Mask-Wearing Behavior at the 2021 NCAA Men’s Basketball Tournament

Large sporting events potentially increase the transmission of SARS-CoV-2 and other communicable diseases due to prolonged close contact and socializing with non–household members.¹ Although rigorous evidence is lacking, requiring face masks, restricting eating/drinking, reducing spectator capacity, encouraging physical distancing, and reserving space between groups of ticketed seats are strategies that attempt to reduce transmission risk. However, even when face masks are required, spectators’ mask-wearing behavior may be inconsistent. To support public health efforts in reducing disease spread and venues’ preparations for future events during COVID-19 or other epidemics, this study quantified spectators’ mask wearing during a high-profile national sporting event.

We conducted repeated cross-sectional aggregated counts of public face mask wearing during the NCAA (National Collegiate Athletic Association) Division I men’s basketball tournament, March Madness, in Indianapolis, Indiana, from March 30 to April 5, 2021. Spectators’ mask-wearing behavior was observed at 5 games in a large indoor stadium limited to 22% capacity. The venue, county, and state had mask requirements. Signage and speaker announcements encouraged mask wearing, physical distancing measures, and empty seating between groups. Ushers enforced compliance.

At multiple points during each game, 6 trained observers collected counts of masking behavior at entrance gates, concession areas, arena seating, upper-deck seating, and exit gates. Observers counted all spectators aged 2 years and older, stratified by sex (based on the observers’ best visual estimate of age and sex). At observation, spectators were classified as correctly masked when following Centers for Disease Control and Prevention guidelines (ie, a cloth face covering or disposable surgical mask that covered the mouth and nose, including the nostrils, and extended below the chin).² Counts were recorded via a locally developed web-based app, MaskCount (Regenstrief Institute Inc). Pretesting using images of public mask-wearing behavior suggested high interobserver agreement (κ = 0.93). Percentages and 95% CIs were calculated after stratifying observations by location and timing and weighting per total people observed. Differences in stratified proportions were assessed by the Wald test at P < .05 (2-sided) with Stata version 16.1 (StataCorp). This study was exempted by the Indiana University institutional review board as observation of public behavior.

The sample included 21 355 spectator observations. The percentage of correctly masked individuals (Table) was 73.9% (95% CI, 71.2%-76.7%). Of female spectators, 81.3% wore masks vs 69.8% of male spectators (difference, 11.5 [95% CI, 9.3-13.7] percentage points). Correct mask wearing was lowest in arena seating (66.7%; 95% CI, 61.9%-71.5%) and upper-deck seating areas (34.5%; 95% CI, 21.7%-47.3%). The highest mask-wearing percentages were in concession areas (83.3%; 95% CI, 81.6%-83.5%) and at entrances (80.3%; 95% CI, 60.8%-99.8%).

This study found that the proportion of spectators who wore masks correctly at 5 national men’s basketball games was 74%, varying by sex and location within the venue. This level was achieved at a venue with multiple enforcement and social messaging efforts and within a county and state with mask requirements. However, this estimate of mask wearing was below the 80% threshold suggested as necessary to reduce the spread of COVID-19.³ Variation between locations could be attributable to differences in enforcement. For example, ushers could not be as prominent in the upper decks due to space limitations, but spectators had to pass near staff at entrances and concession areas. Additionally, in seating areas, masks could be removed for eating/drinking or cheering and spectators may have felt less at risk when seated with members of their own group. Limitations of the study include generalizability to other large indoor gatherings. This study occurred before widespread vaccination availability, and many COVID-19 restrictions were in place owing to high community spread. However, it provides insight on masking behavior at a mass indoor event similar to other large indoor gatherings that have the potential to become superspreader events.⁴ When vaccinations are not available or coverage is low, ensuring adherence to key transmission mitigation strategies such as mask wearing during mass events is critical during the current, and any future, pandemic.

Corresponding Author: Joshua R. Vest, PhD, MPH, Indiana University Richard M. Fairbanks School of Public Health, 1050 Wishard Blvd, RG 6155, Indianapolis, IN 46202 (joshvest@iu.edu).

Accepted for Publication: August 4, 2021.

Published Online: August 16, 2021. doi:10.1001/jama.2021.14057

Author Contributions: Dr Vest had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: All authors.

Acquisition, analysis, or interpretation of data: Vest, Blackburn, Embi.

Drafting of the manuscript: Vest, Blackburn, Peters Bergquist, Embi.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Vest.

Obtained funding: Cash-Goldwasser, Embi.

Administrative, technical, or material support: Vest, Blackburn, Cash-Goldwasser, Peters Bergquist.

Supervision: Vest, Blackburn.

Conflict of Interest Disclosures: Dr Vest reported receiving grants from Resolve to Save Lives to the Indiana University Richard M. Fairbanks School of Public Health during the conduct of the study; receiving personal fees from New York eHealth Collaborative and Pima County for consulting services outside the submitted work; and being a founder and equity stakeholder in Uppstroms, a technology company. Dr Embi reported he is the inventor of the MaskCount web-based app used in this study, developed with support from the Regenstrief Institute, but has no personal financial interests in the product. No other disclosures were reported.

Funding/Support: This work was made possible through funding from Resolve to Save Lives, an initiative of Vital Strategies. Resolve to Save Lives is funded by Bloomberg Philanthropies, the Bill & Melinda Gates Foundation, and Gates Philanthropy Partners, which is funded with support from the Chan Zuckerberg Initiative. Resolve to Save Lives received additional funding for the COVID-19 response from the Stavros Niarchos Foundation and the Start Small Foundation.

Role of the Funder/Sponsor: Resolve to Save Lives had no role in the design and conduct of the study, or in collection, management, analysis, or interpretation of the data. Drs Cash-Goldwasser and Peters Bergquist from Resolve to Save Lives were involved in the preparation, review, and approval of the manuscript and decision to submit the manuscript for publication.

Additional Contributions: We thank Virginia Caine, MD, and Karen Holly, MBA, of the Marion County Public Health Department for logistical support; Katie Allen, BS, Amber Blackmon, MPH, Jyotsna Gutta, MPH, Richard Jones, BS, Nicholas Kiehl, Whitney Seeley, Yvette Tran, MS, Erin Vest, Valerie Yeager, DrPH, MPhil, and Lauren Yoder from Indiana University for data collection support; and Kate Elliott, MPH, MHSMA, and Sasha Litvinov, BA, from Resolve to Save Lives for editorial assistance. Ms Elliott and Ms Litvinov received financial compensation for their contributions.