The videofluoroscopic swallow study (VFSS) is the gold standard diagnostic tool to evaluate oropharyngeal dysphagia. Although objective measurements on VFSS have been described, there is no universal method of analysis, and the majority of clinicians use subjective interpretation alone. The purpose of this investigation was to evaluate the accuracy of subjective VFSS analysis.Study Design
Tertiary care laryngology center.Subjects and Methods
Seventy-six de-identified videos from VFSS evaluations of patients with dysphagia were presented to blinded, experienced speech-language pathologists and laryngologists individually. Evaluators rated each video as normal or abnormal for hyoid elevation (HE), pharyngeal area (PA), pharyngeal constriction ratio (PCR), and pharyngoesophageal segment opening (PESo). A blinded investigator assessed evaluators’ inter- and intrarater agreement and compared their responses to objectively measured results for these parameters to examine accuracy.Results
Evaluators correctly classified only 61.5% of VFSS videos as normal or abnormal, with moderate interrater agreement (κ = 0.48, P < .0001). Intrarater agreement was highly variable (κ = 0.43-0.83). Accuracy was greatest for PCR (71.6%), with poorer performance for HE (61.3%), PESo (59.2%), and PA (45.3%). Interrater agreement was moderate for all parameters, with greater concordance for PCR (κ = 0.59) and PESo (κ = 0.54) and less for HE (κ = 0.40) and PA (κ = 0.44). Evaluators unanimously agreed on a correct interpretation of a VFSS only 28% of the time.Conclusion
Subjective assessment of VFSS parameters is inconsistently accurate when compared with objective measurements, with accuracy ratings ranging from 45.3% to 71.6% for specific parameters. Inter- and intrarater reliability for subjective assessment was moderate and highly variable.