Van Maanen 2012.

Methods	Randomised cross‐over trial
Participants	POSA Participants randomised: 30 Inclusion criteria Age > 18 years Referred to the Department of Otorhinolaryngology, Head and Neck Surgery of the Saint Lucas Andreas Hospital (Amsterdam, Netherlands) Diagnosed with positional sleep apnoea, using full overnight in‐hospital polysomnography defined as AHI > 5, AHI supine ≥ 2 times AHI in other positions, percentage of total sleep time in supine position ≥ 10 and ≤ 90% Exclusion criteria: no exclusion criteria given
Interventions	Vibrating device worn on the back of neck that senses supine position in the on and off position, on randomly assigned nights within 3 months of each other
Outcomes	AHI Supine AHI Non‐supine AHI Total supine sleep time Sleep efficiency SaO2 sleep efficiency Arousal index
Notes	Settings: outpatient ENT services, Saint Lucas Andreas Hospital, Netherlands Duration of follow‐up: 3 sleep studies over 3‐month period Adherence: overnight PSG studies in hospital, in 3 however device malfunctioned in 'on state' Drop out: nil Funding: Not mentioned
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Study authors have not stated the randomisation procedure in clear terms. There was a mere mention that the device was put on or off 'randomly'. We tried to contact the study author for clarification, but were unsuccessful
Allocation concealment (selection bias)	Unclear risk	The methods are not properly stated
Blinding of participants and personnel (performance bias) Objective outcomes ( AHI)	Unclear risk	Device worn on both nights, however due to nature of intervention participants cannot be blinded, but all reported outcomes are objective
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Quote: "The data were reviewed manually for analysis by an experienced sleep investigator, blinded for the activity state of the device" page 323. Comment: outcome assessors were blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	No attrition
Selective reporting (reporting bias)	Unclear risk	As no specific outcomes or comparisons were listed a priori, difficult to ascertain whether the listed outcomes are those planned or not
Other bias	Unclear risk	Statistically analysis is probably not appropriate. They have 3 group comparison, but used t‐tests individually. They have conceded that the distributions were not normal, and sometimes Wilcoxon rank sum test would be used. But it is not clear how and where it was used. It is also not clear whether paired t‐test was used

AHI: Apnoea‐Hypopnoea Index; BMI: body mass index; CPAP: continuous positive airway pressure; ESS: Epworth Sleepiness Scale; FOSQ: Functional Outcome of Sleep Questionnaire; MOS: Medical Outcome Study; OSA: obstructive sleep apnoea; POSA: positional obstructive sleep apnoea; PSG: polysomnography; REM: rapid eye movement; SaO2: oxygen saturation; SD: standard deviation; SOSQ: Symptoms of Sleep Questionairre; SF‐36: Short Form Health Survey; SPT: sleep position trainer; WHO MONICA: World Health Organization Monitoring Trends and Determinants in Cardiovascular Disease