Predatory Publications Bias the Findings of Orthodontic Meta-analyses

This study quantified how frequently questionable publications by predatory or suspect journals appear in orthodontic evidence syntheses and examined whether inclusion of these studies influences pooled estimates in orthodontic systematic reviews (SRs) with meta-analyses (MAs). Orthodontic SRs with MAs published in English between January 1, 2020, and December 31, 2023, were identified through multiple databases. One MA per SR (prioritized primary outcome/earliest time point) was selected when data were extractable from forest plots. Journals were classified as legitimate, suspect, or predatory using a prespecified approach. For MAs containing at least 1 predatory or suspect study, pooled effect estimates were recalculated under 2 scenarios using random-effects models (restricted maximum likelihood), including 1) all studies and 2) legitimate studies only. Changes were assessed using differences in the pooled estimates between scenarios. From 10,905 records, 340 orthodontic SRs with accessible meta-analytical data were included, most commonly addressing treatment modalities (32.4%). Among the 1,992 primary studies identified, most were published in journals indexed in Medline (n = 1,494; 75%). Overall, 16.2% of SRs were published in predatory or suspect (questionable) journals. One-third of selected MAs (114/340; 33.5%) included at least 1 questionable primary study published in a predatory or suspect journal, and 21.4% (73/340) included at least 1 primary study published in a predatory journal. The inclusion of questionable primary studies was similar whether the SR was published in a legitimate (33.7%), suspect (30%), or predatory journal (33.3%). In the final reanalysis set (n = 106), the pooled estimate decreased in 55% and increased in 45% of MAs after removal of suspect/predatory studies. Predatory and suspect publications are present in 1 in 3 orthodontic meta-analyses and may influence the pooled estimates. Routine legitimacy screening, study quality assessment, and sensitivity analyses excluding questionable studies should be integrated into evidence synthesis and editorial assessment in orthodontics.