Utility of the EAT-10 in the detection of dysphagia in high-risk hospitalisation units at a university hospital: a cross-sectional study.

Introduction: the objective was to assess the utility of the Eating Assessment Tool (EAT-10) in hospitalisation units with patients at high risk of dysphagia. Patients and methods: a cross-sectional study was conducted in the Neurology and Internal Medicine wards; patients with admission < 24 hours and in a terminal stage of disease were excluded. In the first 24-48 hours of admission the presence of dysphagia as assessed with the EAT-10, the risk of malnutrition as assessed with the Malnutrition Universal Screening Tools (MUST), and comorbidities using the Charlson index were screened. Results: a total of 169 patients were recruited (76.0 years, 52 % women); 19.5 % were at risk of malnutrition. The EAT-10 instrument could be administered in 80.6 % of the patients, and was positive in 26.6 % (women 34.1 % vs. men 18.4 %; p = 0.025). When comparing patients with higher comorbidity with those with a lower Charlson index, a lower response rate to EAT-10 was observed (78.4 % vs. 93.9 %; p = 0.038), without differences in screening positivity (28.3 % vs. 19.4 %; p = 0.310). The prevalence of dysphagia risk was higher in the Internal Medicine unit than in the Neurology unit (30.4 % vs. 19.6 %; p = 0.133), as was the percentage of cases in which screening could not be performed (21.1 % vs. 11.1 %; p = 0.011). There were no significant differences in risk of malnutrition, mortality, hospital stay, or readmission according to the EAT-10. Conclusions: The EAT-10 has limited utility in the studied hospitalisation units due to a high rate of unfeasible tests, especially among patients at higher risk of dysphagia.