Influenza Vaccination and Non-Pharmaceutical Measure Effectiveness for Preventing Influenza Outbreaks in Schools: A Surveillance-Based Evaluation in Beijing

Abstract

Although schools are known to play a major role in the spread of influenza virus, few studies have evaluated the effectiveness of vaccination and non-pharmaceutical measures for preventing influenza outbreaks in schools. We investigated all febrile illness outbreaks in primary and secondary schools in Beijing reported between August 2018 and July 2019. We obtained epidemiological information on febrile illness outbreaks and oral pharyngeal swabs from students in the outbreaks to test for influenza virus. We surveyed schools that did not report febrile illness outbreaks. We developed multi-level models to identify and evaluate factors associated with serious influenza outbreaks and explored the association of vaccine coverage and outbreaks using multi-stage regression models. We identified a total of 748 febrile illness outbreaks involving 8176 students in Beijing; 462 outbreaks were caused by influenza virus. Adjusted regression modeling showed that large class size (odds ratio (OR) = 2.38) and the number of days from identification of the first case to initiation of an intervention (OR = 1.17) were statistically significant and positively associated with serious outbreaks, and that high vaccination coverage (relative risk (RR) = 0.50) was statistically significant and negatively associated with outbreaks. Multi-stage regression modeling showed that RR decreased fastest when vaccination coverage was 45% to 51%. We conclude that high influenza vaccination coverage can prevent influenza outbreaks in schools and that rapid identification of febrile children and early initiation of non-pharmaceutical measures can reduce outbreak size.

Keywords: febrile outbreak; influenza vaccination; multi-level model; multi-stage regression; non-pharmaceutical measure.

Figure 1

Flow chart and basic information of febrile outbreaks identified in Beijing from August 2018 to July 2019.

Figure 2

(A) Joinpoint model of relative risk (RR) values for different vaccination coverage. There are five joinpoints for the model. The symbol * is shown if the annual percent change (APC) is significantly different from zero at level 0.05. (B) RR values for vaccination coverage from 10% to 90% with 95% confidence interval.