Predicting Need for Escalation of Care or Death From Repeated Daily Clinical Observations and Laboratory Results in Patients With Severe Acute Respiratory Syndrome Coronavirus 2.

Abstract

We compared the performance of prognostic tools for SARS-CoV-2 using parameters fitted either at time of admission or across all time points of an admission. This cohort study used clinical data to model the dynamic change in prognosis of SARS-CoV-2 in a single hospital in England including all patients admitted from 1st February 2020 until 31st December 2020, and then followed up for ICU admission, death, or discharge from hospital for 60 days. We incorporated clinical observations and blood tests into two-time varying Cox proportional hazards models predicting daily 24–48-hour risk of admission to ICU for those eligible, or death for those ineligible for escalation. To develop the model 491 patients were eligible for ICU escalation and 769 were ineligible for escalation. Our model had good discrimination of daily risk of ICU admission in the validation cohort (n = 1141, C statistic = 0.91 (95% CI 0.89 -0.94)) and performed better than other scores (NEWS2, ISCARIC 4C) calculated using only parameters on admission, but overestimated escalation (calibration slope 0.7). A bespoke daily SARS-CoV-2 escalation risk prediction score can predict need for clinical escalation better than a generic early warning score or a single estimation of risk calculated at admission.