Finding the Right Distribution for Highly Skewed Zero-inflated Clinical Data

Abstract

Discrete, highly skewed distributions with excess numbers of zeros often result in biased estimates and misleading inferences if the zeros are not properly addressed. A clinical example of children with electrophysiologic disorders in which many of the children are treated without surgery is provided. The purpose of the current study was to identify the optimal modeling strategy for highly skewed, zeroinflated data often observed in the clinical setting by: (a) simulating skewed, zero-inflated count data; (b) fitting simulated data with Poisson, Negative Binomial, Zero-Inflated Poisson (ZIP) and Zero-inflated Negative Binomial (ZINB) models; and, (c) applying the aforementioned models to actual, highly
skewed, clinical data of children with an EP disorder. The ZIP model was observed to be the optimal model based on traditional fit statistics as well as estimates of bias, mean-squared error, and coverage.