Kengo Nagashima



Brief description

This web application is an implementation of sample size calculation methods for one sample non-parametric survival test/confidence interval (based on the Kaplan–Meier estimator) in JavaScript.

A large sample approximation to the variance of the Kaplan–Meier estimator, an exponential survival distribution, and a uniform entry over [0, accural time] are assumed. For theoretical background, see Fleming & Harrington (1991) and Andersen, Borgan, Gill & Keiding (1993). Moreover, various transformations for the Kaplan–Meier estimator are supported in this application. Log-minus-log, logit, and arcsine square-root transformed confidence intervals have better performance than linear and log transformed confidence intervals (Bie et al., 1987; Borgan & Liestøl, 1990). The required sample size and the performance depend on the method of the transformation. The well known SWOG's calculator (One Sample Nonparametric Survival) use the log transformation, but a sample size formula different form this application is used. We submitted a research paper to a journal about this results with numerical evaluations via simulations. As a result, empirical power of the sample size formula with the arcsine square-root transformation is close to the nominal power than the other transformations. Therefore, this application uses the arcsine square-root transformation as default. When performing analysis, it is reccomended to use the arcsine square-root transformation or more conservative (i.e., log-minus-log or logit) transformation.

Application

Inputs

Results

Conditions

References

  1. Fleming TR, Harrington DP. Counting Processes and Survival Analysis. New York: Wiley, 1991, 236–237, Example 6.3.1.
  2. Andersen PK, Borgan Ø, Gill RD, Keiding N. Statistical Models Based on Counting Processes. New York: Springer-Verlag, 1993, 176–287, Section IV.1–3.
  3. Bie O, Borgan Ø, Liestøl K. Confidence intervals and confidence bands for the cumulative hazard rate function and their small sample properties. Scandinavian Journal of Statistics 1987; 14(3): 221–233.
  4. Borgan Ø, Liestøl K. A note on confidence intervals and bands for the survival function based on transformations. Scandinavian Journal of Statistics 1990; 17(1): 35–41.

Update history

To cite this page