Using geospatial data to eliminate disparities in covid-19 testing

Heath, G. W.; Moreland, T.; Stephenson, S.; Houser, J. A.; Schmitt, C.

Introduction: The first case of COVID-19 in Chattanooga/Hamilton County, Tennessee (CHC) was identified on March 13, 2020. By early April, 51 RT-PCR confirmed cases were identified, with white, non-Hispanic males and females representing 82% (41/51) of positive cases and remaining cases representing black residents (18%;9/51). That few people from racial/ethnic minorities were being tested became a key public health concern. We hypothesized that local mapping of health-related data would identify regions where individuals at greater risk for COVID-19 live and work and have limited access to testing and healthcare services. Methods: The CDC 500 Cities data was used to generate layered maps of prevalence estimates for cardiovascular disease, type 2 diabetes mellitus, chronic lung disease, and the behavioral risk factors of physical inactivity and obesity. Layers also included the CDC Social Vulnerability Index, age distribution, gender, race, ethnicity, and zip codes. Maps were shared with intersectoral collaborators representing the black and Hispanic/Latinx communities who provided specific neighborhood information to the maps. Collaborators included hospital systems, the local health department, community health centers, the private sector, and non-profit organizations. Maps were used to identify geographic sites for mobile and strategic testing within communities at higher risk for the spread of the coronavirus. Specific diverse neighborhoods along with worksites were then provided with testing beginning in early May and ongoing. Results: Strategic and mobile testing beginning in early May increased three-fold the number of identified new cases of COVID-19. Seventy percent (652/932) of these positive tests were amongethnically Hispanic/Latinx and 16% (149/932) among black residents. Positive tests continued to increase at a greater rate among Hispanic/Latinx and black residents compared with white residentsthrough the months of May-July (68/10K vs. 2.6/10K, OR = 4.85, 95% CI 2.66, 9.02). Conclusions: This example of inter-sectoral collaboration, data sharing, and data use throughstrategic mapping of vulnerable populations for COVID-19 was an effective means to enhanceCOVID-19 testing and identification of positive cases throughout CHC. This expanded partneringresulting in targeted testing may be a useful approach among similar communities and subsequentoutbreaks.