Clinical Nutrition 30 (2011) 430e435

Contents lists available at ScienceDirect

Clinical Nutrition
journal homepage: http://www.elsevier.com/locate/clnu

Original Article

Performance of the novel Paediatric Yorkhill Malnutrition Score (PYMS)

in hospital practiceq
Konstantinos Gerasimidis a, b, *, Isobel Macleod a, Anne Maclean a, Elaine Buchanan a, Paraic McGrogan a,
Isabel Swinbank a, Mary McAuley c, Charlotte M. Wright b, Diana M. Flynn a
a
Women and Children Directorate, NHS Greater Glasgow and Clyde, Glasgow, UK
b
Human Nutrition Section, Division of Developmental Medicine, University of Glasgow, Royal Hospital for Sick Children, Glasgow G3 8SJ, UK
c
Children and Young Combined Health, NHS Grampian, Aberdeen, UK

a r t i c l e i n f o s u m m a r y

Article history: Background & aims: Nutritional screening in paediatric inpatients is important. However, there is a lack of
Received 21 July 2010 validated screening tools for this population. In this study the development of a nurse administered
Accepted 31 January 2011 paediatric malnutrition screening tool is described and its performance evaluated.
Methods: The Paediatric Yorkhill Malnutrition Score (PYMS) rate BMI, weight loss, dietary intake and
Keywords: predicted effect of the current condition on nutritional status, with a score of 0e2 for each element.
Nutritional screening tool
Patients with total score of 2 or more are referred for dietetic review. A four month pilot phase was
Malnutrition
conducted in three medical and one surgical wards of a tertiary hospital and the general paediatric ward
of a district general hospital. Performance of the tool was assessed by auditing completion rates, yield,
impact on dietetic workload, and by evaluating dietitians’ feedback.
Results: 1571 patients (72% of admissions) were screened of whom 158 (10%) scored at high risk. Non-
screened children were younger and had a shorter length of hospital stay. Of the 125 patients who scored
at high risk, between the 2nd and 4th month of the pilot, 66 (53%) were assessed by a dietitian of whom
86% were judged to be at true risk of malnutrition and 50% of these were new to the dietetic service.
Dietetic workload did not increase significantly during the pilot phase although the proportion of
referrals from the acute receiving wards increased. Dietitians’ feedback was positive, with recognition
that PYMS identified patients at risk of malnutrition who may not have otherwise been referred.
Conclusions: Nutrition screening by nurses using the new PYMS score is feasible for paediatric inpatients,
identifies children at risk of malnutrition and uses available resources efficiently.
! 2011 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.

1. Introduction with significant potential impact on patient’s health4,5,6,7 and

While disease associated malnutrition is a common problem in
paediatric hospitals, it often remains undetected and untreated1,2,3
Abbreviations: PYMS, Paediatric Yorkhill Malnutrition Score; SGNA, Subjective
Global Nutritional Assessment; STAMP, Screening Tool for the Assessment of
Malnutrition in Paediatrics; TPH, Tertiary Paediatric Hospital; DGH, District General
Hospital.
q Conference Presentation: Part of the data has already been presented at the
BSPGHAN (British Society of Paediatric Gastroenterology, Hepatology and Nutri-
tion) Annual Winter Meeting 2009 in Sheffield, UK in the 43rd conference of the
European Society of Paediatric Gastroenterology, Hepatology and Nutrition
(ESPGHAN 2009) in Budapest, Hungary and at the winter meeting of the British
Association of Parenteral and Enteral Nutrition (BAPEN 2009), Cardiff, UK.
* Corresponding author. Human Nutrition Section, Division of Developmental
Medicine, University of Glasgow, Royal Hospital for Sick Children, Glasgow G3 8SJ,
UK. Tel.: þ44 141 201 6969; fax: þ44 141 201 9275.
E-mail address: k.gerasimidis@clinmed.gla.ac.uk (K. Gerasimidis).
health expenditure.8 Thus identification of children at risk of
malnutrition should be an important part of any program for
providing optimal hospital care. This has been highlighted in recent
national guidelines which state that all patients should be screened
for risk of malnutrition on admission and periodically during
hospital stay.9
Anthropometry and growth charts have traditionally been used
to assess growth and to screen for chronic undernutrition, but this
will not identify children in the early stages of malnutrition or those
at risk of deterioration as a result of an acute medical condition. In
addition, plotting data is time consuming for nurses and requires
training. It has been proposed that nutritional screening tools may
be more appropriate to identify nutritional risk and could be uti-
lised by ward nursing staff.10 These are usually presented as
a questionnaire proforma outlining factors that put patients at high
malnutrition risk, and a number of such screening tools have been

0261-5614/$ e see front matter ! 2011 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.
doi:10.1016/j.clnu.2011.01.015
 K. Gerasimidis et al. / Clinical Nutrition 30 (2011) 430e435
developed for adults.7,11,12 However these tools are inappropriate
for children and validated paediatric screening tools are scarce.
In paediatrics four validated tools have been developed to date:
The first of these, the Subjective Global Nutritional Assessment
(SGNA) for children, as proposed by Secker and Jeejeebhoy,13 is
a comprehensive nutritional assessment method rather than
a rapid screening tool. The other two are shorter screening tools;
one assesses pain, disease condition and food intake in the hospital
and has been validated in a tertiary hospital in France.14 The other,
the Screening Tool for the Assessment of Malnutrition in Paediatrics
(STAMP) has been developed in the UK for use by the nursing staff,
although detailed validation data is as yet not available and it has so
far only been described in abstract form.15 STAMP uses a compar-
ison of weight and height centiles, recent changes in nutritional
intake, and the impact of disease diagnosis on nutritional status.
A recent Dutch tool which combines subjective clinical assessment,
high risk disease, nutritional intake and weight loss was developed
for use by paediatricians and therefore may not be suitable for
nursing use.16
While seeking a tool for use by nursing staff in our tertiary
children’s hospital we considered STAMP and the French tool which
were available at the time, but while the STAMP tool was attractive,
none appeared to fully meet the needs of our hospital population,
leading us to develop our own: the Paediatric Yorkhill Malnutrition
Score (PYMS). We have demonstrated elsewhere that PYMS has
acceptable levels of sensitivity and specificity compared to full
dietetic assessment and to have a more useful profile than other
screening tools already in use.17 In this paper we describe the
process of developing the tool and how it functioned as performed
by nursing staff in day to day practice.
2. Patients and methods
2.1. Development of the Paediatric Yorkhill Malnutrition Score
A multidisciplinary health professional team (medical, dietetic,
nursing, academic, practice development staff) developed
a screening tool that would fulfil the following criteria:
a. To identify the majority of children ("1 years) at the highest
risk of malnutrition (sensitive) whilst not misclassifying too
many patients at low risk.
b. To be quick and easy to use at ward level using information
regularly collected on admission by nursing staff.
c. To be feasible to integrate in regular hospital admission
procedure without significantly increasing nursing and dietetic
workload or staffing levels.
d. To lead to a simple explicit action plan.
The tool (Paediatric Yorkhill Malnutrition Score-PYMS) was
developed based on nutritional screening guidelines of the Euro-
pean Society of Clinical Nutrition and Metabolism (ESPEN).18 It
assesses 4 elements which are all recognised predictors/symptoms
of malnutrition risk: STEP 1: Body Mass Index (BMI) below the 2nd
centile (#2 SD) on UK 1990 growth chart; STEP 2: history of recent
weight loss; STEP 3: recent change in nutritional intake for at least
the past week; STEP 4: the likely effect of the current medical
condition on the nutritional status of the patient for at least the
next week (Fig. 1). Each step bears a score of 0e2 and the total score
reflects the degree of the nutrition risk of the patient. A score of
0 indicates a patient at low risk of malnutrition, a score of one
moderate risk, and a score of two or above high risk (Fig. 1). The
scoring system was the consensus of a multidisciplinary health
profession group in accordance with ESPEN guidelines17 and was
devised to reflect the clinical significance of factors associated with
431
malnutrition risk and to raise awareness of malnutrition risk. The
scoring system was further revised after assessment of the diag-
nostic validity of the tool at different scoring levels/combinations.17
The tool was not to be used before the age of 1 year as we felt that
a generic tool would probably not be valid during the first year of
life. However further work is planned to develop and testing
a related tool for use in infancy.
2.2. Introduction of the PYMS in clinical practice
Following the development of the tool, the PYMS was piloted in 4
paediatric wards (3 general medical, 1 general surgical) of a Tertiary
Paediatric Hospital (TPH) and the general paediatric ward of
a District General Hospital (DGH) with high patient turnover, to
evaluate its performance. All patients (1e16 years) admitted to these
wards over a 4 month period were eligible for screening within 24 h
of admission. Patients from other specialist wards (cardiology, renal,
orthopaedic), critical care and those seen only in the day assessment
unit were not included in the TPH. Prior to the launch of the project,
nursing staff (nw180) attended a 1 h awareness session by a nutri-
tion nurse specialist and received training on anthropometric
measurements of weight and height by a research dietitian and the
hospital auxologist. Infants’ weight were measured with Grade 3
electronic baby scales with accuracy down to 20 g. Older children
were measured on Grade 3 electronic standing scales with accuracy
down to 50 g. Babies were measured nude and children with
minimal clothing. In children aged 1e2 years length was measured
with a supine infantometer, and older children with wall mounted or
portable standiometers, all without shoes. Measurements were
recorded to the nearest millimetre. Each ward had its own equip-
ment and maintenance and calibration of the equipment was carried
out regularly by the department of biophysics. Non weight-bearing
children were measured either with sitting or sling scales. For chil-
dren unable to stand or in those where measurements are imprac-
tical (e.g. major trauma/operations) or unfeasible (e.g. cerebral palsy,
muscular dystrophy, chronically ventilated) measurements of height
were not carried out. In these cases the first PYMS step (BMI) was
replaced by weight and the cut offs defined as being below the 2nd
centile on the UK 1990 growth chart.
In the first month, nurses completed PYMS but did not utilise
these to initiate dietetic referrals, but in months 2e4 they were
asked to refer all children with scores of 2 or more to the dietetic
service. On-site support by the research team was available the first
3 months of the project at the TPH, with only limited support
available at the DGH.
2.3. Performance of PYMS in clinical practice
This was evaluated in four ways:
2.3.1. Audit of completion rates
The rate of PYMS completion was audited as the percentage of
eligible admissions (all admitted patients "1 year) that were
screened successfully by ward nursing staff. On a daily basis the
patient discharge list was retrieved from the hospital internal
network and the proportion off discharged patients who had
a PYMS form fully completed was audited. Self carbonated forms
were introduced to facilitate the auditing process. Differences in
demographics (e.g. gender age), and disease characteristics (length
of hospital stay) were assessed between screened patients and
those who were not. Serological markers of disease activity (e.g.
serum albumin, C-reactive protein, haemoglobin) were retrieved
from the hospital internal network for those patients who had
appropriate blood tests carried out during their hospital stay in
the TPH.
432 K. Gerasimidis et al. / Clinical Nutrition 30 (2011) 430e435
Fig. 1. The Paediatric Yorkhill Malnutrition Score (PYMS).
2.3.2. Assessment of yield
This was defined as the proportion of the patients with PYMS
score " 2 who on review by dietitians were rated as being at true
risk or to require dietetic intervention (positive predictive value).
The gold standard of malnutrition risk in this study was the global
clinical judgement of the dietetic staff, on the principle that what is
needed is a tool that identifies the majority of the patients which
need dietetic referral for malnutrition issues in a clinical setting.
The 6 dietitians in this study were registered senior specialist
dietitians and none of them had been involved in the development
of the PYMS.
2.3.3. Impact on dietetic service
The impact of introducing PYMS on the dietetic service was
audited one month before and during the last month of PYMS
launch. The number of dietetic referrals, means of referral (i.e.
established hospital electronic system, verbal, phone) and whether
these were appropriate or not were recorded daily by the hospital
dietitians using audit sheets.
2.3.4. Dietetic feedback on PYMS performance
Dietitians’ feedback on the performance of the tool was assessed
semi-anonymously using a predefined questionnaire at the end of
the pilot. This assessed the impact of PYMS introduction on work-
load, its ability to identify patients at risk of malnutrition, its
feasibility/functionality within clinical practice and within current
resources. (Questionnaire available on request).
2.4. Statistical analysis and ethics
Differences between two groups were assessed with the
ManneWhitney test and between categories with chi-square test.
Data analysis was conducted with Minitab 15 (Minitab Ltd,
Coventry, UK).
This study was conducted according to the guidelines laid down
in the Declaration of Helsinki. Ethical permission to conduct the
study was not required as this was a clinical audit not involving
active patient participation.19
3. Results
3.1. Audit of completion rates
Between 23rd June and 28th October 2008, 2174 children were
admitted to the 5 pilot wards (Table 1). Of these patients 1571
(72.3%) were screened, with a higher proportion screened in the
TPH [TPH n ¼ 1208 (73.7%) vs. DGH n ¼ 363 (68%), p ¼ 0.011].
Screened patients came from a variety of medical and surgical
specialties (Table 2). In the TPH only, those children who were not
screened were significantly younger and had a shorter length of
hospital stay than those who were screened (Table 1). However
during the last month of the pilot these differences were no longer
significant. There was no difference in gender or serological
markers (TPH patients only) between patients who were screened
for malnutrition and those were not (Table 1).
 K. Gerasimidis et al. / Clinical Nutrition 30 (2011) 430e435 433

Table 1
Demographics, anthropometry, length of hospital stay and serological markers of disease severity between patients screened and those not screened in a TPH and DGH.

Setting TPH DGH

Screened Not screened Total Screened Not screened Total

N (%) 1208 (74%) 432 (26%) 1640 363 (68%) 171 (32%) 534
N (%) male 706 (58%) 265 (61%) 971 (59%) 181 (50%) 94 (55%) 275 (51%)

N Median (IQR) N Median (IQR) N Median (IQR) N Median (IQR) N Median (IQR) N Median (IQR)
Age (years) 1208 6.3 (7.8) 432 4.7 (7.8)a 1640 6.0 (7.9) 357 7.6 (7.8) 171 5.7 (10) 528 7.2 (8.6)
BMI z-score (SD) 1095 0.0 (2.1) N/A 1095 0.0 (2.1) 270 0.2 (1.8) N/A 270 0.2 (1.8)
BMI % #2 SD N (%) 88/1095 8% N/A 88/1095 8% 14/270 5.2% N/A 14/270 5.2%
Albumin (g/L) 293 39 (7.0) 88 38 (6.0) 381 38 (7.0) N/A N/A N/A
CRP (mg/L) 495 8.0 (24) 145 7.0 (23) 640 7.0 (24) N/A N/A N/A
Haemoglobin (g/dL) 533 12.6 (2.0) 172 12.4 (2.5) 705 12.6 (2.1) N/A N/A N/A
Length of stay (days) 1202 1.0 (2.0) 432 1.0 (2.0)a 1624 1.0 (2.0) 340 1.0 (2.0) 162 1.0 (1.0) 502 1.0 (2.0)
Length of stay " 1 day N (%) 1022 85% 320 74%a 1342 82% 246 72% 119 73% 365 73%

N/A: not available; TPH: Tertiary Paediatric Hospital; DGH: District General Hospital; IQR: Inter-quartile range.
a
p < 0.001 from screened in the same hospital for ManneWhitney test or chi-squared test where appropriate.

The rate of PYMS completion increased during the first three
months in both hospitals but showed a slight decline in the last
month (Fig. 2). The completion rates were slightly higher in the
acute receiving wards than in the specialist wards (acute receiving
wards vs. specialist wards: 75% vs. 70%, p ¼ 0.05) in the TPH.
3.2. Malnutrition risk
One hundred and fifty eight patients (9% in DGH vs. 10.5% in
TPH) were scored as at high risk and 147 (10.4% in DGH vs. 9% in
TPH) at medium risk of malnutrition. High risk of malnutrition was
more prevalent in the specialist wards compared with the acute
receiving wards of the TPH (18% in specialist vs. 8.3% in acute
receiving). The majority of the patients at high risk of malnutrition
were scored as at risk on two or more of the steps, with only 27%
scoring at high risk on the basis of a single step of the tool. Seventy
three (46%) of those children who scored at high risk of malnutri-
tion had a normal BMI and 64 (41%) had a low BMI. Of those with
a low BMI 30 (19% of all high scorers) scored zero on all other steps.
For 21 patients (13%) at high risk of malnutrition anthropometric
measurements could not be obtained, so step 1 (i.e. BMI) could not
be computed, but these patients scored high on the basis of others
steps.
3.3. Yield of PYMS
Screened children at high risk of malnutrition were only
referred to dieticians in months 2e4 of the pilot. Of the 125 patients
scored as high risk (PYMS " 2) during those months, 66 (53%) were
assessed by dietetic staff (Table 3). A higher proportion was seen by
the ward dietitians in the TPH than in the DGH (TPH: 57% vs. DGH:
29%; 0.009). No formal information was collected on why patients
were not assessed. Informal discussion with dietetic and nursing
staff suggested that this mainly related to patients passing though
the wards at weekends when no dietetic support was available or
exceptionally busy periods. However there were no differences in
demographic characteristics, length of hospital stay, serological
markers of disease severity (serum albumin, haemoglobin, CRP)
between patients who had been reviewed by the dietitians and
those who had not. More than 50% of these patients were new cases
to the dietetic service (Table 3).
The ward dietitians judged 49 (86%) of the patients seen in the
TPH and 8 (89%) of the patients in the DGH to be at true risk of
malnutrition (Table 3). In the TPH both the majority of the new true
cases identified by PYMS as at high risk and all of the patients
judged to be inappropriate referrals, had been admitted to the acute
receiving medical and surgical wards (Table 3). All of the patients
admitted to the specialist wards and screened at high risk of
malnutrition were judged to be at true risk of malnutrition, but the
majority of them were already under dietetic care. Of those judged
to be at true risk of malnutrition 21% (12/57) did not report recent
weight loss and did not have a low BMI (Step 1; Fig. 1). Of the
patients rated high risk because of decreased dietary intake and/or
due to the effect of the current condition 71% (12/17) were judged
at true risk by the dietitians, but the 5 rated false positives made up
over half of all those judged to be inappropriate referrals.
3.4. Impact on dietetic service

Table 2 During the one month audit period there was no increase in the
Number and percent of all admitted patients screened for malnutrition risk in each percentage of total admissions referred to the dietetic service,
specialty at the TPH.

Specialty Number and percent Total admitted

of patients screened patients

N (%) N
General medical 376 (75%) 502
Respiratory 127 (73%) 174
Diabetes and endocrinology 28 (74%) 38
Gastroenterology 57 (90%) 63
Haematology/Oncology 38 (64%) 59
Neurology 41 (64%) 64
Dermatology 15 (79%) 19
General surgical 457 (73%) 627
Plastics/Burns 19 (70%) 27
Ear/Nose/Throat 9 (60%) 15
Others 35 (74%) 47
Fig. 2. Percent of total admitted patients screened for malnutrition with the PYMS
TPH: Tertiary paediatric hospital. during the four month of the pilot.
434 K. Gerasimidis et al. / Clinical Nutrition 30 (2011) 430e435

Table 3 ideal tool identifies the majority of children at risk of malnutrition

Characteristics of patients at high risk of malnutrition who were reviewed by the and does not misclassify those at low risk, assessment of its
clinical dietitians. Values are % (N) unless otherwise specified.
performance in a clinical setting is essential before introducing it
TPH (n ¼ 57) DGH (n ¼ 9) for routine use. In a previous study we have reported the results of
Boys 51% (29) 55% (5) assessing the diagnostic validity of the PYMS, which demonstrated
Age (years) mean (SD) 7.3 (5.1) 7.8 (5.7) its usefulness as a screening tool. It was found to have acceptable
New identified patients 52% (29) 56% (5)
levels of sensitivity and specificity which were better in compar-
Patients at true risk 86% (49) 89% (8)
Acute receiving wards 78% (28) N/A ison with other similar tools.17 This study complements our
Specialist wards 100% (21) N/A previous work and assessed the performance, feasibility, and
New cases at true risk 76% (22) 80% (4) impact on dietetic workload of PYMS in a real life clinical setting.
Acute receiving wards 73% (19) N/A Introducing PYMS in a TPH and DGH over a 4 month pilot phase
Specialist wards 100% (3) N/A
achieved high completion rates, indicative of its feasibility for
N/A: not applicable; TPH: Tertiary paediatric hospital; DGH: District general routine clinical use within the current resources and without
hospital.
necessitating extra staff. An increase in dietetic workload had been
anticipated, but there was no objective evidence from the acute
receiving wards that this was the case. However one ward dietitian,
compared to before PYMS began (Period 1: 15% vs. during last from the acute receiving ward, felt that PYMS had impacted on her
month of PYMS-Period 2: 11%). In fact there were fewer patients workload. Interestingly there has been a decrease in referrals from
referred from the two specialist wards in the TPH during the PYMS the specialist wards, the reasons for which are not clear.
audit period, while there was no difference in the number and PYMS completion rates were good overall, but those most
proportion of patients referred from the acute receiving wards commonly not screened had shorter stays, were younger, and
(Table 4). Before PYMS introduction 66% of the dietetic referrals tended to come from the acute receiving wars. These may be
were made electronically (hospital established dietetic referral associated with the difficulty in applying the anthropometric
pathway) which increased to 87% during PYMS period. Four (5%) of measurements (both weight and height) in infants and some
the referrals in the survey period before PYMS were inappropriate patients in the specialist wards (severe cerebral palsy, long term
compared to 7 (11%) while PYMS was operating, although only 3 of ventilated patients). Also as many of these patients are already
these came through PYMS screening (Table 4). under regular dietetic care, screening for malnutrition risk in these
patients may have been seen by the nurses as less important.
3.5. Dietetic feedback However PYMS completion rates decreased in the last month, as on
site support was withdrawn suggesting that successful imple-
All 6 dietitians completed and returned the PYMS feedback mentation requires ongoing support and supervision.
questionnaire. All agreed that PYMS identified patients at risk of A weakness of this study is that approximately half of the
malnutrition, but two of them felt that PYMS overestimated risk in patients who were screened at high risk of malnutrition were never
some patients. Both of these dietitians were from the acute assessed by the hospital dietitians. Although specific reasons why
receiving wards in the TPH and DGH. All dietitians agreed that these patients were not assessed by the dietitians were not
PYMS identified patients at risk of malnutrition who may not have collected as part of this study it is believed that some of these
otherwise been referred and improved patient care, and only one patients were admitted and discharged during times where dietetic
(the dietitian from the acute medical and surgical wards) felt that it support was not available (night shifts, weekends). Thus if all of
increased dietetic workload significantly. Half felt the general these patients had been referred this implies that the actual dietetic
introduction of PYMS was feasible within current resources. All the workload could have doubled. Considering an average prevalence
dietitians in the TPH found the PYMS action plan appropriate. of malnutrition risk of 9e11% and an admission number of
approximately 100 patients per week in the TPH and 38 in the DGH
4. Discussion this would result in 10 additional patients per week referred to the
dietetic service in the TPH and 3 in the DGH in the pilot wards
Screening for malnutrition risk on hospital admission is rec- alone. Since approximately 50% of these patients are likely to be
ommended as an integral part of patient’s standard care. While an already under dietetic care, this would mean a net increase of 5
patients per week in the four pilot wards of the TPH and by 1.5
patients per week in the DGH. PYMS has now been in routine
Table 4
clinical use in every ward of the TPH and DGH for 2 years with
Number and percentage of dietetic referrals one month before and during the last
one month of the Paediatric Yorkhill Malnutrition Score pilot phase. a compliance of more than 75% and thus far there has been no
major issue regarding a significant increment in workload. Another
Before During last month of p-valuec
PYMS PYMS
limitation of this study is that the performance of the PYMS was not
evaluated in cardiology, renal, orthopaedic specialties and critical
N % N %
care. However the majority of the patients in these specialist wards
Patient admissions 586 559 are already under dietetic care, in line with the clinical manage-
Acute receiving wards 419 433 0.021
Specialist wards 167 126
ment pathways and malnutrition screening may actually have little
Dietetic referrals importance in such settings.
Acute receiving wards 48 11%a 47 11% There is no universally established definition or method to
Specialist wards 39 23%a 14 11% 0.004 determine nutritional status so our pragmatic benchmark was the
Total 87 61
dietitians’ assessment of whether the referral was appropriate. This
(14 through PYMS)
Inappropriate-false positives 4 5%b 7 11% 0.116 ensures that the tool is relevant for clinical use and would thereby
utilise resources appropriately. PYMS appears to be quite specific in
PYMS: Paediatric Yorkhill Malnutrition Score.
a
Of all ward admissions.
identifying patients whom dietitians then judge to be risk of
b
Of all dietetic referrals. malnutrition. However it is possible that there were more false
c
For difference between “Before PYMS” and “During last month of PYMS”. positives cases in the group who were not referred to dietitians at all.
 K. Gerasimidis et al. / Clinical Nutrition 30 (2011) 430e435
Although we had asked the nursing staff to refer any patient who
screened at high risk of malnutrition, it is possible that nurses chose
not to refer patients where they believed them not, in fact, to be at
risk. Similarly we cannot know whether the reduced rate of referrals
in the specialist wards resulted in true cases being missed. As the
dietitians did not assess patients who scored at low risk of malnu-
trition on PYMS, this study cannot evaluate the sensitivity of the
score, but this has been addressed in a linked study which compared
PYMS to a complete research dietetic assessment17 and we showed
that PYMS identifies 59% of children at risk of malnutrition without
misclassifying patients’ malnutrition risk (negative predictive value:
92% & specificity: 95%). Approximately half of the patients rated as
high risk had a normal BMI z-score, which indicates that although
the use of anthropometric cut offs is important, it does not identify
all patients at risk of future disease associated malnutrition. Indeed
even within those patients who were assessed by the dietitians as
being at true risk, 21% had a normal BMI (Step 1; Fig. 1) and did not
report recent weight loss (Step 2; Fig. 1). However 56% of the false
positives came from patients who scored high at the third (dietary
intake) and forth (disease effect) step on the PYMS chart (Fig. 1)
which implies that the inclusion of these two steps can increase
sensitivity but may also compromise its positive predictive value.
It can be anticipated that with any screening tool some patients
will be misclassified and indeed PYMS generated a few false posi-
tive referrals. All these arose in the acute receiving wards, but so did
the majority of previously unknown true cases whom the ward
dietitians felt may not have been identified without the use of
PYMS. Although inappropriate referrals may increase the dietetic
workload these patients should usually be relatively quickly rec-
ognised and discharged from dietetic services.
5. Conclusions
Regular clinical use of PYMS appears to be feasible in a DGH and
TPH, producing a high yield of patients at risk of malnutrition and
without requiring significant increase in staffing levels or workload.
Statement of authorship
All authors have made substantial contributions to all of the
following: (1) the conception and design of the study, or acquisition
of data, or analysis and interpretation of data, (2) drafting the article
or revising it critically for important intellectual content, (3) final
approval of the version to be submitted. KG designed the study,
undertook the main research activities, conducted data analysis,
drafted the manuscript; IM contributed to the research activities
and revised the manuscript; AM, EB, PM, IS, MM, CMW, DMF
contributed to the study design, coordinated the project and
revised the manuscript.
Conflicts of interest
None to declare.
435
Acknowledgements
The authors of the study would like to acknowledge the
contribution of Mr Toby Mohammed, Dr Graham Stewart, Ms
Christina McGuckin on the development of the PYMS and the
hospital auxologist Ms Wendy Paterson for providing training to
the nursing staff. They are particularly grateful to the nursing and
dietetic departments at the Royal Hospital for Sick Children, York-
hill, Glasgow, and the Royal Alexandra Hospital, Paisley. The project
was funded by the National Health Service Greater Glasgow &
Clyde, Food Fluid & Nutritional Care Planning Implementation
Group, Women and Children Directorate, Glasgow, UK. A small
grant was provided generously by the Yorkhill Children Foundation
to cover consumable expenses. No conflict of interest to declare.
References
1. Hendrikse WH, Reilly JJ, Weaver LT. Malnutrition in a children’s hospital. Clin
Nutr 1997;16:13e8.
2. McWhirter JP, Pennington CR. Incidence and recognition of malnutrition in
hospital. BMJ 1994;308:945e8.
3. Pawellek I, Dokoupil K, Koletzko B. Prevalence of malnutrition in paediatric
hospital patients. Clin Nutr 2008;27:72e6.
4. Kac G, Camacho-Dias P, Silva-Coutinho D, Silveira-Lopes R, Marins VV,
Pinheiro AB. Length of stay is associated with incidence of in-hospital malnu-
trition in a group of low-income Brazilian children. Salud Publica Mex
2000;42:407e12.
5. Klein PS, Forbes GB, Nader PR. Effects of starvation in infancy (pyloric stenosis)
on subsequent learning abilities. J Pediatr 1975;87:8e15.
6. Lucas A, Morley R, Cole TJ. Randomised trial of early diet in preterm babies and
later intelligence quotient. BMJ 1998;317:1481e7.
7. Stratton RJ, Hackston A, Longmore D, Dixon R, Price S, Stroud M, et al.
Malnutrition in hospital outpatients and inpatients: prevalence, concurrent
validity and ease of use of the ‘malnutrition universal screening tool’ (‘MUST’)
for adults. Br J Nutr 2004;92:799e808.
8. Elia M. Nutrition and health economics. Nutrition 2006;22:576e8.
9. NHS Quality Improvement Scotland. Food, fluid and nutritional care in hospitals.
NHS Quality Improvement Scotland; 2003.
10. Green SM, Watson R. Nutritional screening and assessment tools for use by
nurses: literature review. J Adv Nurs 2005;50:69e83.
11. Gerasimidis K, Drongitis P, Murray L, Young D, McKee RF. A local nutritional
screening tool compared to malnutrition universal screening tool. Eur J Clin
Nutr 2007;61:916e21.
12. Kondrup J, Rasmussen HH, Hamberg O, Stanga Z. Nutritional risk screening
(NRS 2002): a new method based on an analysis of controlled clinical trials. Clin
Nutr 2003;22:321e36.
13. Secker DJ, Jeejeebhoy KN. Subjective global nutritional assessment for children.
Am J Clin Nutr 2007;85:1083e9.
14. Sermet-Gaudelus I, Poisson-Salomon AS, Colomb V, Brusset MC, Mosser F,
Berrier F, et al. Simple pediatric nutritional risk score to identify children at risk
of malnutrition. Am J Clin Nutr 2000;72:64e70.
15. McCarthy H, McNulty H, Dixon M, Eaton-Evans MJ. Screening for nutrition risk
in children: the validation of a new tool. J Hum Nutr Diet 2008;21:395e6.
16. Hulst JM, Zwart H, Hop WC, Joosten KF. Dutch national survey to test the
STRONG(kids) nutritional risk screening tool in hospitalized children. Clin Nutr
2009;29:106e11.
17. Gerasimidis K, Keane O, Macleod I, Flynn DM, Wright CM. A four stage eval-
uation of the Paediatric Yorkhill Malnutrition Score in a tertiary paediatric and
district general hospital 2010. Br J Nutr 2010;104:751e6.
18. Kondrup J, Allison SP, Elia M, Vellas B, Plauth M. ESPEN guidelines for nutrition
screening 2002. Clin Nutr 2003;22:415e21.
19. NRES. Defining research. National Research Ethics Service, http://www.nres.
npsa.nhs.uk/applications/apply/is-your-project-research/?locale¼en; 2009
[accessed 07. 10. 10].