Virtanen 2012a.

Methods	Study design: single‐centre, 2‐group, parallel‐design RCT Duration of the study: August 2004‐October 2007 Protocol was published before recruitment of participants: not reported, but no protocol was published Details of trial registration: Main ID: NCT01199653. Date of registration: 10 September 2010 Funding sources: none of the study authors received payments or services, either directly or indirectly. The study was supported by Helsinki University Central Hospital research funds. The founder had no role in the study design, data collection and analysis, decision to publish or preparation of the manuscript.
Participants	Place of study: Uusimaa, Finland Number of participants assigned: 60 participants (28 surgical; 32 conservative) Number of participants assessed: 51 participants (26 surgical; 25 conservative) Inclusion criteria Aged 18‐70 years A completely displaced middle third clavicle fracture, no cortical contact between main fragments Fresh fracture, treatment within 7 days after injury Provided informed consent Exclusion criteria People with multiple injuries Associated neurovascular injury, or suspicion of it Cancer or any severe illness impairing health Pathological fracture Treatment 7 days after injury Open fracture Corticosteroid or immunosuppressive medication Upper extremity fracture at same time An earlier clavicle or shoulder region fracture Pregnancy Lack of consent Age: Surgical group (mean/SD): 41/10.8 years Conservative group (mean/SD): 33/12.0 years Gender of participants assigned (male/female) Surgical group: 24/4 Conservative group: 28/4 Classification of injury: fractures were classified according to the AO/OTA fracture classification and displacement and shortening of the fragments.
Interventions	Timing of intervention: not specified; however, participants with > 7 days after the injury were excluded Type of surgical intervention: open reduction and anterior plate fixation with stainless steel reconstruction plate and screws Type of conservative intervention: simple sling for 3 weeks Rehabilitation: Surgical group: sling was used for comfort for 3 weeks; the postoperative exercise protocol was similar to that in the conservative group. No implant removal was scheduled. Conservative group: pendulum motion was permitted during the 1st 3 weeks, followed by active abduction and flexion up to the horizontal plane from 3‐6 weeks. The full range of active motion was permitted after 6 weeks, and return to full activities was permitted after 3 months. Any co‐interventions: not reported
Outcomes	Length of follow‐up Follow‐up was 1 year Participants were evaluated at 3 and 6 weeks, and at 3 and 12 months Loss to follow‐up: 9 participants were lost to follow‐up at 1 year Surgical group: 2 participants were lost at 1 year Conservative group: 7 participants were lost at 1 year Primary outcomes Function or disability measured by Constant score and DASH score Pain measured using VAS Failure of treatment measured by hardware failure, symptomatic non‐union or malunion Secondary outcomes Other treatment failure measured by: asymptomatic non‐union and symptomatic malunion Adverse events measured by: hardware irritation requiring removal refracture
Notes	Composite adverse events: surgery: brachial plexus irritation (n = 1); refracture (1). Conservative: mild plate irritation (n = 1); symptomatic malunion (n = 2); refracture (2). In Analysis 1.14 we used event rates 1 for surgery and 2 for conservative. Cosmetic result events: not reported
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Block randomisation was used, and block size varied randomly between 4 and ten. Participants were randomly assigned to 2 parallel groups, initially at a 1:1 ratio, to receive either conservative or surgical intervention.
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.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Missing outcome data were not balanced in numbers across intervention groups; however, intention‐to‐treat analysis was performed.
Selective reporting (reporting bias)	Low risk	The study protocol was available and all of the study's pre‐specified (primary and secondary) outcomes that were of interest for this review were reported in the pre‐specified way.
Other bias	Low risk	The study appears to be free of other sources of bias.