Application of restricted mean survival time in clinical follow-up study.
10.3760/cma.j.issn.0254-6450.2019.02.023
- Author:
Z J YANG
1
;
J J LYU
1
;
Y W HOU
2
;
Z CHEN
1
Author Information
1. Department of Biostatistics, School of Public Health, Southern Medical University, Guangzhou 510515, China.
2. Department of Statistics, School of Economics, Jinan University, Guangzhou 510632, China.
- Publication Type:Journal Article
- Keywords:
Hazard ratio;
Proportional hazards assumption;
Restricted mean survival time;
Survival analysis
- MeSH:
Clinical Trials as Topic;
Endpoint Determination/methods*;
Follow-Up Studies;
Humans;
Kaplan-Meier Estimate;
Prognosis;
Proportional Hazards Models;
Regression Analysis;
Survival Rate/trends*;
Treatment Outcome
- From:
Chinese Journal of Epidemiology
2019;40(2):247-250
- CountryChina
- Language:Chinese
-
Abstract:
In clinical follow-up studies, hazard ratio (HR) is routinely used to quantify the differences between-groups, however, it is being estimated by the Cox procedure. HR, the ratio of two hazard functions has abstract meaning only and is in lack of the context to give an intuitive explanation of the survival of patients and the assumption of proportional hazards (PH) must be satisfied. Under this context, the restricted mean survival time (RMST) can be used as a relatively effective measure or index of statistics. This paper introduces the RMST-based statistical analysis methods, including estimation of RMST and its difference, hypothesis testing and regression analysis. The application of RMST in data analysis is also introduced. All the evidence demonstrates that RMST can be used as an effective analytical tool with straightforward interpretation. RMST is also more effective than HR in comparing differences between groups, when non-PH is observed. Therefore, RMST is suggested to be stated along with HR in the process of disease efficacy evaluation and prognosis analysis. Cooperation and complement of the two, a precise reflection on the characteristics of data can be expected.
