RESUMEN
BACKGROUND & AIMS: Different nutrition screening tools have been proposed to screen inpatients, although, there is still no consensus regarding the reference method for identifying patients at nutritional risk. This retrospective cohort study aimed to evaluate the concurrent and predictive validity of the Malnutrition Screening Tool (MST) and Malnutrition Universal Screening Tool (MUST) to identify inpatients at nutritional risk. METHODS: Data from electronic medical records were collected from adult and elderly inpatients in wards of a public tertiary hospital (Hospital de Clínicas de Porto Alegre, Brazil) over one year. Only the first nutritional evaluation was considered for each patient. GLIM criteria was used for malnutrition status (concurrent validity) and hospital stay length and mortality were outcomes to predictive validity. RESULTS: A total of 5270 patients were included in this study. The mean patient age was 59 ± 16 years old, and 50.7% were males. More than 60% of the patients (65.8%) were admitted to a surgical unit, 63.8% had mild disease-related metabolic stress, 50.7% experienced prolonged hospital stays (more than ten days), and 1.9% of the patients died. Considering the availability of nutritional data, it was possible to perform nutritional screening of 98.1% of patients by MST and 96.7% of patients by MUST. A higher proportion of patients at risk were identified by MUST (53.6%) as compared to MST (21.3%; P < 0.001). MST had 89.9% of Specificity, 60.4% of PPV, 94.6% of NPV, and a moderate agreement with malnutrition by GLIM criteria (κ = 0.591; P < 0.001). A prolonged hospital stay was positively associated with the classification of nutritional risk based on both screening tools. Death was positively associated with nutritional risk by MST (hazard ratio [HR] 2.09; 95% confidence interval [CI] 1.37-3.16) and by MUST (HR 1.79; 95% CI 1.00-3.18) after adjustment based on sex, age, admission type (surgical or clinical), and disease-related metabolic stress. CONCLUSIONS: MST may be good concurrent validity to malnutrition by GLIM criteria as compared to MUST. However, both risk nutritional tools were positively associated with predicting a prolonged hospital stay and mortality.
