Reliability and Validity of the Greek Version of Patient Generated-Subjective Global Assessment in Cancer Patients

Abstract

The authors sought to determine the reliability and validity of psychometric properties of the Greek Patient-Generated Subjective Global Assessment (PG-SGA) in cancer patients attending a palliative care unit. The scale was translated into Greek. The study took place in an outpatient palliative care unit and 238 cancer patients completed the nutrition assessment. For treatment effect, the index was administered to 164 patients, whereas for the test of the instrument's robustness to 100 patients. Both assessments took place 15 days after the baseline assessment. Ninety-three patients were well-nourished, 104 were moderately malnourished, and 41 were severely malnourished. Reliability was assessed by the internal consistency and test-retest of the instrument. Validity was assessed with construct validity using the PG-SGA items, treatment effect, and criterion validity. Weight loss, physical examination, activities and function, and symptoms were predictors to the classification. Mean values in male patients classified as severely malnourished had the lowest values in all objective parameters except serum albumin, whereas severely malnourished female patients had the lowest values in all objective parameters except serum albumin and arm muscle circumference (AMC). Values of serum transferrin, AMC, and body mass index were significantly lower in patients classified as severely malnourished. The Greek PG-SGA is a psychometrically sound assessment in Greek cancer patients.