Key baseline scenario parameters | Value | Source |
---|---|---|
Number of residents and staff at the start of each simulation | Residential care home: 29, nursing care home: 47 | Average care home sizes in England [4] |
Visitors | Number of visitors per resident per day ~ \(Beta(\mu =0.24, se=0.1)\) | Personal communication from expert opinion trustee of the Residents and Relatives Association & CEO of Social Care Forum (see Additional file 1: Table S1) |
Admission of residents | Admissions only from hospital, no new admissions from the community | Assumed |
Previous exposure to SARS-CoV-2 of care home residents and staff at the start of the simulation | No previous exposure in residents, previous exposure in staff according to community prevalence assumed (baseline scenario: medium community prevalence: 13% recovered) | Mathematical model of SARS-CoV-2 transmission in the community[14] (see Additional file 1: Table S1) |
Testing probabilities and frequencies | p(symptomatic residents tested) ~ \(Beta(\mu =0.9, se=0.0{5)}\); p(residents without symptoms tested every 28 days) ~ \(Beta(\mu =0.85, se=0.0{5)}\); p(staff without symptoms tested every 7 days) ~ \(Beta(\mu =0.95, se=0.0{5)}\) All symptomatic hospitalised residents were tested in hospital upon discharge | Testing probabilities assumed. Frequency of testing from previous national policy [9, 15] |
Test characteristics | PCR p(false negative test in a care home) ~ \(Beta(\mu =0.2, se=0.0{5)}\); PCR p(false negative test in a hospital) ~ \(Beta(\mu =0.1, se=0.0{1)}\); PCR test 100% specificity | Approximately 90% PCR tests are positive in individuals with early stage COVID-19 clinical infection [10, 11]. Sensitivity in care homes assumed |
Infection prevention and control | p(residents isolated following symptom onset or positive test) ~ \(Beta(\mu =0.8, se=0.0{8)}\); relative infectiousness of isolated residents vs not ~ \(Beta(\mu =0.25, se=0.1{5)}\) | Assumed |
Delay to isolation (in residents) or absence (in staff) | delay to isolation/absence in residents/staff with symptoms ~  \(Gamma(\mu =1, k=4)\) (days); delay to isolation/absence in residents/staff without symptoms ~ \(Gamma(\mu =2, k=4)\)(days) | Assumed |
Transmission rates | The transmission rates between staff and residents, residents and residents and staff and staff are the same \({{R0}_{a}}\)(\(R0\) pathway a) ~ \(Gamma(\mu =2, k=8)\); \({{R0}_{b}}\)(\(R0\) pathway b) = \({{R0}_{a}}\times {mR0}_{b}\); \({mR0}_{b}\)~ \(Beta(\mu =0.5, se=0.1)\) All transmission rates are halved at timepoints when a case is detected (i.e. there are one or more care home residents isolated or staff absent) | Assumed. In line with the R0 estimated for SARS-CoV-2 in healthcare facilities with 5–10 average contacts per day [16] |
Proportion of hospital discharges to the care home by infectious state (if admitted as non-COVID-19 hospitalisation) | Baseline (medium prevalence) scenario: % infected ~ \(Beta(\mu =0.01, se=0.0{01)}\), of which 69% exposed, 13% infectious preclinical, 18% asymptomatic; % recovered ~ \(Beta(\mu =0.05, se=0.0{3)}\); % susceptible = 1-(% recovered + % infected) All residents are tested upon discharge, and, if positive, are isolated upon their return | Mathematical models of SARS-CoV-2 transmission in the community [14] and in hospital [17] (see Additional file 1: Table S1) |
Care home staff working at another site | p(original staff work at an additional care home) ~ \(Beta(\mu =0.01, se=0.00{5)}\); p(replacement staff work at an additional care home) ~ \(Beta(\mu =0.2, se=0.07{)}\) | Personal communication from lead of Thames Valley care home survey (see Additional file 1: Table S1) |