The diagnostic test accuracy of clinical swallow assessment for oropharyngeal aspiration: a systematic review.

Abstract

Background: Oropharyngeal aspiration, the recurrent entry of food and/or fluids below the level of the vocal cords, can result in a range of complications including: chronic lung diseases, aspiration pneumonia, malnutrition and/or dehydration. Video fluoroscopic swallow study is the Gold Standard assessment of oropharyngeal aspiration but is resource intense, exposes the patient to radiation and is not available in all hospitals and centres. The Clinical Swallow Assessment is a bedside swallow assessment widely used to screen and/or assess for oropharyngeal aspiration. The evidence base behind the diagnostic test accuracy of the Clinical Swallow Assessment has not previously been synthesised. Objectives: To synthesise the best available evidence on the diagnostic test accuracy (sensitivity and specificity) of clinical swallow assessment compared with Video Fluoroscopic Swallow Study in diagnosing oropharyngeal aspiration in children and adults with dysphagia. Inclusion criteria: Types of participants Any patients referred for swallowing assessment, specifically assessed for oropharyngeal aspiration were included and there was no exclusion based on age or gender. Study results were excluded for head and neck cancer patients, patients with a tracheostomy in situ and patients with craniofacial anomalies. Focus of the review The focus of the review was to examine the diagnostic test accuracy of clinical swallow assessment, as compared with Video Fluoroscopic Swallow Study. Types of studies This systematic review considered any relevant cross sectional study that measured diagnostic test accuracy. Types of outcomes Outcomes of interest were the sensitivity and specificity of the clinical swallow, as compared with the video fluoroscopic study and the positive and negative predictive values. Where this data was not reported in the studies, these measures were calculated from the reported raw data. Search strategy: Thirteen major databases were searched from their inception until April 31st 2012. There were no limits during the search stage as relevant studies were omitted if search filters such as ‘English’ and ‘Human’ were applied. Methodological quality: Methodological quality was assessed using the QUADAS checklist. Data was collected using the STARD checklist. Sensitivity and specificity measures were combined in meta-analysis to generate a summary receiver operator characteristic plot. Results: There were 1787 titles initially identified. Following duplicate removal and screening against inclusion criteria, 37 papers were retrieved for detailed examination and 24 papers were excluded as they did not meet the inclusion criteria. The most common reason for exclusion was that the paper was not a study of diagnostic test accuracy. There were 13 studies included in the systematic review and found to have high methodological quality. Data extracted from individual studies was statistically combined in meta-analysis to produce a forest plot and summary receiver operating characteristic (sROC) plot. Heterogeneity was evident in the forest plot, particularly for sensitivity as evidenced by the wider confidence intervals for sensitivity compared with specificity. The test sensitivity varied from 21% to 93%, the specificity from 46% to 93%. The summary mean sensitivity and specificity was calculated as 71% and 76% respectively. Positive predictive value was calculated as 60% and negative predictive value was 81%. The scatter of points around the curve on the sROC plot also indicated heterogeneity. Sources of heterogeneity were identified and explored. The shape of the sROC curve strongly supported the finding of a threshold effect, which is expected for studies in which there is a strong interpretative component such as the clinical swallow assessment. This occurs as clinicians may vary in their criteria for what constitutes a positive or negative test result. The overall prevalence of aspiration in the included studies was calculated as 35%. Results are based predominantly on adult, acute post stroke patients. Conclusion: This thesis provides good evidence for an overall estimate of the sensitivity and specificity of clinical swallow assessment compared with video fluoroscopic swallow study for the assessment of oropharyngeal aspiration. In this population, a clinician can be much more confident in a negative test result than a positive test result. A false positive test result may lead to unnecessary patient care and costs, including with-holding oral medications and prescription of modified diets and/or fluids. A false negative test result may lead to compromised lung health and/or pneumonia. Implications for practice: Using calculations of the positive predictive values and negative predictive values, 60% of patients who test positive for aspiration are truly aspirating and 81% of patients who test negative for aspiration are truly not aspirating. Positive and negative test results are affected by the prevalence of the condition in the population. To summarise, the PPV increases and the NPV decreases as prevalence increases and the PPV decreases and NPV increases as the prevalence decreases. For example if the prevalence is much lower (e.g. 10%) the NPV rises to 96% and the PPV decreases to 24%. This thesis provides data for centres where VFSS is not available regarding the diagnostic test accuracy of clinical swallow assessment for oropharyngeal aspiration. Implications for Research: Only one of the included studies provided data for infants and children. None of the included studies addressed infants, children or adults without a neurological aetiology. Further research is needed for infants and children with dysphagia as well as neurologically intact and normally developing infants, children and adults.