116 V O L U M E 6 I S S U E 4 2 0 1 0 infant

E A R LY W A R N I N G S C O R E S © 2010 SNL All rights reserved

It is generally accepted in adult andpaediatric practice, that prior to acute

deterioration and subsequent transfer tointensive care, patients often show signs ofdeterioration which are eitherunrecognised or not acted upon by nursingand medical staff1,2. Early warning scoresbased on physiological observations (heartand respiratory rate etc) whichautomatically trigger medical review havebeen validated as useful ways of detectingdeterioration and prompting interventionto reduce morbidity in both adult3 andpaediatric4 populations.

To our knowledge, no such tools havebeen developed or fully evaluated in thenewborn population. A Medline andEmbase search found no studies directlyrelated to newborn infants. One reason forthis may be the lack of well establishednormal ranges for biophysical variables.Published studies are sparse and not solelyconfined to the perinatal period5-7. Even thestandard textbooks have differences

The Newborn Early Warning (NEW)system: development of an at-risk infantintervention systemThe use of early warning systems is widespread but their use in the neonatal age group hasbeen under-investigated. This article describes the development of a Newborn Early Warning‘traffic-light’ coded observation chart to enable early identification of adverse changes inphysiological parameters. Much work remains to be done but the aim of this initial project is toallow other maternity units to consider how they can improve the safety of at-risk newborninfants on their postnatal wards.

Damian Roland BMed Sci, BMBS, MRCPCHAcademic Clinical Fellow in PaediatricEmergency Medicine, Leicester University

John MadarMRCP, FRCPCH, FHEAConsultant Neonatologist, Neonatal Unit,Derriford Hospital, Plymouthjohn.madar@phnt.swest.nhs.uk

Glenys ConnollyBsc(Hons), RGN, RSCNAdvanced Neonatal Nurse Practitioner,Neonatal Unit, Derriford Hospital, Plymouth

Keywords

neonatal scoring systems; riskstratification; observation chart

Key points

D. Roland, J. Madar, G. Connolly. TheNewborn Early Warning (NEW) system:devel-opment of an at-risk infantintervention system. Infant 2010; 6(4): 116-20.1. The NEW observation chart facilitates

observation of babies deemed at riskand prompts earlier review in thosedemonstrating clinical deterioration.

2. There was an increase in retrievableobservations from 48% in theretrospective audit to 72% in theprospective audit.

3. The NEW chart threshold criteriaprompted management decisions innine (47.3%) of 19 infants who requiredintervention.

4. The chart was considered beneficial bya majority of midwives questionedabout its use.

between chapters in the same book, whichmay result in different clinical approaches(TABLE 1). The absence of, or variance inpublished normal values illustrates adifficulty in establishing responseparameters for newborns who requireobservation. Early warning criteria shouldnot be so brittle as to be over sensitive andthus devalue the tool.

The majority of newborn infants arehealthy and not at risk of significantmorbidity. A second group are clearlyunwell or compromised and declarethemselves as justifying enhanced levels ofcare. Between these are those well babieswhose perinatal circumstances identifythem as at risk of potentially significantmorbidity. These include, for example,those babies at risk of infection throughstreptococcal carriage, or prolongedrupture of membranes, or those babiesborn through meconium. In addition arethose manifesting behaviour slightly out ofthe normal range, but not so far as to

TABLE 1 Normal ranges for newborn infant’s heart rate and respiratory rate as published instandard paediatric texts.

Source Heart rate bpm Respiratory rate

Examination of the Newborn and Neonatal Health. 110-160: 80-90 if 40-60 A multidimensional approach. Ed Lorna Davies, asleep, 160+ if non-distressedSharon McDonald8 distressed

Examination of the Newborn. A Practical Guide. 90-140 - resting 40-60 Helen Baston, Heather Durward9 breaths/min

Roberton’s Textbook of Neonatology. 120-160 usually 35-45Ed Janet M Rennie10

Avery's Diseases of the Newborn. 40-50 newborn, Taesch, Ballard, Gleason11 35-60 thereafter

Advanced Paediatric Life Support – manual12 110-160 30-40

E A R LY W A R N I N G S C O R E S

V O L U M E 6 I S S U E 4 2 0 1 0 117infant

overtly identify them as clearly unwell.Such babies fall into a group where theability to generate a series of structuredobservations with evolving trends inphysiological parameters permits staff ofvarying experience to more clearly deter-mine health, stability and the potentialneed for further intervention. The rapiditywith which a baby can become unwelldrives a need for clear pointers towardsmore aggressive intervention. The earlyidentification of an unwell infant may, forexample, prompt attention to airway orbreathing support, or the earlyadministration of antibiotics and preventsignificant morbidity and even mortality.

The aims of this study were:■ To categorise observations on newborn

infants in order to formulate prompts forassessment/intervention – the ‘earlywarning score’.

■ To develop a recording tool for observa-tions to help generate such a score andprompt appropriate action – theNewborn Early Warning (NEW) chart

■ To assess the chart’s effectiveness in clinical practice

Materials and methodsTwo studies were carried out:■ A retrospective review of observations on

babies admitted to the neonatal unit tocompare key observations with proposedearly warning criteria and determinewhether assessment against these criteriawould have altered management.

■ A prospective study of at-risk babiesobserved using the NEW chart to deter-mine effectiveness of the chart as a clini-cal tool.Derriford hospital is a network neonatal

intensive care unit (previously termed alevel 3 unit) within the Peninsula Neonatalnetwork with around 4,400 deliveries ayear. Babies are looked after on thepostnatal wards, but can be admitted to a15 bedded transitional care ward (TCW)with their mothers (around 900admissions/year) or to the neonatalintensive care unit (NICU) if more unwell

maternal health characteristics, markers ofperinatal stress (eg poor cord pH values)or the need for significant resuscitation atbirth (TABLE 2). These indicators are basedon well established physiological principleswith an incomplete evidence base to backthem up13,14.

Retrospective review

Using the NICU admission records themedical notes of term infants over 2.5kgwho presented to the neonatal unit fromeither the postnatal wards or thetransitional care ward over a two yearperiod were identified. These notes wereexamined for general demographic data,whether the infant had been correctly

(around 450 admissions/year). Internal guidelines include those for the

identification of so called “At-RiskNewborn Infants” or ‘ARNIs’. These areinfants who are deemed to be at increasedrisk of postnatal morbidity by virtue ofpre-identified indicators such as adverse

TABLE 2 At-Risk Newborn Infant (ARNI) criteria.

FIGURE 1 Revised Newborn Early Warning Observation Chart containing some sample entries.

Prenatal Perinatal Postnatal

AT RISK INFANT OBSERVATION CHART

Neonatal Early Warning (NEW)

Indicate any associated features/symptoms present (CNS or airway) using letters.

All observations in Green No action. Continue four hourly observations.

One observation in Yellow Contact neonatal team or senior midwife. Verbal management plan or review to be implemented. Repeat observations in 30 mins.

Two observations in Yellow or One in Red Immediate review required.

Seizures, apnoeas or obvious cyanosis Immediate review required.

Date

Time

Airway/Breathing(RR

(.)HR

(x)

Temp

(o)

CNS

85 200 38.2

O2

Sats

<

90%

80 190 38.0

75 180 37.8

70 170 37.6

65 160 37.4Gru

ntin

g (G

)or

O2

Sats

90-

94%

Irri

tabl

e (I

)

60 150 37.2

55 140 37.0

50 130 36.8

45 120 36.6

40 110 36.4

35 100 36.2

Pin

k

Wak

es t

o fe

ed

30 90 36.0

25 80 35.8

Jitt

ery

(J)

20 70 35.6Dusky

(D) 60

AffixLabel Here

G

Pathologicalcardiotocograph

Scalp pH<7

Group B Streptococcus risk

Premature rupture ofmembranes (PROM)

Thick meconium

Venous cord pH <7.1

Ventilatory support >3minutes

Five minute APGAR <8

Grunting

Abnormal movements

Any ongoing concerns

At the request of reviewingmedical or neonatal staff

identified as an ARNI at birth (TABLE 2) andwhether observations had been recorded.

A pilot NEW observation chart wasdeveloped providing prompts to aid in theidentification of ARNIs and permit therecording of the observed physiologicalvariables of these infants using symbols,highlighting values of concern. The chartwas approved by the Hospital ClinicalRecords and Knowledge Service committee.As well as physiological observations suchas temperature, pulse and respiratory rate,comments about the infant’s work ofbreathing or conscious level wereaccommodated. Observation values wereclassified into red (significantly abnormal),amber (abnormal) or green (normal)ranges. The values used were an amalgamof those found in standard neonataltextbooks selected to ensure chart scaleswere not unwieldy. Values in the chart’samber band were in keeping with the upperrange of normal physiologicalmeasurements.

Clinical observations from the group ofARNIs were then plotted on the NEWchart to see whether the pre-identifiedtrigger criteria would have promptedearlier medical review.

Based on the results of this retrospectiveaudit a revised chart was generated for thesubsequent prospective study withmodified trigger values (FIGURE 1).

Prospective study

The results of the retrospective review wereused to inform an educational programmeincluding presentations and writtenmaterial. It was aimed at midwifery,nursing and medical staff in the maternityunit and designed to raise awareness of the

NEW programme, familiarise staff with theNEW chart and the structure of theproposed study.

NEW charts were made available on thelabour suite and postnatal wards. Thecriteria for using the NEW charts weredisseminated among the midwives andposters highlighting the process placedwidely around the obstetric and neonataldepartment. Any child who was on a NEWchart had their observations recorded fourhourly or more frequently if deemednecessary.

Babies were excluded from the study ifthey were admitted directly to the NICU/TCW or fulfilled automatic admissioncriteria such as being <37 weeks’ gestationor <2.5kg.

All NEW charts had an envelope attachedso brief details of the infant could be sent tothe study administrator as soon asobservations were commenced. All infants’notes were collated when the study wascompleted. Ethical approval was granted bythe local relevant ethical committee.

An intervention was defined as an infantreceiving an investigation (blood test orCXR), treatment (antibiotics) or transferto another care environment.

A questionnaire was sent to all midwivesto obtain qualitative data on their thoughtson the process.

Results

Retrospective review

The initial audit identified 122 terminfants, 51% of these infants fulfilled ARNIcriteria. Eighty-four per cent were correctlyidentified as such (TABLE 4). Only 48%(25/52) of those infants recognised asbeing ARNIs had observations recorded,

but half would have been reviewed earlier(13/25) by a neonatal doctor or nursepractitioner if their observations had beencharted on the NEW chart. Of the babiesadmitted not classified as ARNIs, few hadobservations recorded (5/55 – 8%). Thisaudit was of infants admitted to the NICUand does not contain data on those infantswho were safely discharged home. Basedon this data the decision to conduct aprospective study was made.

Prospective studyOver a three month period informationwas collected on 117 infants who had been

E A R LY W A R N I N G S C O R E S

118 V O L U M E 6 I S S U E 4 2 0 1 0 infant

Total infants122

Fulfil ARNI criteria62 (51%)

Recognised at the time52 (84%)

Not recognised10 (16%)

Observations recorded25

NEW triggers activation13

NEW chart no action12

Observations not recorded27

Not ARNI criteria60 (49%)

Observations recorded5 (8%)

NEW triggers activation4

NEW chart no action1

Observations not recorded55 (92%)

TABLE 3 Details of term babies admitted to the NNU/TCW from postnatal wards.

TABLE 4 At-Risk Newborn Infant (ARNI)criteria for enrolled infants: prospective study.

Reason Totals

PrenatalCTG 9Scalp pH<7 0GBS 6PROM 29

PostnatalMeconium 15Cord pH <7.1 2Ventilatory support 1APGAR <8 1

PostnatalGrunting 14Abnormal movements 0Concern 15Request 0

Unclear 9

TCW (child admitted directly to the transitional care ward because of gestational age) 15

Other (infant readmitted at five days of age) 1

Total 117

E A R LY W A R N I N G S C O R E S

V O L U M E 6 I S S U E 4 2 0 1 0 119infant

NEWBORN EARLY WARNING

Surname

First Name

Hospital No.

NHS Number

D.O.B. Affix patient label hereOBSERVATION CHART FOR NEWBORN INFANTS

DATE TIME

39.0

38.0

37.0

36.0

TE

MP

35.0

85

80

75

70

65

60

55

50

45

40

35

30

25

RE

SP

IRA

TIO

N

20

GRUNTING

200

190

180

170

160

150

140

130

120

110

100

90

80

70

HE

AR

T R

AT

E

60

PINK (>94%)

90-94% COLOUR

(SpO2)DUSKY/BLUE (<90%)

ACTIVE/WAKES TO FEED

JITTERY/IRRITABLE

FLOPPY/DIFF TO ROUSE NEURO

SEIZURES

REDSCORE

AMBER

ALL OBSERVATIONS IN WHITE CONTINUE OBERVATIONS 4 HOURLY OR AS REQUESTED.

ONE IN AMBER CONTACT SHO/ANNP/SENIOR MIDWIFE. VERBAL MANAGEMENT PLAN OR REVIEW. REPEAT OBSERVATIONS IN 30 MINUTES. RESPONSE

TWO IN AMBER OR ONE IN RED IMMEDIATE REVIEW

© PHNT – NEONATAL UNIT 2009 File alongside other observation charts VERSION 6 – March 2010

FIGURE 2 Final version of the NEW observation chart separating out each variable to improve clarity.

recognised as being ARNIs. Based on anaverage of 4,600 deliveries per year,approximately 10% (468/4600) ofdeliveries at Derriford hospital result in anARNI being born. The breakdown of thespecific criteria for this are shown in TABLE

4. Of 117 identified, only 84 charts wereavailable for review (71.2%). Nineteeninfants received an intervention as per thepredefined criteria and in nine thisoccurred as a result of the NEW chart. One infant was admitted to the NICUdirectly from the postnatal wards whodeveloped ABO incompatibility on day 2of life. A chart had been provided for thisinfant although the reasons for this areunclear. The chart did not affect theinfant’s management.

A sample of midwives’ views on theNEW system were obtained viaquestionnaire. Notable responses included:■ A majority felt the chart was beneficial.■ Many commented that the chart made

them more aware of the normal parame-ters for a newborn.

■ Around half felt the chart was overcom-plicated and suggested changes. It was felta different style of might be easier tointerpret.

ConclusionOur study indicated that many infantsachieved ‘at risk’ criteria, often promptingintervention in terms of investigations,anti-microbial management or transfer toa higher level of care. It is important robustprocedures are instituted to avoidunnecessary morbidity and perhapsmortality through inadvertent delay. Thebenefits of early identification of instabilityand of necessary intervention are obviousand an early warning chart with clearprompts for action is one tool forfacilitating this. Our locally designed andimplemented chart appears to have hadsome success in identifying infants at anearlier stage than would have occurred intheir absence. The chart itself may havebeen the arbiter of the increased detectionrate, but the very exercise of introducingthe charts, and the educational packagesurrounding this may also have had aneffect in raising awareness.

The true effect of earlier detection onlonger term morbidity and mortality isdifficult to define with the small numbersof babies involved in this study. However,intuitively, earlier management might beconsidered a positive outcome, unlessprompting unnecessary investigations and

interventions on babies who were deemedunstable by virtue of transgressing thepredefined criteria. On a pragmatic basisthe chart identified nearly 50% of thoseinfants where intervention was deemedclinically appropriate. No direct feedbackwas given about the chart producingunnecessary intervention apart from thedifficulties with the temperature scale.Ultimately however, it is not the chart, orthe highlighting of a set of observationsthat should prompt intervention, but thefull clinical evaluation of the baby thatsubsequently follows. The ability to clearlyassess trends in observations may form animportant part of that evaluation and isone of the attributes of the observationchart. The NEW chart itself is but onecomponent of a system of care and cannotfunction effectively without the otherelements. Having adequate numbers ofstaff able to undertake accurateobservations is a pre-requisite, with cleararrangements for subsequentcommunication of concern and an abilityto respond effectively to those concerns.

Also of note is the fact that direct entrymidwifery students may have had verylimited exposure to or training in the careof the newborn baby and little on therecognition of the unwell infant. Hardpressed staff on labour ward and postnatalwards need effective tools to help them inthe identification and observation of thesevulnerable babies.

It is vital to address any staff reservationsabout the format of the chart. In theoriginal version, the temperature scale wasfelt to be over sensitive, prompting reviewand potential intervention whenunnecessary. The format of the chart withdifferent symbols for each variable was alsofelt to confuse and produce anovercrowded display which was difficult toread. These problems were exacerbated bystaff using poor quality photocopies of theoriginal chart, rather than high qualityreproductions. Budgetary constraints alsocompromised the original charts by the useof grey scale rather than colour banding.

Further work and greater numbers areneeded in order to evolve a working modelwhich is acceptable to all staff andvalidation of the results in a differentclinical setting should take place. As aresult of feedback a further version of thechart has been designed which in pilottesting has proved more popular withmidwifery staff (FIGURE 2). This chart isbased on an obstetric early warning system

from Liverpool developed as a result of theConfidential Enquiry of Maternal andChildhood Health review. A similar chartis being used at the Royal Free inHampstead, UK (personal communicationVivienne van Someren, 2009). This chartseparates out the clinical variables,arguably making it easier to determineindividual trends. No single chart is likelyto cover the needs of all units, butestablishing the principle and providing aneffective template may help others developsimilar tools.

The NEW observation chart is but onecomponent of the systems that need to bein place to ensure optimal care for thesebabies. This work has demonstrated thatsuch charts can help those looking aftersuch babies target at risk newborn infantsmore effectively.

References1. Franklin C., Mathew J. Developing strategies to

prevent inhospital cardiac arrest: analyzing

responses of physicians and nurses in the hours

before the event. Crit Care Med 1994; 22: 244-47.

2. Schein R.M., Hazday N., Pena M. et al. Clinical

antecedents to in-hospital cardiopulmonary arrest.

Chest 1990; 98: 1388-92.

3. Stubbe C.P., Kruger M., Rutherford P., Gemmell L.

Validation of a modified Early Warning Score in

medical admissions. Q J Med 2001; 94: 521-26.

4. Duncan H., Hutchison J., Parshuram C. The pediatric

early warning system score: A severity of illness

score to predict urgent medical need in hospitalized

children. J Crit Care 2006; 21(3): 271-78.

5. Montague T., Taylor P., Stockton R., Roy D., Smith E.

The spectrum of cardiac rate and rhythm in normal

newborns. Pediatr Cardiol 1982; 2: 33-38.

6. Rusconi F., Castagneto M., Gagliardi L., Leo G. et al.

Reference values for respiratory rate in the first 3

years of life. Pediatrics 1994; 94: 350-55.

7. Hooker E., Danzl D., Brueggmeyer M., Harper E.

Respiratory rates in pediatric emergency patients. J

Emerg Med 1992; 10: 407-10.

8. Davies L., McDonald S., eds. Examination of the

Newborn and Neonatal Health. A multidimensional

approach. Churchill Livingstone/Elsevier. 2008.

9. Baston H., Durward H. Examination of the

Newborn. A Practical Guide. Routledge. 2001.

10. Rennie J.M., ed. Roberton’s Textbook of

Neonatology. Fourth Edition 2005. Churchill

Livingstone/Elsevier. 2005.

11. Taesch H.W., Ballard R., Gleason C.A. Avery’s

Diseases of the Newborn 8th edition. Elsevier

Saunders. 2004.

12. Mackway-Jones, Molyneux E., Phillips B., Wieteska S.,

eds. Advanced Paediatric Life Support: The Practical

Approach 4th Edition. Blackwell Publishing. 2005. 7-14.

13. Madar J. Clinical risk management in the newborn

and neonatal resuscitation. Semin Fetal Neonatal

Med 2005; 10: 45-61.

14. Victory R., Penava D., Da Silva O., Natale R.,

Richardson B. Umbilical cord pH and base excess

values in relation to advers outcome events for

infants delivering at term. Am J Obstet Gynaecol

2004; 191: 2021-28.

E A R LY W A R N I N G S C O R E S

120 V O L U M E 6 I S S U E 4 2 0 1 0 infant