Doubly Robust Nonparametric Instrumental Variable Estimators for Survival Outcomes

Instrumental variable (IV) methods allow us the opportunity to address unmeasured confounding in causal inference. However, most IV methods are only applicable to discrete or continuous outcomes with very few IV methods for censored survival outcomes. In this work we propose nonparametric estimators for the local average treatment effect on survival probabilities under both nonignorable and ignorable censoring. We provide an efficient influence function-based estimator and a simple estimation procedure when the IV is either binary or continuous. The proposed estimators possess double-robustness properties and can easily incorporate nonparametric estimation using machine learning tools. In simulation studies, we demonstrate the flexibility and efficiency of our proposed estimators under various plausible scenarios. We apply our method to the Prostate, Lung, Colorectal, and Ovarian Cancer Screening Trial for estimating the causal effect of screening on survival probabilities and investigate the causal contrasts between the two interventions under different censoring assumptions.

[1]  P C Prorok,et al.  Design of the Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial. , 2000, Controlled clinical trials.

[2]  T. Speed,et al.  On the Application of Probability Theory to Agricultural Experiments. Essay on Principles. Section 9 , 1990 .

[3]  P. Austin The use of propensity score methods with survival or time-to-event outcomes: reporting measures of effect similar to those used in randomized experiments , 2013, Statistics in medicine.

[4]  M. Baiocchi,et al.  Instrumental variable methods for causal inference , 2014, Statistics in medicine.

[5]  J. Robins,et al.  Correcting for non-compliance in randomized trials using rank preserving structural failure time models , 1991 .

[6]  Jason P Fine,et al.  Nonparametric binary instrumental variable analysis of competing risks data. , 2017, Biostatistics.

[7]  Mei-Cheng Wang,et al.  Estimating Propensity Scores and Causal Survival Functions Using Prevalent Survival Data , 2012, Biometrics.

[8]  Joseph Kang,et al.  Demystifying Double Robustness: A Comparison of Alternative Strategies for Estimating a Population Mean from Incomplete Data , 2007, 0804.2958.

[9]  C. Schooling,et al.  Bias from competing risk before recruitment in Mendelian Randomization studies of conditions with shared etiology , 2019, bioRxiv.

[10]  Gang Li,et al.  A Bayesian approach for instrumental variable analysis with censored time‐to‐event outcome , 2015, Statistics in medicine.

[11]  Elizabeth L. Ogburn,et al.  Doubly robust estimation of the local average treatment effect curve , 2015, Journal of the Royal Statistical Society. Series B, Statistical methodology.

[12]  Stephen R Cole,et al.  Adjusted survival curves with inverse probability weights , 2004, Comput. Methods Programs Biomed..

[13]  James M. Robins,et al.  Unified Methods for Censored Longitudinal Data and Causality , 2003 .

[14]  T. Jaki,et al.  Instrumental variable estimation in semi‐parametric additive hazards models , 2018, Biometrics.

[15]  Zhiqiang Tan,et al.  Regression and Weighting Methods for Causal Inference Using Instrumental Variables , 2006 .

[16]  Iván Díaz,et al.  Statistical Inference for Machine Learning Inverse Probability Weighting with Survival Outcomes , 2017 .

[17]  Marie Davidian,et al.  Comment: Demystifying Double Robustness: A Comparison of Alternative Strategies for Estimating a Population Mean from Incomplete Data. , 2008, Statistical science : a review journal of the Institute of Mathematical Statistics.

[18]  J. Gohagan,et al.  The Prostate, Lung, Colorectal and Ovarian (PLCO) Cancer Screening Trial of the National Cancer Institute: history, organization, and status. , 2000, Controlled clinical trials.

[19]  Dylan S. Small,et al.  Robust causal inference with continuous instruments using the local instrumental variable curve , 2016, Journal of the Royal Statistical Society: Series B (Statistical Methodology).

[20]  Eric J. Tchetgen Tchetgen,et al.  The GENIUS Approach to Robust Mendelian Randomization Inference , 2017, bioRxiv.

[21]  Edward H. Kennedy Semiparametric theory and empirical processes in causal inference , 2015, 1510.04740.

[22]  Edward H. Kennedy Nonparametric Causal Effects Based on Incremental Propensity Score Interventions , 2017, Journal of the American Statistical Association.

[23]  T. Tosteson,et al.  Using instrumental variables to estimate a Cox’s proportional hazards regression subject to additive confounding , 2014, Health Services and Outcomes Research Methodology.

[24]  M Alan Brookhart,et al.  Instrumental variables I: instrumental variables exploit natural variation in nonexperimental data to estimate causal relationships. , 2009, Journal of clinical epidemiology.

[25]  Erik T Parner,et al.  Causal inference in survival analysis using pseudo‐observations , 2017, Statistics in medicine.

[26]  S. Thompson,et al.  Use of allele scores as instrumental variables for Mendelian randomization , 2013, International journal of epidemiology.

[27]  Iv'an D'iaz,et al.  Non-parametric causal effects based on longitudinal modified treatment policies , 2020 .

[28]  J. Robins,et al.  Instruments for Causal Inference: An Epidemiologist's Dream? , 2006, Epidemiology.

[29]  Stefan Wager,et al.  Estimation and Inference of Heterogeneous Treatment Effects using Random Forests , 2015, Journal of the American Statistical Association.

[30]  J. Fine,et al.  Instrumental variable additive hazards models , 2015, Biometrics.

[31]  Iván Díaz,et al.  Machine learning in the estimation of causal effects: targeted minimum loss-based estimation and double/debiased machine learning. , 2019, Biostatistics.

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

[33]  N. Timpson,et al.  The Effect of Elevated Body Mass Index on Ischemic Heart Disease Risk: Causal Estimates from a Mendelian Randomisation Approach , 2012, PLoS medicine.

[34]  Joshua D. Angrist,et al.  Identification of Causal Effects Using Instrumental Variables , 1993 .

[35]  M. J. van der Laan,et al.  Increasing Power in Randomized Trials with Right Censored Outcomes Through Covariate Adjustment , 2009, Journal of biopharmaceutical statistics.

[36]  Paul J Rathouz,et al.  Two-stage residual inclusion estimation: addressing endogeneity in health econometric modeling. , 2008, Journal of health economics.

[37]  J. Robins,et al.  Double/Debiased Machine Learning for Treatment and Structural Parameters , 2017 .

[38]  Menggang Yu,et al.  Instrumental variable based estimation under the semiparametric accelerated failure time model , 2019, Biometrics.

[39]  Jung In Kim,et al.  Causal Proportional Hazards Estimation with a Binary Instrumental Variable. , 2019, Statistica Sinica.

[40]  Edward H Kennedy,et al.  Instrumental variable methods using dynamic interventions , 2018, Journal of the Royal Statistical Society: Series A (Statistics in Society).

[41]  M. Glymour,et al.  Instrumental variable estimation in a survival context. , 2015, Epidemiology.

[42]  Dylan S. Small,et al.  Bias in estimating the causal hazard ratio when using two‐stage instrumental variable methods , 2015, Statistics in medicine.

[43]  S. Vansteelandt,et al.  On doubly robust estimation of the hazard difference , 2018, Biometrics.

[44]  Markus Frölich,et al.  Nonparametric IV Estimation of Local Average Treatment Effects with Covariates , 2002, SSRN Electronic Journal.

[45]  Dylan S. Small,et al.  A general approach to evaluating the bias of 2‐stage instrumental variable estimators , 2018, Statistics in medicine.

[46]  Shu Yang,et al.  Semiparametric estimation of structural failure time models in continuous-time processes. , 2018, Biometrika.

[47]  A. J. O’Malley,et al.  Adjusting for bias introduced by instrumental variable estimation in the Cox proportional hazards model. , 2017, Biostatistics.