Predictive performance of National Early Warning Score 2 in detecting clinical deterioration in GW

Year Submitted: 2025
Published Date:

Tags: Technology, Care Process ＆ Redesign, Video/Video Conferencing

About this Content

Aims

The study aims to: (1) externally validate the performance of NEWS2 in Singapore ; and (2) analyse the temporal dynamics of NEWS2 to assess its predictive performance across different time windows preceding clinical deterioration.

Background

The National Early Warning Score 2 (NEWS2) is widely used in the hospital to identify clinical deterioration. However, the evidence of its effectiveness in Asian populations is not well-studied.

Methods

This retrospective study analysed all adult patient encounters between July 2022 and July 2023 from a Singapore tertiary hospital. We evaluated NEWS2's effectiveness in predicting clinical deterioration events, which include death, unanticipated intensive care unit /high dependency (ICU/HD) admission, code blue activation within 24 hours and across various timeframes (4-72 hours).

Results

Of the 58,809 patient encounters, 2798 (4.8%) experienced a clinical deterioration event. The 24-hour event rates were 0.51% for mortality, 0.24% for unanticipated ICU/HD admissions, and 0.05% for code blue activation. NEWS2 showed strong overall discrimination (AUROC 0.898, 95% CI: 0.881-0.915) for clinical deterioration. Performance was excellent for mortality prediction (AUROC 0.963, 95% CI: 0.952-0.974), moderate for code blue activation (AUROC 0.795, 95% CI: 0.714-0.876) and modest for unanticipated ICU/HD admission (AUROC 0.761, 95% CI: 0.715-0.808). Temporal analysis showed consistent mortality prediction across different timeframes, with varying performance for unanticipated ICU/HD admission and code blue activation. Prediction efficiency curve supported our institutions practice of using NEWS2 7 as emergency threshold trigger that balanced deterioration event capture (67.1%) and alert burden (9.2%).

Conclusion

Our findings support the use of NEWS2 as a reliable tool for identifying patients at risk of clinical deterioration and death. However, its predictive accuracy for unanticipated ICU/HD admission and code blue activation remains limited. The consistent performance of NEWS2 across various time windows highlights its utility for both immediate risk assessment and ongoing patient monitoring in the general ward to facilitate timely interventions when used alongside clinical judgement.

Lessons Learnt

First, the single-center design may limit generalisability of our findings. Second, our ability to capture deterioration events was limited by the frequency of NEWS2 recordings by nurses on the ground, reflecting real-world practice. Third, we were unable to account for clinical interventions that might have altered the natural course of deterioration.

Additional Information

SHBC Singapore Nursing Award (gold)

Keywords

NEWS2, early warning score, clinical deterioration, prediction, validation

Innovators' Details

Connect with this contributor!

Chen Li - chen.li@nhghealth.com.sg

Project Attachment