Validation Data-Based Adjustments for Outcome Misclassification in Logistic Regression

• Activity Number: 252
• Type: Contributed
• Date/Time: Monday, August 2, 2010 : 2:00 PM to 3:50 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Robert H. Lyles*+ and Li Tang and Hillary M. Superak and Caroline C. King and David Celentano and Yungtai Lo and Jack Sobel
• Companies: Emory University and Emory University and Emory University and CDC and The Johns Hopkins University and Montefiore Medical Center and Wayne State University
• Keywords: bias ; errors-in-variables ; odds ratio

Abstract

Misclassification of the outcome in logistic regression is a known source of bias when estimating adjusted odds ratios. Although various methods related to this problem exist, little guidance has been provided to practitioners wishing to restore validity through misclassification-corrected analyses that make use of external or internal validation data. We provide accessible likelihood-based methods that can be implemented via standard statistical software. Our approach allows for differentiality of sensitivity and specificity parameters according to the values of subject-specific covariates. Using main study and internal validation data from the HIV Epidemiology Research Study, we demonstrate this flexibility in an analysis that adjusts for misclassification of bacterial vaginosis status. We discuss how these methods can be applied under both case-control and cross-sectional sampling.