Abstract TP367: Dysphagia Screening Tool Validation Study

Abstract

Introduction: Dysphagia occurs in roughly 30 to 67% of stroke patients and can lead to significant complications, including aspiration with subsequent pneumonia. There has been a lack of agreement on a single effective dysphagia screening tool. The tool used in our institution was developed internally based on extensive literature review, and includes dysphagia screening items followed by a water challenge. Hypothesis: We suspected this dysphagia screening tool was performing well based on low rates of aspiration pneumonia in our stroke patients. The purpose of this study was to validate the dysphagia screening tool against gold standard video fluoroscopy (VF). Methods: This was a prospective study to determine the sensitivity and specificity of the dysphagia screening tool. Our enrollment goal was 300 patients with suspected ischemic stroke between December 2016 and January 2018. The primary nurse completed the routine dysphagia screen, and the sub-investigator performed a blinded simultaneous screen, followed by VF evaluation. We used a group sequential design with interim analyses every 30 patients to allow for early stopping. The primary hypothesis was sensitivity >85%, evaluated using a one-sided test of proportion at a study alpha level of 0.025. Results: At the second interim analysis, 34/61 patients were positive for dysphagia on VF, while only 19 were positive on the screening tool. Estimated sensitivity was 55.9% (95% CI 39.5-71.1%), with a p-value below the second-stage threshold of 0.000002. Therefore, we terminated the study at this point. Estimated specificity was 51.9% (95% CI 34.0-69.3%), and the concordance between examiners for dysphagia screening as 91.8%. Conclusions: This is a routinely used screening tool across multiple hospitals within this organization. The tool did not show adequate sensitivity or specificity compared against the gold standard. Despite the results, we continue to have low rates of aspiration pneumonia. It is possible that VF may overestimate dysphagia, or that subjective interpretation of dysphagia criteria with VF may complicate the results. In conclusion, due to inadequate sensitivity and specificity of the dysphagia screening tool, we terminated the study and recommended modification of the tool.