The following is a summary of “Statistical power and sample size calculations for time-to-event analysis,” published in the December 2023 issue of Surgery by Zurakowski, et al.

For a study, researchers sought to give lung and cardiovascular doctors the tools to figure out sample size and power for studies that look at results over time. Power and sample size estimates improve the quality of research studies by giving users confidence and information about how many patients were involved in the study. A 5-step plan was given for figuring out sample size when comparing groups based on the time it takes for an event to happen. 

Following were the steps: Figure out the main result of interest; Define the size of the effect and the level of power; Choose the right statistical test; Figure out how many samples are needed; and Write an official power and sample size statement. This method is shown with 5 real-life cases of time-to-event studies in cardiovascular surgery. 

These examples use Cox regression, a 2-sample log-rank test, a 1-sample log-rank test, and competitive risks analysis. Statistical power is an important part of planning studies to ensure that the sample numbers are big enough to find treatment benefits or changes between groups in the time it takes for patients to reach a certain result. Research papers with power and sample size explanations are more logical and credible when they use statistics. It also gives the reader confidence that the results and conclusions are correct and based on a large enough group of patients.

Source: sciencedirect.com/science/article/abs/pii/S0022522322010108