COVID-19 in Three Linked Seattle Homeless Shelters, Spring 2020

Summary

• COVID-19 can spread fast in congregate housing such as homeless shelters, but outbreaks there had not been well described.
• On April 1, 2020, an outbreak was found at three affiliated homeless shelters in King County, Washington. Testing offered to every resident and staff member turned up cases no one had recognized; with further surveillance and repeat testing, 43 people were confirmed infected.
• Stopping spread in shelters is hard. Where there is an outbreak, it takes help enforcing shelter-in-place orders, testing residents and staff, and promptly isolating residents who are sick or test positive.

How it started

The first U.S. COVID-19 case was confirmed in Snohomish County, Washington, on January 20, 2020. The governor issued stay-at-home orders on March 23, and by March 28 neighboring King County had 2,307 confirmed cases.

On March 30, Public Health – Seattle & King County learned that a resident of shelter A — a 67-year-old man with underlying conditions, hospitalized the day before with acute encephalopathy after 2 days of cough, shortness of breath, fever, sore throat and runny nose — had tested positive. He stayed stable without intensive care and after 5 days went to isolation housing, an individual room with clinical support provided by the county.

The shelters

All three served people aged 50 and older and were about 2–5 miles (3–8 km) apart. Residents of shelters B and C used shelter A's day center.

To ease crowding, about half of shelter B's residents had moved to shelter C on March 13; B and C became 24-hour shelters on March 13 and 26. All had indoor bathrooms with sinks and soap.

What the site visits found: staff rotated among the three shelters; residents could come and go before curfew; mats were 3 feet or less apart; shelter C had no hand sanitizer and no showers, so residents took shuttles or public transit to public showers; staff wore cloth face coverings or masks only intermittently, and residents had none.

Testing everyone

Round 1 (March 30–April 1): testing was offered to all available residents and staff at the three shelters, with no symptom screening first; 62.8% of residents who had slept there the night before were tested.

Twelve positive residents went to isolation housing and three to the hospital; positive staff isolated at home.

A CDC team arrived on April 1 to support the county's rapid response teams, which screened 122 residents and staff for COVID-like illness over 3 days, assessed each shelter, and made recommendations.

Changes: on April 5, facing staff shortages, the county recommended closing shelter A — its women went to individual isolation rooms and its men to shelter C, which got thermometers for temperature checks and portable showers. At all shelters: fewer staff rotations, physical distancing, less movement in and out, staff training in cleaning and disinfection, mats moved so residents' heads were at least 6 feet (2 m) apart, and disposable face masks for everyone.

Round 2 (April 7–8), for everyone who had tested negative or missed round 1: 16 of 103 residents (15.5%) and 2 of 15 staff (13.3%) were positive. Fourteen symptom screenings from April 1 to 11 found two more cases among residents, and two residents and two staff tested positive after seeking care on their own.

The toll

By April 11, COVID-19 had been confirmed in 35 of 195 residents (18%) and 8 of 38 staff (21%). Of the 43 cases, 37 (86%) were found by testing everyone, two (5%) by symptom screening, and four (9%) when people sought care themselves. Infected residents had a median age of 61 (range 50–73); staff, 39 (28–57). 96% of residents tested, and 89% of infected residents, were men. Seven residents (20%) were hospitalized; none had died. No staff were hospitalized or died.

What it shows

Likely drivers of spread: residents moving among several shelters; crowded, shared sleeping; the difficulty of physical distancing; possible spread without symptoms; and no face coverings for residents before public health stepped in.

The response — testing everyone, moving the sick and infected to isolation quickly, adding on-site showers to cut trips out, distancing and infection control — took resources and collaboration among the health department, the county's Department of Community and Human Services, shelter staff, community health care providers and federal partners.

Testing regardless of symptoms found most cases here, and because people can spread the virus before or without symptoms, it likely caught more infectious cases than symptom screening would have. Shelters, like long-term care facilities, can amplify outbreaks, and the usual controls — testing, contact tracing, distancing, limiting movement — are hard to apply for people without homes, for whom stay-at-home orders are impractical. CDC recommends that homeless service providers apply infection control, keep residents' heads at least 6 feet apart while sleeping, and promote cloth face coverings for all residents; help enforcing shelter-in-place may be needed; and in outbreaks, infected residents and those who are older or have underlying conditions should be moved into individual housing first.

Limits: not all residents were present, so some infections may have been missed; the response was resource-intensive and may not be sustainable; screening and testing were run separately, so symptoms and results could not easily be linked; and how well the measures worked could not yet be judged.

Sources

Based on Farrell A. Tobolowsky, Elysia Gonzales, Julie L. Self and others, "COVID-19 Outbreak Among Three Affiliated Homeless Service Sites — King County, Washington, 2020," MMWR, Centers for Disease Control and Prevention (Early Release, April 22, 2020); a work of the United States government in the public domain.