Santamaria 2015.

Methods	Trial design: RCT Trial grouping: parallel groups Ethics and informed consent: yes Follow‐up period: not stated Sample size estimate: yes, required a total of 220 people per group ITT analysis: yes; number randomised: 440; number analysed: 440 Funding: funded by Molnlycke Health Care, the manufacturer of the intervention product
Participants	Inclusion criteria: ED and ICU admission for critical illness and/or major trauma > 18 years of age Exclusion criteria: suspected or actual spinal injury precluding the patient being turned pre‐existing sacral or heel PU trauma to sacrum and/or heels Pretreatment age, mean years (SD years): intervention: 54 (20.8); control: 56 (20.5) gender, M/F: intervention: 126/89; control: 132/82 Braden, mean (SD): intervention: 12 (4.2); control: 12 (3.9) Apache 11, mean: intervention: 19; control: 19.5
Interventions	Intervention characteristics: soft silicone multi‐layered foam dressing the dressing was applied bilaterally to the trochanteric and sacral regions the dressings were changed only if there was loss of adhesiveness, shear, excessive moisture, friction, presence of wrinkles, or the combination of these factors Control: no dressing applied
Outcomes	PU incidence Outcome type: dichotomous outcome Reporting: fully reported Unit of measure: numbers Direction: lower is better Data value: endpoint Cost benefit Outcome type: continuous outcome Reporting: partially reported Unit of measure: Australian dollars Direction: lower is better Data value: endpoint
Identification	Sponsorship source: funded by Molnlycke Health Care, the manufacturer of the intervention product. Country: Australia Setting: ICU Comments: no comments Author's name: N Santamaria N Institution: Royal Melbourne Hospital Email: nick.santamaria@mh.org.au Address: Level 6 Office for Research, Royal Melbourne Hospital, Grattan Street, Parkville VIC 3050, Australia
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Participants were randomised in the ED to either the intervention group or to the control group by retrieving the next envelope in a pre‐prepared series of envelopes that had been randomised using a computer‐generated set of random numbers to determine group allocation
Allocation concealment (selection bias)	Unclear risk	Whether the envelopes were opaque and sealed was not reported
Blinding of participants and personnel (performance bias) All outcomes	High risk	Cannot blind as one group has a dressing and the other doesn't
Blinding of outcome assessment (detection bias) All outcomes	High risk	The outcome assessor was not independent of the study team
Incomplete outcome data (attrition bias) All outcomes	Low risk	The analysis was based on ITT (33) where all participants randomised to the intervention were analysed regardless of protocol violations
Selective reporting (reporting bias)	Low risk	PU incidence, PU location, time to PU development and cost were reported
Other bias	Unclear risk	The sample size calculation was based on a control event rate of 4.0% (presumably this was known from existing hospital data). In the study, the control event rate was 13.1%, raising questions about the accuracy of PU diagnosis in the control group during the study. The study was registered on clinicaltrials.gov but only after data collection had begun.