Instructions for using CI-2sample.exe :
This program calculates the 95% confidence interval for 100(exp{β}-1)%, the percentage increase in treatment survival time over control survival time under the accelerated failure time model. The results are output into the file result.txt. In each instance the CI using the Skovgaard
saddlepoint approximation and the normal approximation are computed for one of the following tests:
log-rank (lgr)
Gehan (Ghn)
generalized Wilcoxon (Gwl)
Tarone-Ware (Twr)
Harrington-Fleming (flh).
From a practical point of view, the 95% confidence interval
for 100(exp{β}-1)% is more meaningful than the 95% confidence interval for
β. To see this, let T1 and T2 be control and
treatment survival times respectively with ε as the error distribution on
the log-scale. The ratio of means (or medians for that matter) is
E(T1))/(E(T2)
= exp{μ+β}E(exp{ ε })/(exp{μ}E(exp{ ε })) = exp{β}
so that 100(exp{β}-1)% is the percentage increase in survival time due to the treatment effect. Ninety-five percent confidence intervals for this percentage are output by this software.
The data are input using the file "data.txt" that should be placed in the same directory as the executable file.
A two sample example with data set n=40, n1=21 is indicated at the bottom of this file.
The format for data entry is as follows:
Each row holds data for one patient and patients are ordered according to time of death/censoring.
Each line has the following 3 entries:
time (space) indicator of death (space) indicator of the treatment group.
Note: For ties involving one death time and censored values, increment the censored times by a small amount,
like 10**(-10), and place them after the death time. This program does not deal with tied data.
Upon execution of the file, the user will be asked to enter the values when so prompted:
enter total number of points
40
enter number of points in treatment group
21
enter the test name : lgr, gwl, ghn, twr, or flh
gwl
Happy Trails!
Contents of the file data.txt:
142 1 1
143 1 0
156 1 1
163 1 1
164 1 0
188 1 0
188 1 0
190 1 0
192 1 0
198 1 1
204 0 1
205 1 1
206 1 0
209 1 0
213 1 0
216 1 0
216 0 0
220 1 0
227 1 0
230 1 0
232 1 1
232 1 1
233 1 1
233 1 1
233 1 1
233 1 1
234 1 0
239 1 1
240 1 1
244 0 0
246 1 0
261 1 1
265 1 0
280 1 1
280 1 1
296 1 1
296 1 1
304 1 0
323 1 1
344 0 1