Hedberg 2012.

Methods	Parallel randomised controlled clinical trial Randomisation ratio: not reported Superiority design
Participants	Inclusion criteria: BMI > 48kg/m2 Exclusion criteria: stated only that of the eligible patients, 9 were excluded on medical grounds or because of language difficulties and a further 43 refused to be randomised. Diagnostic criteria: BMI > 48 kg/m2
Interventions	Number of study centres: not reported but appears to be a single centre Treatment before study: for both groups stated that after initial evaluation of the internist, dietician and psychologist, education provided on post‐operative diet and [unspecified] abnormalities were treated [using unspecified methods] before surgery. Titration period: n/a Interventions: 1: Open (laparotomic) biliopancreatic diversion with duodenal switch (BPD/DS) 2: Open (laparotomic)  Roux‐en‐Y gastric bypass (RYGB) All patients received post‐operative multivitamin supplements
Outcomes	Outcomes reported in abstract of publication: weight loss, co‐morbidities, complications
Study details	Run‐in period: not reported Study terminated before regular end: yes (recruitment was terminated before planned sample size was reached)
Publication details	Language of publication: English Funding: not reported Publication status: peer reviewed journal
Stated aim for study	Quote: "to compare the weight loss (primary outcome), perioperative results, and complications in the long and short term, as well as the gastrointestinal symptoms and biochemical profiles (secondary outcomes) in a prospective, randomised controlled trial of [BPD/DS] versus [RYGB] in patients with a BMI > 48 kg/m2" (Hedberg 2012, p. 339)
Notes	BMI: body mass index; BPD/DS: open (laparotomic) biliopancreatic diversion with duodenal switch; n/a: not applicable; RYGB: open (laparotomic)  Roux‐en‐Y gastric bypass
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Comment: no information provided
Allocation concealment (selection bias)	Unclear risk	Comment: randomisation was achieved with sealed envelopes, which were opened after the patient had been anaesthetised – however, it was not stated whether envelopes were opaque
Blinding of outcome assessment (detection bias) All outcomes	High risk	Comment: study was not blinded, except that “the type of procedure was unknown to the patient and staff until 2 days postoperatively”
Incomplete outcome data (attrition bias) Weight	Unclear risk	Comment: attrition is reported and overall was balanced across the groups (3 patients in each group declined follow‐up or did not reply); however, the timing of these dropouts is not stated and no sample sizes are provided for outcomes (the timings of which were also unclear in many cases). It was not stated whether any of the patients who did not drop out failed to provide weight‐loss data.
Incomplete outcome data (attrition bias) Comorbidities	Unclear risk	Comment: as for incomplete outcome data ‐ weight
Selective reporting (reporting bias)	High risk	Comment: there is considerable scope for selective reporting as some outcomes were reported only at baseline whilst others were reported only post‐surgery; also the timing of assessments is unclear in many cases. Overall, many of the measurements stated in the methods appear to be missing from the results. For the patient reported questionnaire, the investigators appear to have chosen specific sets of outcomes to report or exclude. The baseline number and % of patients with diabetes was higher in the BPD/DS than RYGB group (29% versus 4%) but this difference was not reported, except indirectly for two diabetes medication subgroups.
Other bias	High risk	Comment: the required sample size was not achieved due to patients declining randomisation because of their own preferences. Instead, an interim analysis of 47 patients showed significant differences between the 2 groups and the inclusion was stopped.  It was stated that for both groups after initial evaluation by the internist, dietician and psychologist, [unspecified] abnormalities were treated [using unspecified methods] before surgery.