×

zbMATH — the first resource for mathematics

Using decision lists to construct interpretable and parsimonious treatment regimes. (English) Zbl 1419.62490
Summary: A treatment regime formalizes personalized medicine as a function from individual patient characteristics to a recommended treatment. A high-quality treatment regime can improve patient outcomes while reducing cost, resource consumption, and treatment burden. Thus, there is tremendous interest in estimating treatment regimes from observational and randomized studies. However, the development of treatment regimes for application in clinical practice requires the long-term, joint effort of statisticians and clinical scientists. In this collaborative process, the statistician must integrate clinical science into the statistical models underlying a treatment regime and the clinician must scrutinize the estimated treatment regime for scientific validity. To facilitate meaningful information exchange, it is important that estimated treatment regimes be interpretable in a subject-matter context. We propose a simple, yet flexible class of treatment regimes whose members are representable as a short list of if-then statements. Regimes in this class are immediately interpretable and are therefore an appealing choice for broad application in practice. We derive a robust estimator of the optimal regime within this class and demonstrate its finite sample performance using simulation experiments. The proposed method is illustrated with data from two clinical trials.

MSC:
62P10 Applications of statistics to biology and medical sciences; meta analysis
PDF BibTeX XML Cite
Full Text: DOI
References:
[1] Baker, Using relative utility curves to evaluate risk prediction, Journal of the Royal Statistical Society: Series A (Statistics in Society) 172 pp 729– (2009) · Zbl 05622765 · doi:10.1111/j.1467-985X.2009.00592.x
[2] Breiman , L. Friedman , J. H. Olshen , R. A. Stone , C. J. 1984
[3] Brusco , M. J. Stahl , S. 2006
[4] Clark, The CN2 induction algorithm, Machine Learning 3 pp 261– (1989) · doi:10.1007/BF00116835
[5] Freitas, Comprehensible classification models: A position paper, ACM SIGKDD Explorations Newsletter 15 pp 1– (2014) · doi:10.1145/2594473.2594475
[6] Gail, Testing for qualitative interactions between treatment effects and patient subsets, Biometrics 41 pp 361– (1985) · Zbl 0614.62140 · doi:10.2307/2530862
[7] Gail, Value of adding single-nucleotide polymorphism genotypes to a breast cancer risk model, Journal of the National Cancer Institute 101 pp 959– (2009) · doi:10.1093/jnci/djp130
[8] Gail, Weighing the risks and benefits of tamoxifen treatment for preventing breast cancer, Journal of the National Cancer Institute 91 pp 1829– (1999) · doi:10.1093/jnci/91.21.1829
[9] Huang, Characterizing expected benefits of biomarkers in treatment selection, Biostatistics 16 pp 383– (2015) · doi:10.1093/biostatistics/kxu039
[10] Keller, A comparison of nefazodone, the cognitive behavioral-analysis system of psychotherapy, and their combination for the treatment of chronic depression, New England Journal of Medicine 342 pp 1462– (2000) · doi:10.1056/NEJM200005183422001
[11] Laber, Tree-based methods for personalized treatment regimes, Biometrika (2015) · Zbl 06519989 · doi:10.1093/biomet/asv028
[12] Laber, Interactive model building for Q-learning, Biometrika 101 pp 831– (2014) · Zbl 1306.62235 · doi:10.1093/biomet/asu043
[13] Letham , B. Rudin , C. McCormick , T. H. Madigan , D. 2012
[14] Marchand, Learning with decision lists of data-dependent features, Journal of Machine Learning Research 6 pp 427– (2005) · Zbl 1222.68257
[15] Marlowe, Adaptive programming improves outcomes in drug court an experimental trial, Criminal Justice and Behavior 39 pp 514– (2012) · doi:10.1177/0093854811432525
[16] Moodie, Q-learning for estimating optimal dynamic treatment rules from observational data, Canadian Journal of Statistics 40 pp 629– (2012) · Zbl 1349.62371 · doi:10.1002/cjs.11162
[17] Moodie, Q-learning: Flexible learning about useful utilities, Statistics in Biosciences 6 pp 1– (2013)
[18] Orellana, Dynamic regime marginal structural mean models for estimation of optimal dynamic treatment regimes, Part I: Main content, The International Journal of Biostatistics 6 (2010)
[19] Qian, Performance guarantees for individualized treatment rules, Annals of Statistics 39 pp 1180– (2011) · Zbl 1216.62178 · doi:10.1214/10-AOS864
[20] Rivest, Learning decision lists, Machine Learning 2 pp 229– (1987) · doi:10.1007/BF00058680
[21] Robins, Correcting for non-compliance in randomized trials using structural nested mean models, Communications in Statistics-Theory and methods 23 pp 2379– (1994) · Zbl 0825.62203 · doi:10.1080/03610929408831393
[22] Robins, Marginal structural models and causal inference in epidemiology, Epidemiology 11 pp 550– (2000) · doi:10.1097/00001648-200009000-00011
[23] Robins, Estimation and extrapolation of optimal treatment and testing strategies, Statistics in Medicine 27 pp 4678– (2008) · doi:10.1002/sim.3301
[24] Shiffman, Representation of clinical practice guidelines in conventional and augmented decision tables, Journal of the American Medical Informatics Association 4 pp 382– (1997) · doi:10.1136/jamia.1997.0040382
[25] Shortreed, Informing sequential clinical decision-making through reinforcement learning: An empirical study, Machine Learning 84 pp 109– (2011) · Zbl 06031592 · doi:10.1007/s10994-010-5229-0
[26] Shortreed, A multiple imputation strategy for sequential multiple assignment randomized trials, Statistics in Medicine 33 pp 4202– (2014) · doi:10.1002/sim.6223
[27] Su, Subgroup analysis via recursive partitioning, Journal of Machine Learning Research 10 pp 141– (2009)
[28] Taylor, Reader reaction to A Robust Method for Estimating Optimal Treatment Regimes, Biometrics 71 pp 267– (2014) · Zbl 1329.62438 · doi:10.1111/biom.12228
[29] Tian, A simple method for estimating interactions between a treatment and a large number of covariates, Journal of the American Statistical Association 109 pp 1517– (2014) · Zbl 1368.62294 · doi:10.1080/01621459.2014.951443
[30] Tibshirani, Regression shrinkage and selection via the lasso, Journal of the Royal Statistical Society: Series B 58 pp 267– (1996) · Zbl 0850.62538
[31] Tsiatis , A. A. 2006
[32] Vansteelandt, Structural nested models and g-estimation: The partially realized promise, Statistical Science 29 pp 707– (2014) · Zbl 1331.62208 · doi:10.1214/14-STS493
[33] Wang , F. Rudin , C. 2014
[34] Zhang, Estimating optimal treatment regimes from a classification perspective, Stat 1 pp 103– (2012) · doi:10.1002/sta.411
[35] Zhang, A robust method for estimating optimal treatment regimes, Biometrics 68 pp 1010– (2012) · Zbl 1258.62116 · doi:10.1111/j.1541-0420.2012.01763.x
[36] Zhao, Reinforcement learning design for cancer clinical trials, Statistics in Medicine 28 pp 3294– (2009) · doi:10.1002/sim.3720
[37] Zhao, Estimating individualized treatment rules using outcome weighted learning, Journal of the American Statistical Association 107 pp 1106– (2012) · Zbl 1443.62396 · doi:10.1080/01621459.2012.695674
[38] Zhao, Reinforcement learning strategies for clinical trials in nonsmall cell lung cancer, Biometrics 67 pp 1422– (2011) · Zbl 1274.62922 · doi:10.1111/j.1541-0420.2011.01572.x
[39] Zhao, Doubly robust learning for estimating individualized treatment with censored data, Biometrika 102 pp 151– (2015) · Zbl 1345.62092 · doi:10.1093/biomet/asu050
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. It attempts to reflect the references listed in the original paper as accurately as possible without claiming the completeness or perfect precision of the matching.