We present strmst2, a new command to implement k-sample comparisons using the restricted mean survival time (RMST) as the summary measure of the survival-time distribution. Unlike model-based summary measures such as the hazard ratio, the validity of which relies on the adequacy of the proportional-hazards assumption, the measures based on the RMST (that is, the difference in RMST, the ratio of RMST, and the ratio of the restricted mean time lost) provide more robust and clinically interpretable results about the between-group differences. strmst2 performs analysis of covariance-type adjusted analyses for k-sample comparisons as well as the corresponding unadjusted analyses. Pairwise comparisons for the adjusted analyses are summarized using the new postestimation command strmst2pw. We briefly describe the issues of the hazard ratio, introduce details of the method for the RMST, and then illustrate how to use the new commands.
AndersenP. K., and GillR. D.1982. Cox's regression model for counting processes: A large sample study. Annals of Statistics10: 1100–1120.
2.
AndersenP. K., HansenM. G., and KleinJ. P.2004. Regression analysis of restricted mean survival time based on pseudo-observations. Lifetime Data Analysis10: 335–350.
3.
CoxD. R.1972. Regression models and life-tables. Journal of the Royal Statistical Society, Series B34: 187–220.
4.
FlemingT. R., and HarringtonD. P.2011. Counting Processes and Survival Analysis. 2nd ed. Hoboken, NJ: Wiley.
5.
GillR. D.1984. Understanding Cox's regression model: A Martingale approach. Journal of the American Statistical Association79: 441–447.
6.
HernánM. A.2010. The hazards of hazard ratios. Epidemiology21: 13–15.
7.
HjortN. L.1992. On inference in parametric survival data models. International Statistical Review60: 355–387.
8.
KleinJ. P., GersterM., AndersenP. K., TarimaS., and Pohar PermeM.2008. SAS and R functions to compute pseudo-values for censored data regression. Computer Methods and Programs in Biomedicine89: 289–300.
9.
LinD. Y., and WeiL. J.1989. The robust inference for the Cox proportional hazards model. Journal of the American Statistical Association84: 1074–1078.
ParnerE. T., and AndersenP. K.2010. Regression analysis of censored data using pseudo-observations. Stata Journal10: 408–422.
12.
RoystonP., and ParmarM. K. B.2011. The use of restricted mean survival time to estimate the treatment effect in randomized clinical trials when the proportional hazards assumption is in doubt. Statistics in Medicine30: 2409–2421.
13.
RoystonP., and ParmarM. K. B.2013. Restricted mean survival time: An alternative to the hazard ratio for the design and analysis of randomized trials with a time-to-event outcome. BMC Medical Research Methodology13: 152.
14.
TianL., ZhaoL., and WeiL. J.2014. Predicting the restricted mean event time with the subject's baseline covariates in survival analysis. Biostatistics15: 222–233.
15.
UnoH., ClaggettB., TianL., InoueE., GalloP., MiyataT., SchragD., TakeuchiM., UyamaY., ZhaoL., SkaliH., SolomonS., JacobusS., HughesM., PackerM., and WeiL.-J.2014. Moving beyond the hazard ratio in quantifying the between-group difference in survival analysis. Journal of Clinical Oncology32: 2380–2385.
16.
UnoH., WittesJ., FuH., SolomonS. D., ClaggettB., TianL., CaiT., PfefferM. A., EvansS. R., and WeiL.-J.2015. Alternatives to hazard ratios for comparing the efficacy or safety of therapies in noninferiority studies. Annals of Internal Medicine163: 127–134.
17.
ZuckerD. M.1998. Restricted mean life with covariates: Modification and extension of a useful survival analysis method. Journal of the American Statistical Association93: 702–709.
