Methods for Conducting Sensitivity Analysis of Trials with Potentially Nonignorable Competing Causes of Censoring

We consider inference for the treatment-arm mean difference of an outcome that would have been measured at the end of a randomized follow-up study if, during the course of the study, patients had not initiated a nonrandomized therapy or dropped out. We argue that the treatment-arm mean difference is not identified unless unverifiable assumptions are made. We describe identifying assumptions that are tantamount to postulating relationships between the components of a pattern-mixture model but that can also be interpreted as imposing restrictions on the cause-specific censoring probabilities of a selection model. We then argue that, although sufficient for identification, these assumptions are insufficient for inference due to the curse of dimensionality. We propose reducing dimensionality by specifying semiparametric cause-specific selection models. These models are useful for conducting a sensitivity analysis to examine how inference for the treatment-arm mean difference changes as one varies the magnitude of the cause-specific selection bias over a plausible range. We provide methodology for conducting such sensitivity analysis and illustrate our methods with an analysis of data from the AIDS Clinical Trial Group (ACTG) study 002.

[1]  Roderick J. A. Little,et al.  Adjusting for Nonignorable Drop-Out Using Semiparametric Nonresponse Models: Comment , 1999 .

[2]  V. De Gruttola,et al.  Modelling progression of CD4-lymphocyte count and its relationship to survival time. , 1994, Biometrics.

[3]  D. Rubin.  Formalizing Subjective Notions about the Effect of Nonrespondents in Sample Surveys , 1977 .

[4]  N M Laird,et al.  Increasing efficiency from censored survival data by using random effects to model longitudinal covariates , 1998, Statistical methods in medical research.

[5]  S G Baker,et al.  Marginal regression for repeated binary data with outcome subject to non-ignorable non-response. , 1995, Biometrics.

[6]  R. Little,et al.  Pattern-mixture models for multivariate incomplete data with covariates. , 1996, Biometrics.

[7]  John P. Klein,et al.  Estimates of marginal survival for dependent competing risks based on an assumed copula , 1995 .

[8]  R F Woolson,et al.  Slope estimation in the presence of informative right censoring: modeling the number of observations as a geometric random variable. , 1994, Biometrics.

[9]  Eric V. Slud,et al.  Dependent competing risks and summary survival curves , 1983 .

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

[11]  R. Little.  Pattern-Mixture Models for Multivariate Incomplete Data , 1993 .

[12]  J. Robins,et al.  Toward a curse of dimensionality appropriate (CODA) asymptotic theory for semi-parametric models. , 1997, Statistics in medicine.

[13]  J. Robins,et al.  Sensitivity Analysis for Selection bias and unmeasured Confounding in missing Data and Causal inference models , 2000 .

[14]  N M Laird,et al.  Model-based approaches to analysing incomplete longitudinal and failure time data. , 1997, Statistics in medicine.

[15]  J. Robins,et al.  G-Estimation of the Effect of Prophylaxis Therapy for Pneumocystis carinii Pneumonia on the Survival of AIDS Patients , 1992, Epidemiology.

[16]  W F Rosenberger,et al.  Closed-form estimates for missing counts in two-way contingency tables. , 1992, Statistics in medicine.

[17]  D. Rubin,et al.  The central role of the propensity score in observational studies for causal effects , 1983 .

[18]  J. Copas,et al.  Inference for Non‐random Samples , 1997 .

[19]  Donald B. Rubin,et al.  Bayesian Inference for Causal Effects: The Role of Randomization , 1978 .

[20]  H. Quan,et al.  Estimation of the mean when data contain non-ignorable missing values from a random effects model , 1994 .

[21]  J M Robins,et al.  Non-response models for the analysis of non-monotone non-ignorable missing data. , 1997, Statistics in medicine.

[22]  Raymond J. Carroll,et al.  Estimation and comparison of changes in the presence of informative right censoring by modeling the censoring process , 1988 .

[23]  K. Bailey,et al.  Erratum: Estimation and comparison of changes in the presence of informative right censoring: Conditional linear model (Biometrics, September 1989, 45, 939-955) , 1990 .

[24]  J P Klein,et al.  Bounds on net survival probabilities for dependent competing risks. , 1988, Biometrics.

[25]  Miss A.O. Penney.  (b) , 1974, The New Yale Book of Quotations.

[26]  Garrett M. Fitzmaurice,et al.  Logistic Regression Models for Binary Panel Data with Attrition , 1996 .

[27]  R D Gill,et al.  Non-response models for the analysis of non-monotone ignorable missing data. , 1997, Statistics in medicine.

[28]  K. Bailey,et al.  Estimation and comparison of changes in the presence of informative right censoring: conditional linear model. , 1989, Biometrics.

[29]  Robin Sibson,et al.  What is projection pursuit , 1987 .

[30]  K. Bailey,et al.  Analysing changes in the presence of informative right censoring caused by death and withdrawal. , 1988, Statistics in medicine.

[31]  Hulin Wu,et al.  Design of viral dynamic studies for efficiently assessing potency of anti-HIV therapies in AIDS Clinical Trials , 2002 .

[32]  Erik V. Nordheim,et al.  Inference from Nonrandomly Missing Categorical Data: An Example from a Genetic Study on Turner's Syndrome , 1984 .

[33]  M. Wulfsohn,et al.  A randomized controlled trial of a reduced daily dose of zidovudine in patients with the acquired immunodeficiency syndrome , 1990 .

[34]  J. Klein,et al.  Identifiability and Estimation of Marginal Survival Functions for Dependent Competing Risks Assuming the Copula is Known , 1996 .

[35]  N M Laird,et al.  Mixture models for the joint distribution of repeated measures and event times. , 1997, Statistics in medicine.

[36]  Nan M. Laird,et al.  Multivariate Logistic Models for Incomplete Binary Responses , 1996 .

[37]  Nicole A. Lazar,et al.  Statistical Analysis With Missing Data , 2003, Technometrics.

[38]  M. Wulfsohn,et al.  Modeling the Relationship of Survival to Longitudinal Data Measured with Error. Applications to Survival and CD4 Counts in Patients with AIDS , 1995 .

[39]  Roderick J. A. Little,et al.  Modeling the Drop-Out Mechanism in Repeated-Measures Studies , 1995 .

[40]  J. Robins,et al.  Analysis of semi-parametric regression models with non-ignorable non-response. , 1997, Statistics in medicine.

[41]  M. Kenward,et al.  Informative Drop‐Out in Longitudinal Data Analysis , 1994 .

[42]  James M. Robins,et al.  Semiparametric Regression for Repeated Outcomes With Nonignorable Nonresponse , 1998 .

[43]  D. Rubin,et al.  Statistical Analysis with Missing Data. , 1989 .

[44]  M. Kenward,et al.  The analysis of longitudinal ordinal data with nonrandom drop-out , 1997 .

[45]  D. Follmann,et al.  An approximate generalized linear model with random effects for informative missing data. , 1995, Biometrics.

[46]  J. Robins,et al.  Adjusting for Nonignorable Drop-Out Using Semiparametric Nonresponse Models , 1999 .

[47]  Yudi Pawitan,et al.  Modeling a Marker of Disease Progression and Onset of Disease , 1992 .

[48]  Roderick J. A. Little,et al.  A Class of Pattern-Mixture Models for Normal Incomplete Data , 1994 .

[49]  M D Schluchter,et al.  Methods for the analysis of informatively censored longitudinal data. , 1992, Statistics in medicine.