Ultrasound: Reliability of a Pocket-Sized System in the Assessment of Swallowing

Abstract

Purpose Ultrasound imaging offers a noninvasive adjunct to clinical swallowing assessment. Published reliability of sophisticated ultrasound systems is promising; however, no data exist for reliability using more affordable, pocket-sized devices. This study explored intrarater, interrater, and test–retest reliability of swallowing measures acquired with pocket-sized ultrasound technology. Method Five participants collected measures of swallowing from 20 healthy individuals using the Clarius ultrasound. Hyoid excursion and thyrohyoid approximation were derived during saliva, liquid, and puree swallowing. The cross-sectional area of the floor of mouth muscles and tongue thickness were obtained at rest. Measures were collected at two occasions minimum 11 days apart. Reliability was assessed for the entire process of data acquisition including scanning and online measurement, and for offline measurement of saved images. Results For most measures, reliability was poor (ICC [intraclass correlation coefficient] < .50) to moderate (ICC = .50–.75) for the entire process of data acquisition and poor to good (ICC > .75) when measuring saved images. Conclusion Further work is needed to elucidate whether our study findings apply to the Clarius system only or the data suggest a general limitation of pocket-sized ultrasound technology.

Swallowing assessment in patients with dysphagia: Validity and reliability of a pocket‐sized ultrasound system

“…Good intra-and interrater reliability of offline measurement corresponds to published measurement reliability data of laptop-driven ultrasound instrumentation in the assessment of hyoid excursion (Chen et al, 2017). Yet it is not in agreement with the reported insufficient reliability of pocket-sized ultrasound technology in healthy swallowing (Winiker et al, 2021b). Interrater reliability of immediate online measurement on the iPad was insufficient for most measures.…”

“…The first of the two measurement occasions took place no sooner than 27 days after online data collection to avoid recall. The second offline measurement was performed 11 days after the first offline rating (Winiker et al., 2021b ). Each participant was blinded to their own and the other's measurements.…”

“…Intra‐ and interrater reliability for the entire process of data acquisition was poor (intraclass correlation coefficient (ICC) < 0.38) for hyoid excursion and thyrohyoid approximation and moderate (ICC = 0.56–0.70) for tongue thickness. Lower image quality of pocket‐sized ultrasound technology compared with larger instrumentation is one of the different potential explanations for these findings (Winiker et al., 2021b ). Data for impaired swallowing are needed since reliability depends, in part, on the homogeneity of the sample (Bartlett & Frost, 2008 ).…”

Translating Ultrasound into Clinical Practice for the Assessment of Swallowing and Laryngeal Function: A Speech and Language Pathology-Led Consensus Study

“…Further, reliable and valid assessment is related to assessment protocols and equipment selection; for example, reliability and validity may be impacted by the choice of US device. While there are some data to suggest that standard-sized US equipment may provide reliable and valid swallowing measures [ 5 , 12 , 16 ], this could not be confirmed for pocket-sized technology [ 24 , 60 ]. The method of data acquisition, such as the use of hand-held versus fixed transducers, may further affect reliability [ 61 , 62 ]; as might variability between different data collection settings, for example, data collected in the clinical versus research environment.…”

“…Operator training has shown to improve the reliability in the measurement of static images [ 67 ] and an association between reliability and operator skills has been documented for procedures including VFSS [ 68 ], high-resolution manometry [ 69 ] and muscle US data acquisition by physiotherapists [ 70 ]. The relationship between reliability and training is, however, reciprocal, and reliability data are required to determine the necessity and content of training protocols [ 60 ].…”