Nutritional assessment: comparison of clinical assessment and objective variables for the prediction of length of hospital stay and readmission.

BACKGROUND: Nutritional assessment commonly includes multiple nutrition indicators (NIs). To promote efficiency, a minimum set is needed for the diagnosis of malnutrition in the acute care setting. OBJECTIVE: The objective was to compare the ability of different NIs to predict outcomes of length of hospital stay and readmission to refine the detection of malnutrition in acute care. DESIGN: This was a prospective cohort study of 1022 patients recruited from 18 acute care hospitals (academic and community), from 8 provinces across Canada, between 1 July 2010 and 28 February 2013. Participants were patients aged ≥18 y admitted to medical and surgical wards. NIs measured at admission were subjective global assessment (SGA; SGA A = well nourished, SGA B = mild or moderate malnutrition, and SGA C = severe malnutrition), Nutrition Risk Screening (2002), body weight, midarm and calf circumference, serum albumin, handgrip strength (HGS), and patient-self assessment of food intake. Logistic regression determined the independent effect of NIs on the outcomes of length of hospital stay (<7 d and ≥7 d) and readmission within 30 d after discharge. RESULTS: In total, 733 patients had complete NI data and were available for analysis. After we controlled for age, sex, and diagnosis, only SGA C (OR: 2.19; 95% CI: 1.28, 3.75), HGS (OR: 0.98; 95% CI: 0.96, 0.99 per kg of increase), and reduced food intake during the first week of hospitalization (OR: 1.51; 95% CI: 1.08, 2.11) were independent predictors of length of stay. SGA C (OR: 2.12; 95% CI: 1.24, 3.93) and HGS (OR: 0.96; 95% CI: 0.94, 0.98) but not food intake were independent predictors of 30-d readmission. CONCLUSIONS: SGA, HGS, and food intake were independent predictors of outcomes for malnutrition. Because food intake in this study was judged days after admission and HGS has a wide range of normal values, SGA is the single best predictor and should be advocated as the primary measure for diagnosis of malnutrition. This study was registered at clinicaltrials.gov as NCT02351661.