P11 COVID 19: Utility of plain chest radiograph scoring system to predict disease severity and outcomes

IntroductionPlain chest radiograph (CXR) is the most common imaging modality used to evaluate respiratory symptoms CXR severity scoring has been shown to be an independent predictor of need for hospital admission and intubation and mechanical ventilation (IMmoderate changes were: bilateral changes zones 2–3;severe changes: changes zones 4–6 CXRs were reviewed and scored independently by 2 reporters: thoracic radiologist and acute medical physician who were blinded to baseline patient characteristics and outcomes Results200 patients (median age: 79 (IQR 63–86) years) were included, 108 of which were females and 92 males 61 (30 5%) died and 139 (69 5%) were discharged During admission, 19 (9 5%) were admitted to ITU, 2 (1%) to the Non-invasive ventilation (NIV) unit and 179 (89 5%) to COVID Medical wards Of the 61 patients (median age: 82 (IQR 73–89) years;27 (44 3%) male, 34 (55 7%) female) who died: 3 (4 9%) were admitted to ITU, 1 (1 6%) to NIV unit and 57 (93 4%) to COVID medical wards;45 (73 8%) received oxygen up to 15L, 1 (1 6%) received nasal high flow oxygen, 2 (3 3%) received CPAP and 3 (4 9%) received IMno CXR changes (p=0 099), Mild CXR changes (p=0 416), Moderate CXR changes (p=0 283), Severe CXR changes (p=0 994) Severe CXR changes was an independent predictor of IM95% CI 1 156–4 566;p=0 018) ConclusionWe conclude that a CXR severity score is an effective tool to predict risk for hospital admission and the need for I&MV Further larger studies will help validate this score by following up repeat CXRs to determine disease trajectory