Why accurate beats impressive in psychedelic-assisted therapy research for PTSD
Research on psychedelic-assisted therapy needs credibility more than impressive findings.
A new review by Deidre A. Marsh and Edwin Bacalso in Current Treatment Options in Psychiatry argues that some of the recent literature on psychedelic-assisted therapy (PAT) for post-traumatic stress disorder (PTSD) has inflated its case through selective comparisons. They identify three reporting patterns to watch for. Below is a summary of their argument, followed by my perspective on why the critique lands and what it means for the field.
What the paper did and found
Marsh and Bacalso ask a focused question. Are some peer-reviewed papers on PAT for PTSD reporting their findings in ways that overstate efficacy compared to existing treatments? The authors group the reporting issues they find into three categories and illustrate each one with a published example. [I’d noticed some of these myself in my reading and am glad someone wrote a paper that found them as well as other issues]
The three reporting patterns are:
Errors in describing outcomes. Terms like "treatment failure," "loss of diagnosis," and "clinically significant reduction in symptoms" get conflated. Their example is a systematic review that describes "up to two-thirds of people failing to respond to first-line interventions" while citing a paper that actually reported 49 to 70 percent of participants achieving a clinically meaningful reduction. Two=thirds retained their PTSD diagnosis, but that’s not the same as “failing to respond to treatment.”
Errors in comparing divergent samples. PTSD treatment outcomes diverge meaningfully between veteran or active-duty samples and civilian samples, with veterans typically showing worse outcomes. Marsh and Bacalso point to the Phase 3 MDMA-assisted therapy trial by Mitchell et al. (2023), whose sample was 85 percent civilian but whose discussion contrasts those outcomes with a veteran-only Prolonged Exposure (PE) and Cognitive Processing Therapy (CPT) trial.
Errors in dose comparisons. PAT protocols generally involve fewer but much longer sessions. Existing PTSD therapies have developed "massed" formats with three to five sessions a week. The massed format is more similar to PAT in that it delivers more treatment in a shorter time period. Comparing PAT to standard weekly CPT or PE while ignoring the massed-treatment literature distorts the comparison. This is important because massed PE generally finds lower dropout rates.
Not every PAT paper makes these errors. They specifically cite a paper I wrote (Luoma et al., 2020—thanks for the recognition!) describing a meta-analysis of PAT as an example of accurate reporting on comparison treatments and on dosing differences, illustrating that the standard is achievable in this literature.
My perspective
PAT for PTSD is a small literature in which most active investigators want the treatment to work, which makes selective citation especially likely and especially costly.
But the cost is steep. PAT's case to clinicians, regulators, payers, and the public ultimately rests on credibility, and the readers who decide what PAT looks like at scale are peer reviewers, meta-analysts, and guideline writers. A finding those readers treat as a fair report is worth more than a finding that strikes them as inflated, even when the inflated version is more impressive in print. The short-run gain shows up as a long-run discount on the literature, applied at the moment it counts most.
The corrective is not to deflate the science; it is to compare like to like. Compare PAT to massed CPT where massed CPT is the appropriate benchmark. Compare civilian samples to civilian benchmarks. Cite the correct version of treatment response.
For working clinicians and peer reviewers, the three-error framework can be used as short checklist for evaluating PAT-for-PTSD studies. For researchers preparing manuscripts in this area, it is worth a careful read before submission.
Credibility in this literature is central. We protect it by comparing fairly and naming the limits of our own designs, including when our findings are less impressive than we hoped.
Read the full paper: Marsh, D. A., and Bacalso, E. (2026). Psychedelic strawmen: a call for greater transparency of PTSD treatment efficacy in psychedelic-assisted therapy research. Current Treatment Options in Psychiatry, 13:11. https://doi.org/10.1007/s40501-026-00384-1
Stay in touch with PIPS: Sign up for the PIPS newsletter for ongoing research translation in psychedelic-assisted therapy.
About the Author
Jason Luoma, PhD, is the Director of Research at the Portland Institute for Psychedelic Science (PIPS) and a co-founder of Portland Psychotherapy. He studies psychedelic-assisted therapy, shame, and Acceptance and Commitment Therapy (ACT). His own writing on PAT was cited in the Marsh and Bacalso review as an example of accurate reporting. He also hosts the Research Matters Podcast, where he interviews researchers about how they do their work.