Impact of quantitative objective videofluoroscopic swallowing measures on analysis and recommendations in a clinical context

Abstract

Background: Quantitative objective measures are available for use in adult videofluoroscopic swallowing study (VFSS) analysis but are yet to be used routinely in clinical settings. The present study describes the introduction of quantitative objective measures to one regional hospital and investigated the impact implementation of objective measurements in VFSS had on diagnosis and recommendations. Methods: One hundred adults (mean 72.3 years, range 34-93, SD 13.7) consecutively referred for a VFSS over 13 months were recruited. Adults presented with varying likely contributors to their dysphagia: 21% nil diagnosis, 17% stroke, 13% gastroenterology, 12% other neurology, 10% respiratory, 8% head and neck cancer, 8% other otorhinolaryngology (ORL), 6% other and 5% cardiology. All adults received a standardised VFSS protocol (1ml, 3ml, 20ml, 100ml liquid barium plus 5ml barium paste) at 30 frames per second. All VFSS were analysed quantitatively using timing and displacement measures developed by Leonard and Kendall (1997) by an independent rater. All VFSS were also analysed separately using a traditional observational approach by a different rater. Both independent raters were provided with age, gender and primary aetiology but no further clinical information. Both independent rater reports were then compared to the reports completed by the patient’s treating speech-language therapist (SLT). Ten percent of VFSS were objectively reanalysed to check intra-rater reliability. Ten percent of VFSS were double-blind rated by a year two masters student to check inter-rater reliability. Results: Identification of impairments in pharyngeal constriction, hyoid maximum displacement and pharyngoesophageal segment maximum opening achieved only slight to moderate inter-rater reliability across raters (Kappa (K)=0-0.457). In comparison, intra-rater and inter-rater reliability between objective raters achieved excellent agreement (K=0.767-0.998, 0.773-0.926 respectively). Incidence of impairment across both displacement and timing measures were lower in treating SLTs compared with objective independent raters (p<.05). The independent subjective rater identified fewer patients with hyoid and pharyngoesophageal segment (PES) impairment than the independent objective rater. Sensitivity for identification of physiological impairment by the independent subjective rater and treating SLTs when calculated against the independent objective rater as gold standard was variable (Pharyngeal constriction=72%, 36%, PES displacement=28%, 13%, hyoid displacement=55%, 26% respectively). Specificity for identification of physiological impairment by the independent subjective rater and treating SLTs was stronger (Pharyngeal constriction=74%, 83%, PES displacement=82%, 90%, hyoid displacement=68%, 82%). Rehabilitation exercises where recommended significantly less often by the treating SLTs than the independent objective rater (p<.05) and the independent subjective rater (p<.05). Percentage agreement was calculated between the independent objective rater and both the independent subjective rater and the treating SLTs (Effortful swallows=69%, 58%, Masakos=69%, 75%, Shaker=54%, 39%, Mendelsohn manoeuvre=54%, 35% respectively). The most common compensatory recommendation provided by the treating SLTs was diet modifications both fluids and diet (151 times) followed by postural changes (50 times). Conclusion: This study demonstrated that the introduction of objective quantitative measures in VFSS increased inter-rater agreement and accuracy of physiological impairment identification. . This influenced recommended rehabilitation exercises and raters using quantitative measures suggested more targeted rehabilitation than those using subjective assessment.