Evidence mapping and quality assessment of systematic reviews in dental traumatology.

Abstract

Evidence mapping of systematic reviews (SRs) systematically and comprehensively identifies, organizes, and summarizes the distribution of scientific evidence in a field. The aims of this study were to delineate domains in dental traumatology (DT), evaluate the existing SRs within the domains, and identify the paucity of evidence for future research. Domains and sub-domains of DT were established according to the methods of qualitative research. The protocol for evidence mapping was prepared as per the guidelines of GEM and PRISMA. The search strategy was formulated using words and MeSH terms in eight databases without restriction of languages and year of publication. Gray literature, protocol registries, and references of selected articles were also searched. Duplicates were removed, and the final selection of SRs was completed. Data extraction and quality analysis using the ROBIS tool and the PRISMA checklist were performed. The overall search resulted in 64 SRs from 1999 to 2020 with 44 published in last six years. The highest number of SRs had been performed in the Prognostic domain (n=19) followed by the domains of Epidemiology (n=15), Therapeutics (n=10), Oral Biology (n=7), Diagnostics (n=6), Preventive (n=5), and Research Methods (n=2). Within each domain, there were variabilities in the number of reviewers, a priori protocols, search limitations, risk of bias methods, and meta-analysis. Of the SRs, including 4 Cochrane reviews, 28.4% were inconclusive. A low risk of bias was found in 48.4% of the SRs. Among the registered and ongoing SRs, six were from the domain of epidemiology, two in the domain of therapeutics, five from prognostics, and one each in the domains of prevention and research methods. The SRs in DT could be mapped in seven domains with variabilities in the methods. The majority had an a priori registered protocol and a low risk of reporting errors. Within the Epidemiology and Preventive domains, SRs were present in all the sub-domains with the majority demonstrating low-risk of bias (ROB). The domain of prognosis had SRs in most sub-domains but with a high ROB. Insufficient numbers of SRs were present in most sub-domains of the Diagnostics, Therapeutics, Research Methods and Oral Biology domains.