![[PDF]](/pb-assets/Icons/adobe-pdf-icon-1498649896243.png){kind=icon}
Abstract
In many clinical trials it is possible for some subjects to cross over between treatment arms. One can evaluate the effect of crossover by modeling it as a missing-data problem, where for subjects who cross over, one treats the unobserved value of the outcome in the original randomization arm as the missing data. The as-treated analysis is invalid if the crossover is nonignorable, in the sense that the crossovers represent a nonrandom sample of the randomized subjects [1,2]. A recent area of general interest is the development of methods for measuring the sensitivity of inferences to nonignorability in the missing-data mechanism; one such approach is that of Troxel et al. [3]. In this paper we apply their method to the problem of measuring sensitivity to nonignorable crossover in randomized trials, extending it to the case where the crossover mechanism may differ between arms. Our method allows us to identify circumstances under which the as-treated analysis may be more or less sensitive to nonignorable crossover. We illustrate it with the example of a randomized clinical trial (RCT) in multiple sclerosis and a study of the effect of military service on income.
|
Heitjan DF . Ignorability and bias in clinical trials. Stat Med 1999; 18: 2421-2434. Google Scholar | Crossref | Medline | ISI | |
|
Heitjan DF . Casual inference in a clinical trial: a comparative example. Control Clin Trials 1999; 20: 309-318. Google Scholar | Crossref | Medline | |
|
Troxel AB , Ma G , Heitjan DF . An index of sensitivityto nonignorability. Stat Sinica, in press. Google Scholar | |
|
Breslow NE . Clinical trials. In Kotz S , Johnson NL eds. Encyclopedia of statistical sciences, Vol. 2. New York: John Wiley & Sons, 1982: 13-21. Google Scholar | |
|
Lee Y , Ellenberg J , Hirtz D , Nelson K .Analysis of clinical trials by treatment actually received: Is it really an option? Stat Med 1991; 10: 1595-1605. Google Scholar | Crossref | Medline | ISI | |
|
Meier P . Comment on ‘Compliance as an explanatory variable in clinical trials’ by Efron and Feldman. J Am Stat Assoc 1991; 86: 19-22. Google Scholar | Crossref | ISI | |
|
Angrist JD , Imbens GW , Rubin DB . Identification of causal effects using instrumental variables. J Am Stat Assoc 1996; 91: 444-455. Google Scholar | Crossref | ISI | |
|
Sommer A , Zeger SL . On estimating efficacy from clinical trials. Stat Med 1991; 10: 45-52. Google Scholar | Crossref | Medline | ISI | |
|
Cuzick J , Edwards R , Segnan N . Adjusting for noncompliance and contamination in randomized clinical trials. Stat Med 2001; 16: 1017-1029. Google Scholar | Crossref | ISI | |
|
Efron B , Feldman D . Compliance as an explanatory variable in clinical trials. J Am Stat Assoc 1991; 86: 9-17. Google Scholar | Crossref | ISI | |
|
Cornfield J , Haenszel W , Hammond EC , Lilienfied AM , Shimkin MB , Wynder EL . Smoking and lung cancer: recent evidence and a discussion of some questions. J Nat Cancer Inst 1959; 22: 173-203. Google Scholar | Medline | |
|
Rosenbaum PR , Rubin DB . Assessing sensitivity to an unobserved binary covariate in an observational study with binary outcome. J Royal Stat Soc B 1983; 45: 212-218. Google Scholar | |
|
Vach W , Blettner M . Logistic regression with incompletely observed categorial covariates-investigating the sensitivity against violation of the missing at random assumption. Stat Med 1995; 14: 1315-1329. Google Scholar | Crossref | Medline | ISI | |
|
Rotnitzky A , Robins JM , Scharfstein DO . Semipara-metric regression for repeated outcomes with nonignorable nonresponse. J Am Stat Assoc 1998; 93: 1321-1339. Google Scholar | Crossref | ISI | |
|
Scharfstein D , Rotnitzky A , Robins JM . Adjusting for nonignorable dropout using semiparametric models (with discussion). J Am Stat Assoc 1999; 94: 1096-1146. Google Scholar | Crossref | ISI | |
|
Copas JB , Li HG . Inference for non-random samples (with discussion). J Royal Stat Soc B 1997; 59: 55-95. Google Scholar | Crossref | |
|
Copas JB , Eguchi S . Local sensitivity approximations for selectivity bias. J Royal Stat Soc B 2001; 63: 871-895. Google Scholar | Crossref | |
|
Verbeke G , Molenberghs G , Thijs H , Lesaffre E , Kenward MG . Sensitivity analysis for nonrandom dropout: a local influence approach. Biometrics 2001; 57: 7-14. Google Scholar | Crossref | Medline | ISI | |
|
Cook RD . Assessment of local influence. J Royal Stat Soc B 1986; 48: 133-169. Google Scholar | |
|
Rubin DB . Estimating causal effects of treatments in randomized and nonrandomized studies. J Educ Psych 1974; 66: 688-701. Google Scholar | Crossref | ISI | |
|
Rubin DB . Bayesian inference from causal effects: the role of randomization. Annals Stat 1978; 7: 34-58. Google Scholar | Crossref | ISI | |
|
Holland PW . Statistics and causal inference. J Am Stat Assoc 1986; 81: 945-960. Google Scholar | Crossref | ISI | |
|
Imbens GW , Rubin DB . Bayesian inference for causal effects in randomized experiments with noncompliance. Annals Stat 1997; 25: 305-327. Google Scholar | Crossref | ISI | |
| Xie H. An index of sensitivity to nonignorability: Extensions and applications. Doctoral dissertation, Department of Biostatistics, Columbia University. Google Scholar | |
|
Abadie A . Bootstrap tests for distributional treatment effects in instrumental variable models. J Am Stat Assoc 2002; 97: 284-292. Google Scholar | Crossref | ISI | |
|
Imbens GW , Angrist JD . Identification and estimation of local average treatment effects. Econometrica 1994; 62: 467-476. Google Scholar | Crossref | ISI | |
|
Hastie TJ , Tibshirani RJ . Generalized additive models. London: Chapman & Hall, 1990. Google Scholar | |
|
Aranda-Ordaz FJ . On two families of transformations to additivity for binary response date. Biometrika 1981; 68: 357-363. Google Scholar | Crossref | ISI |
