Restricted accessResearch articleFirst published online 2025-3
Multicategory matched learning for estimating optimal individualized treatment rules in observational studies with application to a hepatocellular carcinoma study
One primary goal of precision medicine is to estimate the individualized treatment rules that optimize patients’ health outcomes based on individual characteristics. Health studies with multiple treatments are commonly seen in practice. However, most existing individualized treatment rule estimation methods were developed for the studies with binary treatments. Many require that the outcomes are fully observed. In this article, we propose a matching-based machine learning method to estimate the optimal individualized treatment rules in observational studies with multiple treatments when the outcomes are fully observed or right-censored. We establish theoretical property for the proposed method. It is compared with the existing competitive methods in simulation studies and a hepatocellular carcinoma study.
MurphySA. Optimal dynamic treatment regimes. J R Stat Soc: Ser B (Statistical Methodological)2003; 65: 331–355.
12.
RobinsJM. Optimal structural nested models for optimal sequential decisions. In: Proceedings of the Second Seattle Symposium in Biostatistics: analysis of correlated data. Springer, pp.189–326.
13.
LuWZhangHHZengD. Variable selection for optimal treatment decision. Stat Methods Med Res2013; 22: 493–504.
14.
LaberEBLinnKAStefanskiLA. Interactive model building for q-learning. Biometrika2014; 101: 831–847.
15.
SongRWangWZengD, et al. Penalized q-learning for dynamic treatment regimens. Stat Sin2015; 25: 901.
16.
ShiCFanASongR, et al. High-dimensional a-learning for optimal dynamic treatment regimes. Ann Stat2018; 46: 925.
17.
ChenJFuHHeX, et al. Estimating individualized treatment rules for ordinal treatments. Biometrics2018; 74: 924–933.
18.
ZhangBTsiatisAALaberEB, et al. A robust method for estimating optimal treatment regimes. Biometrics2012; 68: 1010–1018.
19.
ZhangBTsiatisAALaberEB, et al. Robust estimation of optimal dynamic treatment regimes for sequential treatment decisions. Biometrika2013; 100: 681–694.
20.
ZhaoYQZengDLaberEB, et al. Doubly robust learning for estimating individualized treatment with censored data. Biometrika2015; 102: 151–168.
21.
CuiYZhuRKosorokM. Tree based weighted learning for estimating individualized treatment rules with censored data. Electron J Stat2017; 11: 3927–3953.
22.
LeeteOEKallusNHudgensMG, et al. Balanced Policy Evaluation and Learning for Right Censored Data. arXiv preprint 2019; arXiv:1911.05728.
23.
XueFZhangYZhouW, et al. Multicategory angle-based learning for estimating optimal dynamic treatment regimes with censored data. J Am Stat Assoc2022; 117: 1438–1451.
24.
XuYWoodAMSweetingMJ, et al. Optimal individualized decision rules from a multi-arm trial: a comparison of methods and an application to tailoring inter-donation intervals among blood donors in the UK. Stat Methods Med Res2020; 29: 3113–3134.
25.
LouZShaoJYuM. Optimal treatment assignment to maximize expected outcome with multiple treatments. Biometrics2018; 74: 506–516.
QiZLiuDFuH, et al. Multi-armed angle-based direct learning for estimating optimal individualized treatment rules with various outcomes. J Am Stat Assoc2020; 115: 678–691.
28.
ZhangCChenJFuH, et al. Multicategory outcome weighted margin-based learning for estimating individualized treatment rules. Stat Sin2020; 30: 1857–1879.
29.
LeeBKLesslerJStuartEA. Improving propensity score weighting using machine learning. Stat Med2010; 29: 337–346.
YangSImbensGWCuiZ, et al. Propensity score matching and subclassification in observational studies with multi-level treatments. Biometrics2016; 72: 1055–1065.
32.
LeeYLinYWahbaG. Multicategory support vector machines: theory and application to the classification of microarray data and satellite radiance data. J Am Stat Assoc2004; 99: 67–81.
ZhangCLiuYWangJ, et al. Reinforced angle-based multicategory support vector machines. J Comput Graph Stat2016; 25: 806–825.
36.
IshwaranHKogalurUBBlackstoneEH, et al. Random survival forests. Ann Appl Stat2008; 2: 841–860.
37.
ZhangNJiangWZhangY, et al. A novel tailored prognosis calculator for the individualized curative treatments allocation of HCC patients within the Milan criteria. J Hepatol2020; 73: S100–S101.
38.
ZhangNJiangWZhangY, et al. Individualised tailored assessment of therapeutic alternatives for HCC patients within the Milan criteria. Gut2020; 69: 1893–1895.
39.
StuartEA. Matching methods for causal inference: A review and a look forward. Stat Sci2010; 25: 1–21.
40.
CrammerKSingerY. On the algorithmic implementation of multiclass kernel-based vector machines. J Mach Learn Res2001; 2: 265–292.
41.
LopezMJGutmanR. Estimation of causal effects with multiple treatments: a review and new ideas. Stat Sci2017; 32: 432–454.
42.
ScotinaADGutmanR. Matching algorithms for causal inference with multiple treatments. Stat Med2019; 38: 3139–3167.
43.
ScotinaADBeaudoinFLGutmanR. Matching estimators for causal effects of multiple treatments. Stat Methods Med Res2020; 29: 1051–1066.
44.
TherneauTMGrambschPM. Modeling Survival Data: Extending the Cox Model. New York: Springer, 2000.
45.
AtheySWagerS. Policy learning with observational data. Econometrica2021; 89: 133–161.
46.
AntonelliJCefaluMPalmerN, et al. Doubly robust matching estimators for high dimensional confounding adjustment. Biometrics2018; 74: 1171–1179.
47.
WuPZengDFuH, et al. On using electronic health records to improve optimal treatment rules in randomized trials. Biometrics2020; 76: 1075–1086.
48.
ZhangCPhamMFuS, et al. Robust multicategory support vector machines using difference convex algorithm. Math Program2018; 169: 277–305.
49.
BaiXTsiatisAALuW, et al. Optimal treatment regimes for survival endpoints using a locally-efficient doubly-robust estimator from a classification perspective. Lifetime Data Anal2017; 23: 585–604.
50.
JiangRLuWSongR, et al. On estimation of optimal treatment regimes for maximizing t-year survival probability. J R Stat Soc: Ser B (Statistical Methodological)2017; 79: 1165–1185.
Supplementary Material
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.
