Skip to main content


The Multicenter REgistry of potential COVID-19 in emERgency care (Project RECOVER) is designed to acquire data from patients who visited a U.S. emergency department and were tested in the next 14 days for SARS-CoV-2. The investigators believe this will be by far the world’s largest repository of clinical features of COVID-19. The registry will include at least 500 patients tested for SARS-CoV-2 from over 40 institutions representing 26 states and the District of Columbia.

The study aims to enroll a minimum of 20,000 patients by the end of June 2020. Approximately 1/3 of the database is expected to be comprised of COVID-19 positive patients. The aims are to develop a scoring system to estimate the probability of a patient having COVID-19 and to develop the COVID Rule-out Criteria (the CORC rule) to permit exclusion of COVID-19 at the bedside. A third aim will produce a prognostic rule for COVID-19 positive patients.

To accomplish these research aims, the primary research tool is a 169-item data collection instrument, in the REDcap format, that collects symptoms, signs, medical history, laboratory and radiographic results, and outcomes up to 30 days after testing. Recognizing the uncertain diagnostic sensitivity of swab testing, our criterion standard to rule out COVID-19 includes negative testing for 30 days. Each data field, including their content, format, and “field note” instructions were carefully considered by Delphi approach and pilot testing a team of over 15 site investigators. An interim analysis is planned after 5,000 entries to examine for possible need for modifications. After 10,000 patients, the derivation of the first two aims will be constructed, and then validated in the second 10,000 patients.

The importance of this project will increase with time, as COVID-19 becomes an endemic problem, and it can present with a vast array of symptoms, and therefore will always be on the mind of millions of patients and their providers in the emergency care setting. However, swab or antibody testing all patients is neither sensible nor feasible, owing to problems of availability, turn around time and questions about diagnostic sensitivity. In addition to these primary aims, ancillary aims will provide predictive criteria for venous and arterial thromboembolism, myocarditis, need for intubation, and neurological sequalae.

Multi-institutional Participation

Join this multi-institutional research initiative by contacting the research team for more information.

Jeffrey Kline, MD

Alfred Wang, MD