An investigation into the prevalence and identification of malnutrition in hospitalised children

Abstract

The adverse clinical and nutritional consequences of malnutrition in childhood are well recognised and have been shown to predispose a child to an increased risk of chronic diseases in adulthood. Despite this, malnutrition in hospitalised children remains a largely unrecognised problem. Current prevalence rates vary significantly due to the lack of universally defined indices or criteria by which to define malnutrition in children. This thesis investigated the prevalence of malnutrition in hospitalised children. Nine hundred and thirty two children (537 boys, 392 girls, mean age 5.7 years) participated in the Children’s Nutrition Survey. Weight measurements were recorded for 865 (93%) children while height was recorded for only 468 (50%). The prevalence of under-nutrition ranged from 2-21% and was largely dependent on the index used to define malnutrition. The prevalence of over-nutrition, defined in terms of BMI z-scores, was estimated at 13%. Those under 1 year of age were at greatest risk of malnutrition. Expert consensus on universally acceptable criteria by which to define malnutrition in hospitalised children and which are shown to predict both clinical and nutritional outcomes are required.

Many organisations now recommend routine nutrition screening of paediatric patients on admission to hospital. In response to these recommendations, a number of paediatric specific nutrition screening tools have been developed. Although previously examined in children aged 2-16 years, the validity of the STAMP© nutrition screening tool required examination in infants under 2 years of age. A pilot study examining nutrition screening in hospitalised infants found moderate agreement between STAMP© and a full nutritional assessment compared to that determined by the full nutritional assessment (κ= 0.808; 95% CI 0.317, 0.633). Sensitivity and specificity were calculatedas 100% and 87% respectively. The results therefore support the use of the STAMP© tool in infants less than 2 years of age; however, further work in a larger cohort is required to substantiate these results.

The measurement of weight and height measurements is a fundamental component of clinical and nutritional care; however, this can prove difficult in situations where a child cannot stand or be measured using traditional methods. In this thesis, the validity of potential proxy measures for weight and height in children were examined. Strong correlation was observed between weight and mid upper arm circumference (R2=0.713). BMI z-score was strongly correlated with mid upper arm circumference (R2=0.677) and waist circumference (R2=0.659). Correlation coefficients between height and segmental proxy measures were were strong: lower leg length (R2=0.712), knee height (R2=0.705), arm span (R2=0.653) and FL (R2=0.605). Mid upper arm circumference was shown to be a quick, unobtrusive, statistically valid predictor of weight in healthy children aged 11-18 years. Knee height was also shown to both a statistically valid proxy for height and has the advantage of being simple to obtain. The prediction equations determined in this study may be used to monitor growth and provide clinical and nutritional care to children who cannot be weighed or measured using traditional methods, to estimate weight in emergency situations or aid the completion of nutrition risk screening tools which require the measurement of weight and height.

This thesis provides the first large scale, multi-centre data collection regarding the nutritional status of hospitalised children across the UK and Ireland. From the results of this body of work it can be concluded that both over and under-nutrition are prevalent among the general paediatric inpatient population. Examination of the long term nutritional and clinical outcomes of defining malnutrition according to different indices in hospitalised children is required. This study provides a benchmark for policy makers and the health service from which to address the issue of malnutrition in hospitalised children. In addition, the high prevalence of under-nutrition reported in this thesis illustrates the need for paediatric specific nutrition screening tools. Examination of the STAMP© nutrition screening tool suggests that is a valid tool in the identification of malnutrition risk in infants under 2 years of age. Future work examining the long term outcomes, benefits and potential detrimental impact of identifying malnutrition in hospitalised patients is required. Furthermore, the main limitation of examining malnutrition at both a both a clinical and research level is the general lack of anthropometric measurements, in particular height measurements. The final chapter of this thesis clearly highlights the potential use of mid upper arm circumference as a proxy for weight and knee height and lower leg length as strong proxies for actual height. These proxies could be used in a number of settings where height cannot be routinely recorded including: general clinical practice, in the emergency setting where the provision of emergency clinical care can be highly dependent on the estimated weight of a child or in the completion of malnutrition screening tools such as STAMP© and PYMS by healthcare professionals. Future research is required to validate the findings of this study in a larger cohort, across a larger age range.

Date of Award	Jul 2013
Original language	English
Supervisor	Helen Mc Carthy (Supervisor) & Helene McNulty (Supervisor)