key: cord-0863706-t8upoq9a
authors: Yoon, Jane C; Montgomery, Martha P; Buff, Ann M; Boyd, Andrew T; Jamison, Calla; Hernandez, Alfonso; Schmit, Kristine; Shah, Sarita; Ajoku, Sophia; Holland, David P; Prieto, Juliana; Smith, Sasha; Swancutt, Mark A; Turner, Kim; Andrews, Tom; Flowers, Kevin; Wells, Alyssa; Marchman, Cathryn; Laney, Emaline; Bixler, Danae; Cavanaugh, Sean; Flowers, Nicole; Gaffga, Nicholas; Ko, Jean Y; Paulin, Heather N; Weng, Mark K; Mosites, Emily; Morris, Sapna Bamrah
title: COVID-19 Prevalence among People Experiencing Homelessness and Homelessness Service Staff during Early Community Transmission in Atlanta, Georgia, April–May 2020
date: 2020-09-08
journal: Clin Infect Dis
DOI: 10.1093/cid/ciaa1340
sha: c01dda807a831b5139f4be4776e8478f73ac80e1
doc_id: 863706
cord_uid: t8upoq9a

BACKGROUND: In response to reported COVID-19 outbreaks among people experiencing homelessness (PEH) in other U.S. cities, we conducted multiple, proactive, facility-wide testing events for PEH living sheltered and unsheltered and homelessness service staff in Atlanta, Georgia. We describe SARS-CoV-2 prevalence and associated symptoms and review shelter infection prevention and control (IPC) policies METHODS: PEH and staff were tested for SARS-CoV-2 by reverse transcription polymerase chain reaction (RT-PCR) during April 7–May 6, 2020. A subset of PEH and staff was screened for symptoms. Shelter assessments were conducted concurrently at a convenience sample of shelters using a standardized questionnaire RESULTS: Overall, 2,875 individuals at 24 shelters and nine unsheltered outreach events underwent SARS-CoV-2 testing and 2,860 (99.5%) had conclusive test results. SARS-CoV-2 prevalence was 2.1% (36/1,684) among PEH living sheltered, 0.5% (3/628) among PEH living unsheltered, and 1.3% (7/548) among staff. Reporting fever, cough, or shortness of breath in the last week during symptom screening was 14% sensitive and 89% specific for identifying COVID-19 cases compared with RT-PCR. Prevalence by shelter ranged 0%–27.6%. Repeat testing 3–4 weeks later at four shelters documented decreased SARS-CoV-2 prevalence (0%–3.9%). Nine of 24 shelters completed shelter assessments and implemented IPC measures as part of the COVID-19 response CONCLUSIONS: PEH living in shelters experienced higher SARS-CoV-2 prevalence compared with PEH living unsheltered. Facility-wide testing in congregate settings allowed for identification and isolation of COVID-19 cases and is an important strategy to interrupt SARS-CoV-2 transmission

In 2019, approximately 570,000 people experienced homelessness on any given night in the United States (U.S.), and 63% used congregate shelters [1] . In Atlanta, Georgia, an estimated 3, 200 people experienced homelessness on any given night in 2019, and approximately one-quarter were living unsheltered (i.e., living in a place not meant for human habitation) [2] . Risk of SARS-CoV-2 infection, the virus that causes COVID-19, may be higher among people experiencing homelessness (PEH) because of challenges in preventing respiratory disease transmission in congregate shelter settings. PEH might also be at increased risk of severe COVID-19 if infected due to a high prevalence of untreated, chronic medical conditions and obstacles to accessing healthcare [3] [4] [5] [6] [7] [8] .

Fulton County, the largest county in Georgia, which includes 90% of the city of Atlanta, reported the first COVID-19 case on March 2, 2020. A sharp increase in cases was recorded in mid-April 2020. A door-to-door household survey conducted in Fulton and neighboring DeKalb counties during April 28-May 3, 2020, found an estimated 2.5% seroprevalence of SARS-CoV-2 antibodies [9] . Reports of high SARS-CoV-2 infection rates and outbreaks within shelters in other metropolitan areas, in parallel with increasing local case-rates, led to concerns for widespread transmission in Atlanta shelters [10] [11] [12] .

To understand SARS-CoV-2 prevalence and prevent transmission among PEH in Atlanta, homeless service agencies partnered with local and federal government agencies to: (1) determine SARS-CoV-2 prevalence among clients living sheltered and unsheltered and homelessness service staff through viral testing; (2) A c c e p t e d M a n u s c r i p t 6 Methods Participants included clients living in shelters, clients living unsheltered, and staff in Atlanta, Georgia, during April 7-May 6, 2020. Testing at homeless shelters was offered facility-wide to all clients and staff. Testing was offered to clients living unsheltered during homeless outreach service events (e.g., meal services). Participation was voluntary but encouraged by service agencies. Written consent was obtained from each adult (≥16 years of age) or parent or guardian (for children <16 years) for administration of a brief, standardized screening questionnaire and testing for SARS-CoV-2 (see Supplemental Material #1). All screening interviews and specimen collections were conducted on-site at shelters or community events serving PEH. At shelters with more than five people with positive SARS-CoV-2 results upon initial testing, clients and staff were re-screened and re-tested 3-4 weeks later, and testing was also offered to any new clients or staff.

All participants (including parents or guardians of children <16 years) self-reported their sex, race, and ethnicity based on fixed-response categories. Race and ethnicity were combined into mutually exclusive categories and were considered missing when race was missing and ethnicity was either non-Hispanic or missing. Participants from a convenience sample of testing events were interviewed using the screening questionnaire to collect information on symptoms, underlying medical conditions, pregnancy status, and tobacco use. Persons interviewed were asked if they had any medical conditions in the following categories: diabetes, cardiovascular disease, chronic lung, kidney, and liver disease, immunocompromising conditions (e.g., HIV, chronic steroid use), and neurological conditions (e.g., seizure disorder). Nasopharyngeal, oropharyngeal, or nasal mid-turbinate specimens were collected by clinical providers or supervised self-collection and tested by contracted commercial laboratories for SARS-CoV-2 by reverse transcription polymerase chain reaction (RT-PCR).

Positive SARS-CoV-2 test results were provided directly to the person (or parent or guardian for children <16 years) or to the shelter via the county public health department's standard notification procedure. Clients living in shelters who had positive SARS-CoV-2 were isolated immediately in a A c c e p t e d M a n u s c r i p t 7 separate housing unit at the shelter or isolated until transported to an isolation hotel. For clients living unsheltered, results were provided via a clinic hotline number or clinical outreach teams, which located the clients with positive SARS-CoV-2 and arranged transport to the isolation hotel. Staff with positive SARS-CoV-2 isolated at home. Clients were not allowed to return to shared spaces in shelters, and isolated staff did not return to work until they met symptom-or time-based criteria in accordance with the Centers for Disease Control and Prevention (CDC) guidelines at the time [13] .

Shelter assessments were conducted in conjunction with screening and testing at a convenience sample of shelters. These shelters were selected based on availability of testing staff on the day of testing. Using a standardized assessment questionnaire (see Supplemental Materials #2), shelter management was interviewed to collect quantitative and qualitative data on shelter characteristics and services offered. Shelter characteristics included number of clients and staff, type (e.g., daytime only, 24 hours per day, transitional housing), services provided, and sleeping space configurations. Measures and policies implemented by the shelter to mitigate SARS-CoV-2 transmission, including standardized client and staff screening, isolation and quarantine protocols, and IPC, were also collected. Shelter staff were counseled on best practices to prevent transmission in the shelter.

Descriptive statistics were used to characterize the population tested and the proportions with current and recent symptoms, underlying medical conditions, and positive SARS-CoV-2. Continuous variables were compared using Student's t-test, and categorical variables were compared using the chi-square test. Shelter characteristics were described in aggregate but were not analyzed in relation to SARS-CoV-2 prevalence due to small numbers. CDC determined this project to be non-research as Demographic characteristics for tested participants are included in Table 1 A c c e p t e d M a n u s c r i p t 9

The same subset of 1,997 people was screened for symptoms; 12 with inconclusive results were subsequently excluded (Table 3) In one shelter housing adult men, testing revealed a widespread, undetected outbreak. This shelter represented 1% (29/2,050) of people tested on the initial round but 20% (8/40) of those with positive SARS-CoV-2. Client census had been decreased by approximately half to facilitate physical distancing, but the shelter did not enforce city or state shelter-in-place orders nor restrict client movement. The shelter only had congregate sleeping spaces; sleeping mats and beds were placed at least six feet apart, but head-to-toe alignment was inconsistently used. Showers upon entry were encouraged, but face coverings and hand sanitizer were not provided due to insufficient supply.

Household cleaning solutions and Environmental Protection Agency-registered disinfectants appropriate for cleaning high-touch surfaces or items were also not available [15, 16] .

70% of all PEH in Atlanta based on the 2019 point-in-time count in Atlanta [2] . Overall SARS-CoV-2

prevalence among PEH and staff tested in Atlanta from April to May 2020 was low compared with reports among PEH in other large, urban settings [10] [11] [12] . Although SARS-CoV-2 prevalence among clients living in shelters was only 2.1%, it was four times higher than the 0.5% prevalence among A c c e p t e d M a n u s c r i p t 11 PEH living unsheltered at the time. To our knowledge, this is the first report of SARS-CoV-2 prevalence in a population of PEH living unsheltered. Although the risk of SARS-CoV-2 was lower for PEH living unsheltered, unsheltered living situations pose an increased risk of morbidity and mortality compared with living sheltered, so linkages to permanent supportive housing should remain a priority [17] .

Testing at shelters in other large, urban settings in the United States has primarily occurred in response to COVID-19 clusters or outbreaks [10] [11] [12] . In early April 2020, Baggett et al. investigated an outbreak in a large homeless shelter in Boston with a SARS-CoV-2 detection rate of 36% [10] .

During an outbreak across three affiliated shelters in Seattle from March-April 2020, 18% of clients and 21% of staff had positive SARS-CoV-2 [11] . In Atlanta, universal screening and testing of PEH and staff was a proactive strategy, rather than in response to known COVID-19 cases. A recent study of long-term care facilities in Fulton County, Georgia, found 1.5% of residents had positive SARS-CoV-2 when testing was proactive, compared to 47.2% after a case was already diagnosed [18] . Only one shelter we tested was found to have a widespread, undetected outbreak. On March 23, 2020, 3 weeks after the first COVID-19 case was identified in Atlanta, the city implemented a 14-day shelterin-place order, which was followed by a Georgia state-wide shelter-in-place order during April 3-30, 2020. Over half of shelters assessed stopped taking clients during these shelter-in-place orders, and all reported a lower-than-average number of clients, which helped facilitate physical distancing. The low SARS-CoV-2 prevalence among PEH and staff in Atlanta compared with other cities may reflect the impact of shelter-in-place orders coupled with low community prevalence at the time of testing and the proactive testing strategy. [19, 20] . Although the specificity was higher (81%-89%), screening would produce a high proportion of false positive results in low prevalence settings.

Most shelters assessed were conducting standardized symptom screening for clients, but only one-third were screening staff. Despite the limitations of symptom screening, CDC recommends that homeless service providers regularly assess both clients and staff for symptoms using a standardized protocol [21] . People with COVID-19 symptoms should immediately be provided with a face covering, isolated, and tested for SARS-CoV-2. Shelters, in coordination with local public health authorities and community coalitions, should have plans to isolate people with suspected or confirmed SARS-CoV-2 infection to prevent spread [21] . However, because an estimated 40-45% of people with SARS-CoV-2 are asymptomatic, shelters should reduce the number of people served or expand to alternative housing sites, increase physical distancing (i.e., >6 feet), and mandate use of face coverings by all staff and clients inside shelters except when in bed or individual rooms [21, 22] .

Our finding of decreased prevalence in four shelters during repeat testing is consistent with reports from skilled nursing facilities and correctional facilities, supporting the use of universal (facility-wide) testing for early identification and isolation of those with positive SARS-CoV-2 as a strategy to interrupt transmission in congregate settings [23] [24] [25] . A potentially important factor in the observed decreased prevalence in these shelters was the ability to move all PEH identified with SARS-CoV-2 infections to separate housing units at the shelter or to offsite locations. In homeless shelters and encampments located in areas where SARS-CoV-2 community transmission is substantial, CDC recommends that initial (baseline) and regular testing be considered, regardless of whether an initial COVID-19 case has been identified [26] . In areas where community transmission is minimal to moderate, CDC provides examples of testing strategies that can be used to identify asymptomatic cases among both clients and staff, such as sentinel surveillance, positive symptom screening thresholds, or random testing (e.g., every third person) on a regular basis [26] . If a COVID-19 case is identified, CDC recommends repeat testing of all previously negative or untested individuals until testing identifies no new COVID-19 cases for at least 14 days [26] .

A c c e p t e d M a n u s c r i p t 13 Our findings are subject to several limitations. SARS-CoV-2 transmission dynamics are complex, and these results represent a single point in time early in the COVID-19 pandemic. We did not screen or test all PEH or shelters in Atlanta; therefore, the results might not be representative of all PEH or shelters; they are not generalizable to other large metropolitan areas. We cannot compare this prevalence to the general population of Fulton County during this time period because representative testing had not been conducted among the general population. Additionally, misclassification of sheltered and unsheltered housing status might have resulted from movement between settings and the difficulty of verifying unsheltered status. Due to shortages of nasopharyngeal swabs, three specimen collection methods were used. Specimen collection methods have different SARS-CoV-2 detection sensitivities; however, nasal and nasal midturbinate specimens were shown to have >90% sensitivity compared with nasopharyngeal samples [27] . Symptom and medical condition screenings were only conducted at a subset of testing events, which oversampled the larger shelters and may not be generalizable. Shelter clients might not have disclosed symptoms due to fears that they would be removed from shelters. Thus, the proportion of people who reported symptoms might be underestimated. Only nine shelters of 24 underwent facility assessments, including two shelters at which assessments were conducted after initial testing. These findings may reflect improved or modified IPC policies as a result of knowledge of positive cases at these two shelters.

The findings provide an early view into the effects of the COVID-19 pandemic on PEH and homelessness service staff in Atlanta. As evidence grows and guidance evolves during the COVID-19 pandemic, shelters should prioritize mitigation strategies and best practices for congregate and highrisk settings to prevent SARS-CoV-2 transmission while continuing to provide essential services for

A c c e p t e d M a n u s c r i p t 14 M a n u s c r i p t M a n u s c r i p t 19 A c c e p t e d M a n u s c r i p t 21 A c c e p t e d M a n u s c r i p t 22 

The 2019 Annual Homeless Assessment Report (AHAR) to Congress. Part 1: Point-in-Time Estimates of Homelessness

State of Homelessness: Atlanta Continuum of Care

america/homelessness-statistics/state-of-homelessness-dashboards/?State=Georgia. Accessed 29

Preventing and controlling emerging and reemerging transmissible diseases in the homeless

Morbidity and mortality in homeless individuals, prisoners, sex workers, and individuals with substance use disorders in high-income countries: a systematic review and meta-analysis

Chronic Disease Burden of the Homeless: A Descriptive Study of Student-Run Free Clinics in Tampa, Florida

Cardiovascular Disease and Homelessness

Out-of-hospital and emergency department utilization by adult homeless patients

Barriers to homeless persons acquiring health insurance through the Affordable Care Act

Estimated Community Seroprevalence of SARS-CoV-2 Antibodies -Two Georgia Counties

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

COVID-19 Outbreak Among Three Affiliated Homeless Service Sites

Assessment of SARS-CoV-2 infection prevalence in homeless shelters -Four U.S. cities

Discontinuation of Isolation for Persons with COVID -19 Not in Healthcare Settings

Outbreak of drug-resistant Mycobacterium tuberculosis among homeless people in

List N: Disinfectants for Use Against SARS-CoV-2 (COVID-19)

Unsheltered Status, and Risk Factors for Mortality

Mass Screening for SARS-CoV-2 Infection among Residents and Staff in Twenty-eight Long-term Care Facilities in Fulton County, Georgia. medRxiv

Symptom screening at illness onset of health care personnel with SARS-CoV-2 infection in King County

Estimated effectiveness of symptom and risk screening to prevent the spread of COVID-19

Interim Guidance for Homeless Service Providers to Plan and Respond to Coronavirus Disease 2019 (COVID-19)

Universal and serial laboratory testing for SARS-CoV-2 at a long-term care skilled nursing facility for veterans

Initial and repeated point prevalence surveys to inform SARS-CoV-2 infection prevention in 26 skilled nursing facilities -Detroit

Serial laboratory testing for SARS-CoV-2 infection among incarcerated and detained persons in a correctional and detention facility -Louisiana

Interim Considerations for Health Departments for SARS-CoV-2 Testing in Homeless Shelters and Encampments

Swabs Collected by Patients or Health Care Workers for SARS-CoV-2 Testing

We would like to thank Anitra Walker (Mercy Care), Catherine Christie (Mercy