Be informed about the statistical group comparison for survival times

 

🎓️ In our last post we looked at Kaplan-Meier Curves. These are useful for getting a first impression.

❓️ But how do we actually test for differences in the survival time between groups? 

 

Statistical Method:

📊 ️Log-Rank Test: The log-rank test tests for a difference in the survival time between two or more groups. It is basically a Χ² test: it looks at the difference between the number of observed events and the number of expected events under the null-hypothesis.

📊 If the event rates of the different groups are not proportional (you can see this by looking at the Kaplan-Meier curves!), the log-rank test is not the best option. In this case you can use for example the generalized Wilcoxon test. 

📊 The log-rank test weights all events equally but there might be situations in which it makes sense that early events have a higher weight. This is done in the generalized Wilcoxon test and the Tarone-Ware test.

 

Take Home Message:

🛠️ Know your tools: there are different tests for comparing the survival times of groups. Choose the one which fits best the situation!

 

 📢 Stay tuned and continue building statistical confidence for market access.