SARS-CoV-2 in Nursing Homes after 3 Months of Serial, Facilitywide Point Prevalence Testing, Connecticut, USA

Abstract

Nursing homes house populations that are highly vulnerable to coronavirus disease. Point prevalence surveys (PPSs) provide information on the severe acute respiratory syndrome coronavirus 2 infection status of staff and residents in nursing homes and enable isolation of infectious persons to halt disease spread. We collected 16 weeks of public health surveillance data on a subset of nursing homes (34/212) in Connecticut, USA. We fit a Poisson regression model to evaluate the association between incidence and time since serial PPS onset, adjusting for decreasing community incidence and other factors. Nursing homes conducted a combined total of 205 PPSs in staff and 232 PPSs in residents. PPS was associated with 41%-80% reduction in incidence rate in nursing homes. Our findings provide support for the use of repeated PPSs in nursing home staff and residents, combined with strong infection prevention measures such as cohorting, in contributing to outbreak control.

Keywords: COVID-19; Connecticut; RT-PCR; SARS-CoV-2; United States; coronavirus disease; diagnostics; epidemiology; long-term care facility; nursing home; outbreak; prevalence; respiratory infections; severe acute respiratory syndrome coronavirus 2; universal testing; viruses; zoonoses.

Figure 5

Equation for each respective time period X and nursing home .

Figure 6

Equation for the number of new cases in nursing home on calendar day offset by person-days at risk on day t was modeled as a Poisson regression.

Figure 1

Coronavirus disease cases detected in consecutive PPSs in residents (A) and staff (B) in nursing homes, Connecticut, USA. The number of participating nursing homes for each survey is listed above each bar. One facility was excluded from staff testing data due to lack of verifiable testing results during PPS surveys. The results of the first PPS in residents, in which 601 cases were detected, were previously reported in (6). The probability of detecting a positive case decreased significantly (p<0.05) through PPS7 for residents and PPS8 for staff, compared with the first PPS, using logistic regression for comparisons. PPS, point prevalence survey.

Figure 2

Cumulative proportion of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infections in individual nursing homes over a 16-week study period relative to the first PPS, Connecticut, USA. Each colored line represents a single nursing home in the ≈4 weeks before first PPS and 12 weeks following first PPS. Data were centered for all nursing homes by the date of receipt of results for the first PPS, signified by the dashed vertical line on day 0. Red line indicates average proportion infected of the total study population on each day. The number of residents infected in each nursing home is based on cumulative case counts out of the number ever susceptible to SARS-CoV-2 in the nursing home, or the maximum census value in the study period, to account for resident deaths and transfers since the start of reporting. PPS, point prevalence survey.

Figure 3

Paired coronavirus disease incidence rate estimates relative to first PPS, Connecticut, USA. Dashed lines represent single nursing homes included in the study. Points represent the incidence in the 4 weeks before the first PPS and 12 weeks following the first PPS, during which additional PPSs were also conducted. Blue indicates significant decreases in incidence for each nursing home over the 2 time periods (α = 0.05); green indicates significant increases; red indicates nonsignificant changes in incidence. PPS, point prevalence survey.

Figure 4

Coronavirus disease incidence rates in nursing homes (cases/1,000 person-days, red) and in towns and cities (cases/100,000 person-days, blue), Connecticut, USA. Incidence rates are aggregated for the 34 nursing homes in this study and 26 towns and cities in which the nursing homes are located; incidence is presented as rolling weekly averages to account for differences in day-of-week reporting. The shaded rectangle shows the time period in which all 34 nursing homes conducted initial PPSs.