Exponential survivals with censoring and explanatory variables

Abstract

SUMMARY Consideration is given to the analysis of survival times, censored and uncensored, arising from a hazard function that varies exponentially with covariates and is otherwise constant. Significance tests and statements of probabilistic inference are derived for covariate parameters and for the hazard function at average values of the covariate. The results are extended to include a Weibull component in the hazard. Clinical trials on advanced lung cancer patients serve to illustrate the suggested procedures.