12. Which cut-off scores for nutritional triage recommendations should I use when using the PG‑SGA Short Form©?
There is no clear indication that the cut-off values for the PG‑SGA Short Form© should differ from those used in the full version of the PG‑SGA. Although theoretically a patient might possibly be slightly underscored if using the PG-SGA SF© only, but not overscored, it is important to note that the majority (generally 80-90% of a score for a given patient) of the total PG‑SGA numerical score is being based on the patient-generated aspects (PG‑SGA Short Form©).
In general, it would be rather unlikely that a patient’s scoring would be low on the PG-SGA Short Form© and high on the professional-generated items (Worksheet 2 to 4).
Research utilizing the PG-SGA is ongoing globally with studies published on the use of the PG‑SGA Short Form©, also known as the ‘abridged’ version of the PG-SGA in the literature.
The study by Vigano et al. (Vigano AL, di Tomasso J, Kilgour RD, Trutschnigg B, Lucar E, Morais JA, Borod M. The Abridged Patient-Generated Subjective Global Assessment Is a Useful Tool for Early Detection and Characterization of Cancer Cachexia. J Acad Nutr Diet 2014;114(7):1088-98) which was prospectively performed in 207 advanced lung and gastrointestinal cancer patients, showed that ≥9 points as generated by the PG-SGA Short Form© was associated with unfavorable biological markers of cancer cachexia, decreased anthropometric and physical measures, such as body mass index, fat mass, handgrip and leg strength, an average 12% greater length of hospital stay, a dose reduction in chemotherapy, and increased mortality.
The study by Gabrielson et al. (Gabrielson DK, Scaffidi D, Leung E, Stoyanoff L, Robinson J, Nisenbaum R, Brezden-Masley C, Darling PB. Use of an abridged scored Patient-Generated Subjective Global Assessment (abPG-SGA) as a nutritional screening tool for cancer patients in an outpatient setting. Nutr Cancer 2013;65(2):234-9) performed in 90 oncology outpatients showed an area under curve of 0.956 by the PG‑SGA Short Form. In this study, ≥6 points, and ≥7 points were studied as cut-off values for sensitivity and specificity of the PG‑SGA SF© compared to the PG‑SGA Categories (A-Well nourished versus B/C-Moderately malnourished or suspected malnutrition/Severely malnourished). Using the cut-off of ≥6 points resulted in 93.8 sensitivity and 77.6 specificity. Using a cut-off value of ≥7 points improved the specificity to 89.7, but lowered the sensitivity to 84.4.
The optimal cut-off values to be used depend on the purpose or the objectives of the team using the instrument in clinical practice, e.g. aiming for prevention or treatment of malnutrition only; to support expanding or maintaining dietitian or nutritional support services; and/or various types of clinical and outcomes research. Therefore, each organization can make its own decisions on defining which cut-off values should result in referral to a dietitian and other disciplines, but in the context of research, until there are data to support changing the triage approach, the triage as described in the full PG‑SGA© should be maintained.