Inter-rater reliability and clinical relevance of subjective and objective interpretation of videofluoroscopy findings

Abstract

<p>BACKGROUND: Dysphagia is commonly evaluated using videofluoroscopy (VFS). As its ratings are usually subjective normal-abnormal ratings, objective measurements have been developed. We compared the inter-rater reliability of the usual VFS ratings to the objective measurement VFS ratings and evaluated their clinical relevance.<br></p><p>METHODS: Two blinded raters analyzed the subjective normal-abnormal ratings of 77 patients' VFS. Two other blinded raters analyzed the objective measurements of pharyngeal aerated area with bolus held in the oral cavity (PAhold), the pharyngeal area of residual bolus during swallowing (PAmax), the pharyngeal constriction ratio (PCR), the maximum pharyngoesophageal segment opening (PESmax), pharyngoesophageal segment opening duration (POD), airway closure duration (ACD), and total pharyngeal transit time (TPT). We evaluated the inter-rater agreement in the subjective ratings and the objective measurements. Clinical utility analysis compared the measurements with the VFS findings of pharyngeal phase abnormality, penetration/aspiration, and cricopharyngeal relaxation.<br></p><p>RESULTS: In the pharyngeal findings, the subjective analysis inter-rater agreement was mainly moderate to strong. The strongest agreements were on the pharyngeal residues and penetration/aspiration findings. The objective measurements had fair to good inter-rater agreement. Clinical utility analysis found statistically significant connections between TPT and pharyngeal phase abnormality, normal PCR and lack of penetration/aspiration, and normal PESmax and normal cricopharyngeal relaxation.<br></p><p>CONCLUSIONS: The subjective analysis had moderate to strong inter-rater agreement in the pharyngeal VFS findings, especially concerning pharyngeal residues and penetration/aspiration detection, reflecting the efficacy and safety of swallowing. The objective measurements had fair to good inter-observer reproducibility and could thus improve the reliability of VFS diagnostics.<br></p><p>LEVEL OF EVIDENCE: 4.<br></p>