Webb 2006.

Methods	Quasi‐randomised trial Study period: April 2002 to December 2003
Participants	The Women’s and Children’s Hospital of Buffalo, Buffalo, New York, USA 127 children with displaced fractures of the distal third of the forearm Exclusion: age under 4 years, open fracture, pathologic fracture, a refracture through pre‐existing fracture lines, closed physes Sex: 85 male (75% of 113) Age: mean 9.8 years, range 4 to 16 years Fracture type: partially or completely displaced fractures of radius only (49 including 17 physeal fractures) or combined radius and ulna (64); 23 complete radius fractures Assigned: 63 (below‐elbow) / 64 (above‐elbow) Analysed: 53 / 60 (at 8 to 10 weeks); 104 (92%) were followed up at 7.7 months (see Notes)
Interventions	Manipulation and reduction (manual method) by orthopaedic resident at emergency department with analgesia and sedation provided. The hand was held by an assistant while a circumferential plaster cast was applied; if assistant not available fingertraps were applied but the arm was not suspended until after manipulation 1. Below‐elbow plaster cast 2. Above‐elbow plaster cast: The short‐arm portion was applied first and moulded and then the plaster was extended above the elbow Strict elevation for first 24 to 48 hours. First follow‐up visit at 7 to 10 days; with the intention of a remanipulation under general anaesthesia if unacceptable alignment. At 4 weeks, cast was removed if radiological and clinical evidence of healing and participants instructed to perform range‐of‐motion exercises at home. Otherwise, casts left in place for another 2 weeks but above‐elbow casts were cut down to below‐elbow casts. Clinical examination at 8 to 10 weeks and physical therapy prescribed if restricted mobility
Outcomes	Length of follow‐up: mean 7.7 months (3.5 to 11 months) (telephone interview); also at 7 to 10 days, 4 weeks and 8 to 10 weeks (questionnaire on impact of cast on ADLs) Function: ADL during cast use (questionnaire at 8 to 10 weeks) Redisplacement (lost reduction in cast) and remanipulation (some criteria reported but not clear if applied) Duration in cast Complications: refractures (none); stiff elbow requiring physical therapy Range of elbow and wrist motion (cast removal around 6 weeks and 8 to 10 weeks) Time to regain range of motion (questionnaire at 8 to 10 weeks) Days missed school Radiological outcomes: displacement, angulation, deviation
Funding and declarations of interest	Funding source: "The authors did not receive grants or outside funding in support of their research for or preparation of this manuscript." Declarations of interest: reported, none
Notes	Of 10 children in the below‐elbow cast group excluded from the analyses, 7 were lost to follow‐up and 3 were excluded because of surgery. Of 4 children in the above‐elbow cast group excluded from the analyses, 3 were lost to follow‐up and 1 was excluded because of surgery. No results explicitly provided for the 104 participants (numbers in each group not reported) followed‐up via telephone interview at 7.7 months
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	High risk	Quote; “Patients were then randomised to be treated with either a short or a long arm cast on the basis of whether the last digit of their medical record number was odd or even.” Quasi‐randomised: sequence generation is not random
Allocation concealment (selection bias)	High risk	Quote: “Patients were then randomised to be treated with either a short or a long arm cast on the basis of whether the last digit of their medical record number was odd or even.” Predictable sequence: no allocation concealment
Blinding of participants and personnel (performance bias) Subjective outcomes	High risk	Blinding of participants, their parents and care providers not practical
Blinding of participants and personnel (performance bias) Objective outcomes	High risk	Blinding of participants, their parents and care providers not practical
Blinding of outcome assessment (detection bias) Subjective outcomes	High risk	Not blinded
Blinding of outcome assessment (detection bias) Objective outcomes	High risk	Not blinded
Incomplete outcome data (attrition bias) Subjective outcomes	High risk	Differences between the 2 groups in losses (exclusions and losses): below‐elbow 10/63 (16%) versus above‐elbow 4/64 (6%) Additionally, greater losses relating to missing questionnaire responses at 10 weeks, e.g. losses for difficulties with ADLs were 14/63 (22%) versus 6/64 (9%)
Incomplete outcome data (attrition bias) Objective outcomes	High risk	Differences between the two groups in losses (exclusions and losses): below‐elbow 10/63 (16%) versus above‐elbow 3/64 (5%). Additionally, greater losses relating to missing questionnaire responses at 10 weeks
Selective reporting (reporting bias)	High risk	No trial registration or published protocol. Insufficent details of the ADL questionnaire at 10 weeks and no details of telephone interview in Methods or Results
Other bias: major imbalance in baseline characteristics	Unclear risk	Baseline characteristics only provided for 113 participants in the analysis, not the 127 randomised participants No major imbalances (upon statistical testing) but notably more radius‐only fractures in the below‐elbow group (27/53 (51%)) than in the above‐elbow group (22/60 (37%)), and thus conversely more combined radius and ulna fractures in the above‐elbow group. This distribution might reflect some selection bias. This issue was highlighted in a letter commenting on this trial (Kumar 2006)
Other bias: performance bias	Unclear risk	All of the “orthopaedic residents,… had been fully trained in the proper application of plaster casts.” In Discussion: Quote: "Our casts were all applied by orthopedic residents in their third or fourth year of training, with varied amounts of experience in pediatric orthopaedics. There is a learning curve in the application of a well‐molded cast, and the majority of poorly molded casts were applied by residents early in their pediatric orthopaedic training." There was no difference between the 2 groups in the mean cast index
Other bias	Low risk	No other source of bias identified