BACKGROUND & AIMS: We evaluated the ability of Nutritional Risk Screening 2002 (NRS 2002) and Subjective Global Assessment (SGA) to predict malnutrition related to poor clinical outcomes. METHODS: We assessed 705 patients at a public university hospital within 48 h of admission. Logistic regression and number needed to screen (NNS) were calculated to test the complementarity between the tools and their ability to predict very long length of hospital stay (VLLOS), complications, and death. RESULTS: Of the patients screened, 27.9% were at nutritional risk (NRS+) and 38.9% were malnourished (SGA B or C). Compared to those patients not at nutritional risk, NRS+, SGA B or C patients were at increased risk for complications (p=0.03, 0.02, and 0.003, respectively). NRS+ patients had an increased risk of death (p=0.03), and SGA B and C patients had an increased likelihood of VLLOS (p=0.008 and p<0.0001, respectively). Patients who were both NRS+ and SGA C had lower estimates of NNS than patients who were NRS+ or SGA C only, though their confidence intervals did overlap. CONCLUSIONS: The concurrent application of SGA in NRS+ patients might enhance the ability to predict poor clinical outcomes in hospitalized patients in Brazil.
BACKGROUND & AIMS: We evaluated the ability of Nutritional Risk Screening 2002 (NRS 2002) and Subjective Global Assessment (SGA) to predict malnutrition related to poor clinical outcomes. METHODS: We assessed 705 patients at a public university hospital within 48 h of admission. Logistic regression and number needed to screen (NNS) were calculated to test the complementarity between the tools and their ability to predict very long length of hospital stay (VLLOS), complications, and death. RESULTS: Of the patients screened, 27.9% were at nutritional risk (NRS+) and 38.9% were malnourished (SGA B or C). Compared to those patients not at nutritional risk, NRS+, SGA B or C patients were at increased risk for complications (p=0.03, 0.02, and 0.003, respectively). NRS+ patients had an increased risk of death (p=0.03), and SGA B and C patients had an increased likelihood of VLLOS (p=0.008 and p<0.0001, respectively). Patients who were both NRS+ and SGA C had lower estimates of NNS than patients who were NRS+ or SGA C only, though their confidence intervals did overlap. CONCLUSIONS: The concurrent application of SGA in NRS+ patients might enhance the ability to predict poor clinical outcomes in hospitalized patients in Brazil.
Authors: Helena Orell-Kotikangas; Pia Österlund; Kauko Saarilahti; Paula Ravasco; Ursula Schwab; Antti A Mäkitie Journal: Support Care Cancer Date: 2014-11-05 Impact factor: 3.603
Authors: L M Kuroki; M Mangano; J E Allsworth; C O Menias; L S Massad; M A Powell; D G Mutch; P H Thaker Journal: Ann Surg Oncol Date: 2014-09-05 Impact factor: 5.344