Correction to: BMC Psychiatry 19, 307 (2019)

https://doi.org/10.1186/s12888-019-2297-y

Following publication of the original article [1], the authors identified a typographical error in the proposed sample size section. The correct sentence is given below.

The sample size calculation (Stata command sampsi) suggests that 161 patients per arm (322 in total) are needed to detect this effect size or a larger one with 90.14% power allowing for a deflation for including baseline measures in the analysis model (factor 0.84 assuming a correlation between baseline and 52 weeks WSAS of 0.4) and an attrition rate of 25%’.