Related Experiment Videos
Using permutation tests and bootstrap confidence limits to analyze repeated events data from clinical trials.
L Freedman1, R Sylvester, D P Byar
1Biometry Branch, National Cancer Institute, Bethesda, Maryland 20892.
Controlled Clinical Trials
|June 1, 1989
Summary
This study introduces a new nonparametric statistical method for analyzing recurrent bladder cancer in clinical trials. This approach helps compare treatment effectiveness when standard methods are insufficient.
Area of Science:
- Oncology
- Biostatistics
- Clinical Trials
Background:
- Superficial bladder cancer patients face high recurrence rates, necessitating robust statistical analysis for treatment comparison.
- Existing statistical methods may be inadequate for analyzing recurrent event data in clinical trials.
Purpose of the Study:
- To present a novel nonparametric statistical approach for analyzing recurrent events in superficial bladder cancer clinical trials.
- To provide methods for comparing recurrence rates and estimating rate ratios between treatment groups.
Main Methods:
- A nonparametric statistical test utilizing randomization distribution to compare recurrence rates.
- Bootstrap distribution for determining confidence intervals of the rate ratio.
- Monte Carlo methods for implementing the statistical tests and confidence intervals.
Main Results:
- Computer simulations indicate the nonparametric methods are reliable with over 60 recurrences per group.
- An illustrative example demonstrates the practical application of the proposed methods.
Conclusions:
- The presented nonparametric approach offers a viable alternative for analyzing recurrent event data in bladder cancer trials.
- This statistical strategy can be adapted for other clinical trials with similar data structures where standard methods fail.