Estimation of Transmission of COVID-19 in Simulated Nursing Homes With Frequent Testing and Immunity-Based Staffing

Abstract

Importance: Nursing homes and other long-term care facilities have been disproportionately impacted by the COVID-19 pandemic. Strategies are urgently needed to reduce transmission in these high-risk populations.

Objective: To evaluate COVID-19 transmission in nursing homes associated with contact-targeted interventions and testing.

Design, setting, and participants: This decision analytical modeling study developed an agent-based susceptible-exposed-infectious (asymptomatic/symptomatic)-recovered model between July and September 2020 to examine SARS-CoV-2 transmission in nursing homes. Residents and staff of a simulated nursing home with 100 residents and 100 staff split among 3 shifts were modeled individually; residents were split into 2 cohorts based on COVID-19 diagnosis. Data were analyzed from September to October 2020.

Exposures: In the resident cohorting intervention, residents who had recovered from COVID-19 were moved back from the COVID-19 (ie, infected with SARS-CoV-2) cohort to the non-COVID-19 (ie, susceptible and uninfected with SARS-CoV-2) cohort. In the immunity-based staffing intervention, staff who had recovered from COVID-19 were assumed to have protective immunity and were assigned to work in the non-COVID-19 cohort, while susceptible staff worked in the COVID-19 cohort and were assumed to have high levels of protection from personal protective equipment. These interventions aimed to reduce the fraction of people's contacts that were presumed susceptible (and therefore potentially infected) and replaced them with recovered (immune) contacts. A secondary aim of was to evaluate cumulative incidence of SARS-CoV-2 infections associated with 2 types of screening tests (ie, rapid antigen testing and polymerase chain reaction [PCR] testing) conducted with varying frequency.

Main outcomes and measures: Estimated cumulative incidence proportion of SARS-CoV-2 infection after 3 months.

Results: Among the simulated cohort of 100 residents and 100 staff members, frequency and type of testing were associated with smaller outbreaks than the cohorting and staffing interventions. The testing strategy associated with the greatest estimated reduction in infections was daily antigen testing, which reduced the mean cumulative incidence proportion by 49% in absence of contact-targeted interventions. Under all screening testing strategies, the resident cohorting intervention and the immunity-based staffing intervention were associated with reducing the final estimated size of the outbreak among residents, with the immunity-based staffing intervention reducing it more (eg, by 19% in the absence of testing) than the resident cohorting intervention (eg, by 8% in the absence of testing). The estimated reduction in transmission associated with these interventions among staff varied by testing strategy and community prevalence.

Conclusions and relevance: These findings suggest that increasing the frequency of screening testing of all residents and staff, or even staff alone, in nursing homes may reduce outbreaks in this high-risk setting. Immunity-based staffing may further reduce spread at little or no additional cost and becomes particularly important when daily testing is not feasible.

Figure 1.. Susceptible–Exposed–Infectious (Asymptomatic/Symptomatic)–Recovered (SEIR) Model

A, Flowchart of movement through the SEIR compartmental model. This model uses a cohorted framework wherein nursing home residents who become infected either symptomatically (I) or asymptomatically (A) with SARS-CoV-2 are moved out of the non–COVID-19 cohort and separated into a distinct COVID-19 cohort after showing symptoms or testing positive. R indicates individuals who have recovered from (and are assumed immune to) COVID-19; res., residents; and int., intervention. Black lines denote cohorting interventions and gray lines show transitions between infection states. B, There are 2 residents per room, and each interacts with 6 staff per day. Staff interact with 2 other staff each day and also have a daily risk of infection from the community. The dashed line indicates sensitivity analyses with additional limited contacts between residents (eFigure 7 in the Supplement). C, LOD indicates limit of detection; PCR, polymerase chain reaction.

Figure 2.. Cumulative Incidence Proportion at 3 Months From First SARS-CoV-2 Introduction Under Different Cohorting, Staffing, and Testing Interventions

Violin plots show simulation results from 100 stochastic simulations. Cases among staff are split out to distinguish between cases arising in the community and those that are a result of transmission within the nursing home. PCR indicates polymerase chain reaction.

Figure 3.. Probability and Size of an Outbreak Under Varying Community Prevalence Settings

0 indicates a single initial case but no additional community introductions (ie, no infection from the community). Error bars indicate interquartile ranges.