research. No date limits were applied. The investigators identified 506 records (492 from databases, 14 from hand searches). After removing 109 duplicates, 397 unique records remained. Screening of titles and abstracts excluded 276 records, and a further 45 were excluded because the team could not obtain abstracts or full texts in English or Spanish. Seventy-six full texts were assessed for eligibility and 12 studies met inclusion criteria. Eligible articles were required to describe group methods, provide demographic and diagnostic information, and include quantified outcome data. The authors note that all reported doses are oral unless otherwise specified, and that a flow figure summarising the selection process is referenced but not reproduced in the extracted text.

Twelve studies met inclusion criteria; LSD was the most commonly used drug, typically at 50-400mcg p.o., with occasional higher doses or combinations with stimulants, and psilocybin appeared in two studies. Study designs were heterogeneous and included case series, uncontrolled open-label cohorts, and a small number of controlled trials; sample sizes ranged from single-digit cohorts up to reports of 166 participants in some clinical summaries, and the discussion notes randomized, active-placebo controlled trials with sample sizes of 30 and 99. A set of studies focused on alcohol use disorders. One case series combined individual and group LSD-assisted psychotherapy in an inpatient alcohol programme (typical dosages 200-400mcg) and reported mixed outcomes across 40 participants plus seven additional cases: 19 were rated "much improved" (interpreted as abstinent), 11 "moderately improved," and 17 "unchanged." A controlled study with 138 inpatients randomised to (a) full programme including a single individual 200mcg LSD session plus group therapy and AA (n=58), (b) group therapy and AA without LSD (n=35), or (c) treatment-as-usual (n=45) reported at 6–18 month follow-up that 34 in the LSD arm remained sober and 7 were improved, compared with 4 sober and 4 improved in the non-LSD group arm and 7 sober and 3 improved in controls; the increase in sobriety or improvement for the LSD condition was reported as statistically significant versus the other two arms. Another controlled, open-label inpatient study (n=28 in the LSD cohort versus n=34 historical controls) that delivered two LSD-assisted group sessions per participant reported abstinence at 3 months in 17/28 (60.7%) of the LSD group versus 15/34 (44.1%) of controls, and at 12 months 8/28 (28.6%) versus 7/34 (20.6%). By contrast, a single-blind, three-arm trial with 30 participants compared a high-dose individual LSD session (800mcg), an active placebo (ephedrine 60mg), and no drug; at 6 months there were no statistically significant differences in abstinence rates or drinking episodes, with average gains reported as 33.7% (LSD), 31.5% (ephedrine), and 19.6% (control). A separate controlled trial comparing high-dose individual LSD sessions plus the Human Relations Training Laboratory (HRTL) group treatment against HRTL alone and HRTL with a low-dose LSD active placebo (25mcg) in a 60-day inpatient VA programme (71 participants received 500mcg LSD in the active arm) found no significant differences between arms on a clinician-rated functioning measure at 1-year follow-up. Seven uncontrolled studies examined heterogeneous clinical populations described as "neurotic" or with various affective and character disorders. These reports typically used weekly or more frequent group sessions, sometimes with repeated dosing, and generally reported that most participants showed some clinical improvement on the authors' measures or ratings; however, the reporting was descriptive and diagnostic and outcome metrics were inconsistent or unvalidated. One report of a partial-hospitalisation cohort (n=75) noted 10 patients "much improved," 20 "improved," 7 "slightly improved," and 20 "not improved," and mentioned two completed suicides in the sample, neither occurring under acute LSD effect. Another cohort summary described 168 patients who received monthly LSD-assisted group sessions as an adjunct to weekly psychoanalytic group therapy; of 166 with outcome data, 93 were rated "cured" or "much improved." A small experiment with 39 participants compared various small-dose arms and an "ex-patient" arm and suggested the best outcomes were in those who had previously had individual LSD sessions and then engaged in the therapeutic community. Gasser (1995) summarised work from three clinicians who provided MDMA and LSD in group formats to 171 patients; follow-up surveys (n=121 responders) indicated that over 90% described good or slight improvement and only one patient reported persistent depression appearing 3 months after treatment. Across all included studies, the authors estimate approximately 700 participants were administered psychedelics in group contexts or with group support. The extracted text indicates inconsistent reporting of adverse events across studies, and no study reported prolonged psychosis, suicide under acute drug effect, or other serious adverse events directly attributable to drug administration, although systematic safety assessment was generally lacking. Overall, the reviewed studies used a wide range of group formats (nondirective group dosing, group preparation/integration, and individual high-dose sessions embedded within group-based programmes), varied dosing strategies, and diverse outcome reporting that preclude pooling of effects in a straightforward manner.

Trope and colleagues characterise the assembled literature as methodologically heterogeneous and limited by the standards of modern clinical research. They highlight several recurring limitations: absence of adequate control groups, lack of blinding, inconsistent diagnostic criteria and treatment protocols, unvalidated or absent outcome measures, and weak or absent statistical analyses. The review was also restricted to English- and Spanish-language publications, though the authors note they examined English translations or publications from Nordic, Dutch and German teams when available. The authors draw three main observations. First, on safety, they state that the available reports do not permit firm conclusions because adverse-event assessment and reporting were inconsistent; nevertheless, no studies reported prolonged psychosis, suicides during acute drug effect, or other serious adverse events directly attributable to psychedelic administration. Most studies excluded chronically psychotic or medically ill patients, but many included treatment-resistant or comorbid cases, which the authors suggest indicates that psychedelics can be administered with safeguards either via individual dosing plus group support or via group dosing supplemented by non-drug psychotherapy. The authors note that many group-dosing protocols used lower doses or required prior individual dosing experience. Second, regarding alcohol use disorders, four of five controlled studies were conducted in inpatient alcohol programmes. Two of these trials reported statistically significant improvements in sobriety and social functioning in their active arms, while two reported null findings. The authors did not identify a clear pattern relating outcome to drug dose, group therapy dose, or individual versus group administration. They observe, however, that the positive studies and the single uncontrolled positive alcoholism study used either a 12-Step model or group therapy specifically adapted for psychedelic administration, whereas the null studies did not incorporate these elements. Third, for heterogeneous "neurotic" populations, the seven uncontrolled studies generally reported clinical improvement in a majority of participants, but the sole controlled study among them suggested superior outcomes for participants who received higher-dose individual LSD sessions (100mcg s.c.) compared with those who received lower-dose (50mcg s.c.) group LSD sessions; this advantage persisted at 1-year follow-up. Looking forward, the authors recommend that new research into psychedelic-assisted group therapy take advantage of contemporary methodological advances in group psychotherapy research. They point to available evidence-based clinical practice guidelines and validated measures of interpersonal functioning and group cohesion, and recommend statistical approaches such as hierarchical linear modelling to account for nested data and the non-independence of observations in group research. Key design variables for future trials include group size, drug dose and sequencing, number of group sessions, and the type of group therapy. Given current resource intensity of modern psychedelic-assisted individual psychotherapy, the authors suggest groups may initially be most useful for preparation and integration sessions to improve cost-effectiveness while also experimentally manipulating set and setting. They conclude that incorporating group components may reveal how interpersonal context shapes responses to psychedelics and that psychedelic-assisted group therapy remains a promising area for further clinical and scientific exploration, while reiterating that the current evidence base does not permit conclusions about efficacy.