Aberration detection in influenza trends in Iran by using cumulative sum chart and period regression

Abstract

ObjectivesThis study aims to determine the alarm thresholds in influenza outbreaks and aberration detection in the influenza trend in Iran by using cumulative sum control chart (CUSUM) and period regression. MethodsWe used the weekly reported influenza-positive (types A and B) cases from Iran between January 2015 and November 2019. The period regression model and CUSUM chart were used as detection algorithms to figure out the alarm thresholds. ResultsThe mean ± SD and the median (95% CI) of the determined threshold per week were 34.85 ± 15.29 and 28.30 (17.67–64.62). According to the period regression, there were nine epidemic periods of influenza from 2015 to 2019. By using the CUSUM and considering a different h (h is an appropriate value that leads to the desired estimation for upper control limit) for the calculation of the upper control limit, 88, 84, 73 and 67 weeks were determined as the epidemic period. ConclusionAccording to the current study, the incidence of influenza showed a cyclic pattern and the epidemic recurred each year. Understanding this cyclical pattern can help health policymakers launch prevention programs such as vaccination during certain months of the year.