Abstract
Count data are subject to considerable sources of what is often referred to as non-sampling error. Errors such as misclassification, measurement error and unmeasured confounding can lead to substantially biased estimators. It is strongly recommended that epidemiologists not only acknowledge these sorts of errors in data, but incorporate sensitivity analyses into part of the total data analysis. We extend previous work on Poisson regression models that allow for misclassification by thoroughly discussing the basis for the models and allowing for extra-Poisson variability in the form of random effects. Via simulation we show the improvements in inference that are brought about by accounting for both the misclassification and the overdispersion.
Highlights
IntroductionEpidemiologic studies often have data that are subject to a wide array of different types of error
Epidemiologic studies often have data that are subject to a wide array of different types of error.Measurement error, unmeasured confounding, and selection bias are all examples of sources of biased estimators and reduced power for hypothesis tests [1,2]
In this paper we focus on the important case of count data with misclassification and provide a cohesive estimating procedure for inference for a range of models of interest, fixed and random effects models and the cases of known and unknown misclassification rates
Summary
Epidemiologic studies often have data that are subject to a wide array of different types of error. From the Bayesian perspective, Stamey et al [9] are able to free the sensitivity and specificity from being fixed or known by making use of validation data or informative priors. In this paper we focus on the important case of count data with misclassification and provide a cohesive estimating procedure for inference for a range of models of interest, fixed and random effects models and the cases of known and unknown misclassification rates.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
