Protocol for Surgery or Cast of the EpicoNdyle in Children’s Elbows (SCIENCE)

Sample size

The primary outcome is the PROMIS Upper Limb Score for Children. Raw scores are translated into standardized T-scores with a population mean of 50 and a SD of 10. The minimal clinically important difference (MCID) for the PROMIS Upper Limb Score among children with milder forms of disability has been demonstrated to be three to four.²⁴ In general, the bank of paediatric PROMIS measures have a MCID of three points, in a range of different diseases including sickle cell disease, asthma, nephrotic syndrome, and cancer.²⁵ During a patient and public involvement event, it was established that while a score of three to four points appeared to be the minimal difference noticeable to parents, the clinically important difference required to justify surgery was five points or more. Parents and children demanded a larger effect size to justify the intervention of surgery. Other studies have similarly highlighted that patients often seek greater effect sizes to warrant surgical interventions than the established MCID.²⁶ We seek to find a difference of four points between the interventions.

The SD of 10 derived by PROMIS was ascertained based on a sample of children with a higher proportion of chronic illness than the general population. It is anticipated that the variation in outcomes in the treatment of acute medial epicondyle fractures is likely to be less than in a chronic illness. Therefore, blinded sample size re-estimation, based on the SD of the outcome tool when patient recovery is beginning to plateau, is planned. We will perform the sample size re-estimation calculation after the first the 50 patients have completed six months’ follow-up (estimated to be month nine of the main trial). If, as expected, the SD of the sample is notably less than the chronic disease population, we will revisit the study timelines to determine the optimal study duration, thereby enhancing the efficiency of the trial. In the unlikely event that the SD is greater than expected, we will discuss the findings with the trial steering committee to formulate a strategy to meet the increased recruitment target required.

In summary, this study will use the PROMIS Upper Limb Score for Children at one year after randomization as the primary outcome measure. The total number of patients required to obtain a power of 90% to detect a four-point difference between groups for the primary outcome measure will be 266, i.e. 133 patients will be required in each treatment group. With an allowance for a conservative 20% loss to follow-up, we plan to recruit 334 patients in total. To maximize trial efficiency, we will re-estimate the sample size based on the SD of the outcome tool at six months’ follow-up of the first 50 children in the trial.

Consenting

Recruitment will take place in a minimum of 30 NHS trusts who treat children with this injury in the UK, plus additional centres in New Zealand and Australia. Potentially eligible patients will be identified by the clinical team. The research associate will present the patient and parents/guardian with age-appropriate participant information sheets or online study information and a verbal explanation of the trial procedures. The patient/parent/guardian will then be given the opportunity to discuss any issues related to the trial with the research associate and members of their family and friends.

The parent/guardian will then be asked to sign an electronic informed consent form. Children will be invited to sign an electronic assent form (UK and New Zealand). In Australia, mature children (i.e. ≥ aged 13 years, or as decided by the research team) will be invited to sign an electronic consent form. Assent should be taken, where appropriate, however the absence of assent does not exclude the patient from the study if consent has been obtained from the parent/guardian. If a child indicates indicates they do not want to take part, they will not be included in the study.

In the UK, contact details will be retained until the long-term follow-up is complete (when the child reaches skeletal maturity at 16 years of age). Consent/assent forms will be kept until the youngest participant reaches 21 years of age.

In Australia and New Zealand, contact details will be retained for a minimum of one year after the 12-month follow-up period is complete. Completed parent consent and child consent and/or assent forms will be kept until the youngest participant reaches aged 25 years (Australia) or aged 26 years (New Zealand).

Randomization

Those patients who consent to take part in the trial will have their treatment allocated using a secure, centralized, online randomization service. All hospital treatment areas have access to the internet, so will access the randomization service in real time, i.e. there will be no delay in patient treatment.

Consented participants will be randomized to one of two intervention groups (1:1) using a computer randomization service provided by the Oxford Clinical Trials Research Unit (OCTRU). Randomization will be performed using a minimization algorithm, including a random element to ensure balanced allocation of participants across the two treatment groups stratified by centre and dislocation status of the elbow at presentation (i.e. dislocated or not dislocated). The first 30 participants will be randomized using a simple randomization schedule produced by the trial statistician, to seed the minimization algorithm, and a non-deterministic probabilistic element will be introduced to prevent predictability of the treatment allocation.

Stratification by centre within the minimization algorithm will help to ensure that any clustering effect related to the centre will be equally distributed in the trial arms. The catchment area (the local population served by the hospital) will be similar for all of the hospitals, each hospital being a children’s injury unit dealing with these fractures on a daily basis. All of the recruiting hospitals use these techniques as part of their normal practice, i.e. staff will already be equally familiar with both forms of treatment. This cannot eliminate the clinician-specific effect of an individual at any one centre. However, since the procedures are commonplace, many clinicians will be involved in the management of this group of patients – likely between five and 20 clinicians at each centre, including consultants and trainee surgeons. Therefore, we anticipate that each individual clinician will only treat a handful of those enrolled in the trial, reducing the risk of a clinician-specific effect upon the outcome in any one centre.

Stratification by dislocation status of the elbow (i.e. not dislocated at presentation to the emergency department, or dislocated at presentation to emergency department (with a subsequent satisfactory reduction)) within the minimization algorithm will help to ensure that the perceived severity of the injuries through additional soft-tissue damage are balanced across the treatment groups, to take account of the potential differences in the outcome measures. Any participants with an elbow dislocation that is unable to be reduced prior to study enrolment will be excluded from the trial.

Post-randomization withdrawals

Children (or their parents/guardians) may decline to continue to take part in the trial at any time without prejudice. A decision to decline consent or withdraw will not affect the standard of care the patient receives. Children (or their parents/guardians) can withdraw by contacting the research team, with contact details on patient information materials and the trial website. Upon withdrawal of the patient, any data collected up until the time of withdrawal will be retained by the research team and included in the final analysis. Contact details for these patients will be destroyed. Withdrawn patients or patients deemed ineligible after consent will not be replaced.

Blinding

Patients and their parents/guardians cannot be blinded to their treatment. The treating clinician will, of course, not be blinded to the treatment they are providing. However, the treating clinical team will take no part in the follow-up assessment of the patients. The outcome data will be collected directly from the patient and/or their parents/guardians. Outcome assessors will be blinded to the participant’s treatment allocation.

Trial treatments

This trial will compare the two common approaches to treat a displaced medial epicondyle of the humerus in children, as follows.

In operative fixation, children are admitted to hospital for surgery, which is typically scheduled on a daytime trauma operating session, though patients can be enrolled irrespective of the time of presentation/surgery. Children undergo a general anaesthetic. After the skin has been covered in antiseptic, an incision will be made over the medial epicondyle paying particular attention to the location of the ulna nerve. The bone fragments will be opposed in the optimal position achievable under direct vision. A record will be made of the type of fixation used. The bone fragments will be fixed using the preferred technique of the surgeon (i.e. screw/wire(s)). Although the basic principles of fixation are inherent in the technique, there are several different options available to the surgeon, with the most common being screw fixation. The type of implant, size, and insertion technique are not believed to affect the outcome, and will be left entirely to the discretion of the surgeon as per their normal practice. At the end of the procedure, a sling, plaster, splint, or bandage will be applied as per the standard surgical practice. The elbow will be allowed to mobilize as per the usual practice of the treating surgeon under the direction of the clinical team, though fixed immobilization in a cast should not be used for more than four weeks post-randomization.

Nonoperative treatment involves immobilization of the elbow to rest it at around 90° of flexion. The immobilization device (i.e. cast, splint, bandage) is not applied with the intention of directly opposing the bone fragments, and therefore the bone fragments will not align perfectly. In this pragmatic trial, the duration and method of immobilization will be left to the discretion of the treating surgeon as per their usual technique, and will be worn as per the standard practice of the treating surgeon. Subsequently, the elbow will be allowed to mobilize as pain allows, under the direction of the clinical team. Fixed immobilization in a cast should not be used for more than four weeks post-randomization.

In this pragmatic trial, any rehabilitation input, including a formal referral to physiotherapy, will be left to the discretion of the treating clinicians. However, a record of any rehabilitation input (type of input and number of additional appointments) together with a record of any other investigations or interventions will be requested as part of the four-week, six-week, three-month, six-month, and 12-month follow-up datasets from both patients and clinical teams.

Adverse event management

Serious adverse events (SAEs) will be entered onto the SAE reporting form and reported to the central study team. Once notified, causality and expectedness will be confirmed by the chief investigator or trial-nominated clinician. Some adverse events that are foreseeable as part of the proposed treatment will not be reported on a SAE reporting form; they will be recorded on a complications reporting form. These events include: a) general complications – pain, pressure areas or elbow stiffness, symptomatic instability, or nonunion of the bone fragments; and b) complications specifically related to surgery – wound infection, injury/irritation to the ulna nerve, implant irritation, screw cut-out, broken or retained metalwork, and the subsequent need to remove metal pins/ screws.

All participants experiencing SAEs will be followed up as per protocol until the end of the trial. All unexpected SAEs or suspected unexpected serious adverse reactions (SUSARs) that occur between the date of consent and six weeks’ follow-up point will be reported to the sponsor and ethics committee (as per the specifications set out by individual countries).

The end of the trial will be defined as the collection or receipt of the last follow-up questionnaire from the last participant.

Statistical analysis

A separate statistical analysis plan (SAP) with full details of all statistical analyses planned for the data has been produced and will be made publicly available. The SAP has been reviewed and received input from the relevant institition’s Trial Steering Committee (TSC) and Data Safety Monitoring Committee (DSMC).

Any changes or deviations from the original SAP will be described and justified in any updated versions of the SAP, protocol, final report, and/or publications, as appropriate. It is anticipated that all statistical analyses will be undertaken using Stata (StataCorp, USA) or other well-validated statistical packages.

All analyses will be carried out on the intention-to-treat population (that is, all patients will be analyzed in the group they were randomized to, regardless of actual treatment received). To test robustness of the results, sensitivity analyses that supplement the primary analysis will include repeating the primary analysis for the per-protocol population (patients excluded from the per-protocol population will be pre-specified in the SAP) and the as-treated population, bearing in mind that this may introduce bias by losing the benefits of randomization.

The PROMIS Upper Limb Score for Children at 12 months is the primary outcome of the study, and the primary analysis will compare this between the treatment groups in a linear mixed effects method including all patients, at all timepoints, and adjusting for the stratification factors. A simple analysis of covariance (ANCOVA) of the primary outcome at 12 months adjusting only for the baseline PROMIS score will be undertaken as a secondary analysis. If the outcome is not normally distributed, non-parametric techniques will be used with no adjustment (for example the Mann-Whitney U test or the Kruskal-Wallis test).

Although we have allowed for up to 20% missing data in the sample size, we hope to minimize this by using data collection techniques appropriate to the age of participating children. Before carrying out the within-trial analysis, we will check the trial data for any missing data. Where possible, the reasons for missing data will be ascertained and reported. The nature and pattern of the ‘missingness’ will be carefully considered — including in particular whether data can be treated as missing at random (MAR). If judged appropriate, missing data will be imputed using multiple imputation. The resulting imputed datasets will be analyzed and reported, together with appropriate sensitivity analyses.

Standard descriptive statistics will be used to describe the demographics between the treatment groups reporting means and SDs, or medians and interquartile ranges (IQRs), as appropriate for continuous variables, and numbers and percentages for binary and categorical variables. All comparative outcomes will be presented as summary statistics and reported together with 95% confidence intervals (CIs), and all tests will be carried out at a 5% two-sided significance level.

Health economic evaluation

An economic evaluation will be conducted as part of the trial to estimate cost-effectiveness. A health economics analysis plan (HEAP), providing full details of the prospective economic analysis, has been produced and will be made publicly available.

Since the trial is recruiting internationally, the analysis will be limited to UK recruiting sites and follow intention-to-treat principles. The main analysis will adopt an NHS and personal social services perspective. Participants are aged seven to 15 years, thus questions will be primarily completed or assisted by parents/carers.

Healthcare resource use will be costed using the most recently available published national reference costs, reflated to a common year. Index hospital procedures and any sequelae procedures will be costed using OPCS-4 codes and applying reference costs via the NHS Healthcare Resource Group (HRG) grouper.^27,28 Participants’ health service contacts, made in connection with their elbow, will be recorded at three, six, and 12 months and costed using national unit costs.^28,29 Personal expenses, parent/carer time from work, and time from school will also be recorded as part of a broader societal perspective.

Generic health-related quality-of-life will be assessed at baseline, six weeks, and three, six, and 12 months using the EQ-5D-Y questionnaire. EQ-5D-Y scores will be converted to health status scores using the most appropriate tariff available at the time of analysis.^30,31 Using the trapezoidal rule, the area under the curve of health status scores will be calculated, providing patient-level quality-adjusted life year (QALY) estimates.

If overall data missingness exceeds 5%, the primary analysis will include multiple imputation using the MI framework in Stata. Mechanisms of missingness of data will be explored, and multiple imputation methods will be applied to impute missing data, following best practice.^32-34 Imputation sets will be used in bivariate analysis of costs and QALYs to generate incremental cost per QALY estimates and CIs. If the level of missingness is low, then a complete-case bivariate analysis will be conducted without imputation.

Findings will be analyzed and visualized in the cost-effectiveness plane, as cost-effectiveness acceptability curves, net monetary benefit, and value of information analysis. A within-trial analysis will use the first 12 months of data, to correspond to the primary analysis. If incremental costs and benefits are non-convergent within the trial follow-up, then extrapolated modelling will be considered.

Trial oversight

The trial will be conducted in accordance with the Medical Research Council’s Good Clinical Practice (MRC GCP) principles and guidelines, the Declaration of Helsinki,³⁵ OCTRU standard operating procedures, relevant UK/Australian/New Zealand legislation, and this protocol. GCP-trained personnel will conduct the trial.

The day-to-day management of the trial will be the responsibility of the trial manager, supported by the OCTRU administrative staff. This will be overseen by the trial management group, who will meet monthly to assess progress. It will also be the responsibility of the trial manager to undertake training of the research staff at each of the trial centres. The trial statistician, health economist, and information specialist will be closely involved in setting up data capture systems, design of databases, and clinical reporting forms. A TSC and DSMC will be set up.

The DSMC will adopt a DAMOCLES charter, which defines its terms of reference and operation in relation to oversight of the trial. They will not be asked to perform any formal interim analyses of effectiveness. They will, however, see copies of data accrued to date, or summaries of that data by treatment group, and they will assess the screening algorithm against the eligibility criteria. They will also consider emerging evidence from other related trials or research, and review related SAEs that have been reported.

Quality control

This study will be coordinated by the UKCRC registered trials unit (OCTRU at the University of Oxford). We will institute a rigorous programme of quality control. The trial management group will be responsible for ensuring adherence to the trial protocols at the trial sites. Quality assurance checks will be undertaken by the central trial team to ensure the integrity of randomization, study entry procedures, and data collection. The clinical trials unit (CTU) has a quality assurance manager, who will monitor this trial by conducting inspections (at least once in the lifetime of the study, more if deemed necessary) of the trial master file. Furthermore, the processes of consent-taking, randomization, registration, provision of information, and provision of treatment will be monitored by the central trial team. Written reports will be produced for the TSC, informing them if any corrective action is required.

Additionally, the study may be monitored, or audited by sponsor or host sites in accordance with the current approved protocol, GCP, relevant regulation,s and standard operating procedures.

Patient and public involvement

Patients and children were involved from the inception of the trial, including in the development of the funding application. During the funding stage, patient representatives were involved in deciding what outcomes to collect, when to collect them, and the change required to determine superiority. The study was also discussed in detail with members of the NIHR YPAG who drafted a logo, and chose the final logo for the study through online design competition. An ongoing commitment has been made to continue to work with this group in the production of patient-facing materials and the study dissemination plan.

To ensure ongoing patient and public involvement, a patient/carer representative will be actively involved in the trial management. In addition, a further independent patient/carer representative will become a member of the steering committee.

Ethics and dissemination

In the UK, a National Research Ethic Committee approved this study on 25 March 2019 with reference number 19/NW/0158. In Australia, ethics was approved on 29 April 2021 (Reference 68948). In New Zealand, the study was approved on 30 September 2021 (Reference 21/NTB/161).

A manuscript for a high-impact peer-reviewed general medical journal will be prepared simultaneously with reports to the funder. The dissemination strategy will target the orthopaedic and emergency medicine community, the wider medical community, the National Institute for Health and Care Excellence, and policymakers. Authorship will be determined in accordance with the International Committee of Medical Journal Editors guidelines, and other contributors will be acknowledged. The results of this trial will substantially inform clinical practice on the clinical and cost-effectiveness of the treatment of this injury. The results of this project will be disseminated to patients via a targeted website, which includes patient explainer videos, patient information leaflets, and a clinical protocol for clinicians.