Abstract
Objectives. To examine shelter characteristics and infection prevention practices in relation to severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection point prevalence during universal testing at homeless shelters in the United States.
Methods. SARS-CoV-2 testing was offered to clients and staff at homeless shelters, irrespective of symptoms. Site assessments were conducted from March 30 to June 1, 2020, to collect information on shelter characteristics and infection prevention practices. We assessed the association between SARS-CoV-2 infection prevalence and shelter characteristics, including 20 infection prevention practices by using crude risk ratios (RRs) and exact unconditional 95% confidence intervals (CIs).
Results. Site assessments and SARS-CoV-2 testing results were reported for 63 homeless shelters in 7 US urban areas. Median infection prevalence was 2.9% (range = 0%–71.4%). Shelters implementing head-to-toe sleeping and excluding symptomatic staff from working were less likely to have high infection prevalence (RR = 0.5; 95% CI = 0.3, 0.8; and RR = 0.5; 95% CI = 0.4, 0.6; respectively); shelters with medical services available were less likely to have very high infection prevalence (RR = 0.5; 95% CI = 0.2, 1.0).
Conclusions. Sleeping arrangements and staffing policies are modifiable factors that might be associated with SARS-CoV-2 infection prevalence in homeless shelters. Shelters should follow recommended practices to reduce the risk of SARS-CoV-2 transmission.
Homeless shelters provide temporary and emergency housing for people experiencing homelessness. People experiencing homelessness disproportionately experience untreated chronic medical conditions, older age, and barriers to accessing medical care,1 increasing the risk of severe disease from severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), the virus that causes COVID-19.2 Furthermore, as shelters are often congregate settings where it can be challenging to adhere to Centers for Disease Control and Prevention (CDC) guidance to reduce SARS-CoV-2 transmission, such as handwashing, wearing face masks, or remaining at least 6 feet apart, exposure risk might also be increased.3
COVID-19 cases have been detected among clients and staff at US homeless shelters.4–6 A previous study of shelters in 4 US cities identified varied SARS-CoV-2 infection prevalence using universal testing, defined as offering testing irrespective of symptoms.5 This strategy can inform isolation efforts and infection prevalence estimations because it identifies more cases, including asymptomatic and presymptomatic cases, compared with symptom-based testing alone.6
Homeless shelters vary widely in access to resources, operational plans, regulations, size, staffing, and populations served, but the extent to which these traits affect SARS-CoV-2 infection prevalence is unknown. To support shelters during the COVID-19 pandemic, it is critical to identify shelter characteristics with potential to mitigate transmission among clients and staff. In this analysis, we examined shelter characteristics and infection prevention practices in relation to SARS-CoV-2 infection point prevalence at shelters where universal testing occurred.
METHODS
As part of the response to the COVID-19 pandemic, CDC requested collaboration from health and housing partners in 7 US urban areas where universal testing had been performed at shelters during April 1 to June 1, 2020. Participating agencies provided data from site assessments and universal testing.
Universal SARS-CoV-2 testing was conducted per local protocols by real-time reverse transcription–polymerase chain reaction, proactively or in response to a positive viral test at the shelter. Although universal testing was uncommon for the general population, all clients and staff were offered testing; participation was voluntary. If multiple testing events occurred at a shelter, only results of the first event were included. Prevalence was estimated by dividing the number of clients and staff positive for SARS-CoV-2 by the total tested.
Health or housing professionals conducted site assessments to collect information about shelter characteristics and infection prevention practices using a template adapted for local needs (Appendix A, available as a supplement to the online version of this article at https://www.ajph.org). Shared data elements included information about client sleeping arrangements, staffing practices and training, and infection prevention practices such as symptom screening and face mask use. County population density was calculated as inhabitants per square mile,7 and COVID-19 incidence (cases per 100 000 per week)8 was estimated as a 7-day average encompassing 3 days before to 3 days after testing.
Infection prevalence greater than 2.9% (the median) was defined as “high” and compared with prevalence of less than or equal to 2.9%; the subset greater than 10% was defined as “very high” and compared with prevalence of less than or equal to 10%. Other continuous variables were converted to binary variables based on the median. We calculated crude risk ratios (RRs) and exact unconditional 95% confidence intervals (CIs) to assess for associations between high or very high prevalence and shelter characteristics. We conducted analyses in SAS version 9.4 (SAS Institute, Cary, NC).
RESULTS
SARS-CoV-2 testing results and site assessments conducted March 30 through June 1, 2020, were reported for 63 shelters in urban areas in 7 states (California, Colorado, Georgia, Illinois, Nevada, Rhode Island, and Washington). Site assessments were conducted a median of 13 days after the testing event (range = 1 day before to 51 days after). Among 20 infection prevention practices included in this report, 19 (95%) were implemented by greater than 50% and 9 (45%) by greater than 80% of shelters (Table 1).
TABLE 1—
Association of High Point Prevalence During Universal SARS-CoV-2 Testing at Homeless Shelters, Selected County and Shelter Characteristics: 7 US Urban Areas, March 30–June 1, 2020
| Total Shelters, No. (%)a or Median (Range) | High SARS-CoV-2 Infection Prevalence,b No. (%) or Median (Range) | RR (95% CI) | Very High SARS-CoV-2 Infection Prevalence,c No. (%) | RR (95% CI) | |
| Total | 63 | 31 (49.2) | 20 (31.7) | ||
| County-level characteristics | |||||
| COVID-19 incidence (cases per 100 000) | |||||
| High, > 15 (Ref) | 30 (47.6) | 17 (60.7) | 12 (42.9) | ||
| Low, ≤ 15 | 33 (52.4) | 14 (40.0) | 0.7 (0.5, 1.2) | 8 (22.9) | 0.6 (0.3, 1.3) |
| Median | 15.0 (2.9–32.8) | ||||
| Population density (inhabitants per square mile) | |||||
| High, > 3 923 (Ref) | 31 (49.2) | 18 (58.1) | 13 (41.9) | ||
| Low, ≤ 3 923 | 32 (50.8) | 13 (40.6) | 0.7 (0.4, 1.2) | 7 (21.9) | 0.5 (0.2, 1.1) |
| Median | 3 923 (247–17 179) | ||||
| Client sleeping arrangements | |||||
| Percentage of beds filled | |||||
| ≥ 74 (Ref) | 31 (49.2) | 14 (45.2) | 8 (25.8) | ||
| < 74 | 29 (29) | 16 (55.2) | 1.2 (0.7, 2.0) | 12 (41.4) | 1.6 (0.8, 3.4) |
| Median | 76.9 (15.4–125) | ||||
| Beds or mats assigned to 1 person (instead of shared or communal use) | |||||
| No (Ref) | 7 (11.1) | 4 (57.1) | 3 (42.9) | ||
| Yes | 53 (84.1) | 26 (49.1) | 0.9 (0.4, 1.7) | 17 (32.1) | 0.7 (0.3, 1.9) |
| Distance between beds in sleeping area ≥ 3 feet | |||||
| No (Ref) | 10 (15.9) | 6 (60.0) | 4 (40.0) | ||
| Yes | 51 (81.0) | 24 (47.1) | 0.8 (0.4, 1.4) | 15 (29.4) | 0.7 (0.3, 1.8) |
| Separation screen or barrier between beds | |||||
| No (Ref) | 50 (79.4) | 25 (50.0) | 15 (30.0) | ||
| Yes | 10 (15.9) | 5 (50.0) | 1.0 (0.5, 2.0) | 4 (40.0) | 1.3 (0.6, 3.2) |
| Clients sleeping head-to-toe | |||||
| No (Ref) | 19 (30.2) | 14 (73.7) | 10 (52.6) | ||
| Yes | 37 (58.7) | 13 (35.1) | 0.5 (0.3, 0.8) | 8 (21.6) | 0.4 (0.2, 0.9) |
| Staff characteristics and training | |||||
| Staff work at other shelter locations | |||||
| No (Ref) | 35 (55.6) | 18 (51.4) | 11 (31.4) | ||
| Yes | 23 (36.5) | 12 (52.2) | 1.0 (0.6, 1.7) | 9 (39.1) | 1.2 (0.6, 2.5) |
| Medical services routinely available at the facility | |||||
| No (Ref) | 29 (46.0) | 17 (58.6) | 13 (44.8) | ||
| Yes | 34 (54.0) | 14 (41.2) | 0.7 (0.4, 1.2) | 7 (20.6) | 0.5 (0.2, 1.0) |
| All staff trained on hygiene measures and standard precautions | |||||
| No (Ref) | 4 (6.3) | 1 (25.0) | 1 (25.0) | ||
| Yes | 58 (92.1) | 29 (50.0) | 2.0 (0.4, 11.1) | 18 (31.0) | 1.2 (0.2, 7.1) |
| All staff received training on how to correctly put on and take off personal protective equipment | |||||
| No (Ref) | 14 (22.2) | 6 (42.9) | 5 (35.7) | ||
| Yes | 42 (66.7) | 21 (50.0) | 1.2 (0.6, 2.3) | 14 (33.3) | 0.9 (0.4, 2.1) |
| Staff routinely wear masks when interacting with clients | |||||
| No (Ref) | 5 (7.9) | 2 (40.0) | 2 (40.0) | ||
| Yes | 54 (85.7) | 27 (50.0) | 1.3 (0.4, 3.8) | 17 (31.5) | 0.8 (0.3, 2.5) |
| All staff screened for symptoms | |||||
| No (Ref) | 23 (36.5) | 11 (47.8 | 8 (34.8) | ||
| Yes | 39 (61.9) | 19 (48.7) | 1.0 (0.6, 1.7) | 11 (28.2) | 0.8 (0.4, 1.7) |
| All staff have their temperature checked | |||||
| No (Ref) | 25 (39.7) | 12 (48.0) | 9 (36.0) | ||
| Yes | 34 (54.0) | 16 (47.1) | 1.0 (0.6, 1.7) | 10 (29.4) | 0.8 (0.4, 1.7) |
| Staff excluded from working when symptomatic | |||||
| No (Ref) | 2 (3.2) | 2 (100) | 1 (50.0) | ||
| Yes | 61 (96.8) | 29 (47.5) | 0.5 (0.4, 0.6) | 19 (31.1) | 0.6 (0.1, 2.6) |
| Other infection prevention practices | |||||
| Limited number of designated entry points | |||||
| No (Ref) | 2 (3.2) | 1 (50.0) | 1 (50.0) | ||
| Yes | 61 (96.8) | 30 (49.2) | 1.0 (0.2, 4.0) | 19 (31.1) | 0.6 (0.1, 2.6) |
| Handwashing available at facility entry | |||||
| No (Ref) | 22 (34.9) | 11 (50.0) | 6 (27.3) | ||
| Yes | 39 (61.9) | 18 (46.2) | 0.9 (0.5, 1.6) | 12 (30.8) | 1.1 (0.5, 2.6) |
| Points of entry monitored to ensure hand hygiene | |||||
| No (Ref) | 14 (22.2) | 6 (42.9) | 4 (28.6) | ||
| Yes | 44 (69.8) | 23 (52.3) | 1.2 (0.6, 2.4) | 15 (34.1) | 1.2 (0.5, 3.0) |
| Clients routinely wear masks or cloth face coverings when not in their sleeping areas | |||||
| No (Ref) | 11 (17.5) | 5 (45.5) | 3 (27.3) | ||
| Yes | 46 (73.0) | 24 (52.2) | 1.1 (0.6, 2.3) | 16 (34.8) | 1.3 (0.4, 3.6) |
| All clients screened for symptoms | |||||
| No (Ref) | 7 (11.1) | 4 (57.1) | 2 (28.6) | ||
| Yes | 54 (85.7) | 26 (48.1) | 0.8 (0.4, 1.7) | 17 (31.5) | 1.1 (0.3, 3.8) |
| All clients have their temperature checked | |||||
| No (Ref) | 19 (30.2) | 9 (47.4) | 6 (31.6) | ||
| Yes | 44 (69.8) | 22 (50.0) | 1.1 (0.6, 1.8) | 14 (31.8) | 1.0 (0.5, 2.2) |
| Designated area for suspected cases or symptomatic clients to isolate | |||||
| No (Ref) | 20 (31.7) | 12 (60.0) | 9 (45.0) | ||
| Yes | 43 (68.3) | 19 (44.2) | 0.7 (0.5, 1.2) | 11 (25.6) | 0.6 (0.3, 1.1) |
| Symptomatic clients provided with a surgical mask | |||||
| No (Ref) | 4 (6.3) | 2 (50.0) | 2 (50.0) | ||
| Yes | 55 (87.3) | 28 (50.9) | 1.0 (0.4, 2.8) | 17 (30.9) | 0.6 (0.2, 1.8) |
Note. CI = confidence interval; RR = risk ratio; SARS-CoV-2 = severe acute respiratory syndrome coronavirus 2. For additional information about recommended prevention practices in homeless shelters, homeless service providers are encouraged to refer to the Centers for Disease Control and Prevention’s Interim Guidance for Homeless Service Providers to Plan and Respond to Coronavirus Disease 2019 (COVID-19)9 and resources to support people experiencing homelessness: https://www.cdc.gov/coronavirus/2019-ncov/community/homeless-shelters.
Numbers may not equal 100% because of missing or unknown responses.
High SARS-CoV-2 infection prevalence is defined as point prevalence above the median (> 2.9%).
Very high SARS-CoV-2 infection prevalence is defined as point prevalence > 10%.
The average number of beds filled ranged from 8 to 320 (median = 74). The number of clients and staff members tested ranged from 7 to 364 (median = 63), and the number that tested positive for SARS-CoV-2 ranged from 0 to 165 (median = 2). Median prevalence by facility was 2.9% (range = 0%–71.4%). Twenty (31.7%) shelters had very high (> 10%) prevalence.
Shelters implementing head-to-toe sleeping (positioning beds in a row so that clients sleep with heads at alternating ends of the bed compared with their neighbors [i.e., each client’s head lines up with their neighbors’ toes, to maximize distance between clients’ heads and minimize sharing air space while sleeping]) were less likely to experience high infection prevalence (35.1%; 13 of 37) compared with shelters that did not implement head-to-toe sleeping (73.7%; 14 of 19; RR = 0.5; 95% CI = 0.3, 0.8). Shelters that excluded symptomatic staff were also less likely to experience high prevalence (47.5%; 29 of 61) compared with shelters allowing symptomatic staff to continue working (100%; 2 of 2; RR = 0.5; 95% CI = 0.4, 0.6). Shelters with on-site medical services were less likely to have very high prevalence (20.6%; 7 of 34) compared with shelters lacking those services (44.8%; 13 of 29; RR = 0.5; 95% CI = 0.2, 1.0). In this analysis, additional characteristics (e.g., staff training, client and staff symptom screening and face mask use, and county COVID-19 incidence) were not associated with prevalence (Table 1).
DISCUSSION
These data show that universal testing at homeless shelters can identify high prevalence of SARS-CoV-2 infection, even in areas with low incidence of COVID-19, and shelters have adopted a range of infection prevention practices. Adopting head-to-toe sleeping arrangements and excluding symptomatic shelter staff from working were associated with lower risk of high SARS-CoV-2 infection prevalence. Positioning beds so clients sleep head-to-toe is recommended to reduce transmission of respiratory pathogens, including SARS-CoV-2.9 Consistent with CDC recommendations for all persons, shelters should require staff who experience symptoms consistent with COVID-19 to isolate at home for at least 10 days since symptoms first appeared and until symptoms have improved, including at least 24 hours with no fever without taking fever-reducing medication.10 Availability of on-site medical services was associated with lower risk of having very high prevalence of SARS-CoV-2 in clients and staff. Routine access to on-site medical services might enable shelters to quickly identify and isolate symptomatic clients or to identify and maintain appropriate infection prevention practices.
Although symptom and temperature screening were not associated with prevalence in this analysis, they remain fundamental measures to identify individuals with suspected COVID-19. Studies have shown that, given the potential for asymptomatic and presymptomatic transmission of SARS-CoV-2,11 additional measures, including the use of face masks, can reduce the spread of SARS-CoV-2.12 A combination of multiple infection prevention strategies is necessary to reduce transmission of SARS-CoV-2.
These findings are subject to several limitations. The high proportion of adopted infection control practices is a positive finding, but low variability in shelter practices and small sample size limit the ability to assess for confounding or detect statistically significant associations between prevention measures and prevalence. Timing of infection control measures and site assessments varied relative to testing events; associations between shelter practices and prevalence might not be correlated. Data were collected early in the pandemic as guidance regarding testing evolved, and prevalence was calculated for clients and staff together because some sites were unable to separate those results. Site assessments varied between locations and included self-reported data. Finally, results are not generalizable and are subject to biases, including selection (i.e., which areas and shelters offered and which individuals accepted SARS-CoV-2 testing), recall, and social desirability (i.e., tendency to respond favorably to site assessment questions).
PUBLIC HEALTH IMPLICATIONS
Congregate settings, such as homeless shelters, can pose risks for spread of communicable diseases such as COVID-19. However, continuing services for people experiencing homelessness is critical. Sleeping arrangements and staffing policies, such as excluding symptomatic staff and having medical services on site, are modifiable factors that might be associated with transmission of SARS-CoV-2 in shelters. Shelters should follow recommended practices9 to reduce the risk of SARS-CoV-2 transmission, and further studies should evaluate environmental risks associated with transmission in congregate settings.
ACKNOWLEDGMENTS
The authors would like to thank the shelter staff who contributed their time and knowledge to the site assessments. Many thanks to the participating state health departments (e.g., Rhode Island Department of Health, Chicago Department of Public Health), public and private laboratories (e.g., Rhode Island Health State Laboratory, the Lifespan Clinical Microbiology Lab), local homelessness coalitions (e.g., Rhode Island Coalition for the Homeless), and medical centers (e.g., Department of Emergency Medicine and Division of Internal Medicine at the Warren Alpert School of Medicine of Brown University, University of Illinois at Chicago, and Rush University Medical Center). We are grateful to the following individuals for their contributions to this project: Karin Huster, Margaret Lukoff, Libby Page, Ashley Scarborough, Deborah Garneau, Amy Ferguson, Jennifer Barrera, Hannah Guarino, Marilyn McAllister, Sari Greene, Utpala Bandy, Ewa King, Philip Chan, James McDonald, and Louis Marchetti.
The COVID-19 Homelessness Response Team consisted of Cathryn Marchman, JD, MSW, Rebecca Karb, MD, PhD, Rahul Vanjani, MD, MSc, M. Catherine Trimbur, MD, MPH, Anthony Napoli, MD, MHL, Rachel R. Yorlets, MPH, Elysia Gonzales, RN, MPH, Ruthanne Marcus, PhD, MPH, Calla Jamison, MPH, Ashley Meehan, MPH, Susan Cha, PhD, MPH, Kristine Schmit, MD, MPH, Sapna Bamrah Morris, MD, Rosemary G. Remus, Kirstin McEachran, MPH, Judson Bemis, MPH, Devin Clark, Elizabeth S. Davis, MD, and Stockton Mayer, DO.
CONFLICTS OF INTEREST
The authors report no conflicts of interest.
HUMAN PARTICIPANT PROTECTION
This activity was determined by Centers for Disease Control and Prevention to be public health surveillance in accordance with 45 CFR Part 46.
Footnotes
See also the COVID-19/Public Health Preparedness and Response section, pp. 842–875.
REFERENCES
- 1.Lebrun-Harris LA, Baggett TP, Jenkins DM et al. Health status and health care experiences among homeless patients in federally supported health centers: findings from the 2009 patient survey. Health Serv Res. 2013;48(3):992–1017. doi: 10.1111/1475-6773.12009. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Hsu HE, Ashe EM, Silverstein M et al. Race/ethnicity, underlying medical conditions, homelessness, and hospitalization status of adult patients with COVID-19 at an urban safety-net medical center—Boston, Massachusetts, 2020. MMWR Morb Mortal Wkly Rep. 2020;69(27):864–869. doi: 10.15585/mmwr.mm6927a3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Centers for Disease Control and Prevention. Coronavirus disease 2019 (COVID-19): how to protect yourself & others. Available at: https://www.cdc.gov/coronavirus/2019-ncov/prevent-getting-sick/prevention.html. Accessed July 14, 2020.
- 4.Tobolowsky FA, Gonzales E, Self JL et al. COVID-19 outbreak among three affiliated homeless service sites—King County, Washington, 2020. MMWR Morb Mortal Wkly Rep. 2020;69(17):523–526. doi: 10.15585/mmwr.mm6917e2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Mosites E, Parker EM, Clarke KEN et al. Assessment of SARS-CoV-2 infection prevalence in homeless shelters—four US cities, March 27–April 15, 2020. MMWR Morb Mortal Wkly Rep. 2020;69(17):521–522. doi: 10.15585/mmwr.mm6917e1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Baggett TP, Keyes H, Sporn N, Gaeta JM. Prevalence of SARS-CoV-2 infection in residents of a large homeless shelter in Boston. JAMA. 2020;323(21):2191–2192. doi: 10.1001/jama.2020.6887. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.US Census Bureau. QuickFacts. Available at: https://www.census.gov/quickfacts/fact/table/US/PST045219. Accessed June 4, 2020.
- 8.USAFacts. Coronavirus locations: COVID-19 map by county and state. Available at: https://usafacts.org/visualizations/coronavirus-covid-19-spread-map. Accessed June 4, 2020.
- 9.Centers for Disease Control and Prevention. Interim guidance for homeless service providers to plan and respond to coronavirus disease 2019 (COVID-19). Available at: https://www.cdc.gov/coronavirus/2019-ncov/community/homeless-shelters/plan-prepare-respond.html#coalition. Accessed June 23, 2020.
- 10.Centers for Disease Control and Prevention. Coronavirus disease 2019 (COVID-19): isolate if you are sick. Available at: https://www.cdc.gov/coronavirus/2019-ncov/if-you-are-sick/isolation.html. Accessed July 22, 2020.
- 11.Furukawa NW, Brooks JT, Sobel J. Evidence supporting transmission of severe acute respiratory syndrome coronavirus 2 while presymptomatic or asymptomatic. Emerg Infect Dis. 2020;26(7):e201595. doi: 10.3201/eid2607.201595. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Leung NHL, Chu DKW, Shiu EYC et al. Respiratory virus shedding in exhaled breath and efficacy of face masks [erratum in Nat Med. 2020;26(6):981] Nat Med. 2020;26(5):676–680. doi: 10.1038/s41591-020-0843-2. [DOI] [PMC free article] [PubMed] [Google Scholar]