ABSTRACT
Background: The COVID19 pandemic has necessitated innovative ways to provide safe healthcare remotely for large numbers of infected patients. We implemented a COVID Virtual Clinic (CVC) in a tertiary referral centre allowing such patients to be monitored in the community. This study describes the CVC's remote monitoring experience and explores the predictors of need for specialist intervention. Methods: We included all patients enrolled in the CVC at the Mater Misericordiae University Hospital, Dublin between March 1st and June 1st 2020. Patients received a Bluetooth-enabled pulse oximeter and smartphone application (Patient-M-Power®) and uploaded twice-daily SpO2 readings, heart rate and dyspnoea score (1-10). A team of 2-14 healthcare providers monitored results. Abnormal or absent data triggered calls from the CVC, with assessments and/or admission as required. We collected data on demographics, calls received from/made to patients, outcomes and readmissions. Descriptive analysis of the CVC was performed as well as simple logistic regression to explore factors associated with the likelihood of readmission. Results: 502 patients were included (179 (36.4%) male, median age 39 (IQR 50-3) years, 360 (73.2%) staff). Outcomes are illustrated in Figure 1. Median time in CVC was 12 days (IQR 13-10). 1902 calls were made to patients by CVC staff prompted by abnormal data: dyspnoea (41 patients, 8.2%), low SpO2 (133, 26.5%), tachycardia, (99, 19.7%), technical issues (81, 16.1%), absent results (255, 50.1%). This resulted in 45 (9%) patients requiring re-assessment and 42 (8.4%) being readmitted. Of those readmitted, 3 (7%) required critical care admission. Median length of stay was 2 (IQR 6.75-1) days. Those readmitted were more likely to be older (odds ratio [OR] per year older 1.03 (1.01, 1.05), P=0.0050, have an abnormal SpO2 (<94%, OR 5.43 [2.93, 11.1], P<0.001), a high dyspnoea score (>7, OR 4.33 (2.04, 9.3), P<0.001) and be staff (OR 6.08 (3.11, 11.87), P<0.001). Neither gender nor abnormal HR were associated with higher likelihood of readmission. 22.2% of presenting patients were hypoxic in the absence of dyspnoea, of which 70% required admission and one patient required intensive care. Conclusion: We describe the largest remotely monitored cohort of COVID19 patients to date. The low frequency of readmissions and value of SpO2 monitoring and dyspnoea scores as predictors of readmission highlights the value of this model in providing safe care whilst minimising unnecessary admissions.