Ahrens 2017.

Methods	Study design: multicentre RCT (20 centres) Duration of the study: July 2008‐December 2014 Protocol was published before recruitment of participants: yes. doi: 10.1186/1745‐6215‐12‐57 ‐ Trials 2011;12:57. Date of trial registration 7 September 2006, recruitment started 18 December 2007 Details of trial registration: yes. Main ID: United Kingdom Clinical Research Network ID: 8665. Date of registration: 7 September 2006 Funding sources: This study was funded with grants from BUPA Foundation and BESS. Dr. Ahrens reports grants from BESS, grants from BUPA Foundation, during the conduct of the study.
Participants	Place of study: 20 acute‐care hospitals in England (UK) Number of participants assigned: 301 participants (154 surgical; 147 conservative) Number of participants assessed to primary outcome (DASH): 204 participants (111 surgical; 93 conservative) Inclusion criteria Aged 18‐65 years A completely displaced middle third clavicle fracture (Robinson type 2B1 or 2B2) Fresh fracture, treatment within 14 days after injury Being medically fit to undergo surgery (ASA grade I, II, or III) Exclusion criteria Patient's refusal to participate Being medically unfit to undergo surgery (ASA grade IV or V) Any other type of clavicle fracture Established non‐union from a previous fracture A previous fracture around the clavicle A previous operation on the shoulder or clavicle Metabolic bone disease Substantial neuromuscular upper‐limb disability Age Surgical group (mean/SD): 36.1/12.3 years Conservative group (mean/SD): 36.4/11.8 years Gender of participants assigned (male/female) Surgical group: 132/22 Conservative group: 130/17 Classification of injury: fractures were classified according to the Robinson's Classification (Robinson 1998).
Interventions	Timing of intervention: not specified; however, patients with > 14 days after the injury were excluded. In the operative group, the surgical procedure could be done within 3 months after randomisation. Type of surgical intervention: open reduction and plate fixation. Fixation was performed using the Acumed clavicle fixation system, consisting of a precontoured titanium plate. Type of conservative intervention: simple sling for 6 weeks Rehabilitation Surgical group: sling was used for comfort; pendulum and elbow exercises were allowed on the 1st day postoperatively, and the subsequent mobilisation and rehabilitation protocol was the same as that for the non‐operative group. Conservative group: participants were allowed to remove the sling for short periods to wash, dress, write, eat, and use a keyboard as soon as comfort allowed. Active‐assisted ROM was permitted starting at 2 weeks as comfort allowed. Full active mobilisation, resistance exercises, and cross‐arm adduction commenced after 6 weeks. Any co‐interventions: not reported
Outcomes	Length of follow‐up Follow‐up was 9 months Participants were evaluated at 3 and 6 weeks, and at 3 and 9 months and 12 months was also planned in the protocol. Loss to follow‐up: 47 participants were lost to follow‐up at 9 months (analysed for fracture union) and 97 participants were not assessed the function (measured by DASH) Surgical group: 23 participants were lost at 9 months (for clavicle union) and 43 (for DASH). Conservative group: 24 participants were lost at 9 months (for clavicle union) and 54 (for DASH) Primary outcomes Function or disability measured by DASH and Constant score Failure of treatment measured by symptomatic non‐union, malunion or other complications (e.g. mechanical failure, refracture and late neurologic complication). Radiographs were performed at 2 weeks, 6 weeks and 3‐month follow‐up. Secondary outcomes Other treatment failure measured by: cosmetic result: scar asymptomatic non‐union Adverse events measured by: infection (none) hardware irritation with removal others
Notes	SDs and means of DASH and Constant scores were obtained by personal contact with the study authors. Composite adverse events: surgery: plate removal (n = 5); frozen shoulder (n = 2). In conservative: frozen shoulder (n = 1). In Analysis 1.14 we used event rates 5 for surgery and 1 for conservative. Cosmetic result events: surgery: minor scar (n = 3); prominent plate (n = 2). Conservative (n = 0). In Analysis 1.11 we used event rates 3 for surgery and 0 for conservative.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated randomisation lists, stratified by centre, were produced using random permuted blocks and equal allocation to the operative and non‐operative groups.
Allocation concealment (selection bias)	Low risk	Sequentially numbered, opaque, sealed envelopes
Blinding of participants and personnel (performance bias) All outcomes	High risk	Participants and personnel were not blinded
Blinding of outcome assessment (detection bias) All outcomes	High risk	Outcome assessors were not blinded so there is potential for bias in radiographic outcomes. Participants were unblinded so there is potential for bias in self‐reported outcomes.
Incomplete outcome data (attrition bias) All outcomes	High risk	Missing function outcome data (DASH) were not balanced in numbers across intervention groups; more participants in the conservative intervention group were lost to follow‐up (43/154) (28%): surgical versus 54/147 (37%) conservative at 9 months. This may have led to an overestimation of the benefits of surgery.
Selective reporting (reporting bias)	High risk	Outcomes of interest in the review are reported incompletely, and the author failed to report any measure of variance for DASH and Constant scores.
Other bias	Low risk	The study appears to be free of other potential sources of bias.