Freeman‐Sanderson 2016b.

Study characteristics
Methods	Parallel‐group randomised controlled trial with longitudinal follow‐up
Participants	30 participants Inclusion criteria: aged > 18 years; formation and placement of a tracheostomy tube > 48 hours; air‐filled cuffed tracheostomy tube in‐situ; actively mechanically ventilated with PEEP ≤ 10 cmH2O; FiO2 ≤ 40%, spontaneously breathing; triggering ventilatory support; voiceless ≥ 48 hours; awake; able to obey verbal commands. Exclusion criteria: people with hearing impairment.
Interventions	Intervention Early use of speaking valve defined as cuff deflation and use of an in‐line Passy‐Muir ventilator speech and swallowing valve during pressure support ventilation via the tracheostomy tube. Control Usual care, i.e. cuff deflation and use of a speaking valve when a patient was able to self‐ventilate.
Outcomes	Time to phonation measured from tracheostomy insertion to ability to count from 1 to 10 using voice. Ability to phonate was assessed daily by an SLP or nurse not involved in the trial. Duration of tracheostomy cannulation. Duration of mechanical ventilation. Length of stay (ICU and hospital). Time to oral intake. Safety. Quality of life as measured by the Visual Analog Self Esteem Scale (VASES) and the EuroQol‐5D.
Notes	Quality of life data reported in second publication (Freeman‐Sanderson 2016a). Trial registration and 2 abstracts identified in screening.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Random allocation using computer‐generated, permuted‐block randomisation.
Allocation concealment (selection bias)	Low risk	Concealed allocation via sealed opaque envelopes.
Blinding of participants and personnel (performance bias) All outcomes	High risk	No blinding of participants or therapists who administered therapy.
Blinding of outcome assessment (detection bias) All outcomes	High risk	No blinding of the primary outcome assessor.
Incomplete outcome data (attrition bias) All outcomes	Low risk	No indication of incomplete outcome data.
Selective reporting (reporting bias)	Low risk	No evidence of selective reporting (trial registration checked).
Other bias	Low risk	No evidence of other bias. Funding source: none stated. Author conflict of interest: none stated.