Selective outcome reporting among randomized controlled trials in leading orthodontic journals: 2018–24

Objectives To evaluate the prevalence and manifestations of selective outcome reporting (SOR) among randomized controlled trials (RCTs) in leading orthodontic journals, and to explore factors that may be potentially related to the presence of SOR. Methods A manual search was conducted to identify eligible RCTs published in eight leading orthodontic journals between 2018 and 2024. Only RCTs with defined primary outcomes (POs) in both publications and corresponding registrations were included. Discrepancies between publications and registrations were compared with respect to the PO. Logistic regression analysis was performed to assess the association between study characteristics and the presence of SOR. Results A total of 139 eligible RCTs were included for analysis, of which 99 (71.2%) were retrospectively registered. SOR was identified in 70 (50.4%) RCTs. The most frequent type of SOR was the omission of registered PO in the corresponding publication (n = 27, 19.4%), followed by the downgrade of registered PO (n = 25, 18.0%) and the introduction of new PO in the publications (n = 23, 16.5%). According to the logistic regression analysis, the prevalence of SOR was significantly correlated with statistician involvement (OR: 0.349, 95% CI: 0.127–0.961, P = .042), and the number of POs reported in the registration (OR: 5.581, 95% CI: 2.126–14.647, P < .001). Conclusion The prevalence of SOR was high among RCTs in leading orthodontic journals. Clinicians, editors along with other stakeholders are expected to make joint efforts to address this issue and improve the transparent reporting of outcomes.