作者: David P. Byar , Donald K. Corle
DOI: 10.1016/0021-9681(77)90037-6
关键词: Statistics 、 Hazard (logic) 、 Selection (genetic algorithm) 、 Optimal treatment 、 Clinical trial 、 Survivorship curve 、 Medicine 、 Covariate 、 Epidemiology
摘要: Abstract We are interested in the question ‘which treatment is best for which kinds of patients?’ rather than classical question, overall?’. A survivorship function hazard a covariates fitted both all treatments combined and separate treatments. Likelihood ratio tests used to detect significant treatment-covariate interactions. If there none, we test overall treatment. Otherwise define an optimal each patient. By studying combinations lead selection various as optimal, make recommendations different patients.