Prediction models for diagnosis and prognosis of COVID-19 infection

Added April 29, 2020

Citation: Wynants L, Van Calster B, Bonten M, et al. Prediction models for diagnosis and prognosis of covid-19 infection: systematic review and critical appraisal BMJ 2020; 369: m1328

What is this? Efficient early detection and diagnosis of patients with COVID-19 disease is important along with methods to evaluate the prognosis of patients with confirmed disease.

In this rapid systematic review, the authors searched for studies that developed or validated multivariable prediction models for diagnosis and prognosis of COVID-19 infection. They did not restrict by type or language of publication but restricted to studies published since 3 January 2020 and did their search up to 24 March 2020. They included 27 studies which described 31 prediction models. Of these, 18 were diagnostic models for detecting COVID-19 infection (13 were machine learning based on CT scans); and 10 prognostic models for mortality risk, progression to severe disease, or length of hospital stay. They also included 3 models relevant to hospital admission from pneumonia and other events (as proxy outcomes for COVID-19 pneumonia).

What was found: At the time of this review, there is uncertainty on the reliability of these prediction models when used in daily practice.

At the time of the review, the available research that was included indicates that existing models were inadequately reported, had used poor methodology and was subject to a high risk of bias, with their reported performance likely to be optimistic.


Disclaimer: This summary has been written by staff and volunteers of Evidence Aid in order to make the content of the original document accessible to decision makers who are searching for the available evidence on the coronavirus (COVID-19) but may not have the time, initially, to read the original report in full. This summary is not intended as a substitute for the medical advice of physicians, other health workers, professional associations, guideline developers, or national governments and international agencies. If readers of this summary think that the evidence that is presented within it is relevant to their decision-making they should refer to the content and details of the original article, and the advice and guidelines offered by other sources of expertise, before making decisions. Evidence Aid cannot be held responsible for any decisions made about the coronavirus (COVID-19) on the basis of this summary alone.