In the February issue of ADC, Joosten and colleagues have shown that nearly one in five children in The Netherlands are malnourished on admission to hospital, and they have reiterated a call for routine screening and treatment of malnutrition in all hospitals.1 In this unique nationwide study, Joosten and his team were able to enlist 44 hospitals in collecting data on 424 children. The response of the Dutch government to these findings has been to oblige their hospitals to screen all children for malnutrition on admission. A further effect of this study will be to reopen the debate about routine nutritional screening in hospitalised children elsewhere in Europe.

The high prevalence of malnutrition in hospitalised children in The Netherlands is, in fact, similar to that reported in other European countries, USA, Brazil and Turkey. Nearly 20 years ago, Moy et al reported that 14% of 273 children in Birmingham Children's Hospital, UK were severely wasted, with a further 20% being considered to be ‘at risk’ of severe nutritional depletion.2 Similarly, last year, Pawelleck and colleagues found that 24.1% of 475 consecutive admissions to Dr von Hauner Children's Hospital in Munich, Germany were malnourished according to the Waterlow criteria.3 Not much has changed apparently in the last two decades.

Although comparisons between these studies are somewhat confounded by slightly different definitions of ‘malnutrition,’ the accumulation of published data does indicate the existence of what Moy et al called an ‘unrecognised reservoir of malnutrition amongst children in hospital.’2,–,9 Weight-for-height is a widely accepted criterion for defining malnutrition, but appropriate reference standards and charts for weight-for-height are not always available, and it is an unreliable measure when calculated ‘by hand.’10 Furthermore, one of the factors contributing to the oversight of a child's nutritional status remains the failure to measure routinely both height and weight in all children admitted to hospital; in the Munich study, for instance, it was noted that combined weight and height data were absent in around 25% of admitted children. I suspect that this situation pertains widely.

Paediatricians may believe that they can recognise a malnourished child ‘from the end of the bed,’ but the fact is that sometimes we get it wrong; the reproducibility in clinical assessment of nutritional status is poor, even among senior paediatricians—especially in the more severely malnourished.11 Clinical evaluation of nutritional status alone is inadequate for accurate assessment, and anthropometry is essential.

Detailed anthropometry will identify those who are malnourished—either under- or overnourished—but may not identify those ‘at risk’ of developing malnutrition. The size of the ‘reservoir’ of children who are not measurably malnourished but who are nonetheless in a precarious nutritional state is hard to determine. Although attempts to develop one have been made,9 a universally accepted nutritional screening tool for hospitalised children is not yet available. Very few available instruments (including those developed to screen hospitalised adults) have had both reliability and validity evaluated,12 13 and none have examined the subsequent effect on nutritional care of their application.14

Sermet-Gaudelus and colleagues at Necker-Enfants Malades Hospital in Paris, using a multivariate analysis in a large paediatric population, showed that the factors most predictive of weight loss in children during a hospital stay were poor food intake, pain and severity of disease; the combination of these factors was the best predictor of whether patients were at risk of nutritional depletion.9 On the basis of these findings, they developed a paediatric nutritional risk score to identify children at risk of nutritional depletion during hospitalisation. This score identified children at mild, moderate or severe risk of nutritional compromise and, importantly, linked this to appropriate nutritional support for the moderate and severe categories. When they published their paper in 2000, they reported that their Paediatric Nutritional Risk Score was ‘currently being used as part of the routine admission procedure in our hospital.’ By 2006, however, the instrument was not in routine use, most paediatricians in the hospital were not aware of it, and although some of the nurses were ready to apply the tool, they could not see the point because they perceived that the doctors were not interested (Virginie Colomb, personal communication). This highlights a crucial issue; convincing paediatricians of the importance of nutritional screening and engaging them in its implementation is one of the major challenges that any programme, including that proposed by the Dutch government, will face.

Such acceptance will be more likely if screening is not another ‘bolt-on’ responsibility of the paediatricians themselves. In fact, the professional group best placed to undertake routine screening is nurses who already assess patients on admission to hospital.15 The screening instrument should be easy to use and easy to interpret, and it should also be quick; some tools (eg, the Nutritional Risk Index developed for use in the older15) can take 40 min to complete, which renders them wholly impractical for use on a busy admissions ward.

The ideal screening instrument then will be one that can quickly and reliably triage the nutritional status of children, so as to identify the high-risk group that need more detailed assessment and intervention. Acceptability of such a tool would then follow from a demonstration that such earlier intervention was cost-effective with improvement in clinical outcomes (eg, length of hospital stay, number of infectious episodes, use of antibiotics and postoperative complications) and in well-being in the patient.

Screening is only of any use if it can reliably detect those who are at greater risk and if it is linked to appropriate subsequent monitoring and intervention. Reports from both The King's Fund and the British Association of Parenteral and Enteral Nutrition (BAPEN) have recommended that in adult medicine, every hospital should have a nutrition team consisting of a senior clinician with experience in clinical nutrition, dietitians, nutritional nurse, pharmacists and a surgeon with a dedicated interest in nutrition.17 18 Such a team would ensure that the monitoring of growth and nutrition is a routine part of patient care. The Committee on Nutrition of the European Society for Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) recently recommended that nutrition support teams be established in paediatric hospitals, to implement screening for nutritional risk, to identify patients who require nutritional support, to provide adequate nutritional management, to educate and train hospital staff and to audit practice.19

On admission to hospital, other features of a child's primary disease are screened routinely and treated (eg, hydration status, fever, etc), and it is unacceptable that nutritional compromise causing significant clinical risk is not identified. Common sense is required, and while the nutritional risk of children with severe disease (cystic fibrosis, Crohn's disease, cerebral palsy, etc) is usually obvious, nutritional assessment is more difficult and often neglected in those with less severe conditions.20 This neglect is also beginning to have medicolegal consequences, since an increasing number of cases of nutritional neglect are being brought to the courts.13 There is every reason, therefore, for hospitals and healthcare organisations to adopt a minimum set of standards in this area.

It is remarkable that neither the Royal College of Paediatrics and Child Health (of which professional body this is the official journal) nor the British Department of Health has developed any policy on the problem of identifying malnutrition in hospitalised children. It is time that they did.