[Diagnostic algorithm for COVID-19 at the ER].

Dofferhoff, A S M; Swinkels, A; Sprong, T; Berk, Y; Spanbroek, M; Nabuurs-Franssen, M H; Vermaat, M; van de Kerkhof, B; Willekens, M H C; Voss, A

OBJECTIVE Evaluation of a diagnostic algorithm for estimating the risk of COVID-19 in patients who are referred to an emergency department for being suspected of having the disease. DESIGN Retrospective study. METHOD Patients with fever with no apparent cause and patients with recently developed respiratory symptoms, whether or not in combination with fever, were routinely given a PCR test, blood tests (lymphocyte count and LDH levels) and a chest CT scan. The CT scan was assessed according to the CO-RADS classification. Based on the findings, the patients were divided into 3 cohorts (proven COVID-19, strong suspicion of COVID-19, and low suspicion of COVID-19) and the appropriate isolation measures were taken. RESULTS In the period from 8 to 31 March 2020, the algorithm was applied to 312 patients. COVID-19 was proven for 69 (22%) patients. COVID-19 was strongly suspected for 151 (48%) patients and suspicion was low for the remaining 92 (29%) patients. The percentage of patients with positive PCR results and the percentage of patients with abnormal laboratory test results increased as the CO-RADS score increased. Among patients with a CO-RADS score of 4 or 5, this percentage increased further when they also had lymphopenia or elevated LDH levels. We have adjusted the flowchart based on our findings. CONCLUSION In case of patients who have been referred to an emergency department for suspected COVID-19, a good COVID-19 risk assessment can be made on the basis of clinical signs, laboratory abnormalities and low-dose CT scans. Even before the results of the PCR test are known and even if the results are negative, patients can be classified as 'proven COVID-19 patients' using the algorithm.