The Use of Clinical Scores to Predict Intensive Care Admission and Mortality in Patients with Covid-19 Disease

Abstract

Introduction: The SARS-Cov2 pandemic worldwide has to date caused 270,000,000 cases and 5,000,000 deaths. Many clinical scores are available for patients who arrive at the Emergency Department with Covid-19 disease and many studies carried out prospectively and retrospectively have demonstrated their effectiveness. In literature no study and no meta-analysis have been performed to evaluate the admission to intensive care unit (ICU) and mortality rates and to establish the best score to be used. Materials and Methods: The keywords used for the investigation were clinical scores, risk stratifications, Covid patients and Emergency Department. Ultimately the selected articles were entirely found using the PubMed database. In our meta-analysis we evaluated scores obtained studying patients with Covid-19 disease and also scores already known in clinical practice and subsequently used in Covid pandemic. Sensitivity, specificity, positive (PPV) and negative predicting value (NPV), diagnostic accuracy, the Receiver Operating Characteristic (ROC) curve and the Area Under the Curve (AUC) were calculated for each score. We assessed the homogeneity of all studies covered by our meta-analysis and at the end we created a Funnel Plot diagram in order to compare the scores. Results: A total of 8 studies, 5 retrospective observational and 3 prospective, were analyzed. When considering meta-analyzes the highest value, LR + 9.1 (4.76 - 17.5), refers to NEWS. When considering ICU admission and mortality, the AUC values referring to NEWS are 0.9 (0.9 - 1.00) and 0.9 (0.72 - 1.00) respectively. The LR +, AUC values for intensive care and mortality are 3.3 (2.4-4.3), 0.8 (0.72-0.88) and 0.8 (0.72-0.88) for NEWS2. Evaluating the fixed effect, the LR + value of q-SOFA is 2.816. The AUC value is 0.8 (0.76 - 0.85) for mortality and 0.8 (0.64 - 0.93) for intensive care admission. Discussions: NEWS and NEWS2 are the most effective clinical scores to predict mortality and admission to ICU and are probably more practical during the first phase of the disease when the patient is at home and monitoring is self-managed.