The application of zero-inflated count regression models for identifying main factors on the number of blood donor deferral in Shahrekord

Blood and its products have a special role in healthy system of any country. The aim of this study was to modeling the number of blood donor deferral and detecting its main factors based on zero-inflated count regression models.
The data used in this study were drawn from a longitudinal study in which 864 first-time donors were followed up for a maximum five years, from 2008 to 2013. The response variable was the number of blood donor deferral during five years. Also, sex, weight, age, marital status, education and job were used as independent variables. For analyzing data, two zero-inflated Poisson and zero-inflated negative binomial models were used by Bayesian technique. Assessment of models was carried done using Marko chain Monte Carlo methods (MCMC) by WinBUGS. Comparison of models was done using deviance information Bayesian criterion (DIC).
Based on the results of DIC, the zero-inflated negative binomial regression model had smaller DIC and was selected as better model. The body weight had a significant positive effect on the number of blood donor deferral.
Donors with higher body weight returned to donation more, so, their deferral number was higher. Therefore, training and informing can reduce their blood deferral numbers.