INTER-RATER AGREEMENT OF THE BRIEF BEDSIDE DYSPHAGIA SCREENING TEST-REVISED IN PATIENTS WITH STROKE

Abstract

Aim: The aim of the study was to investigate the inter-rater agreement (IRA) of the 8-item Brief Bedside Dysphagia Screening Test-Revised (BBDST-R). Design: An observational IRA study was conducted. Methods: Forty-six patients with stroke were independently assessed by two nurse raters using the BBDST-R. Rater agreement was described and analysed using descriptive statistics, Cohen’s kappa (κ), the proportion of observed agreement (Po), the prevalence index (p-index), and the bias index (b-index). Results: Kappa ranged from -0.046 (p = 0.641; 95% CI = -0.126–0.034) for item “ability to clench the teeth” to 0.784 (p < 0.001; 95% CI = 0.377–1.000) for “thick liquid: cough”. For the overall BBDST-R result, κ was 0.241 (p = 0.114; 95% CI = 0.000–0.558). The Po was ≥ 0.80 for five items; the p-index was ≥ 0.80 for three items. The b-index ranged from 0.03–0.31. Conclusion: The IRA, as expressed by κ, was low for the overall result and variable for the individual items. However, for some items, κ may have misrepresented the IRA due to the high pindex and bindex. Several strategies are recommended to improve the IRA of the instrument. This should enhance the instrument’s capacity to produce consistent results across raters.