Assessment of reactive school closures and optimizing the effects in mitigating influenza epidemics in Hong Kong.

Abstract

<p><strong>Background & aims of study</strong></p><p>Reactive school closures are the non-pharmaceutical intervention, often implemented to mitigate the influenza transmission in community by increasing social-distancing (avoiding effective contacts). In Hong Kong, school closures have been used to mitigate seasonal influenza in 2008, pandemic influenza A(H1N1)pdm09 in 2009, seasonal influenza B in 2018, seasonal influenza A(H1N1) in 2019 and most recently for COVID-19 in 2020.</p><p><br></p><p><strong>Methods & results</strong></p><p>We analyzed virological and syndromic surveillance data on influenza activity in Hong Kong during these 5 seasons across the types and sub-types of influenza viruses (in total 13 epidemics scenarios). We first estimated transmissibility by evaluating the effective reproduction number ( ) and examined changes in transmissibility during the school closure periods. Using multivariate regression model and state-space (compartmental) models, we quantified the impact of these closures on peak incidence and overall incidence of infections (attack rate) and hospitalizations by comparing the simulated counterfactual epidemics with and without school closures. Further we explored data driven frameworks to simulated possible counterfactual/experimental scenarios to optimize the timing and duration of the school closures. We estimated a 16% (95% CI: 10%, 26%) to 41% (95% CI: 35%, 46%) reduction in transmissibility for these school closure interventions in Hong Kong. The simulated incidence under the counterfactual scenario of no school closures during the implemented school closure, estimating that closures led to a reduction by 4.2% (95% CI 1.5%, 6.7%) to 21.0% (95% CI 10.7, 26.7%) in the attack rates. There was considerable variation in the impact of closures depending on the timing of implementation (before/around/after peaks). We found that for a given length of school closure and health capacity, the implementation timing around the end of exponential phase to peak could return the optimum impact on these outcomes.</p><p><br></p><p><strong>Implications</strong></p><p>School closures implemented around the end of exponential phase to peak had higher impact in reduction overall infections compare to other timing of epidemic. Reductions in incidence of infections should have translated to reduced hospitalisations and deaths by a similar fraction based on severity, with the caveat that most infections occur in children while most deaths occur in older adults.<br></p>