Reporting of the methodological quality of search strategies in orthodontic quantitative systematic reviews

This study aimed to assess the reporting of the methodological quality of search strategies undertaken in orthodontic quantitative systematic reviews (SRs) and hence their reproducibility.

A search of a single electronic database (Medline via PubMed) was undertaken to identify interventional orthodontic SRs with meta-analysis published within a 10-year period. The Cochrane Library of Systematic Reviews was also sourced. Full articles were reviewed by two assessors against the eligibility criteria. The reporting quality of each search strategy was assessed using a previously validated checklist with a score of 1 or 2 given for each of the eight items. Cumulative totals were calculated. Guided by previous research, the authors agreed the following cut-offs to categorize the overall level of quality: 8-10 (poor), 10-12 (fair), and greater than 13 (good).

A total of 127 SRs were analysed. The overall median quality score for the reporting of the search strategy was 14 [interquartile range (IQR): 13-15]. Cochrane SRs and those originating in Europe received higher aggregate scores, whereas no difference was evident based on Prospero registration. The continent of the corresponding author predicated the overall score. Non-Cochrane reviews achieved lower overall scores compared to Cochrane reviews (-1.0, 95% confidence interval: -1.65, -0.34, P = 0.003). The most frequently searched database was EMBASE (N = 93) and the median number of authors was 5 (IQR 4-6). Authors of 26.8% of SRs searched the grey literature. Language restrictions were applied to the search strategies of 88 (69.3%) SRs.

The reporting quality of search strategies undertaken in orthodontic SRs is at a good level but differences between Cochrane and non-Cochrane reviews currently exist. The reporting of searching of the grey literature and application of no language restrictions can be improved.