Ballard 2007.

Methods	Randomised, double‐blind, parallel group trial
Participants	N = 104 participants (67 M/37 F); Age 52 years; BMI 33 kg/m2; AHI 42; ESS not reported Inclusion criteria: non‐adherent with CPAP based on 14‐day run‐in (< 4 hours/day); previous diagnosis of OSA Exclusion criteria: compliant with CPAP during run‐in
Interventions	Bi‐PAP with flexible pressure setting versus fixed CPAP setting (identical machine) Run‐in: phase 1 of study identified non‐adherent CPAP users (those using CPAP < 4 hours/day) Study duration: 12 weeks
Outcomes	Machine usage (average hours used and using therapy > 4 hrs) Quality of life (FOSQ)
Funding & conflicts of interest statements	Quote: "This study was supported by a grant from Respironics, Inc. Respironics reimbursed National Jewish Medical and Research Center for part of Dr. Ballard’s time. Dr. Gay has received research support from ResMed. Dr. Strollo has indicated no financial conflicts of interest."
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "Patients were then provided a modified positive airway pressure device (BiPAP Pro, Respironics Inc.) randomly set to either CPAP or BiFlex mode at appropriate pressure(s)"
Allocation concealment (selection bias)	Unclear risk	Quote: "Patients were then provided a modified positive airway pressure device (BiPAP Pro, Respironics Inc.) randomly set to either CPAP or BiFlex mode at appropriate pressure(s)"
Blinding of participants and personnel (performance bias) Machine usage, symptoms, quality of life, withdrawal, adverse effects	Low risk	Quote: "Both patients and the investigators were blinded as to the specific mode assigned to each patient."
Blinding of participants and personnel (performance bias) AHI, blood pressure, treatment pressure	Low risk	Quote: "Both patients and the investigators were blinded as to the specific mode assigned to each patient."
Blinding of outcome assessment (detection bias) Machine usage, symptoms, quality of life, withdrawal, adverse effects	Low risk	Identical machine appearance unlikely to affect measurement of these outcomes.
Blinding of outcome assessment (detection bias) AHI, blood pressure, treatment pressure	Low risk	Identical machine appearance unlikely to affect measurement of these outcomes.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	Study reported to have been analysed on ITT principles. Unlikley to bias machine usage data but quality of life collected from completers. There was differential dropout and this may have influenced the results.
Selective reporting (reporting bias)	Unclear risk	Insufficient information available
Other bias	Low risk	No concerns identified