Research

Psychedelic Therapy Statistics

Psychedelic therapy statistics measure several different things: clinical-trial outcomes, reported harms and use of regulated services. This reference keeps them separate, with original sources and clear denominators. It does not calculate a single success rate for psychedelic therapy.

Sources checked November 17, 2026. The clinical table contains selected psilocybin and MDMA trials. Oregon figures are administrative service-use data, not a clinical trial or evidence that a condition was treated successfully.

Download the selected trial data (CSV) · Download the Oregon service-data extract (CSV)

Selected psychedelic clinical-trial results

Each row describes one study's participants, treatment and endpoint. The outcomes are too different to average or use as a ranking of treatments.

Original studyParticipants and comparisonSelected findingInterpretation limit
Davis et al., published online 2020Major depressive disorder; 27 randomized to immediate or delayed psilocybin-assisted treatment; 24 completedAt four weeks after treatment, 17/24 (71%) met response criteria and 13/24 (54%) met remission criteriaSmall waiting-list study; these percentages describe the treated completer sample
Carhart-Harris et al., 2021Major depression; 30 assigned to psilocybin and 29 to escitalopram; psychological support in both groupsThe primary QIDS-SR-16 depression-score comparison at six weeks was not statistically significant (P=.17)Favorable secondary outcomes do not overturn a nonsignificant primary endpoint
Goodwin et al., 2022Treatment-resistant depression; 233 randomized across three psilocybin dose groups with psychological supportAt week three, the adjusted MADRS change difference for 25 mg vs 1 mg was −6.6 points (95% CI −10.2 to −2.9)A short-term symptom difference is not evidence of sustained remission
Raison et al., 2023Major depressive disorder; 51 assigned to psilocybin and 53 to niacin; psychological support in both groupsAt day 43, the adjusted MADRS change difference was −12.3 points (95% CI −17.5 to −7.2)Selected participants; short follow-up; no individual prediction
Bogenschutz et al., 2022Alcohol use disorder; 95 randomized; 93 received study medication and entered the primary analysisHeavy drinking days averaged 9.7% with psilocybin vs 23.6% with diphenhydramine across 32 weeksA percentage of drinking days is not a percentage of participants cured
MAPP2, Mitchell et al., 2023Moderate-to-severe PTSD; 53 assigned to MDMA-assisted therapy and 51 to placebo with therapyAt 18 weeks, adjusted CAPS-5 symptom change was −23.7 vs −14.8 points; treatment difference −8.9 (95% CI −13.70 to −4.12)A clinical-trial result does not establish routine availability or results outside the protocol

MADRS and QIDS-SR-16 measure depressive symptoms. CAPS-5 measures PTSD symptoms. Their point scales and cutoffs differ, so a ten-point change on one scale is not the same result as a ten-point change on another.

What Oregon's psilocybin data show

Oregon's data describe its regulated service system. They should be labeled by place and reporting period rather than presented as national psychedelic-treatment totals.

The Oregon Health Authority archive listed the second quarter of 2026 as its latest downloadable quarter when checked. We also include the official 2025 year-end summary as a separate annual reference. The two columns cover different lengths of time. They should not be read as a year-over-year growth comparison, and neither is a national count.

Oregon administrative measureJanuary–December 2025April–June 2026
Reported clients served5,9351,220
Individual administration sessions4,628966
Group administration sessions74791

Sources: 2025 year-end summary, page 3; Q2 2026 official data file.

“Clients served” is the agency's reported measure. Do not silently relabel it as unique people across all centers or periods. A group session may involve several clients, and an individual may receive services more than once. Adding individual and group session counts does not yield a client count.

Why the service counts are not efficacy statistics

A recorded service visit does not tell us whether someone had a diagnosis, improved, remained well or would have improved without that service. Administrative usage and clinical outcomes have different purposes.

Oregon also collects reported reaction and demographic fields, with confidentiality protections. Some small values are suppressed and appear as -99 in downloadable data. Treating that code as a negative count or zero would corrupt an analysis. OHA data definitions.

We have not calculated a statewide harm rate from the extract. A usable rate would require careful work on reporting completeness, category overlap, repeated clients and the correct denominator. The extract contains service volumes only.

Does a trial result mean a treatment is proven for everyone?

No. A trial result applies most directly to the people, treatment and setting studied. It does not automatically transfer to a different diagnosis, an unsupported session, a retreat or a different substance.

The 2021 psilocybin–escitalopram trial is a useful check on overstatement: its primary depression-score comparison was not significant, although several secondary outcomes favored psilocybin. Those secondary analyses were not corrected for multiple comparisons. Original trial.

The Goodwin trial reported adverse events in 179/233 participants (77%) across all dose groups; suicidal ideation or behavior or self-injury occurred in every dose group. These observations need their clinical context and do not by themselves establish what caused an event. Original trial.

In Raison's trial, no serious treatment-emergent adverse events were reported, but overall and severe adverse events were more frequent with psilocybin than niacin. “No serious events in this trial” is a study observation, not a guarantee of safety. Original trial.

Why some quoted success rates use smaller denominators

A study may randomize one number of people but have fewer people at the final assessment. Both numbers matter.

MAPP2 randomized 104 participants. Its exploratory “no longer met PTSD criteria” result was 37/52 (71.2%) in the MDMA-assisted group versus 20/42 (47.6%) in the placebo-with-therapy group at study end. Those are endpoint denominators, not the original randomized group sizes of 53 and 51. The study's primary outcome was symptom-score change. Original report.

That distinction is why the downloadable trial dataset includes both study size and analysis-denominator notes. When quoting a percentage, keep the assessment time, comparison and denominator with it.

How to use these figures accurately

Use statistics to describe evidence, not to promise a result. A useful citation includes five elements:

  1. The study: author, year and original source.
  2. The people: diagnosis and sample size.
  3. The intervention: substance, setting and accompanying therapy or support.
  4. The endpoint: what was measured and how improvement was defined.
  5. The time: the exact assessment point or observation window.

For service data, replace the clinical diagnosis and endpoint with the jurisdiction, administrative measure and reporting period. Keep “reported clients” and “sessions” distinct.

Avoid “psychedelic therapy has a 71% success rate.” It combines a broad treatment category with a number that came from one defined study and one particular endpoint.

Downloadable data and field definitions

The trial CSV contains one selected result per trial, plus the separately identified MAPP2 exploratory endpoint. It records study design, endpoint, time point, numerical units, source and interpretation limits.

The Oregon extract contains six service-volume rows: three measures for 2025 and the same three for Q2 2026. It preserves official units and links each value to its source. It is an editorial extraction, not an OHA publication or a complete copy of OHA's dataset.

See the data dictionary for missing-value rules and reuse notes. The original publications and agency files remain the authority for full definitions and methods.

Ketamine and esketamine have different evidence and treatment settings. Their figures are kept in the separate ketamine therapy statistics reference.

How this reference was prepared

We selected original clinical reports that illustrate major differences in population, comparison, endpoint and follow-up. We read their reported findings and checked the linked Oregon agency files on November 17, 2026. We did not contact trial participants, collect clinical records or perform a new clinical study.

The selection is not exhaustive or a systematic review. It is not a meta-analysis, market forecast, clinical guideline or estimate of the odds that an individual will improve. Older research retains its publication year. Administrative data retain their reporting periods. The page's check date describes when we checked the sources.