key: cord-1043367-ebyvjyj9
authors: Ahmed, S. M.; Shah, R. U.; Bale, M.; Peacock, J. B.; Berger, B.; Brown, A.; Mann, S.; West, W.; Martin, V.; Fernandez, V.; Grineski, S.; Brintz, B. J.; Samore, M. H.; Ferrari, M. J.; Leung, D. T.; Keegan, L. T.
title: Comprehensive Testing Highlights Racial, Ethnic, and Age Disparities in the COVID-19 Outbreak
date: 2020-05-09
journal: nan
DOI: 10.1101/2020.05.05.20092031
sha: 070aba42665dc6aed90a5b914d3877ea3a3ffe0c
doc_id: 1043367
cord_uid: ebyvjyj9

The United States (US), which is currently the epicenter for the COVID-19 pandemic, is a country whose demographic composition differs from that of other highly-impacted countries. US-based descriptions of SARS-CoV-2 infections have, for the most part, focused on patient populations with severe disease, captured in areas with limited testing capacity. The objective of this study is to compare characteristics of positive and negative SARS-CoV-2 patients, in a population primarily comprised of mild and moderate infections, identified from comprehensive population-level testing. Here, we extracted demographics, comorbidities, and vital signs from 20,088 patients who were tested for SARS-CoV-2 at University of Utah Health clinics, in Salt Lake County, Utah; and for a subset of tested patients, we performed manual chart review to examine symptoms and exposure risks. To determine risk factors for testing positive, we used logistic regression to calculate the odds of testing positive, adjusting for symptoms and prior exposure. Of the 20,088 individuals, 1,229 (6.1%) tested positive for SARS-CoV-2. We found that Non-White persons were more likely to test positive compared to non-Hispanic Whites (adjOR=1.1, 95% CI: 0.8, 1.6), and that this increased risk is more pronounced among Hispanic or Latino persons (adjOR=2.0, 95%CI: 1.3, 3.1). However, we did not find differences in the duration of symptoms nor type of symptom presentation between non-Hispanic White and non-White individuals. We found that risk of hospitalization increases with age (adjOR=6.9 95% CI: 2.1, 22.5 for age 60+ compared to 0-19), and additionally show that younger individuals (aged 0-19), were underrepresented both in overall rates of testing as well as rates of testing positive. We did not find major race/ethnic differences in hospitalization rates. In this analysis of predominantly non-hospitalized individuals tested for SARS-CoV-2, enabled by expansive testing capacity, we found disparities in both testing and SARS-CoV-2 infection status by race/ethnicity and by age. Further work on addressing racial and ethnic disparities, particularly among Hispanic/Latino communities (where SARS-CoV-2 may be spreading more rapidly due to increased exposure and comparatively reduced testing), will be needed to effectively combat COVID-19 in the US.

symptoms), and gross underestimates of the true number of fatalities [13] [14] [15] [16] [17] [18] . In contrast, Utah is 72 one of only two states (as well as Tennessee) that has maintained high per capita testing rates 73 (testing 188 per 100,000 persons per day) as well as a stable proportion of positive tests 74 (approximately 5.5% of all tests were positive since 12 March) (Figure 1 ) 19 . As such, testing in 75 Utah has not been restricted to the most critical patients, and the majority of SARS-CoV-2 tests 76 are administered in outpatient settings to community-dwelling, social distancing persons. 77 78

The objective of this study is to compare demographic characteristics of positive and negative 79 patients in a population for which testing is more readily available than previously published 80 cohorts. In this article, we present the clinical, epidemiological, and demographic characteristics 81 of all persons (hospitalized and community-dwelling) tested for SARS-CoV-2 infection at 82

University of Utah Health (UHealth) from 10 March to 24 April 2020. UHealth primarily serves 83

Salt Lake County, a diverse, medium-density, medium-sized metropolitan area of just over 1 84 million people. Due to the expanded testing capacity for SARS-CoV-2 in Utah, these data 85

provide an alternative picture of the full spectrum of SARS-CoV-2 epidemiology relative to 86 studies of hospitalized patients. 87 88

Methods 89

All patients tested for SARS-CoV-2 in the UHealth system were eligible for our study. We 90 developed an electronic, near real-time registry of all patients tested for SARS-CoV-2 at UHealth 91 prior to 30 April 2020, for a total sample size of 20 Figure 2 ). Likewise, when 152

comparing the location where individuals presented for testing (e.g., ED, outpatient clinic, 153 mobile testing site), we found variation by race/ethnicity (p<0.001), but Hispanic/Latino persons 154

were only slightly more likely to be tested in the ED compared to non-Hispanic White/Caucasian 155 (6.3% versus 4.9%). We did not find a difference in reported symptoms between 156

Hispanic/Latino and non-Hispanic White/Caucasian persons. 157 158

Age Structure 159

The population tested at UHealth is not similar to the age structure of Salt Lake County ( Figure  160 3). However, the age distribution of those tested at UHealth is consistent between individuals 161 testing negative versus positive. Younger individuals (aged 0-19), were underrepresented in 162 overall rates of testing and rates of testing positive (Table 1, Figure 3 ). Without adjusting for the 163 underlying age distribution of the county population, we found that persons aged 20-59 had the 164 highest odds of testing positive compared to those 0-19 years old (adjOR=1.6, 95% CI: 0.7, 3.6, 165 Table 1).  166 167

Sex 168

Persons who tested positive for SARS-CoV-2 were more likely to be male. Among individuals 169 who tested positive for SARS-CoV-2, over 56% were male (44% were female), while 44% of 170 individuals who tested negative were male (56% of female). Excluding individuals for which 171 gender was unknown or non-binary (11 people), males had 1.4 times the odds of testing SARS-172

COV-2 positive compared to females (95% CI: 1.0, 1.9). 173 174

Clinical Characteristics 175

Among manually extracted charts, 136 (6.7%) individuals tested positive for SARS-CoV-2 176 between 10 March 2020 and 31 March 2020 (Table 2 ). This was similar to the proportion of 177 individuals testing positive for SARS-CoV-2 among all those tested 1229 (6.5%).

Clinical symptoms 180

Persons testing SARS-COV-2-positive and SARS-COV-2-negative had similar vital signs and 181

rates of cough (89%), but varied in presentation of other symptoms. Symptom profiles were 182 different in persons who tested SARS-COV-2 positive compared to those who tested SARS-183 COV-2 negative ( Table 2 ). The median duration between symptom onset and presentation for 184 testing was 4 days (IQR=2.1, 7.0) for both SARS-COV-2 positive and SARS-COV-2 negative 185 persons (Table 2) . 186 187

Hospitalization rates 188

Using all-cause hospitalization occurrence within 14 days after testing, we found that 5.1% of 189 persons tested positive were hospitalized, and 5.1% of persons testing negative were 190

hospitalized. Rates of hospitalization were similar across racial ethnic groups, with 6% of people 191

identifying as Black/African American, 5% of people identifying as non-Hispanic 192

White/Caucasian, and 5% of people identifying as Hispanic/Latino being hospitalized within 14 193

days of SARS-CoV-2 testing. Those identifying as American Indian and Alaskan Native had 194 higher rates of 14-day hospitalization (9%), but the sample size was limited. 195 196 We found that hospitalization differed by age, and older adults had the highest odds of all-cause 197 hospitalization (adjOR=6.9, 95% CI: 2.1, 22.5 for adults 60+ compared to 0-19-year olds; Table  198 1, Figure 3 ). Additionally, we found all-cause hospitalization rates increased with age in both 199 SARS-COV-2-positive and SARS-COV-2-negative persons (Figure 3d ).

Epidemiological Risk Factors 202

Persons testing SARS-CoV-2 positive were more likely to have known exposure to another 203 SARS-CoV-2 person or a history of travel (Table 2) . Among SARS-COV-2-positive persons, 204 57% had been in contact with a confirmed SARS-CoV-2 case and 46% had travelled before their 205

own SARS-CoV-2 testing, compared to 29% and 24% in SARS-COV-2-negative persons, 206 respectively ( Table 2 ). 9% of SARS-COV-2-positive persons and 11% of SARS-COV-2-207 negative persons were healthcare workers (Table 2) . 208 All rights reserved. No reuse allowed without permission.

was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint (which this version posted May 9, 2020. The high-level of SARS-CoV-2 testing availability in Utah has allowed for a higher proportion 211 of the outpatient, community dwelling (and usually social distancing) population to be tested. 212

This expands on previous studies that were limited to more severe or critically-ill patients, as a 213 result of limited test availability. These data provide insight into the epidemiological and clinical 214 characteristics of COVID-19 over a much broader range of disease severity, as well as highlight 215 disparities in COVID-19 response. While the data presented in this study represents only 20% of 216

the total number of people tested in Utah, based on publicly available state-level data, our 217

findings are largely representative of the overall state-level outbreak. 218 219

While the racial/ethnic makeup of the population tested for SARS-CoV-2 is broadly similar to 220 overall county demographics (Figure 2) , important discrepancies remain. We found that persons 221 who identified as non-White had elevated odds of testing positive compared to non-Hispanic 222

White/Caucasians (adjOR=1.1, 95% CI: 0.8, 1.6). The discrepancy was especially pronounced 223

for persons identifying as Hispanic or Latino, who had 2.0 times the odds (95% CI: 1.3, 3.1) of 224

testing SARS-CoV-2 positive compared to those who identified as non-Hispanic White ( Figure  225 2). Though our study is limited to those within a single health system, our results are 226

representative to those of the county and the state as a whole. Persons identifying as non-Hispanic White/Caucasian and Hispanic/Latino were tested at each 248

facility at approximately equal rates -approximately 5% were tested in the ED, and 90% in 249 outpatient clinics. Likewise, using symptom presentation or hypoxia (SpO2 < 90%) as a proxy 250

for severity, we find no difference between persons identifying as non-Hispanic White/Caucasian 251 and as Hispanic/Latino. The duration of time from symptom onset to seeking testing was also 252 similar across races. This suggests racial disparities in SARS-COV-2 burden are not simply due 253

to Hispanic/Latino persons delaying care until symptoms are critical and presenting to the ED, 254 All rights reserved. No reuse allowed without permission.

was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint (which this version posted May 9, 2020. . https://doi.org/10.1101/2020.05.05.20092031 doi: medRxiv preprint nor is it as straight forward as preferentially presenting to mobile testing centers. Identifying the 255 underlying cause of these health disparities is a critical step towards containing SARS-CoV-2. 256 257

Unlike previous studies 6-11 that focus only on severe or critical patients, we are also able to 258 identify differential patterns of testing by age. Promisingly, we find most adults (aged 20-79) in 259 our study have high testing rates (over 100 tests per 10,000 population), which indicates that we 260 are capturing large numbers of mild and moderate illness. On the other hand, we also show that 261 testing rates remain lower in children (under 20 years old), and of those under 20 years old who 262 are tested, rates of SARS-CoV-2 positivity are disproportionately lower. This supports previous 263 studies which highlight risk of asymptomatic SARS-CoV-2 in younger individuals, as inclusion 264 in our study was predicated by having at least one symptom (cough, fever, or shortness of breath) 265 20 . 266 267

In addition to differences due to race, we also identified differences due to age. We found the 268 highest hospitalization rates were among middle-aged SARS-COV-2-positive patients (Figure 3 ). 269

This is broadly consistent with Myers et al. 6 who found hospitalized and ICU admitted patients 270

in California tended to be middle-aged. Like similar studies 8 , we found hospitalization rates 271

were smallest among younger age individuals (aged 0-19). Indeed, we found persons under 20 272 years old who tested positive for SARS-CoV-2 were far less likely to be hospitalized than their 273 peers who were tested for SARS-CoV-2 but tested negative. Compared to other studies, we 274

found age-adjusted lower hospitalization rates than previously reported among those who test 275

positive. 276 277

Given that Utah has the youngest population in the country 33 , a disease that is more likely to be 278 asymptomatic in younger people is especially salient for the overall state burden and the ability 279 to control the outbreak as social distancing measures are loosened. Given our finding that 280 younger individuals are underrepresented in testing due in part to testing criteria, state-level 281 strategies should be re-examined to consider strategies to target testing of younger age groups 282 based on geospatial and epidemiological risk factors, regardless of symptomatology. 283 284

A benefit of expanded testing locations in Salt Lake County was that we were able to assess the 285 demographic characteristics of testing locations. We found that those tested in outpatient 286 facilities tended to be younger and that the proportion of persons tested in the ED increased with 287 age. We found no difference in race/ethnicity by testing location. If testing were to be expanded 288 to lift symptom requirements in an effort to identify more mild and asymptomatic infections, our 289 study highlights the benefit of outpatient and mobile facilities to reach underrepresented 290 demographics. 291 292

This study offers an important contribution to existing COVID-19 literature, which has been 293 primarily focused on hospitalized patients. In contrast to previous work, over 90% of people 294

included in our study were tested in outpatient or community settings. However, our data is 295 limited to persons who met testing criteria; very few asymptomatic infections are captured in this 296 study. Further work understanding the demographic and epidemiological characteristics of 297 asymptomatic infections remains vital to understanding and controlling SARS-CoV-2. By 298 highlighting critical gaps in testing, particularly among Hispanic/Latino communities, where 299 SARS-CoV-2 may be spreading more rapidly due to increased exposure and comparatively 300 reduced testing, this study takes a critical first step towards reversing these health disparities. Our 301 findings, based largely on those who received testing as outpatients, highlights the potential gaps 302 in control of SARS-CoV-2 infection related to age, race, and ethnicity. 303 304 Table 1 Not Calculated All rights reserved. No reuse allowed without permission.

was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint (which this version posted May 9, 2020. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint (which this version posted May 9, 2020. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint (which this version posted May 9, 2020. was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint (which this version posted May 9, 2020. . https://doi.org/10.1101/2020.05.05.20092031 doi: medRxiv preprint was not certified by peer review) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint (which this version posted May 9, 2020. . https://doi.org/10.1101/2020.05.05.20092031 doi: medRxiv preprint

An interactive web-based dashboard to track COVID-19 in real 353 time

Epidemiologic features and clinical course of 355 patients infected with SARS-CoV-2 in Singapore

Characteristics of and important lessons from the coronavirus 357 disease 2019 (COVID-19) outbreak in China: summary of a report of 72 314 cases from 358 the Chinese Center for Disease Control and Prevention

Risk of acquiring antibiotic-resistant bacteria from prior room 360 occupants

Epidemiological characteristics of COVID-19 362 cases in Italy and estimates of the reproductive numbers one month into the epidemic

Characteristics of Hospitalized Adults With 365 COVID-19 in an Integrated Health Care System in California

Characteristics and outcomes of 21 critically ill patients 367 with COVID-19 in Washington State

370 MMWR Morb Mortal Wkly Rep. 2020;69. 371 9. Gold JAW. Characteristics and Clinical Outcomes of Adult Patients Hospitalized with 372 COVID-19-Georgia

Racial and Ethnic Disparities in SARS-CoV-2 376 Pandemic: Analysis of a COVID-19 Observational Registry for a Diverse US 377 Metropolitan Population. medRxiv. 2020. 378 12. Project CT. Most recent data. COVID Track Proj

Geographic differences in COVID-19 cases, deaths, 380 and incidence-United States

Why Only Test Symptomatic Patients? Consider Random Screening for 382 COVID-19

Substantial undocumented infection facilitates the rapid 384 dissemination of novel coronavirus (SARS-CoV2)

Diagnostic testing for the novel coronavirus

Prevalence of SARS-CoV-2 Infection in 391 Residents of a Large Homeless Shelter in Boston

Coronavirus Disease 395 2019 in Children-United States

Disease Control TC, Prevention. COVID-19 in Racial and Ethnic Minority Groups

Labor Force Statistics from the Current Population 401

Accessed May 1, 2020. 403 23. of Labor Statistics USB

Determinants of rural Latino trust in the federal 407 government

Racial/ethnic differences in physician 409 distrust in the United States

Untangling the associations among distrust, race, and neighborhood 411 social environment: A social disorganization perspective

Thiede M. Information and access to health care: is there a role for trust?

Patients' race, ethnicity, 417 language, and trust in a physician

Census US. Median Age in 2010 and 2018 for the United States, Regions, and States. 426 United States Census Bureau