Yale Swallow Protocol於住院病人之可應用性及短版建立

Abstract

研究背景：吞嚥篩檢是安全進食的第一步，Yale Swallow Protocol雖已被臨床使用，但其於跨不同科別的高風險住院病人之可應用性仍未被測試。此外，藉由發展短版Yale Swallow Protocol或可增加其臨床可應用性。 研究目的：一、探討Yale Swallow Protocol使用於跨科別、高風險病房中住院病人（包含：頭頸部癌症、腦中風、帕金森氏症、年齡≥65歲）之可應用性（含不良反應、通過率及花費時間）；二、建立短版Yale Swallow Protocol。 研究方法：以橫斷性研究設計，採連續型收案方式於2024年2月至9月在台大醫院的高風險病房（含：耳鼻喉科、神經內科、復健科及老年醫學科病房）進行收案。所有入院病人符合頭頸部癌症、腦中風、帕金森氏症及年齡≥65歲者都詢問其參與意願。排除條件為目前使用鼻胃管或胃造廔餵食、醫囑禁止由口進食、具飛沫或空氣隔離之疾病者。同意書簽署後，由受過訓練之研究護理師以標準化之Yale Swallow Protocol篩檢所有個案的吞嚥障礙風險，並記錄篩檢後五分鐘內是否有不良反應出現（含血氧濃度低於90%或呼吸速率大於30次/分）及操作Yale Swallow Protocol所需時間。短版建立以項目反應理論檢視各題項的心理計量特性及品質：困難值（代表正確作答該題項所需的能力值）、配適度（評估題項是否符合模型之結構）及測量不變性（檢驗各題項在不同高風險病房之間的難度表現是否一致），保留品質良好之題項。另外，檢視刪除3盎司喝水試驗是否影響未通過率。短版Yale Swallow Protocol的效度評估則以二乘二列聯表及Cohen’s kappa (κ)計算與原始版判讀的一致性。 研究結果：共有502位住院病人同意參與，其中137位來自耳鼻喉科病房、197位來自神經內科病房、39位來自復健科病房及129位來自老年醫學科病房。受試者平均年齡為71.0歲（標準差＝13.22）、男性佔59.8%。502位受試者接受標準化之Yale Swallow Protocol篩檢，不良反應的發生率為零、平均完成時間約為2分鐘45秒。總樣本的未通過率為41.8%，各病房未通過率以老年醫學科病房最高（53.5%）、神經內科病房最低（34.0%）。刪除3盎司喝水試驗後的未通過率僅12.4%，與原始版41.8%有顯著差異，代表3盎司喝水試驗有其必要性。IRT各題項分析結果顯示Yale Swallow Protocol的15個題項中共有九題因困難值相近、配適度不佳或缺乏測量不變性而刪除。據此，提出【短版六題Yale Swallow Protocol】，包含：無法保持警覺、醫囑禁止由口進食、限制臥床時頭部抬高不可超過30度、是否可正確回答地點、配合吐出舌頭及3盎司喝水試驗。比較短版與原始版的未通過率，502位受試者於短版的篩檢結果未通過率為36.1%，略低於原始版的41.8%；進一步比較457位完成所有題數的受試者，發現短版與原始版判讀結果有完美的一致性(κ =1.00)。題項大幅縮減後僅需1547次步驟即可篩檢完502位受試者，相較於原版之6910次步驟，大幅降低臨床人員使用之負荷。 結論：Yale Swallow Protocol為安全、快速、且無天花板效應的吞嚥障礙篩檢工具。【短版六題Yale Swallow Protocol】僅保留原始版半數的題項，可大幅提升篩檢效率，建議未來研究進一步實測其可應用性。

Background: Swallow screening is the first step to ensure safe oral intake. Although the Yale Swallow Protocol has been used in clinical practice and research, its feasibility among high-risk hospitalized adults across different hospital units has not yet been assessed. Furthermore, exploring shortened versions of the Yale Swallow Protocol may enhance screening efficiency and improve its clinical feasibility. Objectives: (1) To examine the clinical feasibility of the Yale Swallow Protocol (i.e., adverse events, failure rates, and administration time) among hospitalized adults in high-risk units (i.e., those with head and neck cancer, stroke, Parkinson’s disease, or aged above 65 years). (2) To develop a shortened version of the Yale Swallow Protocol. Methods: This cross-sectional study applied consecutive sampling and enrolled participants from February to September 2024 in high-risk units (otolaryngology, neurology, rehabilitation, and geriatric units) at National Taiwan University Hospital. Admitted patients diagnosed with head and neck cancer, stroke, Parkinson’s disease, or aged above 65 years were invited to participate. Exclusion criteria included patients with feeding-tube dependency (nasogastric tube or gastrostomy), a nil per os order, and the need for droplet or airborne precautions. After informed consent, a trained research nurse administered the Yale Swallow Protocol to screen all participants for dysphagia risk. Adverse events (defined as oxygen saturation < 90% or respiratory rate > 30 breaths/min) occurring during and within five minutes after screening and administration time were recorded. To explore the shortened version of the Yale Swallow Protocol, the Rasch model was applied to examine item difficulty (the ability level required to endorse an item), fit statistics (item alignment with the model), and measurement invariance (whether item difficulty was consistent across high-risk units). Only items with good psychometric properties would be retained. Additionally, the water challenge was excluded to assess the impact of screening outcomes. The shortened Yale Swallow Protocol was validated through 2×2 contingency tables and Cohen’s kappa (κ) to assess agreement with the original Yale Swallow Protocol. Results: A total of 502 hospitalized adults participated, 137 from otolaryngology, 197 from neurology, 39 from rehabilitation, and 129 from geriatric unit. The mean age was 71.0 years (SD = 13.22), and 59.8% were male. No adverse events occurred during or after administration of the Yale Swallow Protocol. The mean administration time was 2 minutes and 45 seconds. The overall Yale Swallow Protocol failure rate was 41.8%, with the highest rate in the geriatric unit (53.5%) and the lowest in the neurology unit (34.0%). To determine if the protocol could be shortened, the impact of removing the 3-ounce water swallow challenge was first assessed. The failure rate for the protocol without the water challenge decreased from 41.8% to 12.4%. These findings indicate that removing the 3-ounce water swallow challenge would substantially compromise the accuracy of screening outcomes, thus underscoring the necessity of retaining the water challenge within the Yale Swallow Protocol. Additionally, the results of the Rasch model of individual Yale Swallow Protocol items indicated that nine out of 15 items should be removed due to identical or similar difficulty values, poor fit statistics, or non-measurement invariance. The six-item shortened Yale Swallow Protocol was proposed, which included the following items: “Altered Alertness,” “Having a NPO Order”, “Head-of-Bed Elevation < 30,” “State Current Location,” “Follow Instruction: Tongue Sticking Out,” and “3-Ounce Water Swallow Challenge.” The failure rate of the six-item shortened Yale Swallow Protocol was 36.1%, which is slightly lower than the 41.8% rate observed with the original version. The agreement between the shortened and original Yale Swallow Protocol was excellent (kappa = 1.00). The shortened version required only 1,547 operational tasks, compared with 6,910 tasks for the original version, substantially reducing the clinical workload. Conclusion: The Yale Swallow Protocol is a safe, quick, and ceiling-free dysphagia screening tool. The six-item shortened Yale Swallow Protocol retains half of the original items while greatly improving screening efficiency. Future research is recommended to further evaluate the clinical feasibility of this shortened version.