AI-enhanced Covid 19 Prognostic Algorithm (HOST) 1 [COVID-19]
Research type
Research Study
Full title
Improved analysis of data and prognosis in patients with Covid 19 using statistical analysis and deep machine learning (retrospective study)
IRAS ID
282670
Contact name
Fergus Gleeson
Contact email
Sponsor organisation
University of Oxford
Duration of Study in the UK
1 years, 0 months, 1 days
Research summary
Covid-19, a new viral illness caused by SARS-CoV-2, appeared in China in late 2019 and has since spread to the rest of the world. It appears to have predominantly respiratory symptoms. The current method of confirming SARS-CoV-2 infection is by using a Polymerase Chain Reaction (PCR) on a nasopharyngeal swab specimen, which has a turnaround time of 24-48 hours in UK centres. The sensitivity of the PCR test is thought to be approximately 70% and the need for equipment and trained operators has limited the scaling of testing facilities. The PCR test does not identify or stratify patients according to severity of symptoms. \nCurrently, all patients being admitted into hospital have chest x-rays (CXRs), and standardised tests such as oxygen measurements and blood tests, with a smaller number having CT scans. We plan to take the results of routine tests and patient outcomes, and develop a scoring system that will allow clinicians to identify patients who may be sent home, require admission and/or require careful monitoring. It may also allow clinicians to identify patients who may require ventilation and those who are at very high risk from the infection and will thus require ventilation to avoid death. We also aim to develop computer derived artificial intelligence programmes that can help with diagnosis and patient monitoring. It is likely that it will be quicker to develop the scoring system than the artificial intelligence models, so we will conducts these in parallel.\n
REC name
N/A
REC reference
N/A