Influence of study characteristics, methodological rigour and publication bias on efficacy of pharmacotherapy in obsessive-compulsive disorder: a systematic review and meta-analysis of randomised, placebo-controlled trials.