Prognostic accuracy of imaging findings for predicting morbidity and mortality in patients with COVID-19

Objectives This is a protocol for a Cochrane Review (prognosis). The objectives are as follows: To evaluate the accuracy of imaging findings (on chest CT, chest X‐ray, and ultrasound of the lungs) to predict the following medical outcomes in people with COVID‐19. Morbidity, including: complications (e.g. development of hypoxia, severe AKI, delirium); escalation of care (e.g. hospital admission, ICU admission, non‐invasive ventilation, intubation, mechanical ventilation, ECMO, need for renal replacement therapy, need for transfusion); and length of hospital admission. Mortality, including disease‐specific mortality; and all‐cause mortality. We will evaluate each outcome separately (e.g. hypoxia alone), or in combinations (e.g. hypoxia and AKI), depending on the granularity of the outcome data reported in primary studies. We will evaluate each outcome according to time horizon, as detailed in the Methods section. Secondary objectives When data are available, we will investigate whether prognostic accuracy varies according to covariates of interest, including imaging technique, timing of imaging test, test used to confirm COVID‐19, duration of symptoms, timing of outcome confirmation, study design, study setting, participant age, and presence or absence of pulmonary embolism (PE) on CT pulmonary angiography (CTPA).