Confidence interval for logrank test

[Dorothea Ekoka Mbassi]

Dear community,

I am performing survival analyses for a clinical trial comparing parasiticidal performance of the drug ivermectin vs. placebo. As per analysis plan, I need to compare survival in the two groups by logrank test and give a p-value and confidence interval. Performing the logrank test by sts test yields:

Log-rank test for equality of survivor functions

           |   Events         Events
rand       |  observed       expected
-----------+-------------------------
ivermectin |        15          14.43
placebo    |        12          12.57
-----------+-------------------------
Total      |        27          27.00

                 chi2(1) =       0.08
                 Pr>chi2 =     0.7706

.

Do I correctly assume that Pr>chi2, thus 0.7706, is my p-value? How do I adequately obtain the corresponding confidence interval? I only found out how to display CIs in my graph, but this is not what I need.

Many thanks in advance for Your advice!

[Rich Goldstein]

yes, that is your p-value

CI's are intervals around estimates and this test has no estimate so there is no CI; you might, however, be able to get an estimate and CI from one of the survival regression commands (e.g., under certain conditions, the log-rank test matches what you would get from -stcox-)

[Dorothea Ekoka Mbassi]

Thank You Rich, I will consider expanding my analysis by cox regression!