External validation of the National Early Warning Score 2 (NEWS2) prediction of in-hospital death in patients with type II respiratory failure: a multi-centre database study [protocol]
Watkinson P., Collins G., Gerry S., Malycha J., Pimentel M., Prytherch D., Redfern O., Schmidt P., Smith G.
Abstract Background: The National Early Warning Score (NEWS), a vital-signs-based early warning score (EWS), has been suggested to perform poorly in patients with hypercapnic respiratory failure. A newly updated version (NEWS2) includes modified peripheral oxygen saturation (SpO2) thresholds for such patients and additional weightings for higher SpO2 values if they are receiving oxygen therapy. NEWS2 has been mandated for use in acute NHS hospitals. However, these changes await validation. Prior to the publication of NEWS2, the Chronic Respiratory Early Warning Score (CREWS) had been proposed as an adjustment to NEWS for these patients. Methods: Retrospective observational cohort study using the HAVEN database of admissions to four Oxford University Hospitals January - December 2016 and Portsmouth Hospitals NHS Trust January 2012- December 2016. Our primary endpoint will be the ability of NEWS2, NEWS and CREWS to predict in-hospital death within 24 hours of an observation set in patients with and without type II respiratory failure. Secondary endpoints will include cardiac arrest, intensive care admission and the first of intensive care admission, cardiac arrest or in-hospital death. We will conduct subgroup analyses using different methods for identifying patients with or at risk of type II respiratory failure. We will use the C-statistic to assess discrimination and report sensitivities, specificities and predictive values at recognized cut-offs. To assess the incremental change between NEWS and NEWS2 of the new thresholds for COPD we will present change in the C-statistic. Conclusion: We will provide the first multi-centre assessment of whether the changes in NEWS2 have led to improved discrimination of adverse outcomes in patients with type II respiratory failure, without the disadvantaging the performance in patients without this condition.