Date of Award


Degree Type


Degree Name

Master of Science (MS)


Nutrition Science and Dietetics


David A. Larsen


NRS-2002, nutrition screening, SGA


Background/Objective: Nutritional Risk Screening 2002 (NRS-2002) and Subjective Global Assessment (SGA) are widely used screening tools, but there is no gold standard for identifying nutritional risk. The purpose of this study was to assess the nutritional risk among hospitalized gastrointestinal disease patients, the agreement between NRS-2002 and SGA, and to compare the clinical outcome predicting capacity of them.

Subjects/Method: This study was an analysis of secondary data including 332 patients collected by gastrointestinal department of Peking Union Medical College Hospital (PUMCH). All questions of NRS-2002 and SGA, complications, length of hospitalization stay (LOS), cost, and death were collected. To assess the agreement between the tools, κ statistic was adopted. Before assessing the performance of NRS-2002 and SGA in predicting LOS and cost using linear regression, full and saturated model was compared via the global f-test. The complications and death predicting capacity of tools was assessed using receiver operating characteristic curves.

Results: NRS-2002 and SGA identified nutritional risk at 59.04% and 45.18%. The agreement between two tools was moderate (κ index >0.50) for all age groups except individuals aged ≤ 20, which was slight agreement (κ index 0.087). The saturated model did not improve the outcomes of LOS and cost. There was no significant difference in the association of one step of NRS-2002 and LOS (B=2.127, p=0.002) and the association of one step of SGA and LOS (B=2.296, p=0.001). One step of SGA was associated with a relatively large increase in cost (B=0.272, p=0.001) compared to one step of NRS-2002 (B=0.086, p=0.000), but the difference was not significant. There was no difference of NRS-2002 (infectious complications: 0.615, death 0.810) and SGA (infectious complications: 0.600, death: 0.846) in predicting infectious complication and death, but NRS-2002 (0.738) had larger areas under ROC curve than SGA (0.552) in predicting non-infectious complications.

Conclusion: The prevalence of nutritional risk of hospitalized patients was high. There was moderate agreement between NRS-2002 and SGA for all ages except ≤ 20 age group. NRS-2002 and SGA have similar capacity to predict LOS, cost, infectious complications and death, but NRS-2002 seems to perform better in predicting non-infectious complications.


Open Access



To view the content in your browser, please download Adobe Reader or, alternately,
you may Download the file to your hard drive.

NOTE: The latest versions of Adobe Reader do not support viewing PDF files within Firefox on Mac OS and if you are using a modern (Intel) Mac, there is no official plugin for viewing PDF files within the browser window.