Selective outcome reporting in randomized clinical trials using the third molar surgery model

Vinicius Paes de Assis Santos; Daniel Isaac Sendyk; Matheus Dantas de Araújo Barretto; Julia Puglia Nunes; Claudio Mendes Pannuti; Maria Cristina Zindel Deboni

doi:10.1016/j.jcms.2024.03.032

Selective outcome reporting in randomized clinical trials using the third molar surgery model

J Craniomaxillofac Surg. 2024 Mar 21:S1010-5182(24)00115-X. doi: 10.1016/j.jcms.2024.03.032. Online ahead of print.

Authors

Vinicius Paes de Assis Santos¹, Daniel Isaac Sendyk², Matheus Dantas de Araújo Barretto¹, Julia Puglia Nunes¹, Claudio Mendes Pannuti³, Maria Cristina Zindel Deboni⁴

Affiliations

¹ Oral Surgery Department, Faculty of Dentistry, University of São Paulo, Brazil.
² Implantology Department, São Leopoldo Mandic Institute and Research Center, Brazil; Stomatology Department, Faculty of Dentistry, University of São Paulo, Brazil.
³ Stomatology Department, Faculty of Dentistry, University of São Paulo, Brazil.
⁴ Oral Surgery Department, Faculty of Dentistry, University of São Paulo, Brazil. Electronic address: mczdebon@usp.br.

PMID: 38582673
DOI: 10.1016/j.jcms.2024.03.032

Abstract

Selective outcome reporting (SOR) can threaten the validity of results found in clinical trials. Some studies in the literature have analyzed SOR in dentistry, but there is no study that has observed SOR in clinical trials in oral and maxillofacial surgery. Impacted third molar surgery is one of the most used models in clinical trials to study mainly analgesic and anti-inflammatory drug interventions. Our study aimed to evaluate the prevalence of SOR in publications employing the third molar extraction clinical trial model, and to verify whether there was an association between the statistical significance of outcomes and other characteristics that could lead to SOR. A systematic search was performed on the ClinicialTrials.gov platform for randomized clinical trial protocols, using the condition of third molar extraction. The corresponding published articles were sourced in PubMed, Scopus, and Embase databases, and compared with the registered protocols regarding the methodological data, in terms of: sample calculation, primary outcome identification, end-point periods, insertion of new outcomes in the publication, and results of outcomes. 358 protocol records were retrieved; 87 presented their corresponding articles. SOR was identified in 28.74% of the publications, and had a significant relationship with changes in the protocol, insertions of new outcomes, and discrepancies in the types of study. General risk of bias was found to be low. There were associations between SOR and the discrepancies in terms of the type of study, the choice of new outcome, and changes in the history of protocol records. The prevalence of SOR in clinical research using the third molar extraction surgery model is moderate. The quality of the scientific reporting of the results and, consequently, the certainty of evidence relating to the intervention tested can be overstated, increasing the chances of misinterpretation by health professionals.

Keywords: Randomized clinical trial; Scientific evidence; Selective outcome reporting; Third molar.

Publication types

Review