RESUMO
INTRODUCTION: The early identification of patients with SARS-CoV-2 infection is still a real challenge for emergency departments (ED). First, we aimed to develop a score, based on the use of the lung ultrasonography (LUS), in addition to the pre-triage interview, to correctly address patients; second, we aimed to prove the usefulness of a three-path organization (COVID-19, not-COVID-19 and intermediate) compared to a two-path organization (COVID-19, non-COVID-19). METHODS: We retrospectively analysed 292 patients admitted to our ED from 10 April to 15 April 2020, with a definite diagnosis of positivity (93 COVID-19 patients) or negativity (179 not-COVID-19 patients) for SARS-COV-2 infection. Using a logistic regression, we found a set of predictors for infection selected from the pre-triage interview items and the LUS findings, which contribute with a different weight to the final score. Then, we compared the organization of two different pathways. RESULTS: The most informative factors for classifying the patient are known nasopharyngeal swab positivity, close contact with a COVID-19 patient, fever associated with respiratory symptoms, respiratory failure, anosmia or dysgeusia, and the ultrasound criteria of diffuse alveolar interstitial syndrome, absence of B-lines and presence of pleural effusion. Their sensitivity, specificity, accuracy, and AUC-ROC are, respectively, 0.83, 0.81, 0.82 and 0.81. The most significant difference between the two pathways is the percentage of not-COVID-19 patients assigned to the COVID-19 area, that is, 10.6% (19/179) in the three-path organization, and 18.9% (34/179) in the two-path organization (p = 0.037). CONCLUSIONS: Our study suggests the possibility to use a score based on the pre-triage interview and the LUS findings to correctly manage the patients admitted to the ED, and the importance of an intermediate area to limit the spread of SARS-CoV-2 in the ED and, as a consequence, in the hospital.