Background. Deep brain stimulation (DBS) is increasingly being applied to psychiatric conditions such as obsessive-compulsive disorder (OCD), major depression and anorexia nervosa. Double-blind, randomized controlled trials (RCTs) of active versus sham treatment have been limited to small numbers. We therefore undertook a systematic review and meta-analysis of the effectiveness of DBS in psychiatric conditions to maximize study power. Method. We conducted a systematic literature search for double-blind, RCTs of active versus sham treatment using Pubmed/Medline and EMBASE up to April 2013. Where possible, we combined results from studies in a meta-analysis. We assessed differences in final values between the active and sham treatments for parallel-group studies and compared changes from baseline score for cross-over designs. Results. Inclusion criteria were met by five studies, all of which were of OCD. Forty-four subjects provided data for the meta-analysis. The main outcome was a reduction in obsessive symptoms as measured by the Yale-Brown Obsessive Compulsive Scale (YBOCS). Patients on active, as opposed to sham, treatment had a significantly lower mean score [mean difference (MD) -8.93, 95% confidence interval (CI) -13.35 to -5.76, p < 0.001], representing partial remission. However, one-third of patients experienced significant adverse effects (n = 16). There were no differences between the two groups in terms of other outcomes. Conclusions. DBS may show promise for treatment-resistant OCD but there are insufficient randomized controlled data for other psychiatric conditions. DBS remains an experimental treatment in adults for severe, medically refractory conditions until further data are available.