/clinical/,/clinical/cckm-tools/,/clinical/cckm-tools/content/,/clinical/cckm-tools/content/questionnaires/,/clinical/cckm-tools/content/questionnaires/related/,

/clinical/cckm-tools/content/questionnaires/related/name-97131-en.cckm

201606168

page

100

UWHC,UWMF,

Tools,

Clinical Hub,UW Health Clinical Tool Search,UW Health Clinical Tool Search,Questionnaires,Related

Validation of a Modified Early Warning Score in Medical Admissions

Validation of a Modified Early Warning Score in Medical Admissions - Clinical Hub, UW Health Clinical Tool Search, UW Health Clinical Tool Search, Questionnaires, Related


Original papers
QJM
Validation of a modified Early Warning Score in
medical admissions
C.P. SUBBE, M. KRUGER
1
,P.RUTHERFORD
2
and L. GEMMEL
1
From the Departments of Medicine, and
1
Critical Care, Wrexham Maelor Hospital, and
2
Department of Nephrology, University of Wales College of Medicine, Wrexham, UK
Received 17 May 2001 and in revised form 9 July 2001
Summary
The Early Warning Score (EWS) is a simple physio-
logical scoring system suitable for bedside applica-
tion. The ability of a modified Early Warning Score
(MEWS) to identify medical patients at risk of
catastrophic deterioration in a busy clinical area
was investigated. In a prospective cohort study,
we applied MEWS to patients admitted to the
56-bed acute Medical Admissions Unit (MAU) of
a District General Hospital (DGH). Data on 709
medical emergency admissions were collected
during March 2000. Main outcome measures
were death, intensive care unit (ICU) admission,
high dependency unit (HDU) admission, cardiac
arrest, survival and hospital discharge at 60 days.
Scores of 5 or more were associated with increased
risk of death (OR 5.4, 95%CI 2.8–10.7), ICU
admission (OR 10.9, 95%CI 2.2–55.6) and HDU
admission (OR 3.3, 95%CI 1.2–9.2). MEWS can
be applied easily in a DGH medical admission
unit, and identifies patients at risk of deteriora-
tion who require increased levels of care in
the HDU or ICU. A clinical pathway could be
created, using nurse practitioners and/or critical
care physicians, to respond to high scores and
intervene with appropriate changes in clinical
management.
Introduction
Catastrophic deterioration of patients in hospital is
frequently preceded by documented deterioration
of physiological parameters.
1,2
Failure of clinical
staff to respond to deterioration of respiratory or
cerebral function and increase levels of medical
intervention will put patients at risk of cardio-
respiratory arrest.
3,4
Inappropriate action in
response to observed abnormal physiological and
biochemical variables might lead to avoidable
death.
5
Suboptimal care prior to admission to a
critical care unit can lead to increased mortality.
6
Because of resource limitations, the number of
patients that can be monitored and treated in
intensive care units (ICUs) and high dependency
units (HDUs) is restricted. The selection of patients
who might benefit from critical care is therefore
crucial. Identifying medical in-patients at risk of
deterioration at an early stage by means of simple
protocols based on physiological parameters may
reduce the number of pre-ICU resuscitations.
7
The Early Warning Score (EWS)
8
is a tool for
bedside evaluation based on five physiological
parameters: systolic blood pressure, pulse rate,
respiratory rate, temperature and AVPU score. The
ability of a modified EWS, including relative
deviation from patients normal blood pressure and
� Association of Physicians 2001
Address correspondence to Dr C.P. Subbe, 77 Brook Lane, Chester CH2 2EE
Q J Med 2001; 94:521–526

urine output, to identify surgical patients who
would potentially benefit from intensive care has
been recently demonstrated.
9
None of the existing
physiological scoring systems has been validated in
patients admitted on an unselected medical take.
The aims of this study were: (i) to evaluate
the ability of a modified EWS (MEWS, Table 1) to
identify medical patients at risk; and (ii) to examine
the feasibility of MEWS as a screening tool to
trigger early assessment and admission to an HDU
or ICU.
Methods
Data were collected for all medical emergency
admissions admitted to the Medical Admissions
Unit (MAU) of a District General Hospital (DGH)
during March 2000. Patients admitted directly to
Coronary Care, Medical HDU or ICU, and patients
re-admitted during the observation period were not
included in this study.
After appropriate training, nursing staff collected
data while performing routine duties. Demographic
details, systolic blood pressure, pulse rate, temper-
ature, respiratory rate and AVPU score (A for ‘alert’,
V for ‘reacting to vocal stimuli’, P for ‘reacting
to pain’, U for ‘unconscious’) were recorded on
admission. Blood pressure and pulse rate were
measured electronically (DINAMAP, Critikon) and
checked manually where appropriate. The tem-
perature was taken orally (Temp-PlusII, IVAC). The
respiratory rate was counted over a full minute.
AVPU scores were scored according to best
response at time of blood pressure measurement.
Nursing staff collected physiological parameters
twice daily (once am and once pm) on a dedicated
data collection sheet for up to 5 days. Complete-
ness of data was checked daily at the bedside by
two of the investigators (MK, CS).
The collected data were used to calculate a
modified Early Warning Score (MEWS) (Table 1). It
was decided from previous experience to define
a MEWS of five or more as a ‘critical score’. The
highest score reached during admission was
labelled ‘ScoreMax’.
Primary endpoints were HDU admission, ICU
admission, attendance of the cardiac arrest team
at a cardiorespiratory emergency and death at
60 days.
HDU and ICU admission were at the discretion
of the attending physicians, who were unaware of
the MEWS of the patient.
Statistics were generated using SPSS (version
10.0). Relative risk ratios were calculated by using
cross-tabulation of results. We regarded p-0.05
as statistically significant.
Results
During the observation period, data were collected
from 709 admissions. Patients with incomplete
epidemiological or discharge data were excluded.
At least one complete set of physiological data
during the first 24 h of admission, epidemiological
data and discharge dates were available for 673
admissions. These were analysed further.
Overall, 45% of patients were male, and 55%
were female. The mean age of patients was 63 years
(SD 20, range 16–100).
During follow-up, seven patients were admitted
to ICU, 23 to HDU, four were resuscitated by the
cardiopulmonary arrest team and 56 patients died.
Physiological parameters on admission
Mean values for measured parameters on admis-
sion were as follows: blood pressure (systolic)
139 mmHg (SD 27 mmHg), pulse rate 86 bpm
(SD 20 bpm), respiratory rate 20 bpm (SD 5 bpm),
temperature 36.6 8C (SD 0.9 8C).
The majority of patients scored 0 on admission
for blood pressure (91%), pulse rate (78%),
temperature (95%) and AVPU score (92%).
Median score for respiratory rate was 1 (55% of
admissions).
Admission scores ranged from 0 to 9 (median
of 1). The number of patients with critical scores
Table 1 Modified Early Warning Score
32 1 0 1 2 3
Systolic Blood pressure
(mmHg)
-70 71–80 81–100 101–199 0200
Heart rate (bpm) -40 41–50 51–100 101–110 111–129 0130
Respiratory rate (bpm) -9 9–14 15–20 21–29 030
Temperature (8C) -35 35–38.4 038.5
AVPU score Alert Reacting to
Voice
Reacting to
Pain
Unresponsive
C.P. Subbe et al.522

()4) was greatest on the day of admission and
gradually decreased over the period of stay from
7.1% on admission to 4.8% on Day 1, 3.9% on
Day 2 and 1.8% on Day 3. In the 81 patients that
stayed in the MAU for a minimum of 3 days, scores
stayed unchanged for 42, deteriorated in 12 and
improved in 28 patients. During the observation
period the mean of the highest score reached was
2.29 (SD 1.51) (Figure 1).
Outcome
A ScoreMax of 5 or more was associated with an
increased risk of death (OR 5.4, 95%CI 2.8–10.7),
ICU admission (OR 10.9, 95%CI 2.2–55.6) and
HDU admission (OR 3.3, 95%CI 1.2–9.2). End-
points happened at a median of 4 days (0–45 days)
after transfer from the MAU; 22 of the endpoints
were reached while patients were in the MAU.
Endpoints were reached by 7.9% of patients
with ScoreMax of 0–2, 12.7% of patients with a
ScoreMax of 3–4 and 30% of patients with a
ScoreMax of 5–9. Figure 2 shows the frequency of
endpoints according to EWS on admission. Patients
who reached predefined endpoints were signific-
antly older and on admission had lower systolic
blood pressure, higher pulse rate and a higher
respiratory rate (Table 2).
Observations on admission and relative
risk of reaching predefined endpoints
Whereas high EWS were associated with increased
risk to reach endpoints, increased scores for single
parameters did not always translate into an
increased overall risk. Table 3 shows relative risk
Figure 1. Distribution of maximum scores (ScoreMax)
expressed as percentage of all ScoreMax.
Figure 2. Frequency distribution of admission EWS (0-9) and clinical outcomes and combined outcome which aggregates
death, cardiopulmonary resuscitation, HDU admission and ICU admission.
Table 2 Physiological parameters on admission of patients reaching or not reaching endpoints
Endpoint not reached Endpoint reached p
n 598 75
Age (years) 62q/�20 74q/�14 -0.0001
Systolic blood pressure (mmHg) 140q/�30 127q/�127 (0.0001
Pulse Rate (bpm) 86q/�19 92q/�23 (0.03
Respiratory Rate (bpm) 20q/�423q/�7 -0.002
Temperature (8C) 36.7q/�0.9 36.5q/�1 0.06
Data presented as means"SD, p value for independent sample t-test.
Modified Early Warning Score 523

ratios of increased scores, compared with a
score of 0.
Systolic blood pressure scores were not asso-
ciated with significantly increased risk to reach
endpoints (Table 3). If the model for generat-
ing blood pressure scores was altered, endpoints
were more frequently reached in patients with a
systolic blood pressure of -100 mmHg than with
100–140 mmHg (OR 0.38, 95%CI 0.17–0.87).
Incidents were even less frequent in patients with
a systolic blood pressure of)140 mmHg (OR 0.49,
95%CI 0.28–0.86). Similarly, high scores related
to raised temperature were not associated with
increased risk but those related to low temperature
were significant (Table 3).
Patients aged)70 years were significantly more
at risk to reach endpoints than patients aged
-50 years (OR 6.5, 95%CI 2.5–16.7). It is possible
to analyse the effect of adding an age score to
EWS by using Receiver Operator Characteristic
curves. By adding an age score, the area under the
curve increased from 0.67 to 0.72 for scores on
admission and combined endpoints (Figure 3).
Discussion
The Modified Early Warning Score is best regarded
as a defined judgement on routinely recorded
physiological data. Using previously published scor-
ing criteria,
8,9
this study has demonstrated that
raised MEWS scores are associated with increased
mortality in a group of medical emergency admis-
sions. Calculation of the MEWS for emergency
admissions might be useful in triage, to identify
patients of highest risk of deterioration. Appropriate
interventions could then be targeted upon a small
number of patients among the 30–40 daily
admissions within the unselected medical take.
Our study is limited by several factors. It is a
single-centre study on a limited number of patients
in a specific local setting. For technical reasons,
we were unable to collect data for longer then
5 days. The majority of patients who were admitted
to critical care areas or died will have had improve-
ments and deteriorations following transfer out of
the MAU. Physiological data leading up to those
events would probably give additional information
of the physiology prior to catastrophic events.
There are few previous data concerning other
scoring systems and patients admitted via a general
medical ‘take’. The Acute Physiology and Chronic
Health Evaluation (APACHE) II Score
10
and
Mortality Prediction Model (MDM)
11
have only
Table 3 Relative risk ratios (RR) for patients with scores of 1,2 and 3 on admission, compared to patients with a score of 0
3 21012 3
Systolic blood
pressure
-70 71–80 81–100 101–199 0200
RR (95%CI) 8.6 (0.5–139) 5.7 (0.9–35) 2.1 (0.8–5.5) 0.5 (0.7–4.1)
Heart rate -40 41–50 51–100 101–110 111–129 0130
RR (95%CI) NA NA 1.6 (0.7–3.2) 1.5 (0.7–3.4) 3.0 (0.9– 9.5)
Respiratory rate -9 9–14 15–20 21–29 030
RR (95%CI) NA 1.6 (0.4–7.2) 4.4 (1.0–19) 7.9 (1.5–42)
Temperature -35 35–38.4 038.5
RR (95%CI) 5.9 (1.8–19) 0.9 (0.2–3.8)
AVPU score Alert Reacting to
Voice
Reacting to
Pain
Unresponsive
RR (95%CI) 2.0 (0.9–4.8) 5.2 (1.5–18.1) NA
NA, not applicable for scores with insufficient data.
Figure 3. Receiver Operator Characteristic curves for
modified EWS and modified EWS with age score
(0 points for -50 years old, 2 points for 50–70 years
old and 3 points for )70 years old).
C.P. Subbe et al.524

been tested for subgroups of medical patients
with acute renal and congestive heart failure.
12,13
The Simplified Acute Physiology Score (SAPS) was
introduced in 1984 to estimate the risk of death for
patients in intensive care,
14
and has since been
improved
15
and tested in patients with myocardial
infarction.
16–19
A reduced version (SAPS.R) has
been shown to predict outcome accurately in ICU
patients but has not been applied to general
medical patients.
20
None of the available scoring systems appears
to be suitable for bedside assessment of ward
patients in a routine fashion. MEWS is likely to
present a more versatile tool in this context, since it
simply collates the results of routinely collected
variables.
Stenhouse
9
suggests use of a blood pressure
score comparing actual blood pressure with pre-
viously measured pressures judged to be ‘normal’
for the patient. We felt that in the emergency
situation, previous recordings would often not be
available and that the calculation of the score
would lose simplicity. Oxygen saturation is often
recorded at the bedside, but can be misleading if
reviewed outside the context of inspiratory oxygen
concentration.
High scores were more likely to occur early
during admission and falling MEWSs in most
patients over time suggests beneficial effects of
treatment. MEWS could therefore act as another
method of assessing the efficiency of medical
interventions.
We did not exclude patients with ‘Do not
attempt resuscitation’ orders from the study, as we
believe that a sensible discussion of possible out-
comes with the patient and relatives should be
part of the management of critically ill patients.
MEWS might be helpful in identifying some of
the patients in whom this discussion would be
indicated.
Examination of the admission data demonstrates
that the banding of values for each parameter as
described from previous studies on surgical patients
may not be generalizable to the patients admitted
on an unselected reception. While the MEWS used
in this study did identify some of the patients at
increased risk, alterations in the scoring table
(e.g. different blood pressure values, respiratory
rate) might improve the fit of the score to clinical
outcomes.
Case mix may be one important factor, and the
score may be more predictive if different banding
criteria are used for broad ranges of conditions
e.g. respiratory disease, cardiac disease. Integration
of the diagnosis into a scoring system might
diminish the utility of the score by making it too
complex.
To prioritize often scarce ICU resources, it
would be valuable to identify patients who would
benefit from ICU admission, as well as those whose
ICU admission could be prevented by changes
in management on the ward. As patients with
critical scores ()4) in this study were at increased
risk of catastrophic deterioration, MEWS might be
a helpful screening tool to triage patients for
intensified treatment on the ward or in HDU or ICU.
It remains to be seen whether identification of
critically ill patients, selection of a higher level of
care and consecutive changes in management
affect outcome in this patient group.
In conclusion, the MEWS is a simple bedside
tool that can be calculated by nursing staff in a
busy clinical area. In the setting of the unselected
medical ‘take’, it might help to identify some of
those patients at risk of deterioration and need
for more active intervention. A prospective multi-
centre study, expanding the case-mix in more
detail, is needed to evaluate the effects of
increased medical intervention in patients with a
high MEWS.
Acknowledgements
This study was supported by a grant of the
North East Wales NHS Trust.
References
1. Sax FL, Charlson ME. Medical patients at high risk for
catastrophic deterioration. Crit Care Med 1987; 15:510–15.
2. Smith AF, Wood J. Can some in-hospital cardio-respiratory
arrests be prevented? A prospective survey. Resuscitation
1998; 37:133–7.
3. Schein RMH, Hazday N, Pena M, Ruben BH, Sprung CL.
Clinical antecedents to in-hospital cardiopulmonary arrest.
Chest 1990; 98:1388–92.
4. Franklin C, Mathew J. Developing strategies to prevent
inhospital cardiac arrest: analysing responses of physicians
and nurses in the hours before the event. Crit Care Med
1994; 22:244–7.
5. McGloin H, Adam SK, Singer M. Unexpected deaths and
referrals to intensive care of patients on general wards: are
some potentially avoidable? J R Coll Physicians Lond 1999;
33:255–9.
6. McQuillan P, Pilkington S, Allan A, Taylor B, Short A,
Morgan G, Nielsen M, Barrett D, Smith G. Confidential
inquiry into quality of care before admission to intensive
care. Br Med J 1998; 316:1853–8.
7. Goldhill DR, Worthington L, Mulcahy A, Tarling M,
Sumner A. The patient-at-risk team: identifying and
managing seriously ill ward patients. Anaesthesia 1999;
54:853–60.
8. Morgan RJM, Williams F, Wright MM. An Early Warning
Scoring System for detecting developing critical illness.
Clin Intens Care 1997; 8:100.
Modified Early Warning Score 525

9. Stenhouse C, Coates S, Tivey M, Allsop P, Parker T.
Prospective evaluation of a Modified Early Warning Score
to aid earlier detection of patients developing critical
illness on a general surgical ward. State of the Art Meeting,
Intensive Care Society, London, December 1999.
10. Knaus WA, Draper EA, Wagner DP, Zimmerman JE.
APACHE II: a severity of disease classification system.
Crit Care Med 1985; 13:818–29.
11. Lemeshow S, Teres D, Pastides H, Avrunin JS, Steingrub JS.
A method for predicting survival and mortality of ICU
patients using objectively derived weights. Crit Care Med
1985; 13:520–5.
12. Fiaccadori E, Maggiore U, Lombardi M, Leonardi S,
Rotelli C, Borghetti A. Predicting patient outcome from
acute renal failure comparing three general severity of
illness scoring systems. Kidney Int 2000; 58:283–92.
13. Poses RM, McClish DK, Smith WR, Huber EC, Clemo FL,
Schmitt BP, Alexander D, Racht EM, Colenda CC. Results of
report cards for patients with congestive heart failure depend
on the method used to adjust for severity. Ann Intern Med
2000; 133:10–20.
14. Le Gall JR, Loirat P, Alperovitch A, Glaser P, Granthil C.
A simplified acute physiology score for ICU patients.
Crit Care Med 1984; 12:975–7.
15. Le Gall JR, Lemeshow S, Saulnier F. A New Simplified
Acute Physiology Score (SAPS II) based on a European/
North American Multicenter Study. JAMA 1993;
270:2957–63.
16. Moreau R, Soupison T, Vauquelin P, Derrida S, Beaucour H,
Sicot C. Comparison of two simplified severity scores
(SAPS and APACHE II) for patients with acute myocardial
infarction. Crit Care Med 1989; 17:409–13.
17. Schuster HP, Schuster FP, Ritschel P, Wilts S, Bodman KF.
The ability of the Simplified Acute Physiology Score (SAPS II)
to predict outcome in coronary care patients. Intensive Care
Med 1997; 23:1056–61.
18. Sarmiento X, Rue M, Guardiola JJ, Toboso JM, Soler M,
Artigas A. Assessment of prognosis of coronary patients:
performance and customization of generic severity indexes.
Chest 1997; 111:1666–71.
19. Reina A, Vazquez G, Aguayo E, Bravo I, Colmenero M,
Bravo M. Mortality discrimination in acute myocardial
infarction: comparison between APACHE III and SAPS II
prognosis systems. PAEEC Group. Intensive Care Med 1997;
23:326–30.
20. Viviand X, Gouvernet J, Granthil C, Francois G.
Simplification of the SAPS by selecting independent
variables. Intensive Care Med 1991; 17:164–8
C.P. Subbe et al.526