Validity and Reliability of the Eating Assessment Tool (EAT-10)

Abstract

Objectives: The Eating Assessment Tool is a self-administered, symptom-specific outcome instrument for dysphagia. The purpose of this study was to assess the validity and reliability of the 10-item Eating Assessment Tool (EAT-10). Methods: The investigation consisted of 4 phases: 1) line-item generation, 2) line-item reduction and reliability, 3) normative data generation, and 4) validity analysis. All data were collected prospectively. Internal consistency was assessed with the Cronbach alpha. Test-retest reliability was evaluated with the Pearson product moment correlation coefficient. Normative data were obtained by administering the instrument to a community cohort of healthy volunteers. Validity was assessed by administering the instrument before and after dysphagia treatment and by evaluating survey differences between normal persons and those with known diagnoses. Results: A total of 629 surveys were administered to 482 patients. The internal consistency (Cronbach alpha) of the final instrument was 0.960. The test-retest intra-item correlation coefficients ranged from 0.72 to 0.91. The mean (±SD) EAT-10 score of the normal cohort was 0.40 ± 1.01. The mean EAT-10 score was 23.58 ± 13.18 for patients with esophageal dysphagia, 23.10 ± 12.22 for those with oropharyngeal dysphagia, 9.19 ± 12.60 for those with voice disorders, 22.42 ± 14.06 for those with head and neck cancer, and 11.71 ± 9.61 for those with reflux. The patients with oropharyngeal and esophageal dysphagia and a history of head and neck cancer had a significantly higher EAT-10 score than did those with reflux or voice disorders (p < 0.001). The mean EAT-10 score of the patients with dysphagia improved from 19.87 ± 10.5 to 5.2 ± 7.4 after treatment (p < 0.001). Conclusions: The EAT-10 has displayed excellent internal consistency, test-retest reproducibility, and criterion-based validity. The normative data suggest that an EAT-10 score of 3 or higher is abnormal. The instrument may be utilized to document the initial dysphagia severity and monitor the treatment response in persons with a wide array of swallowing disorders.