Is the Best Evidence Good Enough: Quality Assessment and Factor Analysis of Meta-Analyses on Depression

BACKGROUND The quality of meta-analyses (MAs) on depression remains uninvestigated. OBJECTIVE To assess the overall reporting and methodological qualities of MAs on depression and to explore potential factors influencing both qualities. METHODS MAs investigating epidemiology and interventions for depression published in the most recent year (2014-2015) were selected from PubMed, EMBASE, PsycINFO and Cochrane Library. The characteristics of the included studies were collected and the total and per-item quality scores of the included studies were calculated based on the two checklists. Univariate and multivariate linear regression analyses were used to explore the potential factors influencing the quality of the articles. RESULTS A total of 217 MAs from 74 peer-reviewed journals were included. The mean score of Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) was 23.0 of 27 and mean score of Assessment of Multiple Systematic Reviews (AMSTAR) was 8.3 of 11. Items assessing registration and protocol (14.2%, 37/217) in PRISMA and item requiring a full list of included and excluded studies (16.1%, 40/217) in AMSTAR had poorer adherences than other items. The MAs that included only RCTs, pre-registered, had five more authors or authors from Cochrane groups and the MAs found negative results had better reporting and methodological qualities. CONCLUSIONS The reporting and methodological qualities of MAs on depression remained to be improved. Design of included studies, characteristics of authors and pre-registration in PROSPERO database are important factors influencing quality of MAs in the field of depression.