A1992
Title: KISS with the standard error: A sample size determination for a feasible sample size
Authors: Takashi Omori - Juntendo University (Japan) [presenting]
Yumi Takagi - Doshisha University (Japan)
Sae Murakami - Hiroshima University Hospital (Japan)
Abstract: KISS - Keep It Simple, but Scientific. Sample size determination, the number of participants required to detect a clinically relevant treatment effect, is essential at the planning stage of clinical trials. Usually, the size is determined based on the assumed effect and the specified statistical power and type I error rate for a statistical test. However, sometimes the feasible sample size is limited. A paradigm shift suggests that for a given sample size, researchers can maximize the amount of information obtained. A calculation formula based on the standard error is developed that does not rely on a statistical test but on a similar logical approach and can be implemented using standard software. The approach determines the sample size based on the width of the treatment effect (that is, the treatment effect plus or minus its standard error) along with the probability that the effect lies within that width. The focus should be on the estimated effect with uncertainty, such as Wald statistics, and departure from the type I error rate when the sample size is limited.
