Revisiting the Question of Fan Attendance at Professional Sporting Events

Abstract

Whether or not to allow fan attendance at collegiate and professional sporting events during the COVID-19 pandemic has been hotly contested by fans, politicians, and public health officials. The decision may have substantial public health and economic consequences, but these consequences are difficult to estimate using traditional investigative epidemiology techniques. Surge capacity of contact tracing is overwhelmed by large crowds-often drawn from a wide geographic area-and the game day experience makes it difficult to determine when infection may have occurred: Was the infection acquired at the event? At a tailgate party beforehand? Traveling to or from the event?Kurland and colleagues 1 have tackled this question using an ecological approach to capturing the broad, population-level association of National Football League games with COVID-19 rates in the immediate and neighboring counties. This approach is in contrast to studies focusing on a smaller, more targeted population such as the players themselves 2,3 or university populations. 4 They find that hosting games with 20 000 or more fans in attendance was associated with COVID-19 infection rate spikes 14 and 21 days after the game. These spikes were over twice the spikes in areas hosting games without fans in the seats. Furthermore, games with fewer than 5000 fans were not associated with increased caseloads.This study is similar in design to a previous report by Toumi and colleagues, 5 but Kurland and colleagues 1 use a broader base of National Football League games and consider a wider geographic area. As a result, this approach has the potential to capture the outcomes of not only the game itself, but ancillary activities surrounding fan participation. Using a relatively new anomaly detectionbased statistical approach based on detecting departures from 7-, 14-, and 21-day moving averages,