Is the primary endpoint clearly defined?
Are the primary endpoint and the statistics used to calculate the sample size consistent with each other?
Is the rationale or reference of the effect size adequately described?
What level of significance and power is selected?
Is the dropout rate appropriate, and is the final sample size properly calculated?
Is the entire calculation process appropriate and without error?
1) For the sake of understanding, the text is based on the comparison of the mean of continuous quantitative variables between two independent groups. Nevertheless, it is necessary to select an appropriate formula to calculate the effect size according to various observations, statistical methods and the number of groups.
2) Although the alternative hypothesis is true, there are also errors that reject the alternative hypothesis, which is called a Type II error.
4) Relevant clause of this journal in the RESEARCH AND PUBLICATION ETHICS section is as follows: “4. Registration of the clinical trial research - Any research that deals with clinical trial should be registered with the primary national clinical trial registration site such as Korea Clinical Research Information Service (cris.nih.go.kr/) or other sites accredited by WHO or International Committee of Medical Journal Editor such as ClinicalTrials.gov (clinicaltrials.gov/)."
6) The formulas given as examples together with the description below are sample size calculations using z-statistics. Many conditions are required for this formula to be true: the standard deviation and sample size of the two groups are the same, the sample size is large enough, and so on. Depending on the nature of the data used to calculate the sample size and the needs of the clinical study, t statistics will often be required, but the details are very complex and are beyond the scope of this paper. Therefore, when designing a clinical study, it is emphasized that the correct sample size requires more accurate formulas and calculations than simply the following formulas. α: significance level, β: probabilities of making type II errors, zα/2 : (α⁄2) th percentile of the standard normal distribution, zβ : (β) th percentile of the standard normal distribution, n1 and n2: number of observations in each group, d: effect size.
Author Contributions
Jae Hong Park (Writing – original draft)
Junyong In (Conceptualization; Writing – review & editing)
Hyun Kang (Methodology; Supervision)
Jong Hae Kim (Data curation)
Tae Kyun Kim (Data curation; Methodology)
Eun Jin Ahn (Data curation)
Dong Kyu Lee (Conceptualization; Data curation)
Sangseok Lee (Data curation; Methodology; Supervision)
No. (n = 44) | |
---|---|
Significance level | |
0.05 | 43 (97.7) |
not applicable | 1 (2.3) |
Power (%) | |
80 | 34 (77.3) |
90 | 8 (18.2) |
99 | 1 (2.3) |
not applicable | 1 (2.3) |