Evaluation of suspected measles incidence rate trend in Iran and the affecting factors: Negative-Binomial mixed model

The identifying incidence rate trend of disease and its changes lead to update response of surveillance system. Syndromic surveillance system is based on the suspected cases, so it has high speed in detecting outbreaks. This study aimed to evaluate trend of fever and rash incidence rates and detect affecting factors.

This study was a retrospective cohort study and the data included the suspected measles cases in provinces of Iran in 1977-2012, which extracted from surveillance system of vaccine preventable diseases. We fitted Poisson and Negative-Binomial regression models with random effect. Modeling and inferences were based on a Bayesian algorithm. We used R and OpenBUGS software. The fitted models were compared based on Deviance and Chi-square goodness of fit statistics.

Interaction effect between year and immunization campaign was statistically significant (95% CrI:1.083,1.737), after immunization campaign, trend was increasing. The variance of random component in model was statistically significant (95% CrI: 0.219,0.430). On the other hand, province-specific characterizes found affecting factor on suspected incidence rate.

In attention to increasing trend of this incidence in Iran, especially in recently years, and affecting of province-specific characterizes on suspected incidence rate, We found that more accurate control and improvement of quality vaccination is essential