### Abstract

Misclassification in a binary exposure variable within an unmatched prospective study may lead to a biased estimate of the disease-expo sure relationship. It usually gives falsely small credible intervals because uncertainty in the recorded exposure is not taken into account. When there are several other perfectly measured covariates, interrelationships may introduce further potential for bias. Bayesian methods are proposed for analysing binary outcome studies in which an exposure variable is sometimes misclassified, but its correct values have been validated for a random subsample of the subjects. This Bayesian approach can model relationships between explanatory variables and between exploratory variables and the probabilities of misclassification. Three logistic regressions are used to relate disease to true exposure, misclassified exposure to true exposure and true exposure to other covariates. Credible intervals may be used to make decisions about whether certain parameters are unnecessary and hence whether the model can be reduced in complexity.

In the disease-exposure model, for parameters representing coefficients related to perfectly measured covariates, the precision of posterior estimates is only slightly lower than would be found from data with no misclassification. For the risk factor which has misclassification, the estimates of model coefficients obtained are much less biased than those with misclassification ignored. Copyright (C) 2005 John Wiley & Sons, Ltd.

Original language | English |
---|---|

Pages (from-to) | 3463-3477 |

Number of pages | 14 |

Journal | Statistics in Medicine |

Volume | 24 |

Issue number | 22 |

DOIs | |

Publication status | Published - Oct 2005 |

### Keywords

- errors in variables
- binary outcome
- measurement error
- misclassification
- odds ratio
- MEASUREMENT ERROR PROBLEMS
- REGRESSION-MODELS
- LOGISTIC-REGRESSION
- IN-VARIABLES
- ODDS RATIOS
- EXPOSURE
- VALIDATION
- EPIDEMIOLOGY
- INACCURACIES
- INFORMATION

### Cite this

*Statistics in Medicine*,

*24*(22), 3463-3477. https://doi.org/10.1002/sim.2192

**A Bayesian approach to prospective binary outcome studies with misclassification in a binary risk factor.** / Prescott, Gordon James; Garthwaite, P. H.

Research output: Contribution to journal › Article

*Statistics in Medicine*, vol. 24, no. 22, pp. 3463-3477. https://doi.org/10.1002/sim.2192

}

TY - JOUR

T1 - A Bayesian approach to prospective binary outcome studies with misclassification in a binary risk factor

AU - Prescott, Gordon James

AU - Garthwaite, P. H.

PY - 2005/10

Y1 - 2005/10

N2 - Misclassification in a binary exposure variable within an unmatched prospective study may lead to a biased estimate of the disease-expo sure relationship. It usually gives falsely small credible intervals because uncertainty in the recorded exposure is not taken into account. When there are several other perfectly measured covariates, interrelationships may introduce further potential for bias. Bayesian methods are proposed for analysing binary outcome studies in which an exposure variable is sometimes misclassified, but its correct values have been validated for a random subsample of the subjects. This Bayesian approach can model relationships between explanatory variables and between exploratory variables and the probabilities of misclassification. Three logistic regressions are used to relate disease to true exposure, misclassified exposure to true exposure and true exposure to other covariates. Credible intervals may be used to make decisions about whether certain parameters are unnecessary and hence whether the model can be reduced in complexity.In the disease-exposure model, for parameters representing coefficients related to perfectly measured covariates, the precision of posterior estimates is only slightly lower than would be found from data with no misclassification. For the risk factor which has misclassification, the estimates of model coefficients obtained are much less biased than those with misclassification ignored. Copyright (C) 2005 John Wiley & Sons, Ltd.

AB - Misclassification in a binary exposure variable within an unmatched prospective study may lead to a biased estimate of the disease-expo sure relationship. It usually gives falsely small credible intervals because uncertainty in the recorded exposure is not taken into account. When there are several other perfectly measured covariates, interrelationships may introduce further potential for bias. Bayesian methods are proposed for analysing binary outcome studies in which an exposure variable is sometimes misclassified, but its correct values have been validated for a random subsample of the subjects. This Bayesian approach can model relationships between explanatory variables and between exploratory variables and the probabilities of misclassification. Three logistic regressions are used to relate disease to true exposure, misclassified exposure to true exposure and true exposure to other covariates. Credible intervals may be used to make decisions about whether certain parameters are unnecessary and hence whether the model can be reduced in complexity.In the disease-exposure model, for parameters representing coefficients related to perfectly measured covariates, the precision of posterior estimates is only slightly lower than would be found from data with no misclassification. For the risk factor which has misclassification, the estimates of model coefficients obtained are much less biased than those with misclassification ignored. Copyright (C) 2005 John Wiley & Sons, Ltd.

KW - errors in variables

KW - binary outcome

KW - measurement error

KW - misclassification

KW - odds ratio

KW - MEASUREMENT ERROR PROBLEMS

KW - REGRESSION-MODELS

KW - LOGISTIC-REGRESSION

KW - IN-VARIABLES

KW - ODDS RATIOS

KW - EXPOSURE

KW - VALIDATION

KW - EPIDEMIOLOGY

KW - INACCURACIES

KW - INFORMATION

U2 - 10.1002/sim.2192

DO - 10.1002/sim.2192

M3 - Article

VL - 24

SP - 3463

EP - 3477

JO - Statistics in Medicine

JF - Statistics in Medicine

SN - 0277-6715

IS - 22

ER -