Background: Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) is believed to spread mainly through respiratory droplets between persons who are in close quarters. Such droplets are produced when an infected person coughs, sneezes, or talks. As a result, infected and exposed persons are instructed to separate themselves from others to limit further spread. The World Health Organization and the Centers for Disease Control and Prevention advise those who are infected with or have been exposed to coronavirus disease 2019 (COVID-19) to isolate or quarantine at home in a separate bedroom and bathroom if possible (1, 2).

Objective: To determine the feasibility of separate rooms for isolation and quarantine for housing units in the United States.

Methods and Findings: We obtained data on residential dwelling units (excluding group quarters) and occupants from the most recently available (2017) American Housing Survey. This survey includes a representative national sample of housing units and was done by the U.S. Census Bureau, with a response rate of 80.5% (3). We determined the number of bedrooms, bathrooms, and occupants per unit; the type and age of units; and respondent demographics.

We estimated the proportion of dwellings where optimal isolation or quarantine was impossible because separate bedrooms and bathrooms were unavailable. We considered a separate bedroom to be available in all single-occupant dwellings and in multioccupant dwellings where allowing 1 occupant a private bedroom would not impose overcrowding (defined as >2 occupants per bedroom) on other household members. A separate bathroom was considered available in single-occupant dwellings with at least 0.5 bathrooms and in multioccupant dwellings with at least 1.5 bathrooms.

We generated descriptive statistics (means, proportions, and SEs) for all variables. Logistic regression was used to determine the bivariate relationship between a separate bedroom or bathroom being unavailable and characteristics of housing units and occupants. We used survey procedures in SAS, version 9.4 (SAS Institute), with weights and replicate weights provided by the American Housing Survey to generate national estimates, SEs, odds ratios, and 95% CIs using the balanced repeated replication method (4).

The 57 984 occupied dwellings in the sample represented 121.57 million dwellings nationwide that housed about 303 million residents; the dwellings averaged 2.80 bedrooms, 1.82 bathrooms, and 2.49 occupants (Table). Single-family detached homes, newer units, and those occupied by higher-income households had more rooms.