×

Evaluating the correlation coefficient between bivariate survival times – a copula-based approach. (English) Zbl 1438.62205

Summary: The analysis of correlations within pairs of survival times is of great interest to many researchers in biology and medicine. The analysis objective is to investigate the association of bivariate survival data under the setting of low-moderate percentage of censoring through Monte Carlo simulations using a copula approach. Here the association of bivariate survival data is estimated using Spearman’s correlation coefficient. The results from simulation studies show that when the percentage of censoring is low, Gumbel-based estimation procedure is much more robust, and the stronger a positive association is, the more accurate estimate can be obtained when the censoring percentage is 0% and 30%. This is true for the Frank, Gumbel and Clayton-based estimation procedures under the condition that the copula assumption made here is the same as the true one.

MSC:

62P10 Applications of statistics to biology and medical sciences; meta analysis
92B15 General biostatistics
62H05 Characterization and structure theory for multivariate probability distributions; copulas
62N05 Reliability and life testing
62N01 Censored data models
62H20 Measures of association (correlation, canonical correlation, etc.)
PDFBibTeX XMLCite
Full Text: DOI