Inter-rater reliability of the Subjective Global Assessment: a systematic literature review

Objectives: Malnutrition continues to be under-recognized and undertreated in the hospitalized setting. Although no "gold standard" for the diagnosis of malnutrition exists, the Subjective Global Assessment (SGA) is a commonly used malnutrition assessment tool. The study aim was to explore the reporting of inter-rater reliability (IRR) of the SGA when used as a nutritional assessment tool in the published literature.

Methods: A comprehensive literature review was performed identifying 119 articles using the SGA once exclusion criteria were applied. Articles were examined for use of IRR and the Cohen κ agreement between examiners.

Results: The IRR of the SGA was reported in only 13% of articles where the SGA was used for nutritional assessment. The κ agreement was highest when examiners were experienced.

Conclusions: The IRR is rarely reported in the published literature and has not yet been reported across experience levels.