RESUMO
Nutrition screening (NS) allows health professionals to identify patients at nutritional risk (NR), enabling early nutrition intervention. This study aimed to systematically review the criterion validity of NS tools for hospitalized non-critical care pediatric patients and to estimate the prevalence of NR in this population. This research was performed using PubMed, Embase, and Scopus databases until June 2021. The reviewers extracted the studies' general information, the population characteristics, the NR prevalence, and the NS tools' concurrent and predictive validity data. Quality evaluation was performed using the Newcastle-Ottawa Scale, adapted Newcastle-Ottawa Scale, and Quality Assessment of Diagnostic Accuracy Studies (QUADAS-2). The primary studies were qualitatively analyzed, and descriptive statistics were calculated to describe the NR prevalence. Of the total 3944 studies found, 49 met the inclusion criteria. Ten different pediatric NS tools were identified; the most frequently used were Screening Tool for Risk on Nutritional Status and Growth (STRONGkids), Screening Tool for the Assessment of Malnutrition in Pediatrics (STAMP), and Pediatric Yorkhill Malnutrition Score (PYMS). The mean NR prevalence was 59.85% (range, 14.6%-96.9%). Among all NS tools analyzed, STRONGkids and PYMS showed the best diagnostic performance. STRONGkids had the most studies of predictive validity showing that the NR predicted a higher hospital length of stay (odds ratio [OR], 1.96-8.02), health complications during hospitalization (OR, 3.4), and the necessity for nutrition intervention (OR, 18.93). Considering the diagnostic accuracy, robust and replicated findings of predictive validity, and studies' quality, STRONGkids performed best in identifying NR in the pediatric population among the tools identified.