Interventions to enable communication for adult patients requiring an artificial airway with or without mechanical ventilator support

Background Inability to communicate in a manner that can be understood causes extreme distress for people requiring an artificial airway and has implications for care quality and patient safety. Options for aided communication include non‐vocal, speech‐generating, and voice‐enabling aids. Objectives To assess effectiveness of communication aids for people requiring an artificial airway (endotracheal or tracheostomy tube), defined as the proportion of people able to: use a non‐vocal communication aid to communicate at least one symptom, need, or preference; or use a voice‐enabling communication aid to phonate to produce at least one intelligible word. To assess time to communication/phonation; perceptions of communication; communication quality/success; quality of life; psychological distress; length of stay and costs; and adverse events. Search methods We searched the Cochrane Library (Wiley version), MEDLINE (OvidSP), Embase (OvidSP), three other databases, and grey literature from inception to 30 July 2020. Selection criteria We included randomised controlled trials (RCTs), quasi‐RCTs, cluster‐RCTs, controlled non‐randomised parallel group, and before‐after studies evaluating communication aids used in adults with an artificial airway. Data collection and analysis We used standard methodological procedures recommended by Cochrane. Two review authors independently performed data extraction and assessment of risk of bias. Main results We included 11 studies (1931 participants) conducted in intensive care units (ICUs). Eight evaluated non‐vocal communication aids and three voice‐enabling aids. Usual care was the comparator for all. For six studies, this comprised no aid; usual care in the remaining five studies comprised use of various communication aids. Overall, our confidence in results regarding effectiveness of communication interventions was very low due to imprecision, measurement heterogeneity, inconsistency in results, and most studies at high or unclear risk of bias across multiple domains. No non‐vocal aid studies reported our primary outcome. We are uncertain of the effects of early use of a voice‐enabling aid compared to routine use on ability to phonate at least one intelligible word (risk ratio (RR) 3.03, 95% confidence interval (CI) 0.18 to 50.08; 2 studies; very low‐certainty evidence). Compared to usual care without aids, we are uncertain about effects of a non‐vocal aid (communication board) on patient satisfaction (standardised mean diff