Validation and Reliability of the Polish Version of the Eating Assessment Tool-10 Questionnaire

Introduction: Early identification of patients at risk of dysphagia is of paramount importance. To date, no dysphagia screening questionnaire has been validated and translated into Polish that can be widely used in a multidisciplinary setting. Our study aimed to validate and adapt the Polish version of the Eating Assessment Tool-10 questionnaire (EAT-10). Materials and Methods: The EAT-10 questionnaire was translated into Polish using a formal forward-backward translation method. The Polish EAT-10 was administered to 109 patients with different dysphagia etiologies and 24 control subjects. Internal consistency, unidimensionality, test-retest reliability and external validity against the Visual Analog Scale (VAS), and Flexible Endoscopic Evaluation of Swallowing (FEES) were performed. Results: The EAT-10PL demonstrated excellent internal consistency (Cronbach's α = 0.958) and confirmed unidimensionality. We found a strong correlation between EAT-10PL with the Visual Analog Scale (VAS) (rs = 0.94, p < 0.001) and a weaker correlation with the PAS (rs = 0.55, p < 0.001). We reported a sensitivity of 79.5% and specificity at the level of 60.0%, using ≥3 as a cut-off point. The statistically chosen cutoff point for PAS ≥ 2 and EAT-10 ≥ 6 indicated optimal specificity (70.0%) and sensitivity (79.5%) of measurements performed using EAT-10PL. The EAT-10PL questionnaire demonstrates high discriminatory ability relative to the control group (F(4, 104) = 16.219, p < 0.001, η² = 0.38 [95%CI: 0.22-0.48]). Conclusions: The Polish EAT-10 is a valid and reliable, self-administered questionnaire for dysphagic patient identification. The Polish EAT-10 ≥ 3 can be considered abnormal; however, it seems that EAT-10PL is not appropriate for patients with dysphagia and a chronic cough background, and further research is required.