Screening performance of a 100-mL water swallowing test in community-dwelling older adults: A receiver operating characteristic analysis.

Abstract

The development of a dysphagia screening test is an urgent issue in the field of frailty prevention among community-dwelling older people. The purpose of this study was to evaluate the screening performance of a 100-mL water swallowing test (WST). The study employed a cross-sectional design. Participants were 304 (65 men and 239 women, mean age = 80 years) Japanese community-dwelling older adults aged over 65 years. We investigated swallowing disorder using the 10-item Eating Assessment Tool (EAT-10), and compared choking signs, swallowing time and number of swallows, and their combination in the 100-mL WST. We calculated the sensitivity and specificity of these indices. Receiver operating characteristic (ROC) curve analysis was used to determine the optimal cut-off value of swallowing time and number of swallows in the 100-mL WST based on the Youden Index among participants without choking signs. The sensitivity and specificity of choking signs in the 100-mL WST were 20% and 91%, respectively. The discriminating ability of swallowing time and number of swallows among participants without any choking signs was .76 and .72, respectively, in the area under the ROC curve. Diagnostic sensitivity and specificity to discriminate dysphagia from normal swallowing ability were 65% and 74% when the cut-off was >10 s based on maximisation of the Youden Index. The 100-mL WST performed best when the indices of choking signs and swallowing time were combined, with a sensitivity and specificity of 72% and 68%, respectively. The 100-mL WST would be an adequate screening tool when compared to the EAT-10.