National Early Warning Score 2 (NEWS2)

Clinical background

The National Early Warning Score 2 (NEWS2) is an aggregate early warning system intended to standardise the assessment of acute illness severity and early deterioration in adult patients. The instrument was produced by the Royal College of Physicians (UK) and published in 2017 as an update of the original NEWS from 2012. It serves two purposes: to act as a common language for communicating a patient's condition between levels of care and professional groups, and to trigger structured escalation when physiological parameters deviate.

The decision the instrument serves is when a patient in the emergency department or on a ward should receive increased monitoring, an urgent medical review or referral to intensive care. Without a standardised scoring system this decision depends heavily on individual clinical judgement, which in turn leads to both under- and over-escalation. NEWS2 addresses two weaknesses of its predecessor in particular: the inability to handle patients with hypercapnic respiratory failure who require lower oxygen targets, and the absence of new confusion as a warning sign.

Calculating NEWS2

NEWS2 is the sum of points from seven physiological parameters:

$$\text{NEWS2} = P_{\text{resp}} + P_{\text{SpO}$$2} + P{\text{O}2} + P{\text{temp}} + P_{\text{SBP}} + P_{\text{HR}} + P_{\text{ACVPU}}

where each variable scores 0 to 3 points on a fixed scale:

• Respiratory rate: 0 points at 12 to 20 breaths/min, up to 3 points at ≤8 or ≥25
• Oxygen saturation: scored according to the SpO2 scale chosen. Scale 1 (standard) gives 0 points at ≥96%, up to 3 points at ≤91%. Scale 2 (target 88–92%, hypercapnic failure) gives 0 points at 88–92%, with points for values both below and above the target range
• Receiving supplemental oxygen: 0 points if no, 2 points if yes
• Temperature: 0 points at 36.1 to 38.0 °C, up to 3 points at ≤35.0 or ≥39.1 °C
• Systolic blood pressure: 0 points at 111 to 219 mmHg, up to 3 points at ≤90 or ≥220 mmHg
• Heart rate: 0 points at 51 to 90 beats/min, up to 3 points at ≤40 or ≥131 beats/min
• Level of consciousness (ACVPU): 0 points for Alert, 3 points for new Confusion, Voice, Pain or Unresponsive (CVPU)

The derivation rests on the original NEWS, developed in a cohort of about 35,000 consecutive acute medical inpatients with nearly 200,000 vital sign observations, in which NEWS outperformed 33 other early warning systems with an AUC of 0.873 for death or transfer to intensive care within 24 hours [4]. NEWS2 was introduced not as a statistically re-derived model but as a clinical modification of NEWS, in which SpO2 scale 2 and ACVPU were added on expert grounds to address identified clinical gaps.

Interpretation in practice

The score translates into three risk bands with specific clinical actions:

The threshold at 5 points is the single most important: it marks the transition from routine monitoring to active medical review. A patient at 4 points who then develops a new parameter scoring 3 formally falls into the medium band even though the total is below 5, since a single parameter scoring 3 is a red flag in itself.

Validation and performance

A systematic review of early warning systems across different patient groups and care settings found that NEWS performed best in medical (AUC 0.74) and surgical (AUC 0.77) settings and in respiratory disease (AUC 0.77), with considerable heterogeneity between studies (I² 72–99%) and generally insufficient methodological quality [1]. Few studies specifically evaluated NEWS2 at the time of the review.

In a large external validation in an Asian population comprising 58,809 patients on general wards in Singapore, NEWS2 achieved an AUROC of 0.898 for the composite outcome of clinical deterioration (death, unplanned transfer to intensive care or a code blue) within 24 hours [2]. Discrimination was excellent for mortality (AUROC 0.963) but only moderate for unplanned intensive care transfer (AUROC 0.761) and code blue (AUROC 0.795). Calibration was excellent for mortality but poor for intensive care transfer and cardiac arrest, suggesting that the score overestimates the risk of these outcomes in the higher risk ranges. At a threshold of ≥5, 82.2% of all deterioration events were captured, but 22.9% of patients triggered an alert, illustrating the trade-off between sensitivity and alarm burden [2].

A prospective validation in an emergency department in Colombia with 3,986 patients reported an AUC of 0.90 for in-hospital mortality, with an optimal threshold at ≥7 points (sensitivity 66%, specificity 96%) [3]. In a smaller Norwegian prospective cohort of 66 patients with COVID-19, a NEWS2 ≥6 on arrival predicted severe disease with an AUC of 0.822 (sensitivity 80%, specificity 84%) and was superior to qSOFA (AUC 0.624) [4].

Overall, NEWS2 shows strong discrimination for mortality but weaker performance in predicting the need for intensive care and cardiac arrest, particularly outside medical settings.

Limitations

NEWS2 is validated for adult patients (≥16 years in the original guidance, ≥18 years in most validation studies) and should not be applied to children or pregnant women. The instrument was developed for acute care and loses predictive value in elective or chronic settings where vital signs are stable but the patient may nonetheless deteriorate for other reasons.

The most important misapplications are:

• The wrong SpO2 scale: Scale 2 (target 88–92%, hypercapnic failure) should be used only for patients with a clinically agreed lower oxygen target, typically in chronic hypercapnic respiratory failure such as COPD. Applying scale 2 to a patient without hypercapnic failure underestimates hypoxaemia, and applying scale 1 to a patient with a target of 88–92% may give falsely high scores for acceptable saturations.
• A static measurement: NEWS2 is a snapshot. A patient scoring 3 may be stable or deteriorating rapidly. The score must be interpreted as a trend, not as a single measurement.
• Supplemental oxygen as a binary variable: receiving supplemental oxygen scores 2 points whether the patient is on 2 or 15 litres. A patient with an increasing oxygen requirement but an unchanged saturation may therefore have an unchanged NEWS2, masking deterioration. The Royal College of Physicians emphasises that any increase in oxygen requirement should trigger review regardless of the score.
• Limited calibration for non-mortality outcomes: as the validation studies show, the score is most reliable for predicting death and least reliable for predicting the need for intensive care, where other factors (decision-making, bed availability, the patient's chronic status) influence the outcome more than the vital signs do.

References

1. Alhmoud B et al. Performance of universal early warning scores in different patient subgroups and clinical settings: a systematic review. BMJ Open 2021. PMID: 36044371
2. Chen L et al. Predictive performance and temporal dynamics of national early warning score 2 (NEWS2) in detecting clinical deterioration in general ward: A large-scale validation study in Singapore. Resuscitation Plus 2025. PMID: 41142216
3. Vergara P et al. Validation of the National Early Warning Score (NEWS)-2 for adults in the emergency department in a tertiary-level clinic in Colombia: Cohort study. Medicine 2021. PMID: 34622831
4. Myrstad M et al. National Early Warning Score 2 (NEWS2) on admission predicts severe disease and in-hospital mortality from Covid-19: a prospective cohort study. Scandinavian Journal of Trauma, Resuscitation and Emergency Medicine 2020. PMID: 32660623