South of the UES: Improving the ability of speech-language pathologists to detect oesophageal abnormalities during videofluoroscopy swallowing studies.

Purpose: With two-thirds of adults presenting for a videofluoroscopy swallow study (VFSS) with oesophageal abnormalities, it seems prudent to include visualisation of the oesophagus, in the context of the entire swallow process, to provide further information to the diagnostic team. This study aims to evaluate the ability of speech-language pathologists (SLPs) to interpret oesophageal sweep on VFSS and the relative improvement in that ability with additional training. Method: One hundred SLPs attended training in oesophageal visualisation during VFSS, based on a previous study. Ten oesophageal sweep videos (five normal, five abnormal) with one 20 ml thin fluid barium bolus (19% w/v) were presented at baseline and following training. Raters were blinded to patient information other than age. Binary ratings were collected for oesophageal transit time (OTT), presence of stasis, redirection, and referral to other specialists. Result: Inter-rater reliability as measured by Fleiss' kappa improved for all parameters, reaching statistical significance for OTT (pre-test kappa = 0.34, post-test kappa = 0.73; p  < 0.01) and redirection (pre-test kappa = 0.38, post-test kappa = 0.49; p  < 0.05). Overall agreement improved significantly ( p  < 0.001) for all parameters except stasis, where improvement was only slight. Interaction between pre-post and type of video (normal/abnormal) was statistically significant ( p  < 0.001) for redirection, with a large pre-post increase in positive accuracy compared with a slight pre-post decrease in negative accuracy. Conclusion: Findings indicate that SLPs require training to accurately interpret an oesophageal sweep on VFSS. This supports the inclusion of education and training on both normal and abnormal oesophageal sweep patterns, and the use of standardised protocols for clinicians using oesophageal visualisation as part of the VFSS protocol.