Dieticians administered prospectively the ingesta-VVAS in 1762 medical oncology patients. The external validity of the ingesta-VVAS was determined against daily energy intake based on a 24-h dietary recall. Patients had to estimate how they currently ate on a scale from 0 \"nothing at all\" to 10 \"as usual\". Area Under the Receiver-Operating Characteristics (ROC) curve served as determine the optimal cut-off and provide the discriminative power of the tool to detect patients who ingested less or more than 25 kcal kg-1 day-1.
The feasibility of the ingesta-VVAS was 97.7%. The scores were significantly correlated with energy intake (ρ = .67, p < .05), whatever the specific situation (i.e. malnutrition or not). With a cut-off of ≤7, the ingesta-VVAS exhibited a good power discrimination (AUC = .804) to detect patients who ingested less or more than 25 kcal kg-1 day-1, with a sensitivity of 80.8%, a positive predictive value of 83.6%, a specificity of 67.5%, and a negative predictive value of 63.3%. Patients with a score ≤7 on the ingesta-VVAS score were at 12-fold higher probability of nutritional risk [OR 12.3; 95% CI (8.7-17.4); p < .001]. Sensitivity to detect patients with a significant weight loss was 71%, and a positive predictive value of 75.9%.
This easy-to-use ingesta-VVAS is well-correlated with energy intake and may be useful in clinical practice. An ingesta-VVAS score is ≤ 7 could be used to detect patients with nutritional risk of weight loss in medical oncology.