Generalized Multilevel Multisource Functional Regression With an Application to Alzheimer's Disease.

The 2019 Super Bowl LIII and its multi-day, pre-game festivities, which were held in the City of Atlanta, Georgia and surrounding municipalities, presented the opportunity to assess the feasibility of using 9-1-1 emergency medical dispatch (EMD) data for improving syndromic surveillance of planned mass gatherings and their host communities. Syndromic surveillance includes public health-related activities that involve data collection and analyses to rapidly identify and, if necessary, respond to injury and disease outbreaks. Although the Super Bowl championship game was played in Atlanta, there is a large event-related entertainment venue in the City of College Park. Furthermore, the necessity of traversing a contiguous third municipality (East Point) to reach College Park from Atlanta provided the rationale for its study inclusion. At the time of Super Bowl LIII, 540,000 citizens of, and visitors to, these municipalities were potentially exposed to and could have been affected by health threats associated with the events festivities. All 9-1-1 medical calls from these three municipalities are routed to the Grady EMS (GEMS) Emergency Communications Center for standardized processing through the EMD call-processing software program. The Super Bowl LIII Game and pre-game festivities were hosted across a nine-day period in the late January–February 2019. Concerts, festivals, parties, awards ceremonies, military aerial exhibitions, permitted rallies, and protests were held preceding the championship game. According to a media-based report, “more than one-half million people went to Super Bowl Experience and Super Bowl Live.” There were no known official estimates of attendance at > 100 related overlapping events held at various times across the nine days, Following this multi-day, multi-festivity mass gathering, the host cities’ 9-1-1 EMD total call volumes, selected specific types, call-time periodicity, and geographic patterns of call origin were investigated to search for data patterns that could have represented threats to community health status. The resulting data were modeled for practical vision-based analysis, in a search for possible roles in conducting mass gathering syndromic surveillance.