I read the article in the March 2017 issue comparing survival time and comfort of 2 clear overlay retainers of different thicknesses with great interest (Zhu Y, Lin J, Long H, Ye N, Huang R, Yang X, et al. Comparison of survival time and comfort between 2 clear overlay retainers with different thicknesses: a pilot randomized controlled trial. Am J Orthod Dentofacial Orthop 2017;151:433-9).
The article was a pilot study with a sample size of 80 participants, and its clear message is that there are significant differences in fracture of clear overlay retainers made of 0.75-mm and 1-mm thickness, with more fractures in retainers of thinner material. The methodology followed satisfied most criteria of a randomized clinical trial. After going through the article, I have thought the following things.
The authors stated that there were no previous data to calculate the sample size necessary to obtain statistically significant results. After the trial, they should have calculated the sample size with the data they obtained and compared whether the sample of 80 was enough for statistically significant results.
The authors stated that the procedure for randomization was by tossing a coin. The 80 participants should be allocated into 2 groups. This tossing of a coin method for randomization is suitable when the participants are taken in pairs to be allocated to each group. But how the pairs are formed by following the randomization method is not clearly mentioned. The authors mentioned random number generation in the heading of a paragraph but did not mention the procedure followed.
In Table II, the authors gave the data of fracture, retainer loss, no longer fitting, and serious abrasions of the retainers. I consider that retainer loss is not related to thickness of the retainers and should not have been considered in the outcome of failed retainers. The total numbers of lost retainers were 4 in 1-mm thickness and 5 in 0.75-mm thickness, which may have significantly influenced the outcome of failure of retainers.
The authors have given the lower and upper limits of the confidence intervals of the pooled data but not of the individual groups of failures in Table II. They stated that there was a statistically significant difference of the incidences of fracture of retainers between the 2 groups. The authors should have given the upper limit of the confidence interval at least for the fracture of retainers to prove that the result they obtained was not by chance, even though P is 0.281. Because the numbers of fractures are 2 in the 1-mm group and 9 in the 0.75-mm group, the chance might be there, even though the P value is significant when the upper limit of the confidence interval is not considered.
∗ The viewpoints expressed are solely those of the author(s) and do not reflect those of the editor(s), publisher(s), or Association.