Premium
A Test for Discriminating between Additive and Multiplicative Relative Risks in Survival Analysis
Author(s) -
Arjas Elja,
Venzon David
Publication year - 1988
Publication title -
journal of the royal statistical society: series c (applied statistics)
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.205
H-Index - 72
eISSN - 1467-9876
pISSN - 0035-9254
DOI - 10.2307/2347489
Subject(s) - multiplicative function , test (biology) , statistics , survival analysis , mathematics , econometrics , biology , ecology , mathematical analysis
SUMMARY The proportional hazards model of Cox (1972) specified a relative risk of the form exp( β'z ) for a covariate vector z. We propose a test to detect true relative risks that are a less convex function of β'z than the exponential, e.g. a linear function. The technique exploits the expected overestimation of conditional probabilities of failure at the ends of the range of values of β'z. Simulations indicate that the test is slightly conservative and illustrate how power is lost with certain covariate distributions and censoring patterns.