Translating Dysphagia Evidence into Practice While Avoiding Pitfalls: Assessing Bias Risk in Tracheostomy Literature.

Critically ill patients who require a tracheostomy often have dysphagia. Widespread practice guidelines have yet to be developed regarding the acute assessment and management of dysphagia in patients with tracheostomy. In order for clinicians to base their practice on the best available evidence, they must first assess the applicable literature and determine its quality. To inform guideline development, our objective was to assess literature quality concerning swallowing following tracheostomy in acute stages of critical illness in adults. Our systematic literature search (published previously) included eight databases, nine gray literature repositories and citation chasing. Using inclusion criteria determined a priori, two reviewers, blinded to each other, conducted an eligibility review of identified citations. Patients with chronic tracheostomy and etiologies including head and/or neck cancer diagnoses were excluded. Four teams of two reviewers each, blinded to each other, assessed quality of included studies using a modified Cochrane Risk of Bias tool (RoB). Disagreements were resolved by consensus. Data were summarized descriptively according to study design and RoB domain. Of 6,396 identified citations, 74 studies met our inclusion criteria. Of those, 71 were observational and three were randomized controlled trials. Across all studies, the majority (> 75%) had low bias risk with: participant blinding, outcome reporting, and operationally defined outcomes. Areas requiring improvement included assessor and study personnel blinding. Prior to translating the literature into practice guidelines, we recommend attention to study quality limitations and its potential impact on study outcomes. For future work, we suggest an iterative approach to knowledge translation.