Table 3.
Summary risk of bias assessment of included studies, based on domains in Cochrane Risk of Bias 2 tool
| Study reference | Design | Experimental condition | Comparator | Domain | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| D1 | DS | D2 | D3 | D4 | D5 | Other | ||||
| Carhart-Harris et al 2021,63 and Barba et al 202264 (secondary analysis) | Parallel | Two doses of 25 mg psilocybin, daily placebo, and psychological support | Two doses of 1 mg psilocybin, daily escitalopram, and psychological support | Low | - | Low | Low | Low | Low | Some concern |
| Goodwin et al 2022,18 and Goodwin et al 202365 (secondary analysis) | Parallel | Single dose (25 mg or 10 mg) psilocybin with psychological support | Placebo-like dose (25 mg or 10 mg) psilocybin with psychological support | Low | - | Low | Low | Low | Low | Some concern |
| von Rotz et al 202366 | Parallel | Single dose (0.215 mg/kg) psilocybin with psychological support | Placebo and psychological support | Low | - | Low | Low | Low | Low | Some concern |
| Grob et al 201115 | Crossover | Single dose (0.2 mg/kg) psilocybin | Niacin (250 mg) | Low | Low | Low | Low | Low | Low | Some concern |
| Griffiths et al 201614 | Crossover | Single dose (22 mg/70 kg or 30 mg/70 kg) psilocybin | Placebo-like (1 mg/70 kg or 3 mg/70 kg) psilocybin | Low | Low | Some concern | Low | Low | Low | Some concern |
| Ross et al 2016,17 and Ross et al 202167 (secondary analysis) | Crossover | Single dose psilocybin (0.3 mg/kg) and psychotherapy | Niacin (250 mg) and psychotherapy | Low | Low | Low | Low | Low | Low | Some concern |
Domain 1 assessed bias arising from the randomisation process, including blinding and randomisation of the allocation sequence, and baseline differences between groups. Domain S assessed the period and carryover effects specific to the crossover trials. Domain 2 assessed bias due to deviations from the intended intervention, including participant and researcher blinding, and the effects of differential adherence between groups. Domain 3 assessed bias due to missing outcome data. Domain 4 assessed bias in measurement of the outcome, including the appropriateness of metrics and questionnaires used and between group differences in outcome assessment. Domain 5 assessed reporting bias arising from the selective reporting of results or data analyses, or both. The “other” criterion assessed bias due to potential conflicts of interest, such as industry funding and associations with pharmaceutical companies.