Walsh 2013.

Study characteristics
Methods	Design: RCT, parallel 2‐arm, multicentre, pilot study Setting: 6 ICUs in university teaching and general hospitals (UK) Recruitment: August 2009 to December 2010 Maximum follow‐up: 6 months
Participants	100 ICU participants aged ≥ 55 years, Hb < 9 g/dL, mechanical ventilation for ≥ 96 hours, who were expected to require ≥ 24 hours of further mechanical ventilation when assessed Liberal: n = 49; mean age (range) = 68 (55 to 83) years Restrictive: n = 51; mean age (range) = 67 (56 to 80) years
Interventions	Liberal group: received transfusion when Hb ≤ 9.0 g/dL, with target Hb of 9.1 g/dL to 11.0 g/dL Restrictive group: received transfusion when Hb ≤ 7.0g/dL, with target Hb of 7.1 g/dL to 9.0 g/dL
Outcomes	Primary feasibility outcome: difference in mean Hb among groups. Clinical outcomes were assessed, including SF‐12 (with physical function scale)
Notes	Trial name: Restrictive versus Liberal Transfusion Strategies for Older Mechanically Ventilated Critically Ill Patients (RELIEVE) Trial registration (prospective): NCT00944112 Trial funding: supported, in part, by the Chief Scientists Office, Scotland (CZB/4/698); the Scottish National Blood Transfusion Service, the NHS Lothian Academic Health Science Centre; and the Transfusion Medicine Education and Research Foundation COI statement by investigators: no statement of interest reported in the paper
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	"Minimisation by centre and the presence of IHD, including a random element, was used" (Walsh 2013 p 2355)
Allocation concealment (selection bias)	Low risk	Trial used telephone randomisation
Blinding of participants and personnel (performance bias) Objective outcomes	Low risk	Blinding of personnel for this intervention is not feasible, but in our view, for objective outcomes such as mortality (the primary outcome used within this review), we graded risk of bias as 'low' Clinicians were not blinded. Most surviving participants stated that they were unaware of group allocation at 180 days (restrictive group: 67%; liberal group: 78%); 23% of participants in the restrictive group and 9% in the liberal group correctly stated their treatment group
Blinding of outcome assessment (detection bias) Objective measures	Low risk	Blinding of mortality (the primary outcome used within this review) is not relevant, and we graded risk of bias as 'low'
Blinding of outcome assessment (detection bias) Subjective measures	High risk	Researchers concealed from group allocation collected questionnaire‐based measures at 60 and 180 days post randomisation. Clinical outcomes were not documented to have been done blindly; however, investigators did ask surviving patients at 180 days which group they thought they had been allocated to
Incomplete outcome data (attrition bias) All outcomes	Low risk	Missing data were rare at long‐term follow‐up At 60 days, of surviving participants in the restrictive group (37/51), there were 29/37 complete RMIs (Rivermead Mobility Index) and 23/37 completed SF‐12s. At 180 days, of surviving participants in this group (33), there were 29 complete RMIs and 28 SF‐12s and 29 HE questionnaires At 60 days, of surviving participants in the restrictive group (27/49), there were 26/27 complete RMIs and 23/27 completed SFIs. At 180 days, of surviving participants in this group (22), there were 21 complete RMIs and 21 SF‐12s and 20 HE questionnaires
Selective reporting (reporting bias)	Low risk	Trial registration was prospective. No reporting bias was apparent
Other bias	Low risk	No other biases were apparent