Malnutrition is a significant public health issue with a high incidence among hospitalized children. It not only impairs their growth and development but also increases their risk of infections and mortality, prolonging hospital stays and adversely impacting their prognosis. Therefore, early screening and assessment of malnutrition risk in hospitalized children can improve clinical outcomes (1–3). Currently, six main tools are available for assessing malnutrition risk in hospitalized children, among which STRONGkids and PYMS are widely used. These tools facilitate the early detection and treatment of malnutrition in hospitalized children. However, existing studies on STRONGkids and PYMS have primarily focused on single-disease categories, lacking comprehensive research across multiple diseases (4, 5). This study aimed to explore the clinical application value of the modified STRONGkids and PYMS tools in assessing malnutrition risk among hospitalized children and to evaluate the awareness of malnutrition risk among medical staff. The findings are reported as follows.

We evaluated two pediatric nutrition screening tools against WHO-defined malnutrition and downstream clinical outcomes. Both the modified STRONGkids and PYMS tools identified nutritional risk to varying degrees, but the modified STRONGkids demonstrated superior discrimination and provided more informative risk stratification for hospital-acquired infection and length of stay. Risk gradients were most pronounced among infants younger than one year and children with digestive diseases. These findings support routine use of the modified STRONGkids tool, with targeted escalation for high-risk strata.

Nutrition is fundamental for maintaining various physiological functions of the body, and children have higher nutritional demands. Therefore, the risk of malnutrition is relatively high among children. Hospitalized children are at an increased nutritional risk due to factors such as digestive and absorption disorders, increased nutritional consumption, and reduced intake. Critically ill children are at an even higher nutritional risk, which not only affects their growth and development but also impacts their physiological development, cognitive development, and immune function, ultimately leading to adverse outcomes such as increased mortality and prolonged hospital stays (10–12). Therefore, it is crucial to clearly identify whether hospitalized children are at nutritional risk and provide timely and appropriate nutritional therapy to improve their health status and reduce the risk of adverse outcomes. Currently, nutritional risk screening methods for adult patients are relatively well developed, whereas pediatric nutritional screening methods remain fewer and outdated. Commonly used tools for screening nutritional risk in hospitalized children include PNRS, STAMP, PYMS, SGNA, and STRONGkids. Each of these tools has its own advantages and disadvantages, but there is still no universally accepted method for assessing the nutritional status of children. Thus, further research is needed to validate the effectiveness and feasibility of the existing screening methods (13, 14). Healthcare professionals play a crucial role in nutritional risk screening. If they lack awareness of the risk of malnutrition, the quality of their work may decline. Therefore, it is imperative to enhance clinical nutrition knowledge and awareness of nutritional risk screening of healthcare professionals. The absence of universally accepted screening tools and insufficient emphasis on the impact of malnutrition on clinical outcomes by healthcare professionals have resulted in limited research on nutritional risk screening for hospitalized children.

While adult nutritional risk screening tools have become more sophisticated in recent years, they are not suitable for pediatric populations. This is primarily because children's height and weight are constantly changing, and their etiological and pathological conditions differ from those of adults. Therefore, it is essential to find more appropriate nutritional risk screening tools for children.

The modified STRONGkids tool is time-efficient, practical, simple, highly correlated with clinical outcomes, and highly sensitive. However, its assessment requires professionals with specialized knowledge, demanding a higher level of expertise from clinicians or nurses. In addition, it requires family members to retrospectively calculate the child's weight changes, which may introduce calculation errors (15). The PYMS tool is a comprehensive screening tool that integrates nutritional issues with clinical assessments of children, offering high repeatability. However, it does not account for nutritional risks posed by potential diseases (16). Therefore, promoting the clinical application of these two nutritional risk screening tools should be tailored to actual circumstances to enhance clinical diagnostic accuracy and treatment levels and improve clinical outcomes for children.

In this study, both the modified STRONGkids and PYMS tools were used to assess the risk of malnutrition in children with different systemic diseases and ages. The results indicated that children with digestive system diseases and those under 1 year of age had a higher risk of malnutrition. The analysis suggested that digestive system diseases are more likely to cause symptoms such as indigestion, vomiting, loss of appetite, and diarrhea compared to other diseases, leading to reduced nutritional intake and increased risk of malnutrition. Younger children, who have not yet fully developed their immune systems, also grow rapidly and therefore have higher nutritional needs and weaker infection resistance, further increasing their risk of malnutrition (17, 18). In addition, children identified as high nutritional risk by the modified STRONGkids and PYMS tools had higher infection rates and longer hospital stays compared with those at moderate to low risk. Their hemoglobin and prealbumin levels were also lower than those of children at moderate to low risk. This is primarily because biochemical indicators can effectively reflect the nutritional status of children. Hemoglobin and prealbumin are commonly recognized indicators for assessing nutritional status, and their levels are inversely related to the risk of malnutrition. Infection rates and hospital stays can increase with the risk of malnutrition (19–21). Currently, there is a growing emphasis on nutritional risk in clinical practice, and most healthcare professionals recognize the importance of nutritional status for risk screening. The results of this study also revealed that among 80 healthcare professionals, eight were very knowledgeable about the risk of malnutrition and 52 were fairly knowledgeable about it.

This study's strengths include a head-to-head evaluation of two pediatric nutrition screening tools within the same cohort, benchmarking against WHO anthropometric criteria, and linking screening strata to clinically relevant outcomes (hospital-acquired infection and length of stay), thereby enabling a pragmatic screening-to-action pathway. However, the observational, single-system design limits causal inference and generalizability, and both selection and spectrum biases are possible. Screening timing and re-screen cadence were not fully standardized, some inputs relied on parent report, and interrater reliability was not prospectively quantified, introducing potential misclassification. Outcome ascertainment relied on routine surveillance and may have underdetected infections; residual confounding by illness severity, treatment intensity, and staffing cannot be excluded despite adjustment. Several subgroup analyses were underpowered, no multiplicity adjustment was applied, and discrimination was estimated in sample without optimism correction or external validation, so performance may be overstated. The comparison was restricted to two tools, and implementation outcomes (time-to-complete, staff burden, workflow, costs) and the proposed decision thresholds were not prospectively tested. These constraints support cautious interpretation and motivate multicenter validation with predefined re-screen intervals, blinded outcome adjudication, reliability testing, and prospective evaluation of workflow and cost.

This observational, single-system study is susceptible to residual confounding and spectrum bias despite statistical adjustment, and generalizability may be constrained by local case-mix and infection surveillance practices. Screening timing and re-screen cadence were not fully standardized, introducing potential misclassification. Interrater reliability for STRONGkids and PYMS was not prospectively quantified. Missing data and reliance on parent-reported weight or intake may have introduced measurement error, and ascertainment of hospital-acquired infection using routine surveillance may have underdetected events. Several subgroup comparisons, including infants and digestive disease, were underpowered. The proposed decision thresholds and pathway were not prospectively tested, and process metrics such as time-to-complete, staffing burden, workflow disruption, and costs were not measured. Use of WHO criteria as the reference standard may affect calibration relative to alternative definitions, and lack of blinding to screening scores may have introduced observer or incorporation bias.

The modified STRONGkids tool demonstrated superior discrimination for WHO-defined malnutrition compared with PYMS and stratified downstream risk for hospital-acquired infection and length of stay. Risk gradients were most pronounced among infants younger than 1 year and children with digestive diseases. We operationalize these findings into a pragmatic screening-to-action pathway: STRONGkids scores of 4–5 prompt same-day dietitian assessment, baseline laboratory evaluation, initiation of nutrition support, and re-screening at 48–72 h; scores of 1–3 accompanied by digestive disease or age younger than 1 year are managed as functionally high risk. These steps are feasible on general pediatric wards and provide immediate guidance for targeted escalation and scheduled reassessment. External validation and prospective impact evaluation are warranted.