Aim
To record the prevalence of a priori power calculations in manuscripts published in three endodontic journals between 2018 and 2020 and detect further associations with a number of study characteristics including journal, publication year, study design, geographic region, number of centres and authors, whether the primary outcome pertained to a statistically significant effect and whether confidence intervals (CIs) were reported.
Methodology
The contents of the three leading endodontic journals with the highest impact factor (International Endodontic Journal, IEJ; Journal of Endodontics, JOE; and Australian Endodontic Journal, AEJ) were assessed from January 2018 to December 2020. The proportion of articles reporting a priori power calculations were recorded, and relevant associations as described above were assessed. Univariable and multivariable logistic regression were used to identify significant predictors, whilst interaction and linear trend effects were also considered.
Results
A total of 716 original research articles were included. The vast majority were published in the JOE (417/716; 58.2%), followed by the IEJ (225/716; 31.4%) and the AEJ (74/716; 10.4%). Overall, a priori power considerations were reported in 243 out of 716 articles (33.9%). The IEJ presented 1.61 times higher odds for including a priori power considerations compared to JOE (adjusted odds ratio, OR = 1.61; 95%CI: 1.11, 2.34), whilst for the AEJ, the corresponding odds were 41% lower in comparison to JOE (adjusted OR = 0.59; 95%CI: 0.31, 1.14). For each additional year indicating more recent publication, the odds for adopting appropriate reporting practices for power considerations were increased by 64% (adjusted OR = 1.64; 95%CIs: 1.32, 2.04). There was strong evidence that interventional research was associated with 10.54 times higher odds for a priori considerations compared to observational study design (adjusted OR = 10.54; 95%CIs: 5.50, 20.19).
Conclusions
The high prevalence of failure to include a priori power considerations was indicative of suboptimal reporting in endodontic research, in the three endodontic journals analysed. Although the condition improved over time, efforts to incorporate a correct determination of the required sample size at the design stage for any future study should be endorsed by journal editors, authors and the scientific community.