P252 Role of chest x-ray in diagnosis and predicting outcome of Covid19 infection in a district general hospital setting

Introduction: We set out to look at the role of chest x-ray (CXR) in diagnosing novel coronavirus (Covid19) infection and to explore if it can predict clinical outcomes We compared the chest imaging and swab results in Covid19 patients Demographics, symptoms and CXR findings were explored as predictors of clinical outcome in COVID19 patients Methods: All adult patients who had CXR and Reverse Transcriptase Polymerase Chain Reaction (RT-PCR) available between March and June 2020 were included Data was collected retrospectively from electronic case notes CXR reported as typical or atypical COVID features1 was considered as positive For predictors of outcomes, regression analysis was conducted Results: 876 patients had both CXR and RT-PCR swabs Their mean age was 64 6 years and age range was 17 to 105 years 324 (37%) were positive and 552 (63%) negative on RT-PCR CXR showed typical COVID19 changes in 217 (24 8%) and atypical COVID19 changes in 148 (16 9%) patients The sensitivity and specificity of CXR in the overall study group was 59 9% and 69% with positive predictive value (PPV) and negative predictive value (NPV) of 53 2% and 74 6% respectively 692 patients were admitted and the sensitivity and specificity of CXR in this group was 66 9% and 65 5% with PPV and NPV of 53% and 77 4% respectively 148 patients (16 8%) received ventilator support (27 received invasive and 121 had non-invasive ventilator support) and the sensitivity and specificity of CXR in this sub-group was 78% and 42 4% with PPV and NPV of 62 7% and 60 9% There were 129 deaths and on multivariate regression analysis, age (> 60 years) and positive CXR remained significant risk factors for the clinical outcome (Table) In contrast to the current literature, our sample did not show gender as a risk factor to predict the outcome Conclusions: CXR has a reasonable specificity and sensitivity for diagnosing COVID19 and it increases with the severity especially in patients needing ventilator support CXR can be used in predicting worse clinical outcome in COVID19 pneumonia