Individualized treatment rules inform tailored treatment decisions based on the patient’s information, where the goal is to optimize clinical benefit for the population. When the clinical outcome of interest is survival time, most of current approaches typically aim to maximize the expected time of survival. We propose a new criterion for constructing Individualized treatment rules that optimize the clinical benefit with survival outcomes, termed as the adjusted probability of a longer survival. This objective captures the likelihood of living longer with being on treatment, compared to the alternative, which provides an alternative and often straightforward interpretation to communicate with clinicians and patients. We view it as an alternative to the survival analysis standard of the hazard ratio and the increasingly used restricted mean survival time. We develop a new method to construct the optimal Individualized treatment rule by maximizing a nonparametric estimator of the adjusted probability of a longer survival for a decision rule. Simulation studies demonstrate the reliability of the proposed method across a range of different scenarios. We further perform data analysis using data collected from a randomized Phase III clinical trial (SWOG S0819).
QianMMurphySA. Performance guarantees for individualized treatment rules. Ann Stat2011; 39: 1180.
2.
ZhaoYZengDRushAJet al. Estimating individualized treatment rules using outcome weighted learning. J Am Stat Assoc2012; 107: 1106–1118.
3.
LinnKALaberEBStefanskiLA. Interactive Q-learning for quantiles. J Am Stat Assoc2017; 112: 638–649.
4.
WangLZhouYSongRet al. Quantile-optimal treatment regimes. J Am Stat Assoc2018; 113: 1243–1254.
5.
MoodieEEDeanNSunYR. Q-learning: flexible learning about useful utilities. Stat Biosci2014; 6: 223–243.
6.
MurphySA. Optimal dynamic treatment regimes. J Royal Stat Soc: Ser B (Stat Methodol)2003; 65: 331–355.
7.
BlattDMurphySAZhuJ. A-learning for approximate planning. Ann Arbor2004; 1001: 48109–42122.
8.
RobinsJM. Optimal structural nested models for optimal sequential decisions. In Proceedings of the second seattle Symposium in Biostatistics. Springer, pp.189–326.
9.
HendersonRAnsellPAlshibaniD. Regret-regression for optimal dynamic treatment regimes. Biometrics2010; 66: 1192–1201.
10.
DavidianMTsiatisAALaberEB. Chapter 9: Value search estimators for optimal dynamic treatment regimes. In Adaptive Treatment Strategies in Practice: Planning Trials and Analyzing Data for Personalized Medicine. SIAM, 2015. pp.135–155.
11.
LaberEBLizotteDJQianMet al. Dynamic treatment regimes: technical challenges and applications. Electron J Stat2014; 8: 1225.
12.
OrellanaLRotnitzkyARobinsJM. Dynamic regime marginal structural mean models for estimation of optimal dynamic treatment regimes, part i: main content. Int J Biostat2010; 6. Article 8.
13.
ZhangBTsiatisAADavidianMet al. Estimating optimal treatment regimes from a classification perspective. Stat2012; 1: 103–114.
14.
ZhangBTsiatisAALaberEBet al. A robust method for estimating optimal treatment regimes. Biometrics2012; 68: 1010–1018.
15.
ZhouXMayer-HamblettNKhanUet al. Residual weighted learning for estimating individualized treatment rules. J Am Stat Assoc2017; 112: 169–187.
16.
ZhaoYQLaberEBNingYet al. Efficient augmentation and relaxation learning for individualized treatment rules using observational data. J Mach Learn Res2019; 20: 1821–1843.
17.
KalbfleischJDPrenticeRL. The statistical analysis of failure time data. Hoboken, New Jersey: John Wiley & Sons, 2011.
18.
CoxDR. Regression models and life-tables. J Royal Stat Soc: Ser B (Methodol)1972; 34: 187–202.
19.
PakKUnoHKimDHet al. Interpretability of cancer clinical trial results using restricted mean survival time as an alternative to the hazard ratio. JAMA Oncol2017; 3: 1692–1696.
20.
KloeckerDEDaviesMJKhuntiKet al. Uses and limitations of the restricted mean survival time: illustrative examples from cardiovascular outcomes and mortality trials in type 2 diabetes. Ann Intern Med2020; 172: 541–552.
21.
PéronJRoyPOzenneBet al. The net chance of a longer survival as a patient-oriented measure of treatment benefit in randomized clinical trials. JAMA Oncol2016; 2: 901–905.
22.
ZhaoYQRedmanMWLeBlancML. Quantifying treatment effects using the personalized chance of longer survival. Stat Med2019; 38: 5317–5331.
GoldbergYKosorokMR. Q-learning with censored data. Ann Stat2012; 40: 529.
25.
ZhaoYQZengDLaberEBet al. Doubly robust learning for estimating individualized treatment with censored data. Biometrika2015; 102: 151–168.
26.
JiangRLuWSongRet al. On estimation of optimal treatment regimes for maximizing t-year survival probability. J Royal Stat Soc: Ser B (Stat Methodol)2017; 79: 1165–1185.
27.
RubinDB. Estimating causal effects of treatments in randomized and nonrandomized studies. J Educ Psychol1974; 66: 688.
28.
Splawa-NeymanJDabrowskaDMSpeedTP. On the application of probability theory to agricultural experiments (engl. transl. by d.m. dabrowska and t.p. speed). Statist Sci1990; 5: 465–472.
29.
RobinsJ. A new approach to causal inference in mortality studies with a sustained exposure period—application to control of the healthy worker survivor effect. Math Model1986; 7: 1393–1512.
30.
Splawa-NeymanJDabrowskaDMSpeedT. On the application of probability theory to agricultural experiments. Essay on principles. section 9. Stat Sci1990.
31.
FanYParkSS. Sharp bounds on the distribution of treatment effects and their statistical inference. Econ Theory2010; 26: 931–951.
32.
FanYGuerreEZhuD. Partial identification of functionals of the joint distribution of “potential outcomes”. J Econom2017; 197: 42–59.
33.
FirpoSRidderG. Partial identification of the treatment effect distribution and its functionals. J Econom2019; 213: 210–234.
34.
FrandsenBRLefgrenLJ. Partial identification of the distribution of treatment effects with an application to the knowledge is power program (kipp). Quant Econom2021; 12: 143–171.
35.
RussellTM. Sharp bounds on functionals of the joint distribution in the analysis of treatment effects. J Bus Econ Stat2021; 39: 532–546.
36.
FanYShiXTaoJ. Partial identification and inference in moment models with incomplete data. J Econom2023; 235: 418–443.
Ben-MichaelEImaiKJiangZ. Policy learning with asymmetric counterfactual utilities. J Am Stat Assoc2024; In press.
40.
AbbringJHHeckmanJJ. Econometric evaluation of social programs, part III: distributional treatment effects, dynamic treatment effects, dynamic discrete choice, and general equilibrium policy evaluation. Handbook Econom2007; 6: 5145–5303.
41.
RobinsJMRotnitzkyAZhaoLP. Analysis of semiparametric regression models for repeated outcomes in the presence of missing data. J Am Stat Assoc1995; 90: 106–121.
42.
NadarayaEA. On estimating regression. Theory Probab App1964; 9: 141–142.
43.
WatsonGS. Smooth regression analysis. Sankhyā: Indian J Stat, Ser A1964.
44.
ShaliziC. Advanced data analysis from an elementary point of view 2013.
45.
KaplanELMeierP. Nonparametric estimation from incomplete observations. J Am Stat Assoc1958; 53: 457–481.
46.
MitchellM. An introduction to genetic algorithms. USA: MIT Press, 1998.
47.
ZhangT. Statistical behavior and consistency of classification methods based on convex risk minimization. Ann Stat2004; 32: 56–85.
48.
LugosiGVayatisN. On the Bayes-risk consistency of regularized boosting methods. Ann Stat2004; 32: 30–55.
49.
HastieTTibshiraniRFriedmanJHet al. The elements of statistical learning: data mining, inference, and prediction, vol. 2. New York: Springer, 2009.
50.
ByrdRHLuPNocedalJet al. A limited memory algorithm for bound constrained optimization. SIAM J Sci Comput1995; 16: 1190–1208.
51.
LandgrafAJLeeY. Dimensionality reduction for binary data through the projection of natural parameters. J Multivar Anal2020; 180: 104668.
52.
HerbstRSRedmanMWKimESet al. Cetuximab plus carboplatin and paclitaxel with or without bevacizumab versus carboplatin and paclitaxel with or without bevacizumab in advanced nsclc (SWOG S0819): a randomised, phase 3 study. Lancet Oncol2018; 19: 101–114.
53.
RedmanMWCrowleyJJHerbstRSet al. Design of a phase iii clinical trial with prospective biomarker validation: SWOG S0819. Clin Cancer Res2012; 18: 4004–4012.
54.
LouZShaoJYuM. Optimal treatment assignment to maximize expected outcome with multiple treatments. Biometrics2018; 74: 506–516.
55.
ZhangCChenJFuHet al. Multicategory outcome weighted margin-based learning for estimating individualized treatment rules. Stat Sin2020; 30: 1857.
56.
TibshiraniR. Regression shrinkage and selection via the lasso. J Royal Stat Soc: Ser B (Methodol)1996; 58: 267–288.
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
