Comparing survival curves based on medians

BACKGROUND: Although some nonparametric methods have been proposed in the literature to test for the equality of median survival times for censored data in medical research, in general they have inflated type I error rates, which make their use limited in practice, especially when the sample sizes are small. METHODS: In this paper, we propose a new nonparametric test with a simple test statistic to compare median survival times. RESULTS: The results from our comprehensive simulation study show the new test controls type I error rate very well under all situations considered, even for small sample sizes. In addition, it has comparable detecting power compared to existing methods. Another advantage of the proposed test is its relatively simple formula that requires less computation. CONCLUSIONS: We propose a new statistical method for comparing survival curves based on their medians. The new method can be easily implemented and applied to censored event time data.