Sawyer 2017.

Methods	Randomised, parallel‐group study.
Participants	N = 118 adults with newly diagnosed OSA Any adult patient referred for a diagnostic PSG was invited to participate in the study. Inclusion criteria: newly diagnosed with OSA (AHI > 10), PAP‐naive, ≥18 years of age, able to read and speak English. Exclusion criteria: previous diagnosis or treatment of OSA; medical record documented new psychiatric diagnosis within previous six months of study enrolment; requirement of supplemental oxygen or bilevel PAP identified on PAP titration PSG suggesting diagnosis other than OSA; diagnosis of another sleep disorder in addition to OSA based on polysomnogram (i.e. periodic limb movement disorder [≥10 limb movements/hr of sleep with arousal], central sleep apnoea [≥ 5/hours central apneas], insomnia, sleep hypoventilation syndrome, or narcolepsy). Baseline characteristics (per‐protocol): 30% female. Mean age 51.3 (±11.1). Mean AHI = 36. Mean ES S= 19.6. Mean BMI=38.0. Country: USA
Interventions	Participants were randomised to receive usual care (UC, n = 57) or a multi‐phased and tailored intervention (TI, n = 61) targeting social cognitive perceptions of OSA–PAP treatment. TI: intervention addressed cognitive perceptions of the diagnosis and treatment, outcome expectancies with PAP treatment, and PAP treatment self‐efficacy, all domains of SCT. Intervention delivered in four phases: prediagnosis, postdiagnosis (i.e. postdiagnostic polysomnogram), immediately post‐PAP titration polysomnogram, and with week 1 of home PAP treatment. Intervention delivery guided by a protocol and script templates for specific exposure phases to minimise a potential interventionist effect. UC: followed current practice standards for the diagnosis and treatment of OSA in adults (Epstein 2009; Kushida 2006). Included sleep centre–provided informational brochures about OSA, diagnostic testing, and PAP prescription. In addition, access by telephone to sleep centre staff for problems, questions, or concerns was provided during daytime and evening. Study duration: 3 months
Outcomes	CPAP usage (hours/night) at 1 week, 1 month and 3* months. N of adherent participants (usage ≥ 4 hours per night) at 1 week, 1 month and 3* months. Withdrawals
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation employing random block sizes assigned participants to the exposure or comparison group and within each assignment level, 50% were randomly assigned to interview–no interview at study termination and debriefing. A randomisation list was generated by the study biostatistician (TSK) and securely maintained at the clinical research site.
Allocation concealment (selection bias)	Low risk	Random assignment was concealed and completed by consecutive sealed envelope, then opened sequentially by interventionist and study research assistant; sealed envelopes were secured, prepared, and monitored by unblinded study personnel
Blinding (performance bias and detection bias) All outcomes	Low risk	No evidence of blinding of participants, personnel or outcome assessors. Study only had objective outcomes, so lack of blinding is not likely affect outcomes.
Incomplete outcome data (attrition bias) All outcomes	High risk	Data not available for participants excluded (intentionally) post‐randomisation. Reported reasons for missing outcome data may be related to health status.
Selective reporting (reporting bias)	Low risk	Published report Methods provides a comprehensive description of the intervention design, intervention protocol, study design, setting and sample, measures and analysis. Additionally, the pre‐specified plan included per‐protocol fidelity measures and a blinding assessment for participants upon study completion.
Other bias	Low risk	No baseline imbalances or reported/suspected deviations from intended intervention.
Bias arising from the randomisation process (ROB2, primary outcome)	Low risk	Authors report: "After pre enrolment screening and informed consent, participants completed a demographic questionnaire and were randomised. Randomization employing random block sizes assigned participants to the exposure or comparison group and within each assignment level, 50% were randomly assigned to interview–no interview at study termination and debriefing. A randomisation list was generated by the study biostatistician (TSK) and securely maintained at the clinical research site. Random assignment was concealed and completed by consecutive sealed envelope, then opened sequentially by interventionist and study research assistant; sealed envelopes were secured, prepared, and monitored by unblinded study personnel (DAS)." Key baseline characteristics (age, gender, BMI, AHI) were reported and differences were insignificant (P values reported), consistent with chance.
Bias due to deviations from intended interventions (ROB2, primary outcome)	Low risk	Authors report, "In order that a participant blind was supported from the outset of the trial, IRB‐approved consent modification was employed that specifically did not differentiate between study groups in terms of study activities or time commitment and a limited description of the overall study objective; this necessitated debriefing at study termination." Regarding investigator blinding, authors report: "The intervention was delivered by one research assistant (study interventionist, unblinded, MV), a registered nurse without sleep specialty care experience, who was extensively trained to provide the tailored intervention." participants were randomised prior to PSG; therefore, some of the inclusion criteria could only be assessed after randomisation. Of the 118 randomised participants, 30 (18 in intervention arm and 12 in control arm) were excluded based on failure to meet AHI = 10 inclusion criteria. Additionally, authors report: "Exclusions or withdrawals occurred for the following reasons: refused PAP (n = 2), referred to other treatment (n = 1), and titrated on other positive airway device for other sleep‐related breathing disorders (n = 10). Specific to the protocol, four participants did not complete titration PSG and eight participants requested to withdraw due to personal (i.e. transportation, familial issues, work‐related issues) or other pressing health problems. The remaining participants (n = 60) completed the protocol; no attrition or loss of data for the PAP use outcome or feasibility and acceptability outcomes occurred. No study‐related adverse or serious adverse events occurred." Non‐OSA or BiPAP‐requiring sleep disorders were planned exclusionary criteria; these exclusions should not be considered protocol deviations. Additionally, CPAP refusal and referral to other treatment (by provider) are also not deviations from intended interventions in this study. Thus, consistent with author's report, only 12 protocol deviations (6 in each arm) occurred. Administrative withdrawal: 2/31, 2/27; subject withdrawal: 4/31 and 4/27. mITT. Outcome data were not available for participants who withdrew (n = 12).
Bias due to missing outcome data (ROB2, primary outcome) All outcomes	Unclear risk	Data not available for participants excluded (intentionally) post‐randomisation. Neither analyses to correct for bias nor sensitivity analyses were conducted. It is possible that withdrawals were related to health status, as per the reasons outlined in the report (transportation, familial issues, work‐related issues, pressing health problems). The proportion of missing outcome data is similar across groups. Reported reasons make it possible that missingness is related to its true value. However, reported reasons do not differ between groups. While continuing symptoms may have made it more likely for participants to drop out, the roughly equivalent withdrawal rates across groups makes it unlikely that this would have resulted in substantial differences in the estimated effect of the intervention.
Bias in measurement of the outcome (ROB2, primary outcome) All outcomes	Unclear risk	Outcome measured using objective CPAP usage data. Each intervention group outcome data ascertained via automated CPAP device monitoring. Outcome "assessor" is CPAP device: no knowledge of allocation possible. Unable to confirm with study author that distribution of CPAP device makes did not differ between intervention arms.
Bias in selection of the reported result (ROB2, primary outcome) All outcomes	Low risk	Published report Methods provides a comprehensive description of the intervention design, intervention protocol, study design, setting and sample, measures and analysis. Additionally, the pre‐specified plan included per‐protocol fidelity measures and a blinding assessment for participants upon study completion. Methods section indicates that multiple time points planned; each planned outcome reported. Methods section indicates that one, commonly‐employed threshold adherence definition was planned; this outcome was reported in Results. No evidence that multiple analyses (e.g. variable adherence 'thresholds') were conducted.
Overall risk of bias (ROB2, primary outcome) Machine usage	Unclear risk	‐