36

4
Exploring the performance of the National
Early Warning Score (NEWS) in a European
emergency department

N. Alam, I.L. Vegting, E. Houben, B. van Berkel, L. Vaughan, M.H.H. Kramer,

P.W.B. Nanayakkara

Resuscitation. 2015 May;90:111-5

37
 Abstract

Background: Several triage systems have been developed for use in the emergency department (ED), how-
ever they are not designed to detect deterioration in patients. Deteriorating patients may be at risk of going
undetected during their ED stay and are therefore vulnerable to develop serious adverse events (SAEs). The
National Early Warning Score (NEWS) has a good ability to discriminate ward patients at risk of SAEs. The
utility of NEWS had not yet been studied in an ED.

Objective: To explore the performance of the NEWS in an ED with regard to predicting adverse outcomes.

Design: A prospective observational study. Eligible patients were those presenting to the ED during the 6
week study period with an Emergency Severity Index (ESI) of 2 and 3 not triaged to the resuscitation room.

Intervention: The NEWS was documented at three time points: on arrival (T0), hour after arrival (T1) and at
transfer to the general ward ICU (T2). The outcomes of interest were: hospital admission, ICU admission,
length of stay and 30 day mortality.

Results: A total of 300 patients were assessed for eligibility. Complete data was able to be collected for 274
patients on arrival at the ED. NEWS was significantly correlated with patient outcomes, including 30 day mor-
tality, hospital admission, and length of stay at all-time points. Conclusion: The NEWS measured at different
time points was a good predictor of patient outcomes and can be of additional value in the ED to longitudi-
nally monitor patients throughout their stay in the ED and in the hospital.

38
1. Introduction
1.1. Background
The triage of patients at the time of presentation
to emergency departments (EDs) is crucial to the
provision of safe patient care. Accurately triaging
patients is difficult, requires experience and may be
subject to inter-observer variability.1-3 Several triage
systems have been developed for use in the ED,
including the Emergency Severity Index (ESI) and
the Manchester Triage Scale (MTS).4-6 These sys-
tems provide a method of categorizing all incom-
ing ED patients by level of acuity ranging from life
threatening to non-urgent and dictate how quickly
patients should be seen. However, while virtually all
EDs employ triage systems to determine treatment
priority, the evidence suggests that less attention
is paid to the longitudinal monitoring of patients
once they are in the department.7 There are no
widely used scores specifically designed to detect
deteriorating patients or to predict the chance of
early intensive care unit (ICU) admission or death in
ED patients. Further, small, single-site studies have
demonstrated that longitudinal measurement of
routine hemodynamic parameters in EDs is poor.7
These factors suggest that deteriorating patients
may be at risk of going undetected during their
ED stay and are therefore vulnerable to develop
serious adverse events,8,9 such as unexpected car-
diac arrest and unnecessary ICU admission, with a
higher consumption of resources through longer
lengths of hospital stay (LoS). First introduced in
1997, early warning scores (EWS) were developed
in response to concerns about the failure to detect
deteriorating physiological parameters in ward pa-
tients.10 They are based on patient’s vital signs and
linked to ‘triggers’, which mandate the escalation
of monitoring or call for assistance. However, while
EWS are now widely used internationally, there are
a lot of different EWS scoring systems, with many
being adapted for use in individual hospitals.11 The
National Early Warning Score (NEWS), developed
in conjunction with the Royal College of Physi-
cians of London, has been more rigorously tested
and performs better than any of the 33 published
systems commonly in use.12 It has a good ability
to discriminate ward patients at risk of cardiac ar-
rest, death or unexpected intensive care unit (ICU)
admission and it is currently being promoted as a
standardized system across the UK. The utility of
NEWS has not yet been studied in an ED. The aim
of this study was to explore the performance of
NEWS with regard to predicting adverse outcomes,
such as ICU admission and death, in adult patients
as well as the ability of NEWS to predict the need for
hospital admission in an ED population. The study
also aimed to assess the feasibility of the use of
NEWS as a structural monitoring tool in a Dutch ED.
2. Methods
2.1. Design and setting
A prospective, observational feasibility study was
performed at the ED of the VU Medical Centre
(VUmc), an academic urban tertiary care centre in
Amsterdam, with approximately 31000 ED visits per
year. The ED of the VUmc uses the Emergency
Severity Index (ESI) for triage.5
2.2. Study population
4
Eligible patients were those of 18 years and older
presenting to the ED of the VUmc during the 6
week study period (7th January till 15th February
2013, between 12.00 and 20.00 hrs) with an ESI
of 2 and 3 not triaged to the resuscitation room.
Excluded were patients with an ESI 4 or 5 and
patients undergoing cardiopulmonary resuscitation
(ESI 1). Patients lost to follow-up were removed
from the study analysis. The standard distribution
of patients according to ESI at our ED is as follows:
ESI-1 4.9%, ESI-2 5.0%, ESI-3 44%, ESI-4 39.3%,
ESI-5 6.7%.13 The eligibility criteria were chosen on
basis of results of an earlier study performed in our
ED.13 The researchers monitored patients during the
time frame of 12.00–20.00 hrs as it mostly is the
busiest time of the day at the ED. Patients triaged
with category ESI 4 or 5 present with minor com-
plaints, the larger percentage (95–99%) of these
patients were discharged home within 4 hrs. The
Medical Ethical Committee of the hospital approved
the study. Informed consent was waived as routine
care was not influenced and no therapeutic inter-
vention was introduced.
39
 2.3. Methodology
The data used for calculation of the NEWS were
collected at three time points, namely at arrival ED
(time point 0: NEWS – T0),1 hour after arrival at
the ED (time point 1: NEWS – T1), and if appli-
cable at transfer to a hospital ward (time point 2:
NEWS – T2). Two trained researchers (EH and
BvB) collected data, which contained demographic
details, medication use, medical history, and vital
parameters needed to calculate the NEWS (see
Table 1). This data collection was in addition to
any observations taken at the time of triage. EWS
scores were calculated using the NEWS-score as
proposed in the report by the Royal College of
Physicians.14 The outcomes of interest for this study
were: hospital admission, ICU admission, length of
stay and 30 day mortality.
2.4. Statistical analysis
For data analysis IBM SPSS 20.0 (Chicago, USA)
was used. Student’s t test was used for inde-
pendent samples to compare means for nor-
mally distributed variables and Mann–Whitney U
test for variables that were not normally distrib-
uted. A p < 0.05 regarded as statistically signifi-
cant. Descriptive statistics were used to determine
patient characteristics (presented as mean ± SD).
Skewed variables were summarized using me-
dian and interquartile range (IQR). Receiver Op-
erating Characteristic (ROC) analysis was used to
identify the NEW score for the highest sensitivity.
The NEW scores were divided into three aggre-
gates, aggregate 0-4 (low clinical risk), aggregate
5-6 (medium clinical risk) and aggregate 7 or
more (high clinical risk) for the purposes of this
study and according to the NEWS thresholds and
triggers.14
40
3. Results
3.1. Patient population
A total of 300 patients were assessed for eligibility.
Complete data was collected for 274 patients at
moment of entry to the ED. The demographic char-
acteristics are shown in table 2. NEWS measured at
moment of entry for these 274 patients is shown in
table 3. The NEWS was calculated an hour later in
247 patients (Table 3a). Only 133 of the 247 pa-
tients could be followed up to calculate the NEWS
at discharge from the ED (admitted to the ward or
discharged home), as shown in table 3b. It was not
possible to collect data for all patients at all-time
points due to organizational reasons.
3.2. NEWS scores
At T0 NEW scores ranged from 0 to 11 with a
median score of 2.0 (IQR 1–4). The distribution
of NEWS measured at T0 is shown in table 3.
The distribution of NEWS measured at different time
points is shown in table 3a and table 3b.
3.3. Admission
A total of 130 patients were admitted to hospital,
142 were discharged and 2 transferred to other
hospitals (Table 2). Admission was significantly
correlated with NEWS at T0, T1 and T2, tested
with chi square (linear by linear association) with
categorized NEWS (p < 0.001), see table 3a and b.
The AUROCs (95% CI) for NEWS for admission at
T0, T1, T2 was respectively 0.664 (0.599–0.728),
0.687 (0.620–0.754), 0.697 (0.609–0.786).
3.4. Length of stay
Length of stay was significantly correlated with
NEWS, at all of the measured time points, p value
for x2 was <0.001. Spearman Rank Correlation
was significant (p < 0.0001). Median length of stay
more than doubled for a score >7 compared with
a score of 0–4. Table 4 demonstrates the NEWS
measured at T0 related to length of stay.
3.5. ICU admission
A total of 10 patients were admitted to the ICU.
We found that ICU admission significantly corre-
lated with NEWS at time points T0 (Fisher’s exact
test with NEWS above or below 7: p = 0.003),
T1 (p < 0.001) and at T2 (p = 0.046). Five of
the patients which were admitted to the ICU had
an aggregate score of 0–4 while, the other five
patients had aggregate scores of seven or more.
In the group of patients with low aggregate
scores, there were four patients which were ad-
mitted for a short while for observation and dis-
charged shortly after (Table 5). The details of the
patients admitted to ICU are given in table 5a.

3.6. Mortality
The 30 day mortality for all patients included in the
study was 4.0% (n = 11). The ESI score for 10
deceased patients was 3 and one had a score of
2. We found that mortality was significantly corre-
lated with NEWS at T0 (p = <0.0001, Table 6). The
significant correlation between NEWS and mortality
remained at all other measured time points.
4
The 30 day mortality was not significantly relat-
ed to ESI, p = 0.816. The AUROCs (95% CI) for
NEWS for mortality at pT0, T1 and T2 was respec-
tively 0.768 (0.618–0.919), 0.867 (0.769–0.964),
0.767 (0.568–0.966). Patients who died were older
(mean age 74, 25 years; SD 11.9; independent
sample t test p = 0.002) and had a higher NEWS
score (mean 6.00; SD 3.6; Mann–Whitney U test
p = 0.002) compared with patients who survived
(mean age 59.08; SD 20.1); NEWS (mean 2.51;
SD 2.56). The mean length of stay was 8.2 days for
patients who died and 4.1 days (Mann–Whitney U
test p = 0.001) for patients who survived. The de-
tails of the patients who died are given in table 6a.
Of the individual physiological measures compris-
ing the NEW score, only the respiratory rate was
significantly associated with mortality at all meas-
ured time points (Table 7). Pulse rate had a strong
correlation with mortality if measured an hour after
entry ED (at T1). No correlations could be found for
all other physiological parameters.
41
42

4. Discussion
To the best of our knowledge this is the first pro-
spective study performed in an ED to evaluate the
performance of the NEWS, a standardized score.
We found that the NEWS was significantly corre-
lated with patient outcomes, including 30day mor-
tality, hospital admission, and length of stay at all
timepoints. ICU admission was significantly corre-
lated with NEWS at T0 and T1. Several studies have
been performed to investigate the value of early
warning scores in identifying patients at risk of de-
terioration15,16, their effect on clinical outcomes17–22,
and their ability to predict clinical outcomes.23–27
Using early warning score may influence clinical
outcomes positively. However studies performed
to date have used different early warning scores
which makes it difficult to draw general conclusions
concerning the utility of these scores.28,29
4
Moreover, no studies have been performed to vali-
date the use of NEWS in de the ED. Our study
indicates a number of potential uses for the NEWS
in the ED. ED triage systems such as ESI and MTS
were explicitly developed to assess the priority
of patients fo clinical intervention. Although some
studies have demonstrated that MTS and ESI may
provide us with some useful prognostic information
as a predictor of admission or 24 hour mortal-
ity,30–32 the validity of these triage scales has not
been evaluated across different triage levels for
outcomes such as early death and hospital admis-
sion. Moreover there was wide variation between
different triage scales as well as studies using the
same scales. While physiological scoring alone is
unable to identify patients at risk of poor outcomes,
the recognition of which was a spur for the devel-
opment of MTS in particular. Our study demon-
strates that NEWS can further risk stratify patients
within higher ESI risk categories, both for death
and need for admission to ICU. Patients with a
high NEW score have not only been identified as
being at risk of a poor outcome, but have already
physiologically deteriorated to the extent where ur-
gent medical review and intervention is required.
ESI and MTS have also been shown to predict the
need for medical admission.
43
44
Again, within higher triage categories, NEWS dem-
onstrated that it was able to predict not only the
need for admission, but length of inpatient stay. This
suggests that not only patients with higher NEW
scores should be seen more urgently by medi-
cal staff, but these patients could be fast tracked
for admission once stabilized. This study stands in
contrast to others on the performance of NEWS,
as mortality was calculated at 30 days, rather than
within 24 hrs. This was a pragmatic choice, due
to the small sample size and, as predicted, no
patients in the study died or experienced an unex-
pected cardiac arrest within 24 hrs of presentation.
NEWS was a reasonable predictor of death at 30
days and all patients who died during the course of
the index admission has NEWS scores that classi-
fied them as being at medium or high clinical risk.
The four patients who were identified as being low
risk who died within 30 days all did so either within
a palliative setting on the index admission or on
a subsequent admission. The study also looked
at patients longitudinally, rather than at single time
points. Therefore NEWS seems to be a good pre-
dictor of mortality at 30 days, even when calculated
at different time points during the stay at the ED.
This study conceived, in part, to explore the utility
of NEWS as a structural monitoring system in the
ED, analogous to its ward use. The study demon-
strated that using NEWS was possible in the ED
setting, although much larger studies are needed
to determine how frequently observations need to
be performed and whether the trigger thresholds
for intervention need to be adjusted in ED patients.
These findings suggest that the use of NEWS in
conjunction with the ESI triage has the potential to
better identify patients in need of urgent attention
and potentially expedite admissions to the medi-
cal wards and intensive care, thereby smoothing
patient flow through the ED. Our study differs from
previous work as we measured the NEWS at dif-
ferent time points during a patients stay at the ED.
Furthermore, motivated and trained researchers
performed the measurements. For a proper uptake
and implementation of the NEWS in future, train-
ing and motivating the staff is important. Quality
of care may be greatly improved by a more col-
laborative approach, where in the NEWS can be
a potential asset. A standardized approach, such
as the NEWS, can help in improving continuity of
care, improve communication between staff and
provide a basis for the management of critically
ill patients. When comparing the individual physi-
ological parameters of the deceased patients with
the patients who survived, respiratory rate had the
most significant correlation with mortality in our
study. This is in keeping with other studies,1, 33, 34
which have found that an elevated respiratory rate
is a strong indicator of acute illness and distress.
However, it is the sign which is documented most
poorly in ED patients, implying that more work need
to be done in the ED to improve longitudinal moni-
toring of the patients during their stay at the ED.7
Our study has several limitations, as it was a single
center study of observational nature in a specific
setting with a convenient patient selection. Patient
observations were collected by trained clinicians
under trial conditions, rather than nursing staff un-
der departmental pressures and so might be more
accurate than routine observations. However, we
could not obtain complete datasets at all consecu-
tive time points; a large percentage of the patients
needed diagnostic tests and were not present at the
ED for further monitoring at time point T1, or were
otherwise directly transported from the ED to the
ward (T2). Although the researchers had request-
ed to be notified when patients were discharged,
nurses or other staff sometimes forgot to notify on
time. Some eligible patients were also missed as
only two researchers were available to calculate
the NEWS during the inclusion period and only
one patient at a time could be monitored. However,
most of the patients who presented at the ED were 4
included in the study at T0. Therefore this sample
is a good representation of the total population of
ESI 2 and 3 patients at our ED. The median age of
our study population was 60 years and the mortality
was relatively low (4%). The number of deceased
patients was too small to perform a further in depth
analysis. The dataset was also too small to detect
the effect of changes on patient outcomes with
each one point increase or decrease in the NEWS.
The patients in our ED stay for a relatively short
period, with most patients (84%) being discharged
within four hours.13 The other additional benefits of
NEWS, such as improving continuity of patient care
and fostering a uniform approach to the unwell
patient, were unable to be explored. The stand-
ardization of the score however offers many op-
portunities of easier implementation, e.g. incorpo-
ration of NEWS education in the nursing curriculum.
5. Conclusion
The NEWS measured at different time points was
a good predictor of patient outcomes. The NEWS
can be of additional value in the ED, although not
specifically as a triage system, but as a means to
longitudinally monitor patients throughout their stay
in the ED and in the hospital. Through its use, clini-
cal staff has a better indication whether a patient is
more at risk. Hereby making it possible to timely
intervene and stabilize a patient before further dete-
rioration occurs. Moreover, it may serve as a useful
tool in managing patients in a complex ED environ-
ment and so enhance the quality of patient care.
45
 References
1. Farrohknia N, Castren M, Ehrenberg A, et al. Emergency depart-
ment triage scalesand their components: a systematic review of the
scientific evidence. Scand JTrauma Resusc Emerg Med 2011;19:42.
2. Storm-Versloot MN, Ubbink DT, Choi V, Luitse JS. Observer
agreement of the Manchester Triage System and the Emergency
Severity Index: a simulation study. Emerg Med J 2009;26:556–60.
3. Fitzgerald G, Jelinek GA, Scott D, Gerdtz MF. Emergency depart-
ment triage revisited. Emerg Med J 2010;27:86–92.
4. Eitel DR, Travers DA, Rosenau AM, Gilboy N, Wuerz RC. The
emergency severity index triage algorithm version 2 is reliable and
valid. Acad Emerg Med 2003;10:1070–80.
5. Gilboy N, Tanabe P, Travers DA. The Emergency Severity Index
version 4: changesto ESI level 1 and pediatric fever criteria. J Emerg
Nurs 2005;31:357–62.
6. Manchester Tirage Group. Emergency triage. Harayana: Blackwell
PublishingInc.; 2006 [ref. type: generic].
7. Ludikhuize J, Smorenburg SM, de Rooij SE, de JE. Identification
of deteri-orating patients on general wards; measurement of vital
parameters and potential effectiveness of the Modified Early Warning
Score. J Crit Care 2012;27:424–513.
8. Kause J, Smith G, Prytherch D, Parr M, Flabouris A, Hillman
K. A comparison of antecedents to cardiac arrests, deaths and
emergency intensive care admissions in Australia and New Zea-
land, and the United Kingdom – the ACADEMIA study. Resuscitation
2004;62:275–82.
9. Hogan H, Healey F, Neale G, Thomson R, Vincent C, Black
N. Preventable deathsdue to problems in care in English acute
hospitals: a retrospective case record review study. BMJ Qual Saf
2012;21(9):737–45.
10. Morgan RJM, Williams F, Wright MM. An Early Warning Scoring
System fordetecting developing critical illness. Clin Intensive Care
1997;8 [ref. type:generic].
11. Griffiths JR, Kidney EM. Current use of early warning scores in
UK emergency departments. Emerg Med J 2012;29:65–6.1
2. Smith GB, Prytherch DR, Meredith P, Schmidt PE, Featherstone PI.
The abilityof the National Early Warning Score (NEWS) to discrimi-
nate patients at risk of early cardiac arrest, unanticipated intensive
care unit admission, and death. Resuscitation 2013;84:465–70.
13. Vegting IL, Nanayakkara PW, van Dongen AE, et al. Analysing
completion times in an academic emergency department: coordina-
tion of care is the weakest link. Neth J Med 2011;69:392–8.
14. Royal College of Physicians. National Early Warning Score
(NEWS): standardising the assessment of acute illness severity in
the NHS. Report of a working party. London: RCP; 2012 [ref. type:
generic].
15. Gardner-Thorpe J, Love N, Wrightson J, Walsh S, Keeling N. The
value of Modified Early Warning Score (MEWS) in surgical in-patients: a
prospective observational study. Ann R Coll Surg Eng 2006;88:571-5.
16. Heitz C, Gaillard J, Blumstein H, Case D, Messick C, Miller C.
Performance ofthe maximum modified early warning score to pre-
dict the need for higher care utilization among admitted emergency
department patients. J Hosp Med 2010;5:E46-52.
17. Bokhari S, Munir T, Memon S, Byrne J, Russell N, Beed M. Im-
pact of critical care reconfiguration and track-and-trigger outreach
team intervention on outcomes of haematology patients requiring
intensive care admission. Ann Hematol 2010;89:505-12.
46
18. Subbe CP, Davies R, Williams E, Rutherford P, Gemmell L.
Effect of introducing the Modified Early Warning score on clinical
outcomes, cardio-pulmonary arrests and intensive care utilisation in
acute medical admissions. Anaesthesia 2003;58:797-802.
19. Paterson R, MacLeod DC, Thetford D, et al. Prediction of in-
hospital mortality and length of stay using an early warning scoring
system: clinical audit. ClinMed 2006;6:281–4.
20. Peris A, Zagli G, Maccarrone N, et al. The use of Modified Early
Warning Score (MEWS) may help anaesthesists in post-operative
level of care selection in emergency abdominal surgery. Minerva
Anestesiol 2012;78:1034-8.
21. Patel M, Jones M, Jiggins M, Williams S. Does the use of a
“track and trigger”warning system reduce mortality in trauma pa-
tients? Injury 2011;42:1455-9.
22. Pittard A. Out of our reach? Assessing the impact of introducing
a critical care outreach service. Anaesthesia 2003;58:882-5.
23. Reini K, Fredrikson M, Oscarsson A. The prognostic value of the
Modified Early Warning Score in critically ill patients: a prospective,
observational study. Eur J Anaesthesiol 2012;29:152-7.
24. Armagan E, Yilmaz Y, Olmez O, Simsek G, Gul C. Predictive
value of the modified Early Warning Score in a Turkish emergency
department. Eur J Emerg Med 2008;15:338-40.
25. Carle C, Alexander P, Johal J, Columb M. Design and validation
of an obstetric modified early warning score: secondary analysis of
the Intensive Care National Audit and Research Centre Case Mix
Programme database. Anaesthesia. 2013;68:354-67.
26. Burch VC, Tarr G, Morroni C. Modified early warning score pre-
dicts the need for hospital admission and inhospital mortality. Emerg
Med J 2008;25:674–8.
27. Merz T, Etter R, Mende L, et al. Risk assessment in the first fifteen
minutes: aprospective cohort study of a simple physiological scoring
system in the emergency department. Crit Care 2011;15:R25. Epub
2011 Jan 18.
28. Subbe C, Williams E, Fligelstone L, Gemmell L. Does earlier
detection of critically ill patients on surgical wards lead to better
outcomes? Ann R Coll Surg Engl 2005;87:226-32.
29. Alam N, Hobbelink EL, van Tienhoven AJ, van de Ven PM,
Jansma EP, Nanayakkara PW. The impact of the use of the Early
Warning Score (EWS) on patient outcomes: a systematic review.
Resuscitation 2014;85:587–94.
30. van der Wulp I, Schrijvers AJ, van Stel HF. Predicting admission
and mortality with the Emergency Severity Index and the Manchester
Triage System: a retrospective observational study. Emerg Med J
2009;26:506–9.
31. Martins HM, Cuna LM, Freitas P. Is Manchester (MTS) more than
a triage system? A study of its association with mortality and admis-
sion to a large Portuguese hospital. Emerg Med J 2009;26:183–6.
32. Boeije T, Frederikse MP, van der Heijden FHWM, Mencl F. Ab-
stract: the Emergency Severity Index (version 3): a good predictor
of admission, length of stay andmortality in a European emergency
department. J Emerg Med 2006;30:241;
33. Barfod C, Lauritzen MM, Danker JK, et al. Abnormal vital signs
are strong predictors for intensive care unit admission and inhospital
mortality in adults triagedin the emergency department – a prospec-
tive cohort study. Scand J Trauma Resusc Emerg Med 2012;20:28.
34. Cuthbertson B, Boroujerdi M, Prescott G. The use of combined
physiological parameters in the early recognition of the deteriorating
acute medical patient. J R Coll Physicians Edinb. 2010;40:19-25.
 4

47