Introduction Many environmental studies often involve the analysis of count data, such as the number of hospitalizations caused by air pollution, where Poisson Regression (PR) is the stan- dard basic technique. However, overdispersion is widely seen in this regression model. Overdispersion in this model occurs when the response variance is greater than the mean. This may cause standard errors of the estimates to be deflat- ed or underestimated. The Negative Binomial Regression model is a generalization of the Poisson regression model that allows for overdispersion by introducing an unob- served heterogeneity term. The Generalized Poisson Regression (GPR) model developed by [1] is used to model dispersed count data to handle the overdispersion problem. Refer to [2-4] for a good overview of the base generalized Poisson model and its derivation. Generalized Poisson is similar to the negative binomial in that it incorporates an extra heterogeneity or dispersion parameter. Another difficulty occurs when there are excess zeros in the data. Zero-inflated count models were first introduced by [5] to provide a method of accounting for excessive zero counts. A popular approach to the analysis of such data is to use zero-inflated Poisson (ZIP) and zero-inflated Negative Binomial (ZINB) regression models. Recent developments have discussed extending the Poisson or negative binomial distributions into models that account for extra zeros [6-8]. Some studies have used extensive ecological datasets to compare the performance of these distributions for a vari- ety of environmental conditions [9-15]. Many of the existing studies have focused on the asso- ciation between environmental pollution and hospital admissions for Chronic obstructive pulmonary disease (COPD), a group of diseases characterized by airflow obstruction that can be associated with breathing-related symptoms (e.g., cough, exertional dyspnea, expectoration, and wheeze). There is increasing interest in the use of hos- pital admission data in studies of short-term exposure effects attributed to air pollutants. Numerous studies have investigated the relationship between air pollution and hos- pital admissions for COPD [16-22]. While adverse effects of exposure to air pollutants and hospital admissions for COPD are well studied, little is known about the effect of air pollutants on COPD symp- toms. This study focuses on modeling air pollution and both Pol. J. Environ. Stud. Vol. 21, No. 3 (2012), 565-568 *e-mail: macengiz@omu.edu.tr Original Research Zero-Inflated Regression Models for Modeling the Effect of air Pollutants on Hospital Admissions Mehmet Ali Cengiz* Department of Statistics, University of Ondokuz Mayıs, Istatıstık Bölümü, 55139 Atakum-Samsun, Turkey Received: 13 May 2011 Accepted: 31 October 2011 Abstract Count regression methods are the fundamental tool used for modeling the association between environ- mental pollution and hospital admissions. Data with many zeros are often encountered in count regression models. Failure to account for the extra zeros may result in biased parameter estimates and misleading infer- ences. Zero-inflated Poisson and zero-inflated negative binomial regression models have been proposed for situations where the data generating process results in too many zeros. Keywords: count regression, zero-inflated models, air pollution