This study aimed to estimate the level of burnout in mental health professionals and to identify specific determinants of burnout in this population. A systematic search of MEDLINE/PubMed, PsychINFO/Ovid, Embase, CINAHL/EBSCO and Web of Science was conducted for original research published between 1997 and 2017. Sixty-two studies were identified as meeting the study criteria for the systematic review. Data on the means, standard deviations, and prevalence of the dimensions of burnout were extracted from 33 studies and included in the meta-analysis (n = 9409). The overall estimated pooled prevalence for emotional exhaustion was 40% (CI 31%–48%) for depersonalisation was 22% (CI 15%–29%) and for low levels of personal accomplishment was 19% (CI 13%–25%). The random effects estimate of the mean scores on the Maslach Burnout Inventory indicate that the average mental health professional has high levels of emotional exhaustion [mean 21.11 (95% CI 19.98, 22.24)], moderate levels of depersonalisation [mean 6.76 (95% CI 6.11, 7.42)] but retains reasonable levels of personal accomplishment [mean 34.60 (95% CI 32.99, 36.21)]. Increasing age was found to be associated with an increased risk of depersonalisation but also a heightened sense of personal accomplishment. Work-related factors such as workload and relationships at work, are key determinants for burnout, while role clarity, a sense of professional autonomy, a sense of being fairly treated, and access to regular clinical supervision appear to be protective. Staff working in community mental health teams may be more vulnerable to burnout than those working in some specialist community teams, e.g., assertive outreach, crisis teams.