PASSReviewer 1· 95% conf
The study design is rigorous with adequate randomization, blinding of assessors, a priori power analysis, and clear inclusion/exclusion criteria.
Evidence
direct quote[Methods, Randomisation and masking]
“Treatment assignment was carried out using stochastic minimisation, controlling for age, sex at birth, and method of self-harm.”direct quote[Methods, Randomisation and masking]
“Researchers conducting follow-up assessments, as well as the trial statistician, remained masked to the treatment allocation.”paraphrase[Methods, Participants]
“Under these assumptions, and with a 5% significance level and 90% power, a study with no clustering would require 158 patients per group. ... assuming an intraclass correlation coefficient of 0.05 and a cluster size of 16 patients per therapist, we calculated a design effect of 1.75, increasing the numbers required to 277 participants per group. Furthermore, with an expected loss to follow-up of 15%, the final numbers recruited were calculated as 326 per group, 652 in total.”PASSReviewer 2· 95% conf
The study design is rigorous, with clear randomization, blinding of assessors and statistician, a priori sample size calculation, pre-specified inclusion/exclusion criteria, and handling of missing data via sensitivity analyses.
Evidence
direct quote[Methods, Randomisation and masking]
“Recruited adolescents were randomly allocated in a 1:1 ratio to one of two trial arms: YCMAP plus enhanced treatment as usual, or enhanced treatment as usual alone. Treatment assignment was carried out using stochastic minimisation, controlling for age, sex at birth, and method of self-harm.”direct quote[Methods, Randomisation and masking]
“Researchers conducting follow-up assessments, as well as the trial statistician, remained masked to the treatment allocation. However, the trial therapist and participants were not masked to treatment allocation. Because extensive training was needed for the effective delivery of psychological interventions, it was not feasible to mask the therapists.”paraphrase[Methods, Statistical analysis (sample size)]
“Under these assumptions, and with a 5% significance level and 90% power, a study with no clustering would require 158 patients per group... Assuming an intraclass correlation coefficient of 0.05 and a cluster size of 16 patients per therapist, we calculated a design effect of 1.75, increasing the numbers required to 277 participants per group. Furthermore, with an expected loss to follow-up of 15%, the final numbers recruited were calculated as 326 per group, 652 in total.”