COTS 2007.

Methods	Study design: multicentre RCT (8 centres) Duration of the study: April 2001‐December 2004 Protocol was published before recruitment of participants: not reported, but no protocol was published Details of trial registration: not registered Funding sources: some of the study authors received grants or outside funding from the Orthopaedic Trauma Association and Zimmer Inc
Participants	Place of study: 8 centres from Canada, including St. Michael's Hospital and Sunnybrook and Women's College Health Sciences Centre, Toronto; McMaster University Medical Center, Hamilton; Brantford General Hospital, Brantford; London Health Sciences Centre, London, Ontario; Royal Columbian Hospital, New Westminster, British Columbia; Montreal General Hospital, Montreal, Quebec; and Foothills Medical Centre, Calgary Number of participants assigned: 132 participants (67 surgical; 65 conservative) Number of participants assessed: 111 participants (1‐year follow‐up: 62 surgical, 49 conservative) and 95 participants (2‐year follow‐up: 52 surgical, 43 conservative). Inclusion criteria Aged 16‐60 years Middle third clavicle fracture (amenable to plate fixation with at least 3 screws in each proximal and distal fragment) Completely displaced clavicle fracture (no cortical contact between the fragments) No medical contraindications to general anaesthesia Provided informed consent Exclusion criteria Fracture in the proximal or distal third of the clavicle Pathological or open fracture Fracture with > 28 days after the injury Neurovascular injury associated Head injury associated (a Glasgow Coma Scale score < 12) Upper extremity fracture distal to the shoulder Inability to comply with follow‐up (a transient ability or an inability to read or complete forms) Medical contraindication to surgery and/or anaesthesia (such as heart disease, renal failure or active chemotherapy) Age: Surgical group (mean/SD): 33.5/12.7 years Conservative group (mean/SD): 33.5/12.1 years Gender of participants assessed (male/female) Surgical group: 53/9 Conservative group: 34/15 Classification of injury: fractures were classified according to displacement, angulation and shortening of the fragments
Interventions	Timing of intervention: not specified; however, participants with > 28 days after the injury were excluded Type of surgical intervention: open reduction and plate fixation (44 participants with limited contact dynamic compression plates; 15 with 3.5‐mm reconstruction plates; 4 with pre‐contoured plates; and 4 with other plates) Type of conservative intervention: standard sling for 6 weeks Rehabilitation: Surgical group: sling was used for comfort for 7‐10 days, physiotherapy was not specified and participants returned to full activities at 3 months Conservative group: rehabilitation was not specified Any co‐interventions: not reported
Outcomes	Length of follow‐up The follow‐up was 2 years Participants were evaluated at 6 weeks, 3, 6 and 12 months (COTS 2007) and 2 years (Schemitsch 2011) Loss to follow‐up: 21 participants were lost to follow‐up at 1 year (COTS 2007) and 37 participants at 2 years (Schemitsch 2011) Surgical group ‐ 5 participants were lost at 2 years Conservative group ‐ 22 participants were lost at 2 years Primary outcomes Function or disability measured by DASH and Constant score Failure of treatment (number of participants who have undergone or are considered for a non‐routine secondary surgical intervention for symptomatic non‐union, malunion or other complications) Secondary outcomes Other treatment failure measured by: cosmetic results (measured by deformity, asymmetrical result, unsightly scar, hardware prominence) Adverse events measured by: short‐term follow‐up: wound infection and/or dehiscence (time point not reported); transient brachial plexus symptoms long‐term follow‐up: hardware irritation requiring removal, skin and nerve problems and stiffness, complex regional pain syndrome (time point not reported)
Notes	SF‐36 questionnaire was assessed in a publication of COTS 2007 (Pearson 2010); however, the results were not available for analysis because the study authors did not report them. DASH and Constant scores presented graphically only We contacted the study authors to request data (i.e. SD for DASH and Constant scores); however, they declined to provide them. We extracted 12‐month Constant and DASH scores as reported in a review by the same authors (McKee 2012), extracted the (presumed) SEM from the graphical presentation of results and calculated the SD from the SEM. Composite adverse events: surgery: wound infection and/or dehiscence (n = 3); hardware irritation requiring removal (n = 5); incisional numbness (n = 18); sensitive and/or painful fracture site (n = 10); transient brachial plexus symptoms (n = 8). Conservative: symptomatic malunion (n = 9); sensitive and/or painful fracture site (n = 10); complex regional pain syndrome (n = 1); transient brachial plexus symptoms (n = 7). In Analysis 1.14 we used event rates 18 for surgery and 10 for conservative Cosmetic result events: surgery: deformity (n = 2); scar (n = 3); hardware prominence (n = 11). Conservative: deformity (n = 3); asymmetrical results (n= 22). In Analysis 1.11 we used event rates 14 for surgery and 22 for conservative.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The sequence generation was performed by a computer random number generator (personal contact)
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	High risk	Missing outcome data were not balanced in numbers across intervention groups; more participants in the conservative intervention group were lost to follow‐up (5/67) (7%): surgical versus 16/65 (25%) conservative at 12 months; 15/67 (22%) surgical versus 22/65 (34%) conservative at 24 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 study author failed to report any measure of variance for DASH and Constant scores.
Other bias	High risk	The trial did not permit investigators to tell whether or not the 2 groups were balanced at baseline for the primary outcomes.