Study Objectives
Rigorous SARS-CoV-2 testing is an important public health measure as it leads not only to early identification and prevention of transmission, but also to optimization of emergency care and resource allocation. Yet, the US has experienced a significant burden of illness, with reports suggesting a disproportionate amount falling on racial/ethnic minorities. Despite the public health importance, little is known about the discrepancies in the testing rate by region and race/ethnicity. In this context, we investigated the differences in and factors associated with per capita testing volumes.
Methods
This is an analysis of population-based data of nine racially/ethnically and geographically diverse states (AL, AZ, DE, FL, IN, NV, OR, TN, TX). We analyzed county-level testing data reported by state health departments and sociodemographic data reported by the U.S. Census Bureau. All data are as of June 7, 2020. The outcome was the number of SARS-CoV-2 testing (PCR and/or serology) per 1,000 individuals at the county-level. To identify factors associated with outcome, we fit a multivariable Poisson regression model including states, county-level death rate, mean household income, and proportion of major races/ethnicities.
Results
We examined data from 646 counties from nine states. The median rate of SARS-CoV-2 testing per 1,000 individuals differed widely, ranging from 15 in Texas to 54 in Delaware (Table). The multivariable model identified factors significantly associated with the rate of testing—state, death rate per 1,000, % non-Hispanic white, % non-Hispanic black, and % Hispanic (all P<0.05). For example, compared to Texas, higher testing rates were observed in Delaware (rate ratio [RR], 2.47) and Tennessee (RR, 2.92). In contrast, the magnitude of race/ethnicity-outcome association was smaller—eg, RR of 0.96 per 10% increase in non-Hispanic black and 0.85 per 10% increase in Hispanic demographics.
Conclusions
There were significant between-state differences in the SARS-CoV-2 testing rate. Counties with a higher proportion of race/ethnicity minorities had significantly lower testing rates while their magnitude of association was relatively small. Our findings should facilitate further investigations into the reasons for discrepancies, which will, in turn, optimize prevention and treatment strategies against this public health emergency.
Table.
Characteristics of Nine U.S. States
| State | Number of Counties, n | Tests Per 1,000,Median (IQR) | Deaths Per 1,000,Median (IQR) | Household Income ($), mean (SD) | Non-Hispanic White (%), Median (IQR) | Non-Hispanic Black (%), Median (IQR) | Hispanic (%), Median (IQR) |
|---|---|---|---|---|---|---|---|
| Alabama | 67 | 41 (35-54) | 0.08 (0.02-0.21) | 57,311 (11,084) | 69.0 (53.4-80.5) | 22.7 (11.1-42.6) | 2.5 (1.6-3.9) |
| Arizona | 15 | 46 (30-58) | 0.07 (0.00-0.20) | 62,191 (10,742) | 54.1 (44.3-57.5) | 1.2 (0.7-2.7) | 29.9 (15.3-36.2) |
| Delaware | 3 | 54 (51-78) | 0.36 (0.34-0.48) | 82,986 (9,856) | 62.2 (60.2-68.6) | 24.3 (18.3-24.4) | 9.1 (8.1-9.4) |
| Florida | 67 | 47 (41-55) | 0.06 (0.03-0.13) | 67,593 (14,754) | 72.0 (60.2-77.3) | 11.0 (8.2-17.7) | 9.4 (5.7-19.1) |
| Indiana | 92 | 31 (24-39) | 0.12 (0.03-0.35) | 67,588 (11,353) | 93.5 (88.0-95.5) | 1.0 (0.5-2.9) | 2.9 (1.7-4.8) |
| Nevada | 17 | 38 (19-63) | 0.00 (0.00-0.05) | 71,692 (13,030) | 72.2 (65.6-79.1) | 1.7 (0.6-2.5) | 16.6 (12.5-24.2) |
| Oregon | 36 | 30 (25-35) | 0.00 (0.00-0.03) | 67,558 (12,403) | 84.6 (76.7-87.6) | 0.6 (0.4-0.9) | 8.6 (6.4-14.2) |
| Tennessee | 95 | 46 (36-59) | 0.00 (0.00-0.04) | 60,840 (12,869) | 90.1 (84.6-93.4) | 3.4 (1.4-8.2) | 2.6 (1.9-4.3) |
| Texas | 254 | 15 (9-25) | 0.00 (0.00-0.05) | 68,689 (14,324) | 58.8 (41.9-73.1) | 3.6 (0.9-8.9) | 26.6 (18.1-50.2) |