Introducing a nutritional risk screening tool in a South African hospital

Background Nutritional screening facilitates the early identification of hospitalised children at risk of malnutrition. Screening tools have scarcely been evaluated in the developing world where the burden of malnutrition is greatest. Methods A retrospective study was undertaken of 113 patients admitted to the general paediatric wards at Rahima Moosa Mother and Child Hospital in Johannesburg, South Africa. Children 6 months to 14 years old were screened for malnutrition using anthropometry and correlating WHO z-scores, and retrospectively assessed for nutritional risk using a modified STAMP (mSTAMP). Results The mSTAMP identified additional patients at nutritional risk. The majority (87%) of children with normal anthropometry scored as medium and high risk using the mSTAMP. Weight loss and length of hospital stay (LOS) were higher in medium and high risk groups: One (5%) low risk child lost weight, compared with 8 (38%) medium and 12 (57%) high risk children (p = 0.021). Low risk children had a median LOS of two and half days (IQR 1–8) compared with medium and high risk groups, with medians of three (IQR 3–8) and six (IQR 4–9) days respectively (p = 0.04). Conclusion The mSTAMP identified more children at risk of malnutrition who may not have been considered for nutritional therapy during the hospital stay using anthropometry screening alone. There is a place for nutritional risk screening in developing world settings, but tools may need to be modified locally. Further studies and validation of these tools in sub-Saharan Africa seem prudent and may result in improved nutrition and outcomes of hospitalised children.