Driven evidence-based nursing to diagnostic accuracy of water swallow test for stroke patients with dysphagia

Background: The ability of nursing staff to determine stroke deglutition disorders is important in care of stroke patients. Purpose: This study aimed to increase diagnostic accuracy of stroke deglutition disorders among nursing staff in stroke care. Method: By comprehensive determination of the most accurate water volume for the diagnosis of stroke deglutition disorders using evidence-based meta-analysis. Results: The diagnostic accuracy of stroke deglutition disorders made by nursing staff was increased from 4.75% to 100%, and the incidence of aspiration pneumonia in patients with stroke was decreased from 2.59‰ to 1.36‰. Conclusion: By applying meta-analysis of diagnostic tests to clinical practice, nursing staff can increase diagnostic accuracy of stroke deglutition disorders. Correspondence to: Ching-Hui Chuang, MSN, RN, Head Nurse, Department of Rehabilitation, Kaohsiung Chang Gung Memorial Hospital, and Assistant Professor, Department of Nursing, Chang Gung University of Science and Technology, No.123, Dapi Rd., Niaosong Dist., Kaohsiung City 83301, Taiwan, ROC, Tel: +8 867731-7123; +8 867731-2965; E-mail: chinhui@cgmh.org.tw

1. Accuracy rate of water swallow tests in patients with stroke deglutition disorders.
2. Completion rate of water swallow test checklist in patients with stroke deglutition disorders. 3. Incidence of aspiration pneumonia in patients with stroke deglutition disorders.

Purpose
Due to the lack of accuracy and knowledge of water swallow tests in caring for patients with stroke deglutition disorders, this study aimed to evaluate the accuracy of water swallow tests in patients with stroke deglutition disorders using meta-analysis. By comparing diagnostic accuracy before and after applying "water swallow tests" in patients with stroke deglutition disorders, clinical healthcare staff will be able to determine high-risk, and provide substantial suggestions to clinical caring practice for reference. Furthermore, the results can be used in early prevention of aspiration pneumonia in patients with stroke deglutition disorders, so these patients can receive prompt and safe care.

Literature review
This study investigated the importance of the water swallow test on diagnostic accuracy of stroke deglutition disorders in patients with stroke using systemic retrospective analysis and meta-analysis. English databases such as Cochrane Library, MEDLINE, CINAHL, ProQuest Dissertations and Theses (PQDT), PubMed, PsycInfo and AgeLine as well as Chinese databases such as CETD were used for literature searching. The publication dates of studies in individual databases were between 1990 and December, 2014, and search keywords were focusing on studies with Population (P): "patients with Stroke Deglutition Disorders [MeSH]", Intervention (I): "water swallow test [Title/ Abstract]", Comparison (C): "Fiber-optic Endoscopic Examination (FEES) and Video-Fluoroscopic swallowing study (VFSE)", Outcome (O): "sensitivity and specificity", and Style (S): "diagnostic research". The inclusion criteria for literature assessment and selection were as follows: articles with publication dates between 1990 and January 2013; patients with Stroke Deglutition Disorders; no specific sex or species limitations; using water swallow test, FEES or VFSE as interventions; and using sensitivity and specificity as index factors. The exclusion criteria consisted of non-Chinese or non-English articles; involving two or more interventions in which individual importance of monotherapy couldn't be determined; review articles or retrospective articles. In addition, duplicate articles among selected literature were only counted once. A total of 171 articles from nine Chinese or English databases were screened, and only 12 articles related to study objectives were selected in the end [7][8][9][10][11][12][13][14][15][16] (Figure 1).

Accurately use QUADAS
The scoring system of the Cochrane handbook Quality Assessment of Diagnostic Accuracy Studies (QUADAS) was applicable for diagnostic studies. There were in total 11 questions, and the respondents could choose "Yes, No, Unclear or Not Applicable" as their answers. "Yes" answers got 1 point, other answers received zero points. Two different reviewers who had received evidence-based training critically and independently reviewed each selected article. Comments from a third reviewer were sought to reach a consensus while different opinions from two reviewers were received (Figures 2 and 3).

Accurately and critically assess "importance"
This study used the statistical analysis software Stata version 13.1 to analyze the accuracy of the water swallow test. The index variables were sensitivity, specificity and diagnostic odds ratio. A bivariate random effect model was applied in this study. A forest plot was used to present effect size estimates of sensitivity and specificity; values of 0.2, 0.5, 0.8 in 95% confidence interval (95% CI) were used to indicate low, moderate or high effect size; and a Funnel plot was applied to show publication bias [17]. The regression analysis showed no significant publication bias between studies (p=.08, Figure 4). Therefore, publication bias did not have impact on the results of this meta-analysis.
One hundred and seventy-one articles from 9 Chinese or English databases were screened, and 12 articles that were related to study objectives were selected in the end. The publication dates of these literatures were between 1990 and December, 2014. Total subjects recruited in those 12 articles were 1010 people. The sensitivity results of the meta-analysis showed that Q=97.92; p=.00; I 2=83.66%, and the specificity results displayed that Q=137.45; p=.00; I 2=88.36%. Hence, heterogeneities between 12 articles were noted ( Figure 5). The source of potential heterogeneity was derived from the volumes used in of the water swallow test using univariate meta-regression analysis. The   Figure 6, with the area under the curve (AUC) = 0.77 (0.73-0.81).    Meta-regression revealed that the sensitivity of water swallow tests with ≧50 ml volume was significantly higher than the test with <50 ml volume (p=.002) (Appendix I).

Sensitivity analysis
The results of anomaly detection model analysis, including Cook's distance and scatter plot, are shown in Figure 7. We found that one

Accurately determine the evidence and integrate knowledge
According to the evidence of the GRADE guidelines written in the handbook for DTA reviews, which was published by the Cochrane diagnostic test accuracy working group, the results of this diagnostic meta-analysis was strong recommendation, high quality evidence. In summary, we conclude that the specificities of water swallow tests with either ≧50 ml or <50 ml volumes were similar. However, the sensitivities were different, i.e. water swallow tests using ≧50 ml volume had significantly higher sensitivity.

From evidence to application Find out the differences between knowledge and actions (7A)
The water swallow test has had difficulty in being considered as an effective swallow test for a long time. According to the study results published by Maeshima et al. [20] and Tohara et al. [21], excessive water volume used in a water swallow test may result in aspiration pneumonia in stroke patients. Our study results show that water swallow tests using ≧50 ml volumes could lead to the most sensitive results in stroke patients. Meanwhile, we also found that water swallow tests using less than 50 ml volume presented lower sensitivity and moderate specificity, which consequently led to more false negative results.
There have been several articles published investigating the accuracy of water swallow tests. However, the evidence level of the study results couldn't be upgraded due to the lack of meta-analysis. This study conducted a model prediction procedure such as sensitivity analysis prior to the meta-analysis to improve the accuracy of the water swallow test in the results of meta-analysis. We also used graded guidelines proposed by Deeks [22], Liberati et al. [23] and Whiting et al. [24]     in our meta-analysis, including a bivariate random effect model and HSROC, to carefully analyze our data. In conclusion, we recommend a water swallow test using ≧50 ml volume in stroke deglutition disorders evaluation.

Importance evaluation
Through conducting a comprehensive diagnosis of stroke deglutition disorders, the diagnostic results reminded us of the risks of aspiration pneumonia. Therefore, early intervention of swallowing therapy performed by a speech therapist may reduce the incidence of aspiration pneumonia in stroke patients, i.e. the incidence before and after the implementation of a water swallow test promotion program was diminished from 2.59‰ to 1.36‰, while the OR was 0.48 (95% CI=0.28-0.8, P=.002).

Conclusion
This study aimed to investigate the diagnostic accuracy of the water swallow test in patients with stroke deglutition disorders. It was a systemic retrospective review. After data collection, screening, encoding, registration and analysis, the study results showed that a water swallow test using ≧50 ml volume results in higher study sensitivity. In subsequent follow-up studies, we used 100 ml in the water swallow test. The results showed that the accuracy and completion rates of water swallow test could reach 100% once the consistency training of the water swallow test was completed. Moreover, the incidence of aspiration pneumonia and the medical expenses for subsequent treatment were effectively reduced. In conclusion, application of accessible water swallow tests and establishment of related healthcare knowledge should be done without delay. Studies regarding the safety of water swallow tests should be pursued in the future. Hopefully by doing so, patients with stroke deglutition disorders can be treated promptly and effectively, and the professional image as well as ability of healthcare staff can also be boosted.