Written by Erin Yeh
Psychedelic-assisted therapy (PAT) has gained a reputation as being more effective than traditional antidepressants (TAD) for treating depression, especially for patients who have not responded well to TADs. However, the effectiveness of these treatments may depend on whether the studies are open, where patients are aware of the treatment they are assigned – or blind, where patients are not aware.
To determine the effectiveness of TAD and PAT for treating major depression, a team of researchers from the University of California, San Francisco, the University of California, Los Angeles, and Imperial College London compared the results of psychedelic treatment trials with the results of an open-label TAD trial in which patients knew they were receiving an antidepressant (Gamma PsychiatryDigital ID: doi:10.1001/jamapsychiatry.2025.4809). By analyzing both groups under unblinded conditions, the potential benefits from effects associated with patients’ knowledge of their treatment are equal.
The difference between blind and non-blind
In clinical trials, patients are ideally “angry” about receiving an active treatment or a placebo. However, some patients can second guess their treatment based on side effects — called functional unblinding — which can make treatment appear more effective than it actually is and possibly affect results. In PAT trials, 90-95% of participants are able to correctly identify their treatment due to strong drug side effects, even when using a placebo. Compared to TAD trials, about 63% of participants guessed correctly.
To date, only one clinical trial has directly compared PAT (psilocybin) with TAD (escitalopram). The study found no difference between the two treatments on the main outcome measure, but the PAT showed significant improvement on secondary measures of depression. This mixed result has led to debate about how reliable this study really is.
Because participants in a TAD trial may not know whether they received an active drug or a placebo, it is difficult to compare them to drugs. Comparing PAT and TAD trials, where patients know what they are taking, provides a more equal comparison.
Database search and comparisons
The team searched the PubMed database to identify experiences of major depressive disorder in adult patients (ages 18 to 65 years). Studies that used either TADs in open-label or PAT-blinded or open-label trials (eg, lysergic acid diethylamide, psilocybin, mescaline, San Pedro, peyote, 5-methoxy-N, N-dimethyltryptamine, or ayahuasca) were included. Studies conducted in inpatient settings or that included patients with psychotic depression or significant comorbidities were excluded, but an exception was given for comorbid anxiety because it frequently co-occurs with depression. Booster and combination trials were also excluded, as well as trials with run-in periods, which is the period between recruitment and randomization in clinical trials. However, data from run-on periods were included as open-label TADs if the run-on period met all inclusion criteria, such as the open-label phase of a discontinuation trial. Articles were independently reviewed for inclusion by two members of the research team, depression scores were extracted, references were screened for additional trials, and study authors were contacted when data were missing.
Of the 619 records retrieved, 38 met the inclusion criteria. Of the 38 trials, only 24 had the required variables, including 16 open-label TAD trials and eight PAT trials. In the PAT trials, six were formally blinded and two were open. Data were analyzed using Bayesian and frequentist models. The researchers focused on the extent to which patients’ depression scores changed from the beginning of the study to its end. To do this, they combined the results of several studies and used methods that account for differences between them, such as using different scales to classify depression.
Contrary to the researchers’ expectations, the results showed that there was no significant difference in effectiveness between TADs and PAT. The difference was only about 0.3 Hamilton Rating Scale for Depression (HAM-D) units, according to Bayesian and frequentist estimates. This finding remained consistent across the different groups of studies analysed.
Open TAD trials have shown better outcomes than blinded treatment. This finding is consistent with the researchers’ hypothesis that blind people can be effectively blinded, and thus blinding should make a difference. But the PAT is always open, so formal blinding doesn’t make a difference. The estimated difference between open and blinded label for TADs was 0.85 HAM-D units using the Bayesian model and 1.29 HAM-D units in the frequentist model. Both models show that blinding was not clinically meaningful.
Meta-analyses of TADs to treat depression showed about a 2.4 point improvement on the HAM-D scale compared to placebo. Given the small difference in improvement, some researchers have questioned the effectiveness of these medications. The PAT showed a mean effect of approximately 7.3 HAM-D points. With a difference of approximately 5 points, PAT appears to be a better treatment option than TADs.
But these results may be misleading. First, open-label TADs tend to work slightly better than blinded treatment. This may be influenced by the patient’s knowledge of the treatment assigned to him, including positive expectancy. Second, placebo responses in PAT trials are much lower — about four points lower than in TAD trials, according to one analysis — making the treatment appear more effective by comparison. The team also expects that the “know SIBO” effect (the disappointment patients feel when they realize they are in the control group) could also influence placebo suppression in PAT trials. Thus, about 55% of the apparent benefit of PAT may come from weak outcomes in the placebo group rather than strong improvements in the treatment groups. In some cases, placebo patients in the PAT trials became worse, while the TAD trials showed improvement in the placebo groups.
Limitations to consider
Some PAT trials have exclusively recruited patients with treatment-resistant depression (TRD), whereas no TAD trial has focused on TRD. Another key difference is that the PAT trials measured outcomes much earlier than the TAD trials, which may have influenced PAT to appear more effective. In addition, PAT participants were more likely to be from a higher education background and there were fewer minorities, suggesting bias in the results.
There were also issues regarding expectations and blinding. Most PAT trial patients can correctly guess whether they have received effective treatment. In contrast, in open TAD trials there was no uncertainty, because patients knew exactly what they were taking. Moreover, expectations vary, even when treatments are not effectively blinded. PAT often receives more positive media attention compared to TADs, which may lead patients to expect better outcomes and thus influence outcomes.
Using two HAM-D models may also have created minor discrepancies, although the overall conclusions remained the same when using only one common measure. In one part of the analysis comparing blinded and open TAD trials, researchers relied on data from a separate meta-analysis with slightly different criteria, meaning there may be hidden differences — such as patient severity or how participants were selected — that might affect the results.
The research team only examined symptom reduction and did not fully consider side effects and improvements in daily functioning. The analysis also only looked at overall changes within patients and did not include direct effects of treatments, placebo effects, and natural improvement over time. As such, the study cannot isolate the amount of improvement resulting from treatment rather than the factors mentioned above.
While blinding made a difference in trials of TADs, it did not for PAT trials, confirming that PAT trials are as effective as open trials. The study results run counter to the growing popularity and hype surrounding the use of psychedelics to treat depression, but that doesn’t mean PAT is ineffective. Importantly, it could still be a treatment option for patients and healthcare providers to consider.
