An international Delphi consensus for reporting of setting in psychedelic clinical trials

Psychedelic drugs are thought to interact strongly with contextual factors commonly referred to as 'set and setting'—the individual's mental state and the environmental, social and procedural context of use. The authors note that despite longstanding recognition of these influences, clinical research has lacked consistent, rigorous reporting of extra‑pharmacological variables, producing uncertainty about which non‑drug factors most affect outcomes and when they exert their effects. A recent systematic review cited by the authors found many trials failed to report even basic contextual features, limiting transparency and comparability across studies. This study set out to produce consensus‑based reporting guidance for extra‑pharmacological variables in psychedelic clinical trials. Using an international Delphi process, the investigators aimed both to generate a parsimonious, implementable checklist of setting variables warranting routine reporting and to probe how experts currently conceptualise the boundaries and temporality of 'set' and 'setting'. The resulting product is the Reporting of Setting in Psychedelic Clinical Trials (ReSPCT) guidelines, derived from input by clinicians, researchers and former trial participants across multiple countries and institutions.

The study employed a four‑round Delphi consensus method, an iterative, quasi‑anonymous survey technique used to derive expert agreement on complex topics. Initial recruitment involved personalised email invitations to 149 experts plus 34 identified by snowballing; 89 experts consented and completed round 1. Panel composition aimed to include psychedelic clinicians, researchers and past clinical‑trial participants with relevant expertise in areas such as psychopharmacology, neuroimaging, psychedelic‑assisted therapy, trial design and harm reduction. Diversity was assessed using self‑reported sociodemographic and professional characteristics, and predefined subgroups were formed by gender, ethnicity, primary expertise, field and prior trial experience. Round 1 asked experts to list up to ten setting variables important to report; 770 free‑text responses were coded by the lead author and iteratively grouped by the lead authors into 49 distinct items. In round 2 (n = 73) experts rated each item for importance and for coherence (clarity of formulation), and provided optional feedback. Importance ratings were evaluated against predefined consensus thresholds (the study reports items reaching 'important' or 'very important' by ≥70% as consensus), while items with less than 90% coherence were revised. Subgroup analyses were used to detect minority consensus that might be missed by whole‑group voting, and items were assessed for uniqueness and implementability; overlapping items were combined, and items judged standard practice or problematic (for example privacy concerns) could be rejected. Round 3 (n = 68) presented modified items and proposed combinations for re‑rating; experts also voted on whether subjective items better reflecting 'set' should be retained, moved to an optional subsection, or excluded. Round 4 (n = 62) solicited feedback on the preliminary guidelines and on the study process, and included targeted questions about the separability, overlap and temporal boundaries of 'set' and 'setting'. Throughout, acceptance criteria included whole‑group importance ratings meeting thresholds, high coherence ratings (≥90% for coherence was used for modification decisions) and subgroup rescue of items where appropriate. Final guidelines were refined collaboratively with study leads and panel feedback and accompanied by an explanatory document and checklist.

The Delphi produced the 30‑item ReSPCT guidelines, organised into four sections: physical and sensory environment; dosing session procedure; therapeutic framework and protocol; and participant subjective experiences. Participation declined across rounds from 89 in round 1 to 62 in round 4 (a 30% attrition rate). Experts represented 17 countries and approximately 50 institutions, although the extracted text states that demographic and professional details are presented in a table not reproduced here. Round 1 generated 770 responses that were coded into 49 unique items; four items were mentioned by at least half the respondents: music and sound (n = 74), objects/decorations/artwork (n = 55), number of people present (n = 52), and access to nature (n = 46). In round 2, 33 of the 49 items (67%) met whole‑group consensus thresholds; one of these was removed because it duplicated standard clinical‑trial reporting. After quality checks for coherence, uniqueness and implementability, 17 items were accepted as written, 7 rejected and 27 were combined into 18. Subgroup analyses rescued several items that had not reached whole‑group consensus. Round 3 yielded further convergence: modified items met the 90% coherence threshold, proposed combinations exceeded 90% acceptance, and re‑ratings of previously subgroup‑consensus items added two items to whole‑group consensus. Experts voted to retain five items judged more subjective (pertaining to 'set') rather than exclude them; 46% supported keeping them as standard, 39% as an optional 'participant experience' subsection, and only 15% voted to exclude them. These steps produced a 30‑item preliminary guideline set. In round 4, at least 95% of respondents reported being satisfied or very satisfied with the preliminary guidelines and judged the process clear and successful. Regarding uptake, at least 74% said they would 'definitely' use the preliminary guidelines and 25% would 'consider' using them; only one expert (2%) said they would not. Responses about conceptualisation of set and setting were heterogeneous: 37% agreed or strongly agreed that set and setting can be studied separately, while 42% disagreed or strongly disagreed. Panel members estimated an average conceptual overlap of 49% between set and setting. On temporality, about 68% agreed that setting spans from recruitment to the end of follow‑up, 16% limited it to the treatment phase, and 10% to the dosing session. A notable item-level change was cultural competence and safety: it initially reached consensus only in some subgroups but, after debate, its whole‑group rating rose from 50% to 73%, securing its inclusion. The final ReSPCT package includes a downloadable checklist and explanatory document with item descriptions, reporting prompts and a literature summary.

Pronovo‑Morgan and colleagues present the ReSPCT guidelines as the first international consensus on which extra‑pharmacological variables should be routinely reported in psychedelic clinical trials. They argue the 30 items—covering environment, session procedures, therapeutic protocol and participant experience—are intended to improve transparency, reproducibility and comparability across studies, and to facilitate data aggregation and dismantling studies. The authors suggest the guidelines are applicable to multisite trials (with site‑specific details appended) and could support more pragmatic study designs that acknowledge and report contextual influences rather than attempting to mask them, a relevant point given the challenges of blinding in psychedelic research. The discussion highlights heterogeneity of expert views about what constitutes set versus setting, and the study team interprets this divergence as a likely contributor to inconsistent reporting practices. Rather than enforcing a strict dichotomy, the guidelines treat the 30 factors as interrelated. The authors note two items—cultural competence and cultural safety—were initially borderline but were included after panel debate emphasised concerns about power imbalances and lack of diversity. They also acknowledge limitations: the Delphi was survey‑based, which may have limited depth of dialogue; the panel was not fully representative of broader perspectives (most experts were relatively young, Western and 75–80% White); and the focus on clinical trials excluded nonclinical viewpoints. The authors emphasise that expert consensus does not equate to empirical truth and call for further empirical research to evaluate which contextual variables most strongly influence clinical outcomes. Finally, they present the ReSPCT materials (checklist and explanatory document) as resources to be used alongside trial reports to strengthen methodological rigour and to guide future research on drug‑context interactions.