Power and Sample Size Calculation for Non-Inferiority Trials with Treatment Switching
Author Information
Author(s): Austin Shih, Chih-Yuan Hsu, Yu Shyr
Primary Institution: Vanderbilt University
Hypothesis
How does treatment switching affect power and sample size calculations in non-inferiority trials using restricted mean survival times?
Conclusion
The proposed simulation-based approach helps understand how treatment switching influences power and sample sizes in non-inferiority trials.
Supporting Evidence
- The simulation study shows that treatment switching can significantly affect power and sample size estimates.
- A shorter switching time and higher switching probability amplify changes in power and sample sizes.
- When the relative effectiveness of treatments is greater than one, power decreases and sample size increases.
Takeaway
This study shows that switching treatments can change how many people you need in a trial and how likely it is to find a real effect.
Methodology
A simulation-based approach was used to calculate power and sample size for non-inferiority trials with treatment switching.
Potential Biases
Participants who switch treatments may have different survival outcomes than those who do not, potentially biasing results.
Limitations
The assumption that treatment effects are the same for all participants may not hold true, especially for those switching treatments.
Participant Demographics
The study involved 1,234 women with early-stage breast cancer and 463 patients with chemotherapy-refractory metastatic colorectal cancer.
Statistical Information
P-Value
0.005
Statistical Significance
p<0.05
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website