Abstract：
Although the emergency department (ED) is the initial care setting for the majority of older adults requiring hospital admission, there is a paucity of ED-based dysphagia research in this at-risk population. This is driven by barriers to dysphagia evaluation in this complex care environment. Therefore, we assessed the reliability of trained, non-clinical ED research staff in administering dysphagia screening tools compared to trained speech pathologists (SLPs). We also aimed to determine perceptual screening discrepancies (e.g. voice change) between clinical and non-clinical staff. Forty-two older adults with suspected pneumonia were recruited during an ED visit and underwent dysphagia (Toronto Bedside Swallow Screening Tool; TOR-BSST©) and aspiration (3-oz water swallow test; 3-oz WST) screening by trained non-clinical research staff. Audio-recordings of screenings were re-rated post-hoc by trained, blinded SLPs with discrepancies resolved via consensus. Cohen\'s kappa (unweighted) revealed moderate agreement in pass/fail ratings between clinical and non-clinical staff for both the TOR-BSST© (k = 0.75) and the 3 oz WST (k = 0.66) corresponding to excellent sensitivity and good specificity for both the TOR-BSST (SN = 94%, SP = 85%) and the 3 oz WST (SN = 90%, SP = 81%). Further analysis of TOR-BSST perceptual parameters revealed that most discrepancies between clinicians and non-clinicians resulted from over-diagnosis of change in vocal quality (53%). These results support the feasibility of non-clinical research staff administering screening tools for dysphagia and aspiration in the ED. Dysphagia screening may not necessitate clinical staff involvement, which may improve feasibility of large-scale ED research. Future training of research staff should focus on perceptual assessment of vocal quality.
摘要：
尽管急诊科（ED）是大多数需要住院的老年人的初始护理环境，在这一高危人群中,很少有基于ED的吞咽困难研究.这是由在这种复杂的护理环境中吞咽困难评估的障碍驱动的。因此,我们评估了受过训练的可靠性，与受过训练的言语病理学家（SLP）相比，非临床ED研究人员使用吞咽困难筛查工具。我们还旨在确定临床和非临床工作人员之间的感知筛查差异（例如语音变化）。在ED访视期间招募了42名疑似肺炎的老年人，并由训练有素的非临床研究人员进行了吞咽困难（多伦多床边燕子筛查工具；TOR-BSST©）和误吸（3盎司吞咽试验；3盎司WST）筛查。放映的录音由受过训练的人重新评估，通过共识解决差异的盲化SLP。Cohen的kappa(未加权)显示,TOR-BSST©(k=0.75)和3ozWST(k=0.66)的临床和非临床工作人员之间的通过/失败评级有适度的一致性,对应于TOR-BSST(SN=94%,SP=85%)和3盎司WST(SN=90%，SP=81%)。对TOR-BSST感知参数的进一步分析显示，临床医生和非临床医生之间的大多数差异是由于对声音质量变化的过度诊断（53％）。这些结果支持非临床研究人员在ED中管理吞咽困难和误吸的筛查工具的可行性。吞咽困难筛查可能不需要临床工作人员参与，这可能会提高大规模ED研究的可行性。未来对研究人员的培训应侧重于对声乐质量的感知评估。
