Non-surgical casting versus surgical reduction for children with severely displaced distal radial fractures (the CRAFFT Study): a multicentre, randomised, controlled non-inferiority trial and economic evaluation.

# Non-surgical casting versus surgical reduction for children with severely displaced distal radial fractures (the CRAFFT Study): a multicentre, randomised, controlled non-inferiority trial and economic evaluation *The Lancet 2026* Articles Non-surgical casting versus surgical reduction for children with severely displaced distal radial fractures (the CRAFFT Study): a multicentre, randomised, controlled non- inferiority trial and economic evaluation Daniel C Perry, Alex Zimmermann, Juul Achten, Nicolas Nicolaou, David Metcalfe, Daphne Kounali, Mark D Lyttle, Samuel Frempong, Duncan Appelbe, James Mason, Rebecca Kandiyali, Phoebe Gibson, Matthew L Costa, on behalf of the CRAFFT Study Investigators (TOTS Research Collaborative and the PERUKI Network)* Summary Lancet 2026; 407: 1538–48 Background Severely displaced distal radial fractures are among the most common and controversial injuries in Published Online children. Despite observational evidence of reliable remodelling with growth in younger children, their alarming April 8, 2026 radiographic appearance—particularly when completely displaced (o-ended)—has driven routine surgical reduction https://doi.org/10.1016/ and fixation. The Children’s Radius Acute Fracture Fixation Trial (CRAFFT) aimed to evaluate the clinical and cost- S0140-6736(26)00409-5 eectiveness of surgical reduction compared with non-surgical casting. See Comment page 1484 *Members are listed in Methods This pragmatic, multicentre, randomised, non-inferiority trial included participants (aged 4–10 years) with a appendix 1 (pp 23–27) severely displaced distal radial fracture from 49 hospitals in the UK. Recruiting centres were secondary or tertiary care Liverpool Institute of Child hospitals providing acute paediatric trauma care. Participants were randomly assigned to either non-surgical casting Health and Well Being, University of Liverpool, or surgical reduction, using a minimisation algorithm with a random element and stratification factors were centre, Liverpool, UK age group, fracture location, and displacement severity. Participants and their parents and carers could not be masked (Prof D C Perry PhD); Alder Hey to treatment. Surgical reduction was performed under general anaesthesia or conscious sedation, to restore anatomical Children’s Hospital, Liverpool, alignment, with fixation permitted at the discretion of the surgeon. Non-surgical care involved immobilisation of the UK (Prof D C Perry); Kadoorie Institute for Trauma, fracture in a plaster cast without general anaesthesia or sedation, and without purposeful manipulation of the fracture Emergency and Critical Care position. Immobilisation of the fracture beyond 6 weeks post-randomisation was not recommended. The primary (Prof D C Perry, J Achten PhD, outcome was upper limb function at 3 months, measured using the Patient Report Outcomes Measurement System D Metcalfe PhD, D Appelbe PhD, (PROMIS) Upper Extremity Score for Children in the intention-to-treat population, which included all participants in Prof M L Costa PhD), and Oxford Clinical Trials Research Unit the groups to which they were randomly assigned, irrespective of treatment received. The non-inferiority margin was (A Zimmermann MSc, conservatively set at –2·5 points for the main trial population. A prespecified subgroup analysis was powered to D Kounali PhD), Nuffield assess whether non-surgical casting could exclude a larger more clinically relevant margin of –5 points among Department of Orthopaedics, children with completely o-ended fractures. Complications and serious adverse events were summarised in a safety Rheumatology and Musculoskeletal Sciences, (as-treated) population defined by treatment received. A within-trial economic evaluation was undertaken from the University of Oxford, Oxford, perspective of the UK National Health Service (NHS) and Personal Social Services over a 12-month time period. The UK; Sheffield Children’s trial was registered with the ISRCTN registry, ISRCTN10931294, recruitment is complete and extended follow-up to Hospital, Sheffield, UK 3-years post-randomisation is ongoing. (N Nicolaou MBBS); Emergency Department, Bristol Royal Hospital for Children, Bristol, Findings Between Aug 11, 2020, and May 30, 2024, 1227 children were screened for eligibility across 49 UK hospitals. UK (M D Lyttle MBChB); 477 children were excluded (54 met exclusion criteria and 423 did not enter the study, the majority for lack of clinical Research in Emergency Care Avon Collaborative Hub or parental equipoise). 750 participants were randomly assigned, 375 to the non-surgical casting group and 375 to (REACH), University of the the surgical reduction group. 456 (61%) participants were male, 294 (39%) were female, and the median age of West of England, Bristol, UK participants was 7·9 years (IQR 6·5–9·5). 329 (44%) of the 750 participants had completely o-ended fractures. (M D Lyttle); Warwick Clinical Primary outcome data were collected from 640 (85%) participants. At 3 months post-randomisation, the mean Trials Unit, University of Warwick, Warwick, UK PROMIS Upper Extremity score was 44·9 (SD 8·7) in the non-surgical casting group and 46·6 (8·8) in the surgical (S Frempong PhD, reduction group (adjusted mean dierence –1·64 [95% CI –2·84 to –0·44]), with the confidence interval favouring Prof J Mason PhD, surgical reduction but extending beyond the prespecified non-inferiority margin of –2·5 points. In children with R Kandiyali PhD); Wirral, completely o-ended fractures, findings were consistent with non-inferiority against the wider, prespecified margin Merseyside, UK (P Gibson BA Hons) for this group. Most complications within 8 weeks occurred in the surgical reduction group, including pressure damage (n=2), wound infections (n=6), scarring (n=5), and nerve irritation (n=1). During the 12-months of follow- Correspondence to: Prof Daniel C Perry, Liverpool up, refracture occurred in 13 participants (nine after non-surgical casting and four after surgical reduction). From an Institute of Child Health and Well NHS and Personal Social Services perspective, non-surgical casting was associated with a significant reduction in Being, University of Liverpool, mean cost per patient of £1665 (95% CI 1487 to 1843) and a marginal incremental reduction in quality-adjusted life- Liverpool L7 8XP, UK danperry@liverpool.ac.uk years (QALYs; –0·023 [95% CI –0·037 to –0·009]). The probability of non-surgical casting being cost-eective at the £20 000 and £30 000 per QALY threshold was 100%, indicating that the small short-term functional advantage of See Online for appendix 1 surgical reduction was not cost-eective. 1538 Articles Interpretation The CRAFFT trial did not demonstrate non-inferiority of non-surgical casting at 3 months against a conservative margin; however, the observed dierence in favour of surgical reduction was small, below thresholds that families considered meaningful, and did not persist beyond early recovery. Surgical reduction was associated with higher costs, early procedural complications, and only a modest improvement in cosmetic appearance, supporting consideration of a cast-first strategy for most children. Funding National Institute for Health and Care Research (NIHR) Health Technology Assessment programme (17/22/02) and the NIHR Oxford Biomedical Research Centre. Copyright © 2026 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4.0 license. Introduction Observational studies of distal radial fractures in Distal radial fractures are the most common bony injury children up to age 10 years suggest that even completely in children, accounting for about half of paediatric displaced fractures—including those termed as long-bone fractures.1 Severely displaced injuries are completely o-ended fractures—can correct over time typically treated by closed reduction under sedation or with restoration of function and little long-term general anaesthesia, often followed by Kirschner wire or deformity.5,6 Although such fractures appear visibly bent plate fixation.2 This procedure restores immediate and alarming to families and clinicians, these sometimes alignment, at the cost of anaesthesia, procedural pain, realign naturally without manipulation or fixation. A and frequent complications, including re-displacement small UK comparative cohort found more complications and wire-related problems.3,4 Despite this established in children undergoing surgical reduction than in those practice, there is a growing uncertainty about whether treated with simple casting.7 The only randomised trial to surgical reduction is necessary. date, published only as an abstract and based on Children’s bones have a remarkable capacity to radiographic outcomes, reported no advantage for remodel, distinguishing them from adult bones. surgical reduction in children younger than 11 years.8 Through this process, deformity corrects as the child This uncertainty led the British Society for Children’s grows, with the greatest potential for remodelling found Orthopaedic Surgery to prioritise the treatment of these close to areas of active growth, particularly the growth fractures for future research, resulting in the development plates (physes). Remodelling is greatest in younger of the Children’s Radius Acute Fracture Fixation Trial children and close to active growth plates. (CRAFFT).9 The trial was designed to determine whether Research in context Evidence before this study confirm non-inferiority of non-surgical casting at 3 months We searched PubMed, Embase, and the Cochrane Central using a conservative margin set at the lower end of the Register of Controlled Trials (CENTRAL) at the time of funding minimally important difference. However, the observed in May, 2017, and repeated the search iteratively throughout difference was smaller than the threshold families considered the trial, most recently on Dec 3, 2025, using combinations of meaningful to justify surgical reduction, and by 6 months and the terms “distal radius fracture”, “children”, “paediatric”, 12 months there was no evidence of any difference between “pediatric”, “surgery”, “operative”, “non-operative”, and treatments. Surgical reduction was associated with higher “randomised trial”, with Boolean operators and truncation as costs, early procedural complications, and only modestly better appropriate. These searches identified no randomised cosmetic appearance, which narrowed over time, highlighting controlled trials addressing the question, a Cochrane review the trade-off between early gains and longer-term outcomes. highlighting the paucity of evidence for the treatment of Implications of all the available evidence children’s wrist fractures, and only case series relevant to the The CRAFFT trial provides the most robust evidence to date for comparison of interest. Although these case series suggested the treatment of these injuries, adding to previous case series reliable remodelling with growth in younger children, the and an earlier small, randomised trial. This is a topic not pronounced visual deformity and radiographic displacement of addressed in any international guidelines. Although surgical these injuries have continued to drive routine surgery. One reduction offered small early advantages, these fell below small randomised trial, published only in abstract form, thresholds families considered clinically meaningful and did not reported no benefit of surgery, leaving a persistent absence of persist. Given the higher costs and avoidable risks associated robust evidence to guide practice. with surgical reduction, the findings support a cast-first Added value of this study strategy as the preferred initial approach for most children with The CRAFFT trial is the first large, multicentre randomised study severely displaced distal radial fractures. to address this long-standing uncertainty. The trial did not Articles non-surgical casting provides comparable functional preventing adherence to trial procedures or follow-up outcomes to surgical reduction for severely displaced (eg, insucient English comprehension, developmental distal radial fractures in children, while reducing delay or abnormality, or no access to mobile data or the treatment-related harms and costs. internet). Eligible children and their families were approached by Methods a local researcher in either the emergency department or Study design orthopaedic fracture clinic. A suite of digital information The CRAFFT study was a pragmatic, multicentre, and animations that had been co-produced with families randomised, controlled non-inferiority trial conducted in were used to explain the study. Families were asked to 49 UK hospitals. Recruiting centres were secondary or provide written informed consent (parents or carers) and tertiary care hospitals providing acute paediatric trauma assent (children). All trial procedures were administered care. The trial was coordinated by the Oxford Clinical electronically throughout the study, including consent, Trials Research Unit, sponsored by the University of randomisation, and data collection. Sex was recorded as Oxford. A trial steering committee and data and safety male or female from parental report at recruitment. monitoring committee oversaw progress, conduct, and Ethnicity was not recorded. participant safety. The trial was approved by the UK National Health Service (NHS) Research Ethics Randomisation and masking Committee (West Midlands, Solihull, reference Enrolment and randomisation was overseen by the local 20/WM/0054) and registered prospectively with the researcher in either the emergency department or ISRCTN registry, ISRCTN10931294; recruitment is orthopaedic fracture clinic. Participants were randomly complete and extended follow-up until 3 years post- assigned (1:1) to surgical reduction or non-surgical See Online for appendices 2–4 randomisation is ongoing. The protocol (appendix 2), casting using a secure web-based system provided by the statistical analysis plan (appendix 3), and health Oxford Clinical Trials Research Unit. Randomisation economics analysis plan (appendix 4) have been used a minimisation algorithm with a random element published.10,11 In response to the COVID-19 pandemic, to ensure balanced allocation across treatment groups. some participating centres temporarily paused Stratification factors were centre, age group (4–6 years vs recruitment during local restrictions. No significant 7–10 years), fracture location (metaphyseal vs physeal), changes were made to the trial protocol, design, and displacement severity (completely o-ended vs not outcomes, or analysis plan. completely o-ended). Stratification by age group and Patient and public involvement was embedded fracture location ensured balance across groups with throughout the trial. Patient and public involvement reduced remodelling potential with increasing age and in contributors played a central role in the study’s design, fractures further from the growth plate. conduct, oversight, and dissemination. They informed Participants and their parents and carers could not be the selection of outcomes most relevant to families, masked to treatment. The treating clinicians could not be guided the definition of a meaningful treatment eect, masked to the treatments they were providing; however, and co-developed participant-facing materials, including the treating clinical team did not take part in the animations used during recruitment. Patient and public follow-up assessment of participants. Outcome data were involvement members also contributed to the design of collected directly from the participant or their parent or the embedded qualitative study to optimise enrolment. carer by the central trial team, using electronic data Families continue to play a key role in dissemination capture. The central trial team and statisticians were eorts. masked to treatment allocation during data cleaning. There were no independent clinician outcome assessors. Participants Children aged 4–10 years inclusive, with a severely Procedures displaced distal radial fracture (metaphyseal or Salter– Surgical reduction was performed under general Harris II), with or without an associated ulnar fracture, anaesthesia or conscious sedation by surgical teams, were eligible if the treating clinician considered that according to local practice. The fracture was manipulated surgical reduction would usually be performed. The age to restore anatomical alignment, and fixation was range reflected the age for which both clinical equipoise permitted at the discretion of the surgeon. The method of and validated patient-reported outcome measurement fixation was recorded. Non-surgical care involved were most appropriate. Eligible fractures included both immobilisation of the fracture in a plaster cast by either completely o-ended fractures and those displaced but orthopaedic surgeons or emergency physicians without not completely o-ended, as both groups reflect common general anaesthesia or sedation, to hold the bones in the presentations in practice. Exclusion criteria were an optimal position without purposeful manipulation of the injury more than 7 days old; a more complex wrist fracture position. The choice of cast type (above-elbow or fracture (eg, open or extending into the joint); additional below-elbow) followed local practice guidelines. Details fractured bones elsewhere in the body; or factors of non-surgical casting were recorded. Analgesia was 1540 Articles prescribed at the discretion of the treating clinician, were assessed can be found in the protocol. This report following hospital guidelines, and details of analgesic use presents outcomes to 12 months. Prespecified long-term were recorded. Rehabilitation followed standard practice outcomes at 2 years and 3 years will be reported separately. at the treating centre. Immobilisation beyond 6 weeks post-randomisation was not recommended. Details of Statistical analysis rehabilitation, including physiotherapy sessions, were The minimally important dierence for PROMIS recorded. No artificial intelligence or machine learning measures in children is typically about 3–4 points, based tools were used in data collection, data management, or on paediatric anchor-based studies.18,19 During study statistical analysis for this study. development, families advising the study team indicated that, in the context of surgery, a dierence greater than Outcomes this minimum would be required to justify operative Outcome data were collected using REDCap electronic intervention, identifying a threshold of approximately data capture tools,12 with data handling procedures 5 points. The non-inferiority margin was therefore prespecified in the trial protocol. Families were prompted conservatively set at –2·5 points, representing half the by e-mail or SMS to complete follow-up questionnaires at value required by families and below published 6 weeks, 3 months, 6 months, and 12 months post- minimally important dierences. Sample size estimation randomisation. The primary outcome timepoint, outcome used a one-sided α of 0·025, reflecting the directional choices, and follow-up schedule were established in nature of the non-inferiority hypothesis. To achieve 90% conjunction with families advising the study. If there was power, with a one-sided α of 0·025, 674 participants no response to the initial and reminder messages, families providing primary outcome data at 3 months (337 were contacted by telephone. per group) were required. Allowing for 10% loss to The primary outcome for this study was functional follow-up, the total planned recruitment was recovery assessed using the Patient Report Outcomes 750 participants. Measurement System (PROMIS Bank version 2.0) Upper The main trial population consisted of children Extremity Score for Children Computer Adaptive Test at with fractures that were either completely o-ended or 3-months post-randomisation. This is a validated patient- those displaced but not completely o-ended. Completely reported outcome system developed with the US National o-ended fractures represent the most severe injuries Institutes of Health.13,14 Measures are reported as and are those most likely to prompt invasive treatment, standardised T-scores derived using item response theory including internal fixation. We therefore prespecified (mean 50 [SD 10]), with lower scores indicating worse a subgroup analysis powered to assess whether upper-extremity function. In this study population, non-surgical casting could exclude a larger dierence observed scores ranged from approximately 13·7 to 57·3. (5 points) that families considered meaningful in this Given the age of participants, the parent-proxy version context. Recruitment to the main population was was used throughout. planned to continue until at least 200 children with A range of secondary outcomes were measured. Pain completely o-ended fractures were enrolled, providing was measured using the Wong–Baker FACES Pain Rating adequate power to evaluate non-inferiority against this Scale, a validated self-reported tool measuring pain wider, clinically relevant margin of –5 points in this intensity on a 0–10 scale.15 Health-related quality of life subgroup. was measured using EQ-5D-Y, a validated tool for All primary and secondary clinical outcomes were assessing childhood health-related quality of life16 analysed in the intention-to-treat population, with all (EQ-5D-Y consists of two components: a descriptive participants analysed according to their randomised system which provides utility scores for economic allocation, irrespective of the treatment received. The evaluation, and a visual analogue scale [VAS]). An primary estimand reflected a treatment policy strategy, EQ-5D-Y value set for UK children is not yet available, estimating the average treatment eect of surgical therefore the reported analysis relates to the adult utility reduction versus non-surgical casting on upper limb value set for the EQ-5D-3L.17 All complications, including function at 3 months, irrespective of adherence. A loss of reduction, infection, refracture, unplanned per-protocol analysis was also undertaken, excluding surgery, and adverse events related to cast or surgery, participants who did not receive their allocated inter- were recorded throughout the study. Health-care vention as intended. Safety outcomes were summarised utilisation data, consisting of details of treatments descriptively according to the treatment received (as-treated received and interventions related to any complications, population), in line with standard reporting guidance. along with all health-care contacts related to the injury, Prespecified subgroup analyses were undertaken were collected. Data were sourced from both families and by displacement severity (completely o-ended vs not hospital records. Parental work absence, out-of-pocket o-ended), age group (4–6 vs 7–10 years), and fracture costs, and school absence were also recorded. Cosmesis location (metaphyseal vs physeal), to examine consistency was assessed using a parent-reported VAS. Further details of treatment eects across key clinical groups. The trial of secondary outcomes and the timepoints at which they was not powered to detect interaction eects. Articles 1227 patients met inclusion criteria Non-surgical Surgical Total casting group reduction group (N=750) (n=375) (n=375) 477 excluded Sex 54 met exclusion criteria Male 225 (60%) 231 (62%) 456 (61%) 340 declined to participate 83 unable to enter study Female 150 (40%) 144 (38%) 294 (39%) Age, years Median (IQR) 7·8 (6·5–9·5) 8·0 (6·5–9·6) 7·9 (6·5–9·5) 750 randomly assigned Age group 7–10 years 249 (66%) 249 (66%) 498 (66%) 4–6 years 126 (34%) 126 (34%) 252 (34%) 375 assigned to non-surgical casting 375 assigned to surgical reduction Mechanism of injury* group group High energy fall 192 (51%) 202 (54%) 394 (53%) 349 received non-surgical casting 336 received surgical reduction 26 received surgical reduction 39 received non-surgical casting Low energy fall 183 (49%) 173 (46%) 356 (47%) Side of injury Left 213 (57%) 224 (60%) 437 (58%) 75 did not complete follow-up 67 did not complete follow-up 3 at baseline 1 at baseline Right 162 (43%) 151 (40%) 313 (42%) 3 withdrew 1 lost to follow-up Injury to dominant arm 19 at week 6 21 at week 6 18 lost to follow-up 20 lost to follow-up No 195 (52%) 208 (55%) 403 (54%) 1 withdrew 1 withdrew Yes 172 (46%) 160 (43%) 332 (44%) 11 at month 3 11 at month 3 7 lost to follow-up 10 lost to follow-up Unsure or 7 (2%) 7 (2%) 14 (2%) 4 withdrew 1 withdrew ambidextrous 16 at month 6 14 at month 6 Missing 1 (<1%) 0 1 (<1%) 16 lost to follow-up 14 lost to follow-up 26 at month 12 20 at month 12 Displacement severity 26 lost to follow-up 18 lost to follow-up Completely off- 165 (44%) 164 (44%) 329 (44%) 2 withdrew ended Not completely 210 (56%) 211 (56%) 421 (56%) off-ended 375 included in intention-to-treat 375 included in intention-to-treat analysis analysis Fracture location Metaphyseal 303 (81%) 302 (81%) 605 (81%) Figure 1: Trial profile fracture Withdrawal indicates that the participant actively declined further study participation from that point onwards. Physeal fracture 72 (19%) 73 (19%) 145 (19%) Data are n (%). *The mechanism of injury was determined by the treating clinician The primary model was a three-level mixed-eects for each participant. model, with repeated measures nested within individuals, and individuals nested within centres. The baseline score Table 1: Baseline characteristics was included as part of the outcome. The model featured a full interaction between treatment and time, enabling estimation of treatment dierences at each follow-up displacement severity, and fracture location). A point point. The model was adjusted for the randomisation estimate of the mean dierence for each statistical test stratification factors. A restricted maximum-likelihood- was presented with corresponding two-sided 95% CIs based repeated measures model with an unstructured and p values. covariance matrix was used. Full model specification was A within-trial economic evaluation was conducted from prespecified in the statistical analysis plan (appendix 3). the perspective of the UK NHS and Personal Social A complier-average causal eect analysis was prespecified Services, over a 12-month time period, in line with the UK for the primary PROMIS Upper Extremity outcome at National Institute for Health and Care Excellence (NICE) 3 months. reference case.20 All participants were included in the All randomly assigned participants contributed to the economic analysis. Health-care resource use was costed primary analyses using all available data across in pounds sterling (£), using 2022 unit costs derived from timepoints. The model operated as an implicit imputation national published sources (appendix 1 pp 2–4).21,22 model under missing at random assumptions, Quality-adjusted life-years (QALYs) were estimated using conditional on observed outcome data at other timepoints the area-under-the-curve method, adjusted for baseline and adjustment factors included in the model. Sensitivity imbalances. analyses for informative missingness of the primary A bivariate regression model of costs and QALYs, outcome data at 3 months were undertaken using adjusted for relevant covariates and combined with non- multiple imputation, incorporating treatment group, parametric bootstrap sampling, was used to estimate time, and randomisation stratification factors (age, total and incremental costs, QALYs, and net monetary 1542 Articles benefits, along with associated confidence intervals. have neurovascular compromise requiring urgent Multiple imputation was employed to address missing reduction), or families declining consent (n=340). data for both costs and EQ-5D utility scores.10 Declined consent was most often due to a treatment The primary analysis followed an intention-to-treat preference (n=246), with 155 preferring surgical approach based on the economic analysis set, reporting reduction and 91 preferring non-surgical casting net monetary benefits at willingness-to-pay thresholds of (appendix 1 p 7). 750 participants were randomly £20 000 and £30 000 per QALY gained. Sensitivity analyses assigned, 375 to the non-surgical casting group and 375 considered scenarios including complete case data, use to the surgical reduction group (figure 1). of the Dutch EQ-5D-Y-3L value set, and exploration of Baseline demographics (age, sex, and time from injury) societal costs to include parental productivity losses, were similar between those screened and those recruited days of purchased childcare, private physiotherapy, and (appendix 1 p 8). Baseline demographic characteristics by purchased medication. We prespecified our analysis in a intervention group were similar and are summarised in health economics analysis plan (appendix 4) and report table 1, and patient-reported outcome measures are findings in accordance with the Consolidated Health presented in appendix 1 (p 9). 456 participants (61%) were Economic Evaluation Reporting Standards statement.23 male and 294 (39%) were female, and the median age of All statistical analyses were undertaken using Stata and participants was 7·9 years (IQR 6·5–9·5). Overall, R (version 4.3.1). No formal interim ecacy analyses or 329 (44%) of the 750 participants had completely stopping guidelines were specified; the Data Monitoring o-ended fractures. Committee reviewed unmasked safety data at prespecified Primary outcome data at 3 months were available for intervals and could recommend modifications or early 640 (85%) of the 750 randomly assigned participants. termination for safety concerns. Across the 750 participants, there were 78 protocol deviations, of which 73 were considered important. Role of the funding source Of these important protocol deviations, there were The funders had no role in study design, data collection, 65 inter vention crossovers, three consent deviations, data analysis, data interpretation, or writing of the report. and five randomisation errors. The intention-to-treat population included all 750 participants and the per- Results protocol population excluded those with major protocol Between Aug 11, 2020, and May, 30, 2024, 1227 children deviations and comprised 677 participants (344 in the were screened for eligibility across 49 UK hospitals non-surgical casting group and 333 in the surgical (appendix 1 pp 5–6). Of these, 477 were not randomised: reduction group). 54 met exclusion criteria and 423 were eligible but did In the surgical reduction group, 336 (90%) of not enter the study (figure 1). The most frequent 375 underwent treatment as assigned. Almost all reasons for exclusion were an injury more than 7 days procedures were performed under general anaesthesia old (n=17) and inability to adhere to trial procedures or (291 [87%] of 336), with 45 (13%) under conscious or follow-up (n=16). Among eligible patients, the most procedural sedation. Stabilisation was achieved with common reasons for non-participation were lack of plaster cast application alone in 269 (72%) participants, clinician equipoise (n=81; of which 12 were shown to Kirschner wires in 100 (27%) participants, plates in Non-surgical casting Surgical reduction PROMIS Upper Extremity Mean difference (95% CI) Score treatment difference All patients 44·9 46·6 –1·64 (–2·84 to –0·44) Fracture location Physeal fracture 45·8 47·4 –1·66 (–2·89 to –0·43) Metaphyseal 44·7 46·3 –1·52 (–3·47 to 0·44) Displacement severity Completely off-ended 44·2 45·8 –1·39 (–2·86 to 0·09) Incompletely off-ended 45·4 47·1 –1·79 (–3·07 to –0·51) Age (years) 6·5 (25th percentile) –1·58 (–2·90 to –0·25) 7·9 (50th percentile) –1·62 (–2·82 to –0·42) 9·5 (75th percentile) –1·67 (–3·06 to –0·29) –3·5 –2·5 –1·5 –0·5 0·5 Favours surgical reduction Favours non-surgical casting Figure 2: PROMIS Upper Extremity Score treatment difference at 3 months by subgroups, including fracture location, displacement severity, and participant age at randomisation Surgical reduction is the reference group. PROMIS=Patient Report Outcomes Measurement System. Articles 57·3 50·0 40·0 30·0 20·0 13·7 Post- Week Month Month Month injury 6 3 6 12 Figure 3: PROMIS Upper Extremity Scale Scores at each trial timepoint in the surgical reduction and non-surgical casting groups Data are PROMIS Upper Extremity Scale Scores (95% CI). Higher scores indicate better upper extremity function. PROMIS=Patient Report Outcomes Measurement System. five (1%) participants, and a TENS nail in one (<1%) participant. 39 (10%) participants received non-surgical casting despite allocation to surgical reduction, most commonly due to family decision (n=21), then clinician decision (n=16), or due to concurrent COVID-19 rendering anaesthesia unsafe (n=2). In the non-surgical casting group, 349 (93%) of 375 underwent treatment as assigned. Most were immobilised in a full cast (198 [57%] of 349), with the remainder in backslabs. 26 (7%) participants received surgical reduction, most due to family preference (n=17), with others due to a clinician decision (n=9), which included loss of fracture position. At 3 months (the primary endpoint), the mean PROMIS Upper Extremity score was 44·9 (SD 8·7) in the non- surgical casting group and 46·6 (8·8) in the surgical reduction group (appendix 1 pp 10–11). The adjusted mean dierence was –1·64 (95% CI –2·84 to –0·44; p=0·0073), favouring surgical reduction (figure 2; appendix 1 p 12). As the confidence interval extended beyond the prespecified non-inferiority margin of –2·5 points, the primary intention-to-treat analysis did not demonstrate non- inferiority of non-surgical casting. The per-protocol analysis gave similar findings, with an adjusted mean dierence of –2·07 (95% CI –3·32 to –0·81; p=0·0013), Against the wider non-inferiority margin of –5 points, again extending beyond the –2·5-point margin (appendix 1 the lower bound of the confidence interval remained p 12). The complier-average causal eect analysis for the well above this threshold, supporting non-inferiority of primary outcome yielded results consistent in direction non-surgical casting for this subgroup. Estimates of and magnitude with the primary intention-to-treat treatment eect across the other prespecified subgroups estimand (appendix 1 p 12). were similar in direction and magnitude to the primary In the subgroup of children with completely o-ended analysis (figure 2). fractures (n=329), the adjusted mean dierence at By 6 months, observed mean scores had improved to 3 months was –1·39 (95% CI –2·86 to 0·09; p=0·066). 48·2 (SD 8·3) in the non-surgical casting group and 1544 erocS ytimertxE reppU SIMORP Intervention Non-surgical Surgical Total Surgical reduction group casting reduction (N=750) Non-surgical casting group (n=388) (n=362) Unplanned surgery 0 31 (9%) 31 (4%) Child or parent decision* 0 17 (5%) 17 (2%) Clinician decision* 0 8 (2%) 8 (1%) Fracture lost position, 0 6 (2%) 6 (<1%) better-fitting cast required, or cast adjustment† Intra-operative complications Nerve injury 0 0 0 Vascular injury 0 0 0 Wire or screw breakage 0 0 0 Other 0 0 0 Cast complications Pressure areas 0 2 (1%) 2 (<1%) Other 0 0 0 Other complications Time since randomisation Wound infection 0 6 (2%) 6 (1%) Scarring problems 0 5 (1%) 5 (1%) Refracture after removal of 3 (1%) 1 (<1%) 4 (1%) cast Nerve irritation 0 1 (<1%) 1 (<1%) *Each of these participants were allocated to non-surgical casting but received surgical reduction within 2 weeks. †Five participants were randomly assigned to surgery, received surgery within 2 weeks, and then had an additional surgery owing to loss of fracture position or cast adjustment; one participant was initially randomly assigned to non-surgical casting and received surgical reduction within 2 weeks. Table 2: Complications up to 8 weeks, by treatment received Non-surgical Surgical Total casting reduction (N=750) (n=388) (n=362) Complications Refracture 6 (2%) 3 (1%) 9 (1%) Adjacent to previous 1 (<1%) 1 (<1%) 2 (<1%) fracture location Through old fracture site 5 (1%) 2 (1%) 7 (1%) Wound infection with 0 1 (<1%) 1 (<1%) antibiotics prescription Unplanned surgery Due to refracture 5 (1%) 2 (1%) 7 (1%) Other 0 0 0 Table 3: Complications between 8 weeks and 12 months, by treatment received Articles 48·6 (8·0) in the surgical reduction group (appendix 1 (255 for non-surgical casting and 243 for surgical p 17). The adjusted dierence was –0·47 (95% CI reduction), and cost data were available for –1·67 to 0·73; p=0·44) by 6 months, consistent with no 492 participants (251 for non-surgical casting and dierence between groups. At 12 months, scores were 241 for surgical reduction). Missing data resulted 50·5 (SD 7·3) in the non-surgical casting group and from incomplete or partially completed participant 50·6 (7·3) in the surgical reduction group, with an questionnaires. There was a significant dierence in adjusted dierence of –0·41 (95% CI –1·52 to 0·70; QALYs, with a mean dierence of –0·024 (95% CI –0·042 p=0·47), confirming no dierence in longer-term to –0·005), favouring surgical reduction. Mean per recovery (figure 3). patient total and incremental costs and QALYs are Analyses adjusting for covariates associated with provided in appendix 1 (pp 14–15). Following multiple outcome and missingness did not materially alter the imputation of missing values, non-surgical casting results, and sensitivity analyses in the intention-to-treat was associated with a significant mean per-patient population exploring informative missingness under a cost saving of £1665 (95% CI 1487 to 1843) and a range of plausible assumptions were consistent with the small incremental reduction in QALYs (–0·023 primary finding (appendix 1 p 18). [95% CI –0·037 to –0·009]). Non-surgical casting was Pain, measured using the Wong–Baker FACES Pain therefore substantially less costly but marginally less Rating Scale, improved rapidly in both groups over the eective than surgical reduction. The results remained first 3 months and then plateaued (appendix 1 p 19). consistent across all sensitivity analyses (appendix 1 Using this scale, small dierences favoured surgical p 16). The incremental cost-eectiveness ratio plane, reduction at 6 weeks, 3 months, 6 months, and 12 months which depicts the bootstrap distribution, shows that (all <0·4 points) and below the 2-point dierence considered clinically meaningful. 2000 Quality of life (assessed via EQ-5D-Y) showed a dierence at 3 months and 6 months favouring surgical reduction, but no dierence at 6 weeks or 12 months (appendix 1 p 20). EQ-5D VAS scores were similar between groups throughout with a final treatment dierence by 12 months of –0·18 (95% CI –2·04 to 1·69; p=0·854) on the 100-point VAS (appendix 1 p 21). Cosmesis scores were 0 higher throughout follow-up in the surgical reduction group, with the largest dierences observed at 6 weeks and 3 months. These dierences narrowed over time, –1000 with a small mean dierence of –6·8 points (95% CI –9·3 to –4·3) on the 100-point VAS by 12 months (appendix 1 p 22). School absence was slightly worse in –2000 –0·05 0 0·05 the surgical reduction group (mean 3·6 days) than in the non-surgical casting group (mean 3·0 days), an absolute dierence of 0·6 days. Parental satisfaction with both treatments was high at 12 months, with no meaningful dierence observed (appendix 1 p 13). Results for secondary outcomes were consistent in the subgroup of completely o-ended fractures, with no systematic dierences compared with the overall population. Within 8 weeks, most complications occurred in the surgical reduction group, including pressure damage (n=2), wound infections (n=6), scarring (n=5), and nerve irritation (n=1). Refracture occurred in four participants (three after non-surgical casting and one after surgical reduction; table 2). Between 8 weeks and 12 months, complications were uncommon. Refracture occurred in nine participants (six after non-surgical casting and three after surgical reduction). Additional unplanned surgery was rare and related to new injuries in all cases (seven participants: five after non-surgical casting and Figure 4: Cost-effectiveness plane and cost-effectiveness acceptability curve two after surgical reduction; table 3). The cost-effectiveness plane displays the distribution of incremental cost-effectiveness ratios from bootstrapped simulations comparing non-surgical casting with surgical reduction. The cost-effectiveness acceptability curve All participants were included in the economic analysis. shows the probability that non-surgical casting is cost-effective across a range of willingness-to-pay thresholds, Complete EQ-5D data were available for 498 participants based on the proportion of simulations falling below each threshold. QALY=quality-adjusted life-years. )£( tsoc latnemercnI Cost-effectiveness plane with willingness-to-pay thresholds Incremental QALY 1·00 0·75 0·50 0·25 0 20 000 400 00 600 00 800 00 100 000 ytilibaborP Non-surgical casting more costly Non-surgical casting more costly and less effective and more effective Non-surgical casting less costly Non-surgical casting less costly and less effective and more effective Bootstraps £200 00 per QALY £300 00 per QALY Cost-effectiveness acceptability curve Willingness-to-pay threshold (£) Articles all cost and eect pairs lie below the £20 000 and avoiding them. Parents consistently emphasised function £30 000 willingness-to-pay thresholds (to the right of the and return to activity over radiographic or cosmetic willingness-to-pay threshold lines; figure 4). This explains perfection and were often impressed by children’s why the cost-eectiveness acceptability curve indicates resilience. These findings provide important context to that non-surgical casting has a 100% probability of being the quantitative outcomes, underscoring that families cost-eective at both the £20 000 and £30 000 willingness- were broadly reassured by recovery trajectories regardless to-pay thresholds. of initial treatment allocation. Although the findings directly apply to children aged Discussion 4–10 years with fractures severe enough to prompt The CRAFFT trial did not demonstrate statistical non- surgical reduction, consideration should be given to inferiority of non-surgical casting compared with surgical how the results can be extrapolated. Younger children reduction at the 3-month primary endpoint. However, were not included because the trial relied on validated the non-inferiority margin was deliberately conservative patient-reported outcome measures appropriate to this and set at the lower end of the minimally important age range; however, extrapolation to younger children is dierence for PROMIS scores. The observed treatment biologically plausible given their greater remodelling eect was small and below the threshold that families potential. By contrast, caution is warranted when had indicated would be required to justify surgical extending these findings to older children, where reduction. Beyond 6 months there was no dierence in remodelling potential is reduced. function between treatments, suggesting that any early Fractures that would not ordinarily prompt surgical dierence was transient and did not persist beyond the reduction were not studied (ie, those managed with initial period of recovery. The same pattern of recovery manipulation under analgesia or minimal sedation); was observed in the subgroup of children with completely however, the findings support a broader shift away from o-ended fractures, indicating consistency among those routine manipulation of distal radial fractures in the with more severe injuries. The results of the complier- emergency setting, where potential harms might average causal eect analyses were similar to those from outweigh anticipated benefits. the per-protocol analyses, with both showing a slightly Despite the trial’s strengths, several limitations should greater dierence at 3 months in favour of surgical be acknowledged. A small number of eligible children reduction. The intention-to-treat analysis was prioritised were not randomly assigned because neurovascular because it better reflects routine clinical practice where compromise led to loss of clinical equipoise requiring crossover and non-adherence are inevitable. From a urgent reduction. Because transient pain and health-system perspective, non-surgical casting oered paraesthesia are common at presentation, neurological substantial cost savings with only a very small and concern was not a formal exclusion criterion; instead, transient loss of health utility. clinician judgement was used to identify cases in which Cosmesis scores were higher in the surgical reduction randomisation was inappropriate, such as threatened group during early follow-up. However, this dierence skin integrity, dense neurological deficit, or vascular narrowed over time. Cosmetic appearance was not compromise. perfect in either group. Notably, no child was reported to Clinicians may seek clarification regarding the degree have undergone a subsequent surgery for a corrective of radiographic angulation or displacement included in osteotomy, suggesting that any residual deformity was the study; however, the trial was deliberately pragmatic, below the threshold for clinical concern. Minor residual reflecting thresholds that typically prompt intervention in deformity appears to be well tolerated, and cosmetic routine practice. The subgroup of completely o-ended dierences diminished with growth and remodelling. fractures provided a clear marker of more severe injury, Longer-term follow-up to 3 years post-randomisation is and findings in this group were consistent with the main ongoing and will provide further insight into whether analysis. Although some variability in fracture severity these cosmetic dierences continue to converge with and rehabilitation protocols across centres is inevitable, time. Embedded qualitative studies supported these the large sample size, number of recruiting centres, and observations, by exploring parents’ experiences of their stratified randomisation make it unlikely that any residual child’s recovery and participation in the trial.24,25 Parents imbalance materially influenced the results. in both groups initially expressed concern about Additional outcomes such as range of motion might cosmesis, describing the arm as looking “bent” or also be of interest; however, follow-up was deliberately abnormal, and some worried about whether the aligned with routine care, without additional mandated appearance would improve over time.24 However, by assessments, to maximise completeness and prioritise 1 year, most reported being reassured and considered validated, patient-centred functional outcomes. Radio- their child’s recovery complete, even if the appearance graphic alignment was not a prespecified outcome, was not perfect.25 Families of children in the surgical because the trial focused on function and patient-centred reduction group sometimes reflected negatively on scars, outcomes; radiographic remodelling is being investigated while those in the non-surgical casting group valued separately as an exploratory analysis. Findings from the 1546 Articles embedded qualitative study support this approach, Contributors indicating that families generally prioritised function DCP conceived the study and led the funding application. DCP led study and appearance over radiographic appearance. oversight and manuscript preparation. JA coordinated the day-to-day management of the study and contributed to study design. NN, DM, and The interventions themselves were not standardised, MDL provided clinical expertise in study design and interpretation. but the techniques used are widely adopted, and the DA developed the data management and randomisation systems. pragmatic multicentre design enhances generalisability. SF, JM, and RK led the health economic analysis. AZ and DK did the Adherence to allocated treatment was high overall, statistical analyses. PG led patient and public involvement activities. MLC contributed to study design, oversight, and interpretation. although crossover occurred slightly more often in the All authors contributed to data interpretation, critically revised the surgical reduction group, reflecting preferences at the manuscript for important intellectual content, and approved the final point of care. Notably, equipoise was acceptable to version for publication. AZ, DK, and DA had access to the study data and take responsibility for verifying the consistency of the underlying families across both groups, as reflected by recruitment data. AZ and DK take responsibility for the accuracy of the data analysis. and the distribution of treatment preferences. All authors had full access to all the data in the study and accept Finally, the trial was conducted within the UK, and responsibility for the decision to submit for publication. clinical practice, pathways, and thresholds for Declaration of interests intervention might dier globally. The economic The University of Oxford received a grant from the National Institute for evaluation reflects UK costs and willingness-to-pay Health and Care Research (NIHR) Health Technology Assessment programme (17/22/02) on behalf of the authors for the conduct of the thresholds within a state-funded health-care system. trial. DCP is an NIHR Research Professor and Chair of the NIHR Nevertheless, the clinical findings are likely to be Health Technology Assessment (HTA) Prioritisation Committee; both generalisable across settings, given the common nature DCP and MLC were previously members of the NIHR HTA General of the injury, the routine techniques studied, and the Board and now serve on the NIHR HTA Prioritisation Committee. DCP was funded through an NIHR Research Professorship pragmatic multicentre design. (NIHR301655). NN is a board member of the British Patellofemoral The adoption of these findings into practice might be Society, the British Society for Children’s Orthopaedics, and the British challenging. An embedded qualitative study of recruiting Orthopaedic Sports Trauma and Arthroscopy Association. NN has clinicians described anxiety about leaving severely undertaken paid lectures for Smith and Nephew and Arthrex. All other authors declare no competing interests. displaced fractures to remodel, diculty aligning the protocol with existing pathways, and so-called territorial Data sharing De-identified participant data will be made available on reasonable tensions between orthopaedic and emergency teams, with request to the corresponding author, subject to approval by the Trial some colleagues expected to resist change regardless of Steering Committee and appropriate data sharing agreements. evidence.26 This aligns with survey data showing that most Acknowledgments surgeons still reduce o-ended distal radius fractures in This study was funded by the National Institute for Health and Care children under anaesthesia, and few would accept non- Research (NIHR) Health Technology Assessment programme (17/22/02) surgical casting despite previous evidence of remodelling.27 and supported by the NIHR Oxford Biomedical Research Centre. We would like to thank all those involved in making the CRAFFT trial Ongoing randomised trials in North America and possible, including the children and families who took part, the research Europe28–30 suggest growing international equipoise but associates and clinical teams at all participating hospitals, and the also suggest that the publication of CRAFFT will mark multidisciplinary sta (emergency clinicians, orthopaedic surgeons, the beginning of the broader shift required to change radiographers, nurses, and physiotherapists) who contributed to the study across the UK. We are especially grateful to Kinzah Abbasi, practice. Successful implementation will require Marta Campolier, Louise Spoors, and the broader Oxford Trauma and multidisciplinary agreement on indications for surgery, Emergency Care Team at the Kadoorie Institute for Trauma, Emergency integration of non-surgical casting into local pathways, and Critical Care, University of Oxford, for their commitment to trial and careful management of clinician and families’ delivery. We acknowledge the oversight provided by the Trial Steering Committee and by the Data Monitoring and Safety Committee: expectations that a child’s wrist must always be Joy Adamson, Lyndon Mason, Simon Scott, Jamie Kirkham, straightened immediately. Richard Wilkin, Charlie Pearman, and Amy Moscrop. This trial has been The CRAFFT trial did not demonstrate non-inferiority conducted as part of a portfolio of trials in the UKCRC-registered Oxford Clinical Trials Research Unit at the University of Oxford. It has followed of non-surgical casting compared with surgical reduction their standard operating procedures ensuring compliance with the at the 3-month primary endpoint, although the observed principles of Good Clinical Practice and the Declaration of Helsinki and dierence in upper limb function was small, below any applicable regulatory requirements. The views expressed in this thresholds families considered clinically meaningful, and report are those of the authors and not necessarily those of the NIHR or the Department of Health and Social Care. The authors used ChatGPT did not persist beyond early recovery. Surgical reduction to improve the grammar and sentence structure during the preparation was associated with modestly better cosmetic appearance, of this manuscript. After using this tool, the authors reviewed and edited although this advantage diminished with longer follow-up the content as needed and take full responsibility for the content of the from injury, and was associated with higher costs and publication. early procedural complications. Non-surgical casting References delivered substantial cost savings with only a small and 1 Moon RJ, Harvey NC, Curtis EM, de Vries F, van Staa T, Cooper C. Ethnic and geographic variations in the epidemiology of childhood transient reduction in health utility. Taken together, these fractures in the United Kingdom. Bone 2016; 85: 9–14. findings support non-surgical casting as the preferred 2 Khan H, Monsell F, Duy S, Trompeter A, Bridgens A, Gelfer Y. initial treatment strategy for most children with severely Paediatric distal radius fractures: an instructional review for the FRCS examination. Eur J Orthop Surg Traumatol 2023; 33: 2169–72. displaced distal radial fractures. Articles 3 Proctor MT, Moore DJ, Paterson JM. Redisplacement after 17 Dolan P. Modeling valuations for EuroQol health states. Med Care manipulation of distal radial fractures in children. 1997; 35: 1095–108. J Bone Joint Surg Br 1993; 75: 453–54. 18 Thissen D, Liu Y, Magnus B, et al. Estimating minimally important 4 Ramoutar DN, Shivji FS, Rodrigues JN, Hunter JB. The outcomes dierence (MID) in PROMIS pediatric measures using the scale- of displaced paediatric distal radius fractures treated with judgment method. Qual Life Res 2016; 25: 13–23. percutaneous Kirschner wire fixation: a review of 248 cases. 19 Morgan EM, Mara CA, Huang B, et al. Establishing clinical Eur J Orthop Surg Traumatol 2015; 25: 471–76. meaning and defining important dierences for Patient-Reported 5 Crawford SN, Lee LS, Izuka BH. Closed treatment of overriding Outcomes Measurement Information System (PROMIS) measures distal radial fractures without reduction in children. in juvenile idiopathic arthritis using standard setting with patients, J Bone Joint Surg Am 2012; 94: 246–52. parents, and providers. Qual Life Res 2017; 26: 565–86. 6 Houshian S, Holst AK, Larsen MS, Torfing T. Remodeling of 20 National Institute for Health and Care Excellence. Guide to the Salter-Harris type II epiphyseal plate injury of the distal radius. Methods of Technology Appraisal 2013. National Institute for J Pediatr Orthop 2004; 24: 472–76. Health and Care Excellence (NICE), 2013. 7 Marson BA, Ng JWG, Craxford S, et al. Treatment of completely 21 Curtis LA. Unit costs of health and social care 2017. Personal Social displaced distal radial fractures with a straight plaster or Services Research Unit. University of Kent, 2017. manipulation under anaesthesia. Bone Joint J 2021; 103-B: 902–07. 22 NHS England. National Schedule of NHS costs 2021/2022. 2023. 8 Hernandez A, Martinez A. Displaced distal radius fractures in https://www.england.nhs.uk/publication/2021-22-national-cost- children younger than 11 years old: a randomised controlled trial collection-data-publication/ (accessed March 13, 2025). comparing two treatment methods. 2018. https://ota.org/sites/files/ 23 Husereau D, Drummond M, Petrou S, et al. Consolidated Health abstracts/2018/OTA%20AM18%20Paper%20147.pdf?utm_ Economic Evaluation Reporting Standards (CHEERS) statement. source=chatgpt.com (accessed Nov 11, 2025). BMJ 2013; 346: f1049–1049. 9 Perry DC, Wright JG, Cooke S, et al. A consensus exercise 24 Phelps EE, Tutton E, Costa ML, Achten J, Moscrop A, Perry DC. identifying priorities for research into clinical eectiveness among Protecting my injured child: a qualitative study of parents’ children’s orthopaedic surgeons in the United Kingdom. experience of caring for a child with a displaced distal radius Bone Joint J 2018; 100-B: 680–84. fracture. BMC Pediatr 2022; 22: 270. 10 Perry DC, Achten J, Mason J, et al. The protocol for a multicentre 25 Phelps EE, Tutton E, Costa ML, et al. Being recovered: a qualitative prospective randomized noninferiority trial of surgical reduction study of parents’ experience of their child’s recovery up to a year versus non-surgical casting for displaced distal radius fractures in after a displaced distal radius fracture. Bone Jt Open 2024; 5: 426–34. children: Children’s Radius Acute Fracture Fixation Trial (CRAFFT) 26 Phelps EE, Tutton E, Costa ML, Achten J, Gibson P, Perry DC. protocol. Bone Jt Open 2025; 6: 560–68. A qualitative study of clinicians’ experience of a clinical trial for 11 ISRCTN. When children up to 11 years old break the bones in their displaced distal radius fractures. Bone Jt Open 2024; 5: 324–34. wrists, do they need surgery to perfectly realign the bones, or will 27 Laaksonen T, Puhakka J, Kosola J, Stenroos A, Ahonen M, nature ‘self-correct’ the bones as they heal without restricting the Nietosvaara Y. Most surgeons still prefer to reduce overriding distal use of the arm? 2025. https://www.isrctn.com/ISRCTN10931294 radius fractures in children. Acta Orthop 2021; 92: 235–39. (accessed Nov 4, 2025). 28 Balmert Bonner L, Janicki J, Georgiadis A, et al. Distal Radius 12 Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Interventions for Fracture Treatment (DRIFT) trial: study protocol Research electronic data capture (REDCap)—a metadata-driven for a multicentre randomised clinical trial of completely translated methodology and workflow process for providing translational distal radius fractures at paediatric hospitals in North America. research informatics support. J Biomed Inform 2009; 42: 377–81. BMJ Open 2025; 15: e088273. 13 Cella D, Yount S, Rothrock N, et al. The Patient-Reported Outcomes 29 Abildgaard KR, Buxbom P, Rahbek O, et al. Is casting of displaced Measurement Information System (PROMIS): progress of an NIH paediatric distal forearm fractures non-inferior to reduction under Roadmap cooperative group during its first two years. Med Care general anaesthesia? Study protocol for a pragmatic, randomized, 2007; 45 (suppl 1): S3–11. controlled non-inferiority multicentre trial (the casting trial). Trials 14 Varni JW, Thissen D, Stucky BD, et al. PROMIS Parent Proxy 2024; 25: 420. Report Scales for children ages 5–7 years: an item response theory 30 Laaksonen T, Stenroos A, Puhakka J, et al. Casting in finger trap analysis of dierential item functioning across age groups. traction without reduction versus closed reduction and Qual Life Res 2014; 23: 349–61. percutaneous pin fixation of dorsally displaced, over-riding distal 15 Wong DL, Baker CM. Pain in children: comparison of assessment metaphyseal radius fractures in children under 11 years old: a study scales. Pediatr Nurs 1988; 14: 9–17. protocol of a randomised controlled trial. BMJ Open 2021; 16 Wille N, Badia X, Bonsel G, et al. Development of the EQ-5D-Y: 11: e045689. a child-friendly version of the EQ-5D. Qual Life Res 2010; 19: 875–86. 1548 --- [PDF原文](https://sci-net.xyz/storage/7932541/671820289278eb3633a25635a0555b9458fb905181ab88fac0c16f87a78877ca/Non-surgical-casting-versus-surgical-reduction-for-children-with-severely-displaced-distal.pdf) DOI: 10.1016/S0140-6736(26)00409-5