SAS procedure for relapses of hepatitis based on a negative binomial model

Objective We used a negative binomial regression model to explore its goodness of fit with the real-world relapses of hepatitis,aiming at providing scientific evidences for analysis of numeric count data.Methods The follow-up data of patients with hepatitis B and hepatitis C between June and December in 2007,delivered by a single physician from a Taiyuan infection hospital,was analyzed using the GENMOD(generalized linear model) process of SAS to fit the negative binomial regression and to look for influencing factors of hepatitis relapse.Results Relapses of hepatitis increased with advanced age,long duration of the disease,poor confidence with treatment and the strong belief that financial factors are related with outcomes.Couclusion The process of SAS/GENMOD based on fitness with negative binomial regresions provided simple and formatted outputs covering common link functions and probability distributions.The results may be useful to explain the relapses of hepatitis.