Reporting Sample Size Calculation in Randomized Clinical Trials Published in 4 Orthodontic Journals

Marialicia Calderon-Augusto, Carlos Flores-Mir, Luis Ernesto Arriola-Guillén

Research output: Contribution to journalArticlepeer-review

Abstract

Objective: The purpose of this study was to describe sample size calculations in randomized clinical trials (RCTs) published in 4 orthodontic journals. Methods: This cross-sectional study evaluated 142 RCTs published between 2015 and 2019 in the 4 journals with the highest impact factor in orthodontics according to the SCIMAGO 2018 ranking. In the study, 2 trained and experienced orthodontists assessed whether the RCTs evaluated reported their sample size calculations, and whether they adequately described the criteria for the calculations, including the level of significance, test power, precision or effect size (clinically relevant difference), and expected variability. The sample size calculation was considered adequately reported when the above 4 criteria were described. Results: We identified 120 publications (84.5%) reporting the sample size calculation, but only 70 (58.3%) fully described the above parameters. Inadequate calculation included failure to report the confidence level (ranging from 0% to 12.9%), test power (ranging from 0% to 20%), effect size (ranging from 0% to 22.5%), and expected variability (ranging from 22.6% to 80%). According to the journal, some parameters of sample size calculation were more frequently reported. Conclusions: RCTs published in the 4 leading orthodontic journals frequently do not report the parameters used for sample size calculations.

Original languageEnglish
Pages (from-to)214-219
Number of pages6
JournalTurkish Journal of Orthodontics
Volume34
Issue number4
DOIs
StatePublished - Dec 2021

Keywords

  • Sample size calculations
  • orthodontic journals
  • randomized clinical trials

Fingerprint

Dive into the research topics of 'Reporting Sample Size Calculation in Randomized Clinical Trials Published in 4 Orthodontic Journals'. Together they form a unique fingerprint.

Cite this