Performance of the National Early Warning Score in Groups of Patients with Differing Diagnoses

  • Connor Price

Student thesis: Doctoral Thesis


The National Early Warning Score (NEWS) is a tool mandated for use within all NHS trusts within the UK and is widely used around the world. It is an early warning score system, and it indicates to clinical staff whether a patient’s physiology is abnormal, and if the patient is at risk of deterioration. This triggers intervention by clinicians and escalates patients in need to more specialised care and more frequent monitoring.
NEWS has been validated on many different cohorts of patients by a multitude of researchers.
Validation in patients with specific medical conditions has not been performed comprehensively, as researchers have typically validated for conditions which are of clinical importance to them. Where these evaluations have taken place, the processes vary. This study standardises the analysis approach and applies it to groups of patients with similar conditions, as established within NHS reporting. It aims to establish whether use of NEWS is appropriate in patients with a variety of medical conditions, and if use of alternative early warning scoring systems could better detect deterioration in some patients.
We used a dataset from Queen Alexandra Hospital, Portsmouth Hospitals University NHS Trust, containing patient admissions occurring between January 2010 and June 2018. Analysis was done on 360,361 admissions containing 9,299,562 observation sets.
Comparison of NEWS performance in 121 patient groups established that while NEWS performs well, performance is not equal in all patients. Performance ranged from outstanding (AUC 0.9 – 1.0) to acceptable (0.7 - 0.8) for the primary outcome (death or ICU admission within 24 hours of an observation). Within patients suffering from cardiac arrest, AUCs were as low as 0.733, compared to an AUC of 0.911 in breast cancer patients. NEWS thresholds used to escalate patients in clinical care were less effective in low performing groups, with greater proportions of patients requiring escalation, and greater numbers of patients at risk of not meeting the threshold which would escalate them to the required care pathways.
Comparison with other early warning scores proved that NEWS is the best performing early warning system of those evaluated. For the primary outcome, NEWS was the best performing early warning score in every patient group. In the sub-outcome of ICU admission within 24 hours, it was outperformed in only 3 out of 121 patient groups. That performance improvement was marginal (AUC difference of <0.05 between NEWS and the highest performer), and therefore insufficient to warrant a change in hospital practice.

The results of this study show that there is no clear benefit to using any of the alternative scoring systems evaluated for patients in any of the diagnostic groups, however there is room for improvement of NEWS in certain patient groups. Further research is required to identify more specific patient diagnoses in which improvements could be made, and whether improvement of NEWS is practical or possible while retaining the benefits of its simplicity and universality.
Date of Award18 Jan 2024
Original languageEnglish
Awarding Institution
  • University of Portsmouth
SupervisorJim Briggs (Supervisor), David Prytherch (Supervisor) & Christina Johanna Fitch (Supervisor)

Cite this