Accuracy of nutritional screening tools in assessing the risk of undernutrition in hospitalized children

OBJECTIVE: To evaluate the predictive accuracy of screening tools for assessing nutritional risk in hospitalized children in developed countries.

METHODS: A systematic review of literature (MEDLINE, EMBASE and Cochrane Central databases until 17-01-2014) of studies on the diagnostic performance of pediatric nutritional screening tools. Methodological quality was assessed using a modified QUADAS tool. Sensitivity and specificity were calculated for each screening tool per validation method. A meta-analysis was performed to estimate the risk ratio of different screening result categories of being truly at nutritional risk.

RESULTS: Eleven studies were included on one or more of the following screening tools: Pediatric nutritional Risk Score (PNRS), Screening Tool for the Assessment of Malnutrition in Paediatrics (STAMP), Paediatric Yorkhill Malnutrition Score (PYMS) and Screening Tool for Risk on Nutritional Status and Growth (STRONGkids). Due to variation in reference standards, a direct comparison of the predictive accuracy of the screening tools was not possible. A meta-analysis was performed on 1629 children from seven different studies. The risk ratio of being truly at nutritional risk was 0.349 (95%CI 0.16-0.78) for children in the low vs the moderate screening category and 0.292 (95% CI 0.19-0.44) in the moderate vs high screening category.

CONCLUSION: There is insufficient evidence to choose one nutritional screening tool over another based on their predictive accuracy. The estimated risk of being at 'true nutritional risk' increases with each category of screening test result. Each screening category should be linked to a specific course of action, although further research is needed.