[Usefulness of COVID-19 patients first analysis made at the Emergency Department as a prediction tool for mortality (PMCovid Score).]

ABSTRACT

OBJECTIVE: The COVID-19 pandemic has generated a high demand for hospital resources taking the national health system to its limit. In order to reduce this over burden and to avoid a potential system collapse, it would be beneficial to generate scientific evidence for the prognosis of the disease and to count with models that are able to forecast the mortality and progression of the disease. Identify mortality risk factors in COVID-19 patients from analytic data obtained from the Emergency Service at our hospital and to elaborate a prognostic score for predicting 30-day mortality (PMCovid Score) that will be included in the report submitted by the Emergency Clinic Laboratory. METHODS: Transversal descriptive study in a population that came to the Emergency Service at the University Hospital of Jaen between March 8th and April 7th 2020. We obtained the variables for the prognosis by a univariate data analysis. On this basis, we applied a multivariate analysis of the logistical regression of the mortality after 30 days in order to generate a prognostic score which was validated subsequently by the TRIPOD method. RESULTS: 298 patients were included. PMCovid Score assigns 1 point to patients age >=77 years old;1 point to patients with a urea level >=49 mg/dL, 1 point to erythrocyte values <4.6x106/microL, 2 points to platelet values <165x103/microL;1 point to patients with a percentage of lymphocytes below 18.1%;1 point to those with a % of monocytes <6.8% and 2 points if the % of eosinophils is <0.4%.Our score had a predictive accuracy of 88.6% (AUC 0.886 IC at 95%;0.842-0.931), with a sensibility of 91.7% (IC at 95% 82.810-100) and a specificity of 69.7% (IC at 95% 63.840-75.680). CONCLUSIONS: PMCovid Score provides the doctor with information on the prognosis of the positive COVID-19 patient along with the usual first analysis data, the necessary parameters for the calculations are available at all Emergency Laboratory Clinics. This information can be very useful for the management of this kind of patients and their classification based on the risk supplied by the PMCovid Score.