Malnutrition is common in hospitals and it is important to implement an appropriate nutrition screening tool to identify patients at risk. The aim of the study was to assess the sensitivity and specificity of the malnutrition screening tool developed by the Malnutrition Advisory Group of the British Association of Parenteral and Enteral Nutrition against subjective global assessment in hospitalised patients with cancer. A cross-sectional study assessing the risk of malnutrition and nutritional status of sixty-five hospitalised patients with cancer, aged 56 +/- 15 years. According to subjective global assessment, 25 % of patients were well nourished and 75% were malnourished (63% were moderately or suspected of being malnourished and 12% severely malnourished). The malnutrition screening tool had a low sensitivity of 59% and a specificity of 75%. The positive predictive value was 88% and the negative predictive value 38%. There were significant linear trends between subjective global assessment classification and percentage weight loss in the previous six months (P < 0.001) and body mass index (P = 0.007). The malnutrition screening tool developed by the Malnutrition Advisory Group of the British Association of Parenteral and Enteral Nutrition is not a suitable screening tool for detecting risk of malnutrition in hospitalised patients with cancer.