Central to personalized medicine and tailored therapies is discovering the subpopulations that account for treatment effect heterogeneity and are likely to benefit more from given interventions. In this article, we introduce a change plane model averaging method to identify subgroups characterized by linear combinations of predictive variables and multiple cut-offs. We first fit a sequence of statistical models, each incorporating the thresholding effect of one particular covariate. The estimation of submodels is accomplished through an iterative integration of a change point detection method and numerical optimization algorithms. A frequentist model averaging approach is then employed to linearly combine the submodels with optimal weights. Our approach can deal with high-dimensional settings involving enormous potential grouping variables by adopting the sparsity-inducing penalties. Simulation studies are conducted to investigate the prediction and subgrouping performance of the proposed method, with a comparison to various competing subgroup detection methods. Our method is applied to a dataset from a warfarin pharmacogenetics study, producing some new findings.
KosorokMRLaberEB. Precision medicine. Annu Rev Stat Appl2019; 6: 263–286.
2.
QianMMurphySA. Performance guarantees for individualized treatment rules. Ann Stat2011; 39: 1180–1210.
3.
ZhangBTsiatisAALaberEB, et al. A robust method for estimating optimal treatment regimes. Biometrics2012; 68: 1010–1018.
4.
ZhaoYZengDRushAJ, et al. Estimating individualized treatment rules using outcome weighted learning. J Am Stat Assoc2012; 107: 1106–1118.
5.
ZhouXMayer-HamblettNKhanU, et al. Residual weighted learning for estimating individualized treatment rules. J Am Stat Assoc2017; 112: 169–187.
6.
FosterJCTaylorJMRubergSJ. Subgroup identification from randomized clinical trial data. Stat Med2011; 30: 2867–2880.
7.
NegassaACiampiAAbrahamowiczM, et al. Tree-structured subgroup analysis for censored survival data: validation of computationally inexpensive model selection criteria. Stat Comput2005; 15: 231–239.
8.
SuXZhouTYanX, et al. Interaction trees with censored survival data. Int J Biostat2008; 4.
9.
DusseldorpEVan MechelenI. Qualitative interaction trees: a tool to identify qualitative treatment–subgroup interactions. Stat Med2014; 33: 219–237.
10.
LohWYHeXManM. A regression tree approach to identifying subgroups with differential treatment effects. Stat Med2015; 34: 1818–1833.
FriedmanJHFisherNI. Bump hunting in high-dimensional data. Stat Comput1999; 9: 123–143.
13.
LipkovichIDmitrienkoADenneJ, et al. Subgroup identification based on differential effect search—a recursive partitioning method for establishing response to treatment in patient subpopulations. Stat Med2011; 30: 2601–2621.
14.
HuangXSunYTrowP, et al. Patient subgroup identification for clinical drug development. Stat Med2017; 36: 1414–1428.
15.
SysoevOBartoszekKEkströmEC, et al. Psica: decision trees for probabilistic subgroup identification with categorical treatments. Stat Med2019; 38: 4436–4452.
16.
YangJDahabrehIJSteingrimssonJA. Causal interaction trees: finding subgroups with heterogeneous treatment effects in observational data. Biometrics2021; 78: 624–635.
17.
CaiTTianLWongPH, et al. Analysis of randomized comparative clinical trial data for personalized treatment selections. Biostatistics2011; 12: 270–282.
18.
TianLAlizadehAAGentlesAJ, et al. A simple method for estimating interactions between a treatment and a large number of covariates. J Am Stat Assoc2014; 109: 1517–1532.
19.
GuoCLiJ. Homogeneity and structure identification in semiparametric factor models. J Bus Econ Stat2022; 40: 408–422.
20.
AltsteinLLiG. Latent subgroup analysis of a randomized clinical trial through a semiparametric accelerated failure time mixture model. Biometrics2013; 69: 52–61.
21.
ShenJHeX. Inference for subgroup analysis with a structured logistic-normal mixture model. J Am Stat Assoc2015; 110: 303–312.
22.
BergerJOWangXShenL. A Bayesian approach to subgroup identification. J Biopharm Stat2014; 24: 110–129.
23.
JonesHEOhlssenDINeuenschwanderB, et al. Bayesian models for subgroup analysis in clinical trials. Clinical Trials2011; 8: 129–143.
24.
LipkovichIDmitrienkoAB D’Agostino SrR. Tutorial in biostatistics: data-driven subgroup identification and analysis in clinical trials. Stat Med2017; 36: 136–196.
LiJJinB. Multi-threshold accelerated failure time model. Ann Stat2018; 46: 2657–2682.
35.
WangJLiJLiY, et al. A model-based multithreshold method for subgroup identification. Stat Med2019; 38: 2605–2631.
36.
FanASongRLuW. Change-plane analysis for subgroup detection and sample size calculation. J Am Stat Assoc2017; 112: 769–778.
37.
WeiSKosorokMR. Latent supervised learning. J Am Stat Assoc2013; 108: 957–970.
38.
WeiSKosorokMR. The change-plane cox model. Biometrika2018; 105: 891–903.
39.
HeYLinHTuD. A single-index threshold Cox proportional hazard model for identifying a treatment-sensitive subset based on multiple biomarkers. Stat Med2018; 37: 3267–3279.
40.
LiJLiYJinB, et al. Multithreshold change plane model: estimation theory and applications in subgroup identification. Stat Med2021; 40: 3440–3459.
41.
HuangYChoJFongY. Threshold-based subgroup testing in logistic regression models in two-phase sampling designs. J R Stat Soc2021; 70: 291–311.
42.
FragosoTMBertoliWLouzadaF. Bayesian model averaging: a systematic review and conceptual classification. Int Stat Rev2018; 86: 1–28.
43.
HansenBE. Least squares model averaging. Econometrica2007; 75: 1175–1189.
44.
LiangHZouGWanAT, et al. Optimal weight choice for frequentist model average estimators. J Am Stat Assoc2011; 106: 1053–1066.
45.
HansenBERacineJS. Jackknife model averaging. J Econom2012; 167: 38–46.
46.
ZhangXWanATZouG. Model averaging by jackknife criterion in models with dependent data. J Econom2013; 174: 82–94.
47.
GaoYZhangXWangS, et al. Model averaging based on leave-subject-out cross-validation. J Econom2016; 192: 139–151.
48.
AndoTLiKC. A model-averaging approach for high-dimensional regression. J Am Stat Assoc2014; 109: 254–265.
49.
YanXWangHWangW, et al. Optimal model averaging forecasting in high-dimensional survival analysis. Int J Forecast2021; 37: 1147–1155.
50.
SunYHongYLeeTH, et al. Time-varying model averaging. J Econom2021; 222: 974–992.
51.
KruseRMSilbersdorffASäfkenB. Model averaging for linear mixed models via augmented lagrangian. Comput Stat Data Anal2022; 167: 107351.
52.
HuangTLiJ. Semiparametric model average prediction in panel data analysis. J Nonparametr Stat2018; 30: 125–144.
53.
LiJLvJWanAT, et al. Adaboost semiparametric model averaging prediction for multiple categories. J Am Stat Assoc2022; 117: 495–509.
54.
LiJXiaXWongWK, et al. Varying-coefficient semiparametric model averaging prediction. Biometrics2018; 74: 1417–1426.
55.
GaoYZhangXWangS, et al. Frequentist model averaging for threshold models. Ann Inst Stat Math2019; 71: 275–306.
56.
DeanL. Warfarin therapy and the genotypes cyp2c9 and vkorc1. Medical Genetics Summaries2012; 2: 257–263.
57.
SconceEAKhanTIWynneHA, et al. The impact of cyp2c9 and vkorc1 genetic polymorphism and patient characteristics upon warfarin dose requirements: proposal for a new dosing regimen. Blood2005; 106: 2329–2333.
58.
LimdiNABrownTMYanQ, et al. Race influences warfarin dose changes associated with genetic factors. Blood2015; 126: 539–545.
59.
ZhangYWangHJZhuZ. Single-index thresholding in quantile regression. J Am Stat Assoc2022; 117: 2222–2237.
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.
