Jackson 2015.

Methods	Randomised controlled parallel‐arm trial
Participants	POSA Participants randomised: 86 Inclusion criteria ≥ 18 years of age, supine OSA (supine AHI at least twice the non supine AHI) on overnight diagnostic PSG, total AHI ≥ 10 ≥ 4 h of sleep with ≥ 30 min sleep in both the lateral and supine recumbent positions and 30 min of REM sleep Exclusion criteria Minimum blood SaO2 < 75% in REM or 80% in non‐REM Clinically significant co‐existing disease (e.g. diabetes, unstable ischaemic heart disease) Sleepiness deemed to be unsafe and requiring urgent treatment (e.g. history of falling asleep while driving or working, or ESS score > 16) Any musculoskeletal condition that precluded moderate exercise (as this was part of the sleep hygiene instructions) or lying on their side while asleep
Interventions	A 10‐point sleep guide to improve OSA with and without sleep position modification device. The sleep position modification device consisted of a band of stretch cotton worn around the chest, just below the nipple line and with straps over the shoulder to hold it in place. The band was secured at the front with buttons and the ball was contained in a pocket at the rear, over the thoracic spine.
Outcomes	Outcomes: supine sleep time, total sleep time, AHI, supine AHI, sleep efficiency, arousal index, BMI, blood pressure, ESS and quality‐of‐life scales e.g. FOSQ, SOSQ, and neuropsychological test battery like psychomotor vigilance test, response inhibition test, Trail Making Task A and B, digit symbol substitution task, digits span test and controlled word association task
Notes	Setting: outpatient department Austin Health, USA Duration of treatment: 4 weeks Adherence not assessed Funding: likely mixed funding "This work was supported by grants from the Institute for Breathing and Sleep, the Austin Health Medical Research Foundation and the Harold and Cora Brennen Benevolent Trust" Confict of interest declared: Dr Howard has received funding from ResMed foundation, Edansafe and Prevention Express
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Quote: "A randomisation sequence at a 1:1 active:control ratio was computer generated by a third party and the investigators were provided with sealed envelopes", page 547 Comment: 1:1 allocation was planned, but numbers in the arms are very dissimilar (37 and 49)
Allocation concealment (selection bias)	Low risk	Quote: "A randomisation sequence at a 1:1 active:control ratio was computer generated by a third party and the investigators were provided with sealed envelopes. These were opened in order of enrolment", page 547 Comment: computer‐generated randomisation sequence was provided by a third party in sealed envelopes to be opened in the order of enrolment
Blinding of participants and personnel (performance bias) Objective outcomes ( AHI)	High risk	Not possible due to the nature of intervention
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Blinding not mentioned
Incomplete outcome data (attrition bias) All outcomes	Low risk	No participants lost to follow‐up
Selective reporting (reporting bias)	Unclear risk	Primary outcome for which sample size was calculated is not reported explicitly
Other bias	Unclear risk	It is not clear whether the study adjusted for multiple comparisons.