Statistics review 12 Survival analysis 英文参考文献.docVIP

Available online /content/8/5/389 Review Statistics review 12: Survival analysis Viv Bewick1, Liz Cheek1 and Jonathan Ball2 1Senior Lecturer, School of Computing, Mathematical and Information Sciences, University of Brighton, Brighton, UK 2Senior Registrar in ICU, Liverpool Hospital, Sydney, Australia Correspondence: Viv Bewick, v.bewick@brighton.ac.uk Published online: 6 September 2004 Critical Care 2004, 8:389-394 (DOI 10.1186/cc2955) This article is online at /content/8/5/389 ? 2004 BioMed Central Ltd Abstract This review introduces methods of analyzing data arising from studies where the response variable is the length of time taken to reach a certain end-point, often death. The Kaplan–Meier methods, log rank test and Cox’s proportional hazards model are described. Keywords Cox’s proportional-hazards model, cumulative hazard function H(t), hazard ratio, Kaplan–Meier method, log rank test, survival function S(t) Introduction The graph of S(t) against t is called the survival curve. The Kaplan–Meier method can be used to estimate this curve from the observed survival times without the assumption of an underlying probability distribution. The method is based on the basic idea that the probability of surviving k or more periods from entering the study is a product of the k observed survival rates for each period (i.e. the cumulative proportion surviving), given by the following: Survival times are data that measure follow-up time from a defined starting point to the occurrence of a given event, for example the time from the beginning to the end of a remission period or the time from the diagnosis of a disease to death. Standard statistical techniques cannot usually be applied because the underlying distribution is rarely Normal and the data are often ‘censored’. A survival time is described as censored when there is a follow-up time but the even