Contact, travel, and transmission: The impact of winter holidays on influenza dynamics in the United States

Abstract

Background: The seasonality of influenza is thought to vary depending on changes in environmental factors, human contact, and travel patterns. During winter holidays, when children are out of school, typical patterns of potential disease-causing contact and travel change. The aim of this research is to understand the impact of winter school holidays on age-specific and spatial flu transmission, and the resulting size and spread of the epidemic. We hypothesize that reduced contacts among children during school holidays will lower their infection rates, infection rates among adults will subsequently decrease as fewer children are sick, and increased travel during the holidays will hasten the spread of flu across the country.Methods: Using aggregated medical claims data from physician offices and hospitals in the United States, we examined influenza dynamics before, during, and after the winter holidays from 2001 to 2009. We quantified the changes to transmission and characterized holiday-associated changes to epidemic trajectories among children and adults in a spatial context. Informed by this empirical evidence, we used a deterministic, age-specific spatial metapopulation model to compare models with and without holiday-associated behavioral changes to understand the effects of holidays on flu season trajectories and spatial spread. Results:In analyzing eight seasons of medical claims data, we observed a mean 15% drop in the effective reproductive number immediately after Christmas. Rates of influenza-like illness among children experienced a sharp dip and recovery during and after the holidays, while adults experienced more gradual dips and recoveries during these periods. The model results demonstrated that both child and adult flu incidence dipped during the holiday period, but trends among adults lagged those in children. The peaks of the child and adult flu epidemics were delayed when the holiday was introduced, but the attack rate did not change significantly for either age group. In the medical claims data and the model-generated data with the holiday, the spatial synchrony of epidemic trajectories before and after the holiday remained similar. In the model-generated data without the holiday, however, epidemic trajectories became spatially asynchronous.Conclusions: Our work suggests that winter holidays delay the epidemic peaks and increase the spatial synchrony of natural trajectories for seasonal influenza in the United States every year. These findings have implications for the design of influenza intervention strategies, such as the proper timing and duration of school closures, and the spatial and demographic reallocation of vaccines.

Double trouble? When a pandemic and seasonal virus collide

“…To understand the process of respiratory disease transmission and mitigation, we developed a metapopulation model that represents age-specific contact between children and adults within metropolitan areas and is spatially divided into metropolitan areas linked through air tra c flows. The model is described in detail in [7]. We simulated a seasonal influenza epidemic in the US, including typical holiday school closure, and seasonality with the season ending at April 30, 2020.…”

“…Influenza dynamics are often evaluated through measuring influenza-like illness (ILI), which tracks cases of individuals with symptoms consistent with influenza, but where influenza is not laboratory confirmed. Typical ILI dynamics in an influenza season exhibit a small pre-holiday peak, followed by a larger post-holiday peak [7]. Recent ILI data for the United States show a markedly lower epidemic peak for the 2019-2020 influenza season compared to what would be expected based on the pre-holiday peak [8].…”

Effects of weather-related social distancing on city-scale transmission of respiratory viruses

“…School closures may reduce rates of medically-attended influenza in school-aged children, although with highly heterogeneous effects (2% to 29% reductions), and with lesser effects on younger children and adults. [1][2][3][4][5] Generalizing to broader social distancing efforts from these studies is difficult, however, as school closures tend to have limited impacts on working-age and older adults, and school-aged children may recongregate outside of schools. 4,[6][7][8] In February 2019, unusually high snowfall in western Washington State led to widespread school and workplace closures and to reduced regional travel.…”

Identifying factors that may improve mechanistic forecasting models for influenza

“…For example, the impact of school 20 closures on transmission has been modeled for cities, such as Hong Kong [12], and 21 Countries, such as France [13], while the contribution of climatic drivers has been 22 assessed for U.S. states [14] as well as individual cities [8]. 23 Meteorological conditions have long been thought to play a role in the transmission 24 dynamics of influenza [15]. Heuristically at least, this is supported by the radically 25 different evolutions of ILI profiles in tropical versus temperate zones [16].…”

“…More 231 recently, [24] showed that school vacations delay epidemic peaks and act to synchronize 232 incidence profiles at different locations. The effects of humidity in modulating influenza 233 transmission have also been well studied [8,14,25], including the benefits of ensemble 234 models incorporating specific humidity, which could be used to provide forecasts in 235 real-time [9].…”