Incorporating prior beliefs about selection bias into the analysis of randomized trials with missing outcomes.

Daniel O Scharfstein, Michael J. Daniels, James M. Robins

Research output: Contribution to journalArticle

Abstract

In randomized studies with missing outcomes, non-identifiable assumptions are required to hold for valid data analysis. As a result, statisticians have been advocating the use of sensitivity analysis to evaluate the effect of varying assumptions on study conclusions. While this approach may be useful in assessing the sensitivity of treatment comparisons to missing data assumptions, it may be dissatisfying to some researchers/decision makers because a single summary is not provided. In this paper, we present a fully Bayesian methodology that allows the investigator to draw a 'single' conclusion by formally incorporating prior beliefs about non-identifiable, yet interpretable, selection bias parameters. Our Bayesian model provides robustness to prior specification of the distributional form of the continuous outcomes.

Original languageEnglish (US)
Pages (from-to)495-512
Number of pages18
JournalBiostatistics
Volume4
Issue number4
StatePublished - Oct 2003

Fingerprint

Selection Bias
Randomized Trial
Research Personnel
Bayesian Robustness
Model Robustness
Bayesian Model
Missing Data
Sensitivity Analysis
Data analysis
Valid
Specification
Methodology
Evaluate
Beliefs
Robustness
Decision maker
Missing data
Bayesian model
Selection bias
Sensitivity analysis

ASJC Scopus subject areas

  • Medicine(all)
  • Statistics and Probability
  • Statistics, Probability and Uncertainty

Cite this

Incorporating prior beliefs about selection bias into the analysis of randomized trials with missing outcomes. / Scharfstein, Daniel O; Daniels, Michael J.; Robins, James M.

In: Biostatistics, Vol. 4, No. 4, 10.2003, p. 495-512.

Research output: Contribution to journalArticle

@article{dbcd24969f7c45e6874c404e91d538e8,
title = "Incorporating prior beliefs about selection bias into the analysis of randomized trials with missing outcomes.",
abstract = "In randomized studies with missing outcomes, non-identifiable assumptions are required to hold for valid data analysis. As a result, statisticians have been advocating the use of sensitivity analysis to evaluate the effect of varying assumptions on study conclusions. While this approach may be useful in assessing the sensitivity of treatment comparisons to missing data assumptions, it may be dissatisfying to some researchers/decision makers because a single summary is not provided. In this paper, we present a fully Bayesian methodology that allows the investigator to draw a 'single' conclusion by formally incorporating prior beliefs about non-identifiable, yet interpretable, selection bias parameters. Our Bayesian model provides robustness to prior specification of the distributional form of the continuous outcomes.",
author = "Scharfstein, {Daniel O} and Daniels, {Michael J.} and Robins, {James M.}",
year = "2003",
month = "10",
language = "English (US)",
volume = "4",
pages = "495--512",
journal = "Biostatistics",
issn = "1465-4644",
publisher = "Oxford University Press",
number = "4",

}

TY - JOUR

T1 - Incorporating prior beliefs about selection bias into the analysis of randomized trials with missing outcomes.

AU - Scharfstein, Daniel O

AU - Daniels, Michael J.

AU - Robins, James M.

PY - 2003/10

Y1 - 2003/10

N2 - In randomized studies with missing outcomes, non-identifiable assumptions are required to hold for valid data analysis. As a result, statisticians have been advocating the use of sensitivity analysis to evaluate the effect of varying assumptions on study conclusions. While this approach may be useful in assessing the sensitivity of treatment comparisons to missing data assumptions, it may be dissatisfying to some researchers/decision makers because a single summary is not provided. In this paper, we present a fully Bayesian methodology that allows the investigator to draw a 'single' conclusion by formally incorporating prior beliefs about non-identifiable, yet interpretable, selection bias parameters. Our Bayesian model provides robustness to prior specification of the distributional form of the continuous outcomes.

AB - In randomized studies with missing outcomes, non-identifiable assumptions are required to hold for valid data analysis. As a result, statisticians have been advocating the use of sensitivity analysis to evaluate the effect of varying assumptions on study conclusions. While this approach may be useful in assessing the sensitivity of treatment comparisons to missing data assumptions, it may be dissatisfying to some researchers/decision makers because a single summary is not provided. In this paper, we present a fully Bayesian methodology that allows the investigator to draw a 'single' conclusion by formally incorporating prior beliefs about non-identifiable, yet interpretable, selection bias parameters. Our Bayesian model provides robustness to prior specification of the distributional form of the continuous outcomes.

UR - http://www.scopus.com/inward/record.url?scp=2942658011&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=2942658011&partnerID=8YFLogxK

M3 - Article

C2 - 14557107

AN - SCOPUS:2942658011

VL - 4

SP - 495

EP - 512

JO - Biostatistics

JF - Biostatistics

SN - 1465-4644

IS - 4

ER -