A study was conducted to validate a machine learning tool designed to predict the risk of dysphagia (difficulty swallowing) in hospitalized patients, comparing its results with clinical evaluations.
A total of 149 inpatients in the ENT department were assessed over three weeks, showing the algorithm's performance with an AUROC score of 0.97 and an accuracy of 92.6%.
Findings indicated that older age, male sex, and oropharyngeal malignancies increased the likelihood of being at risk for dysphagia, suggesting the tool could greatly aid in identifying at-risk patients in clinical settings with low dysphagia awareness.