Are Zero-Modified Models the Panacea for Epidemiological Data with Excess Zeroes?

• Activity Number: 672 - Methods for Infectious Disease Epidemiology
• Type: Contributed
• Date/Time: Thursday, August 2, 2018 : 10:30 AM to 12:20 PM
• Sponsor: Section on Statistics in Epidemiology
• Author(s): Ali Arab* and Frederic Mortier
• Companies: Georgetown University and CIRAD, UPR Forests and Societies/Forests and Societies, Univ Montpellier, CIRAD
• Keywords: Zero-Inflated Models; Hurdle Models; Lyme Disease; Model Choice; Spatial; Spatio-Temporal

Abstract

Epidemiological data often include excess zeroes, in particular, for rare or emerging diseases or diseases that are not common in specific areas, specific time periods, or are hard to detect. A common approach to modeling data with excess zeroes is to use zero-modified models (i.e., hurdle and zero-inflated models). Here, focusing on spatial and spatio-temporal count data, we first explore if zero-modified modeling is systematically the most effective approach for data with excess zeroes. Also, we discuss potential links between spatial or temporal structure of the data, zero-inflation, and model choice. To demonstrate our work, we provide a case study on five-year counts of confirmed cases of Lyme disease in several states of the United States.