背　景：吞水測驗（water swallow test）是咀嚼吞嚥困難之評估工具之一，護理人員能以吞水測驗以了解腦中風吞嚥困難病人吸入性問題的風險。綜觀國內外的研究，並沒有針對吞水測驗的精確度分析一致性的結果。目　的：本研究進行腦中風吞嚥困難病人接受吞水測驗之統合分析，以探討吞水測驗之精確度。方　法：採系統性回顧法，使用「腦中風吞嚥困難病人（stroke with deglutition disorders）」、「吞水測驗（water swallow test）」、「內視鏡吞嚥檢查（fiber-optic endoscopic examination）、錄影螢光吞嚥檢查（video-fluoroscopic swallowing study）」、「敏感度（sensitivity）」、「特異度（specificity）」、「診斷型（diagnostic research）」為關鍵字，搜尋Cochrane Library、MEDLINE、CINAHL、ProQuest、PubMed各資料庫，2000年至2014年12月符合納入準則進行文獻搜尋。研究品質以Cochrane handbook Quality Assessment of Diagnostic Accuracy Studies(QUADAS)tool做評讀，使用Stata13.1進行統合分析。結　果：評讀後納入11篇品質良好的研究，共787位個案。匯集敏感度為0.65（95%信賴區間［0.57, 0.73］），匯集特異度為0.78（95%信賴區間［0.70, 0.85］），合併診斷比值比（diagnostic odds ratio）為6.56（95%信賴區間［4.89, 8.80］），≥ 50 ml吞水測驗測試比＜ 50 ml之敏感度更顯著（p=.002）。結　論：吞水測驗在腦中風吞嚥困難病人的吞嚥評估發揮了重要作用，≥ 50 ml之吞水測驗帶來更高的敏感度。

Background: The water swallow test (WST) is the easiest bedside screening tool for nurses to use to detect aspiration risk in patients with stroke. No meta‐analysis has yet been conducted of the results of the various studies that have examined the efficacy of the WST. Objective: This study summarizes the findings of previous studies and performs a systematic review with meta-analysis in order to assess the diagnostic accuracy of the WST in detecting aspiration risk in patients with stroke. Methods: Keywords that included stroke with deglutition disorders, water swallow test, fiber‐optic endoscopic examination, video‐fluoroscopic swallowing study, sensitivity, specificity, and diagnostic research design were used to query eleven electronic databases: Cochrane library, Medline, CINAHL, ProQuest, PubMed database. Articles indexed on these databases that met the inclusion criteria and that were published between 1990 and December 2014 were extracted for analysis. A standardized critical appraisal tool from the Quality Assessment of Diagnostic Accuracy Studies (QUADAS) tool by Cochrane handbook was used to assess methodological quality. Stata 13.1 software was used to conduct the meta-analysis. Results: The main outcome was the diagnostic accuracy of the WST in detecting aspiration risk in patients with stroke. There were 787 stroke patients in the 11 articles that were included in the meta‐analysis. The pooled sensitivity was 0.65 (95% confidence interval [CI]=[0.57, 0.73]); the pooled specificity was 0.78 (95% CI [0.70, 0.85]); and the pooled diagnostic odds ratio was 6.56 (95% CI [4.89, 8.80]). A significant difference in sensitivity (p=.002) was detected between more (≥ 50 ml) WST and less (＜ 50 ml) WST. Conclusions: The diagnostic accuracy of the WST for detecting aspiration risk in patients with stroke plays an important role in screening tests, with the ≥ 50 ml volume WST providing higher levels of sensitivity.