Learning Treatment-Response Models from Multivariate Longitudinal Data

Treatment effects can be estimated from observational data as the difference in potential outcomes. In this paper, we address the challenge of estimating the potential outcome when treatment-dose levels can vary continuously over time. Further, the outcome variable may not be measured at a regular frequency. Our proposed solution represents the treatment response curves using linear time-invariant dynamical systems---this provides a flexible means for modeling response over time to highly variable dose curves. Moreover, for multivariate data, the proposed method: uncovers shared structure in treatment response and the baseline across multiple markers; and, flexibly models challenging correlation structure both across and within signals over time. For this, we build upon the framework of multiple-output Gaussian Processes. On simulated and a challenging clinical dataset, we show significant gains in accuracy over state-of-the-art models.

[1]  A. TUSTIN,et al.  Automatic Control Systems , 1950, Nature.

[2]  Benjamin C. Kuo,et al.  AUTOMATIC CONTROL SYSTEMS , 1962, Universum:Technical sciences.

[3]  D. Rubin.  Estimating causal effects of treatments in randomized and nonrandomized studies. , 1974 .

[4]  D. Rubin.  INFERENCE AND MISSING DATA , 1975 .

[5]  D. Rubin,et al.  Reducing Bias in Observational Studies Using Subclassification on the Propensity Score , 1984 .

[6]  J. Robins.  A graphical approach to the identification and estimation of causal parameters in mortality studies with sustained exposure periods. , 1987, Journal of chronic diseases.

[7]  Jorge Nocedal,et al.  A Limited Memory Algorithm for Bound Constrained Optimization , 1995, SIAM J. Sci. Comput..

[8]  S. Greenland.  Dose‐Response and Trend Analysis in Epidemiology: Alternatives to Categorical Analysis , 1995, Epidemiology.

[9]  David Card.  The Causal Effect of Education on Learning , 1999 .

[10]  James M. Robins,et al.  Causal inference for complex longitudinal data: the continuous case , 2001 .

[11]  J. Robins,et al.  Marginal Structural Models and Causal Inference in Epidemiology , 2000, Epidemiology.

[12]  Ana Ivelisse Avilés,et al.  Linear Mixed Models for Longitudinal Data , 2001, Technometrics.

[13]  R G Mark,et al.  MIMIC II: a massive temporal ICU patient database to support research in intelligent patient monitoring , 2002, Computers in Cardiology.

[14]  D. Cutler.  Linear systems analysis in pharmacokinetics , 1978, Journal of Pharmacokinetics and Biopharmaceutics.

[15]  J. Robins,et al.  Doubly Robust Estimation in Missing Data and Causal Inference Models , 2005, Biometrics.

[16]  Brent A. Johnson,et al.  Semiparametric inference in observational duration-response studies, with duration possibly right-censored , 2005 .

[17]  Yee Whye Teh,et al.  Semiparametric latent factor models , 2005, AISTATS.

[18]  C. Chiou,et al.  Physiological changes during hemodialysis in patients with intradialysis hypertension. , 2006, Kidney international.

[19]  Guanglei Hong,et al.  Effects of kindergarten retention on children's social-emotional development: an application of propensity score method to multivariate, multilevel data. , 2008, Developmental psychology.

[20]  Michalis K. Titsias,et al.  Variational Model Selection for Sparse Gaussian Process Regression , 2008 .

[21]  J. Lok.  Statistical modeling of causal effects in continuous time , 2004, math/0410271.

[22]  S. Dube,et al.  Renal replacement therapy in intensive care unit. , 2009, The Journal of the Association of Physicians of India.

[23]  Carl E. Rasmussen,et al.  Gaussian processes for machine learning , 2005, Adaptive computation and machine learning.

[24]  J. Pearl.  Causal inference in statistics: An overview , 2009 .

[25]  H. Chipman,et al.  BART: Bayesian Additive Regression Trees , 2008, 0806.3286.

[26]  Neil D. Lawrence,et al.  Computationally Efficient Convolved Multiple Output Gaussian Processes , 2011, J. Mach. Learn. Res..

[27]  Jennifer L. Hill,et al.  Bayesian Nonparametric Modeling for Causal Inference , 2011 .

[28]  D. Roden,et al.  The Emerging Role of Electronic Medical Records in Pharmacogenomics , 2011, Clinical pharmacology and therapeutics.

[29]  E. Moodie,et al.  Estimation of dose–response functions for longitudinal data using the generalised propensity score , 2012, Statistical methods in medical research.

[30]  D. Green,et al.  Modeling Heterogeneous Treatment Effects in Survey Experiments with Bayesian Additive Regression Trees , 2012 .

[31]  Jessica G. Young,et al.  The parametric g‐formula to estimate the effect of highly active antiretroviral therapy on incident AIDS or death , 2012, Statistics in medicine.

[32]  Joaquin Quiñonero Candela,et al.  Counterfactual reasoning and learning systems: the example of computational advertising , 2013, J. Mach. Learn. Res..

[33]  Neil D. Lawrence,et al.  Gaussian Processes for Big Data , 2013, UAI.

[34]  Lihong Li,et al.  Counterfactual Estimation and Optimization of Click Metrics in Search Engines: A Case Study , 2015, WWW.

[35]  James Hensman,et al.  Scalable Variational Gaussian Process Classification , 2014, AISTATS.

[36]  Peter Szolovits,et al.  A Multivariate Timeseries Modeling Approach to Severity of Illness Assessment and Forecasting in ICU with Sparse, Heterogeneous Clinical Data , 2015, AAAI.

[37]  E. Moodie,et al.  Optimal individualized dosing strategies: A pharmacologic approach to developing dynamic treatment regimens for continuous‐valued treatments , 2016, Biometrical journal. Biometrische Zeitschrift.

[38]  Ricardo Silva,et al.  Observational-Interventional Priors for Dose-Response Learning , 2016, NIPS.

[39]  Milos Hauskrecht,et al.  Learning Adaptive Forecasting Models from Irregularly Sampled Multivariate Clinical Data , 2016, AAAI.

[40]  Suchi Saria,et al.  A Bayesian Nonparametic Approach for Estimating Individualized Treatment-Response Curves , 2016, ArXiv.

[41]  David C. Kale,et al.  Directly Modeling Missing Data in Sequences with RNNs: Improved Classification of Clinical Time Series , 2016, MLHC.

[42]  Suchi Saria,et al.  Integrative Analysis using Coupled Latent Variable Models for Individualizing Prognoses , 2016, J. Mach. Learn. Res..

[43]  Martín Abadi,et al.  TensorFlow: Large-Scale Machine Learning on Heterogeneous Distributed Systems , 2016, ArXiv.

[44]  Suchi Saria,et al.  A Non-parametric Bayesian Approach for Estimating Treatment-Response Curves from Sparse Time Series , 2016, MLHC.

[45]  Thorsten Joachims,et al.  Recommendations as Treatments: Debiasing Learning and Evaluation , 2016, ICML.

[46]  Yebin Tao,et al.  Semiparametric Regression and Machine Learning Methods for Estimating Optimal Dynamic Treatment Regimes. , 2016 .

[47]  Suchi Saria,et al.  Reliable Decision Support using Counterfactual Models , 2017, NIPS.

[48]  Suchi Saria,et al.  What-If Reasoning with Counterfactual Gaussian Processes , 2017, NIPS 2017.