Study setting {9}
This pRCT will be conducted at three academic trauma units at three different urban centres in Ireland. Each centre serves a referral population of >500,000 and receives all grades of trauma from both urban and rural environments on a 24-hour basis. A trauma team is on call daily and includes two trainee surgeons and a consultant orthopaedic surgeon. Surgeries are performed in part, in total or supervised by consultant orthopaedic surgeons. Regional and general anaesthesia is used at the discretion of the anaesthetist. Ankle fracture surgery is conducted on both a day case or overnight basis. Ward-based physiotherapy is provided daily to facilitate early discharges. All hospital personnel contributing to the recruitment and patient pathways in this trial will undergo training in the objectives and methodology of the study.
Eligibility criteria {10}
Inclusion Criteria
All skeletally mature (closed distal tibial physis), acute ankle fractures treated with anatomical reduction and stable internal fixation, including AO/OTA 44A1.3 to 44A3.3, 44B and 44C.
- Isolated lateral malleolar fractures
- Isolated medial malleolar fractures
- Bimalleolar fractures
- Trimalleolar fractures
- Syndesmosis injuries that have been surgically fixed with either screw or tightrope.
- Closed, grade I, or grade II open fractures.
Exclusion Criteria
- Skeletal immaturity (open distal tibial physis)
- Gustilo grade-III open fractures
- Tibial plafond fractures
- Polytraumatised patients
- Non-ambulatory status before injury
- Expected insufficient stable fracture fixation with standard surgical technique
- Pre-existent cognitive disability, neurological disease or inability to comply with non-weight-bearing mobilisation
- Grossly comminuted fragility fractures
Patient recruitment and consent: {26a}{15} (Who will take informed consent? And recruitment)
All patients admitted to the hospital with ankle fracture (AO/OTA 44A1.3 to 44A3.3, 44B and 44C) deemed appropriate for surgical intervention will be asked by the admitting trainee or consultant surgeon to participate in the trial and provided with a patient information leaflet (Appendix 1), given time to read the document and ask questions. If the patient agrees to enter the trial, they will sign the consent form in the presence of the admitting doctor on the morning of or the night before their surgery.
Interventions
Explanation for the choice of comparators {6b}
Traditional non-weight bearing (NWB) cast immobilisation is a common practice in many centres, and this protective protocol might not be necessary.
Intervention description {11a}
In accordance with a pragmatic study, the surgical approach and choice of the implant will be at the surgeon's discretion. Surgeons may or may not be authors in the study. Surgical practice at the three institutions is to achieve anatomical reduction and rigid fixation. The commonly used osteosynthesis system for fixation is the small fragment system, with a one-third tubular plate commonly used. The use of locking mode is not routinely used. Other systems are also available. All patients will be assessed by a physiotherapist for gait stability and provided with walking aids according to randomisation. Patients in the walking boot group (Group A) will be instructed to weight-bear as tolerated immediately with or without walking aids for balance. Patients in the NWB group (Group B) will be instructed to strictly bear non-weight using crutches or frames for a total of six weeks. Group A will be instructed to remove the walking boot four times a day at minimum to perform ankle range of motion exercises until they attend outpatient physiotherapy following their first postoperative visit. All patients will receive a postoperative care information sheet according to their grouping (Appendix 2 and 3).
Patients will be allocated randomly to one of two groups;
Group A:
- Will receive a walking boot orthosis postoperatively in theatre and allowed weight bearing as tolerated and range of motion (ROM) exercises immediately.
- Elevation of the affected foot in the first two weeks will be encouraged to reduce swelling.
- The first follow-up appointment will be after two weeks. This visit is for surgical site inspection, removal of sutures, check X-ray and referred to physiotherapy to continue ROM exercises and weight-bearing as tolerated progressing to full weight-bearing.
Group B:
- Will receive full below-knee cast postoperatively in theatre and prevented weight-bearing for six weeks.
- Elevation of the foot in the first two weeks will be encouraged to reduce swelling.
- The first follow-up appointment was after two weeks. This visit is for surgical site inspection, removal of sutures, check X-ray and re-application of a full below-knee cast.
- The second follow-up is after six weeks, for removal of cast and referral to physiotherapy to commence ankle ROM exercises and weight-bearing as tolerated progressing to full weight-bearing.
Criteria for discontinuing or modifying allocated interventions {11b}
The trial will be terminated early if a 20% complication rate (wound complications and fixation failure) is detected in either of the treatment groups (10,14).
Strategies to improve adherence to interventions {11c}
Surgeon:
Surgical treatment for all participants will remain unchanged in both groups, and all surgeons have agreed to participate in the trial.
Participant:
In the postoperative setting on the ward and before discharge, a physiotherapist will reinforce the patient's role in the trial and provide them with information leaflet appropriate to their grouping. At subsequent outpatient follow-up visits, patients will be reminded of the trial. The trial Case Report Form (Appendix 4) will record if the patients have received outpatient physiotherapy.
Relevant concomitant care permitted or prohibited during the trial {11d}
The choice of and duration of DVT prophylaxis will be at the surgeon's discretion.
Provisions for post-trial care {30}
None.
Outcomes {12}
The primary outcome measure is the functional Olerud-Molander Ankle Score (OMAS) at six weeks postoperatively. This score ranges from 0 to 100, with 100 representing normal ankle function (15). Secondary outcome measures include complication rate (infection and fixation failure), the full arc of ankle motion (plantar flexion and dorsal-flection) measured in degrees using a goniometer, RAND 36-Item Short Form Survey (SF-36) scoring, the time needed to return to work in days and postoperative hospitalisation length in days.
Follow up:
Patients will be followed up in an outpatient setting at two weeks, six weeks, twelve weeks, six months and one year postoperatively. At each visit, the OMAS and RAND-36 health questionnaire will be collected. Surgeons who may or may not be authors in the study reviewing patients in either group will also, at each follow-up appointment, complete a case report form that was created by the trial team (Appendix 4). This record included surgical site assessment, any complication, X-ray evaluation, ankle range of motion (using goniometry), information regarding return to work, confirmation of physiotherapy referral and confirmation of collection of OMAS and RAND-36 questionnaire.
Participant timeline {13}
Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) Figure of enrolment, interventions and assessments.
Key: IWB: immediate weight-bearing, NWB: non-weight-bearing, OMAS: Olerud Molander Ankle Score, RAND-36: 36-item Short-Form Health Survey.
Sample size {14}
An a priori power analysis for the superiority of treatment with immediate weight-bearing and ROM will be conducted for this hypothesis. To detect a clinically significant 10-point difference in the Olerud and Molander Ankle Score (OMAS) at six weeks, with a standard deviation of 19(16–19). Alpha = 0.05 and β=0.20 (80% power), two-sided test and a maximum loss of follow up of 20% (N=145), this was rounded to 160 to support the block randomisation (20 block size), and a sample size of 80 per group is necessary.
Assignment of interventions:
Allocation
Sequence generation {16a}
An online computer-generated block randomisation list (20 patients per block) will be created at the start of the trial via the website http://www.randomization.com. This list has a unique number that is stored safely and can be double-checked.
Concealment mechanism {16b} Implementation of randomization {16c}
Upon skin closure, a circulating theatre nurse, who is not part of the trial team, will consult the randomised block database, which is kept secure and password protected. The nurse will inform the surgical team that a walking boot (Group A) or a cast (Group B) is to be applied. The patient's details will be entered into the database, and they will be assigned a trial number. The surgeons will be blinded to the intervention until fracture fixation is complete.
Assignment of interventions:
Blinding
Who will be blinded and procedure for unblinding if needed {17a/b}
We understand there is some unavoidable risk of bias to this particular type of RCT where the intervention is impossible to blind as both cast and boot are visible. This has been a known weakness in all previous similar RCTs. Furthermore, in line with the pragmatic trial design, the patient's care must follow the routine hospital follow-up procedure. It was not feasible to introduce independent assessors to mitigate such bias. On the other hand, patient-reported outcome measures (OMAS and RAND-36) are completed solely by patients, and the assessor has no role in these data. Additionally, to reduce the risk of bias, the surgeon will be blinded until the surgical procedure is complete.
Data collection and management
Plans for assessment and collection of outcomes {18a}
At each follow-up visit, the OMAS and SF-36 Health questionnaire will be collected from the participants by outpatient clinic nurses. The attending orthopaedic consultant or NCHD fills up a case report form, which includes documentation of the following information: surgical site assessment and complication, X-ray evaluation, ankle full arc measure (goniometry), information regarding return to work, confirmation of physiotherapy referral and confirmation of collection of OMAS and RAND-36 Health survey.
Plans to promote participant retention and complete follow-up {18b}
We have developed a patient tracking system to allow researchers to monitor follow-up carefully. As part of this system, a weekly list of expected patients is provided to the research nurse in the OPD, and this list is reviewed daily to record attendance. In case of a patient being absent from the clinic, another appointment for the following week will be arranged, the RAND-36 and the OMAS score will be posted to the patients with prepaid envelope enclosed, and the patient will be contacted to encourage follow-up.
Data management {19}
Three forms are collected and checked at each follow-up visit and then stored securely in the trial locker: two patient-reported outcomes measure (PROM), the OMAS and RAND-36, and the case report form. Every two weeks, all data are transferred to a temporal database located within the hospital computer system by two researchers, and a read-only copy is stored in a separate folder. This is then cross-checked before data are transferred to the statistical software for statistical analysis and reporting by the statistician and the research team.
The RAND-36 requires multiple steps analysis; this will be performed with oblique scoring and the orthogonal-factor analytic model (20). Normative data for the Irish population will be used as a reference (21).
Confidentiality {27}
Data management will be in accordance with the General Data Protection Regulation
(GDPR) Health Service Executive (HSE) and Health Research regulations (22,23). Data will be kept anonymously in the database within the HSE local hospital computer system in protected folders to ensure confidentiality. Paperwork will be stored in the trial locker in a locked researcher office within the hospitals.
Plans for collection, laboratory evaluation and storage of biological specimens for genetic or molecular analysis in this trial/future use {33}
Not applicable as there is no laboratory evaluation or biological specimen collection in this study.
Statistical methods
Statistical methods for primary and secondary outcomes {20a and b}
Continuous outcomes and other key variables will be described by their means and SDs, medians and IQRs, and total range. Categorical variables will be described by counts and respective proportions.
The primary outcome, OMAS score at six weeks, will be analysed using multiple linear regression with fixed effects for study arm and centre, from which we will report the centre-adjusted difference in mean OMAS scores between the two arms with its 95% confidence interval (CI) and exact p-value. The complete set of longitudinal OMAS scores, from 2 weeks to 1 year, will be analysed with the corresponding linear mixed-effects model (with the identity link function), with fixed effects for the centre, arm, and time (dummy coded) and an interaction term between arm and time. Time-specific differences in mean OMAS scores between arms will be calculated from the model's results, along with 95% CIs, and the exact p-value from the time by arm interaction.
Longitudinally measured secondary outcomes will be similarly analysed and reported: ankle range of motion with the same linear mixed-effects model; and RAND-36 scores with an ordinal mixed-effects model with a logit link function (i.e. the "proportional odds model). Time to return to work (in the subset working at baseline) and length of stay will be analysed using Cox proportional hazards models, with fixed effects for centre and treatment arm. Infection (any) and fixation failure at any point during study follow-up will be similarly analysed using logistic regression.
There will be no formal adjustments for multiplicity, but we will report all estimated treatment effects alongside exact p-values, allowing the reader to make whatever adjustments they prefer. Any deviations from the above plan will be fully described and justified in the final report of the trial results. All analyses, as well as the production of tables and plots, will be conducted using R (version 3.6.3, R Project for Statistical Computing) and Stata16 software. All analyses will be conducted or supervised by the Principal Statistician of the HRB Clinical Research Facility Cork under their quality system and relevant standard operating procedures and following regulatory guidance (e.g., ICH E9).
Because outcomes at week six are collected as part of routine care, we do not anticipate and have intentionally aimed to avoid any missing values at that time point. However, if there are missing values, then we will carefully consider the reasons why data are missing and employ appropriate methods, which could range from complete case analysis in the presence of very little missing data that is deemed MCAR, to multiple imputations under assumptions of MAR. For missing longitudinal data, we will take advantage of full-information maximum likelihood estimation for the mixed-effects models to retain all patients in the models. Irrespective of the actual approach taken (which cannot be optimally decided on without consideration of the actual study data), we will explore choices and their potential impact on inferences with sensitivity analyses.
Interim analyses {21b}
No planned interim analysis.
Methods in analysis to handle protocol non-adherence and any statistical methods to handle missing data {20c}
Any protocol non-adherence will be disclosed and handled accordingly. An effort will be made to prevent missing data as much as possible. Unavoidable missing data, such as withdrawals from the study or loss to follow-up data, will be analysed on an intention-to-treat basis, including sensitivity analysis. Multiple imputations using Chain Equations MICE will be used for missing data. (24–26).
Plans to give access to the full protocol, participant level-data and statistical code {31c}
The full protocol is available at the registry website and will be published in one of the trial protocol journals.
Oversight and monitoring
Composition of the coordinating centre and trial steering committee {5d}
Four authors at the coordinating centre (University Hospital Waterford) take responsibility for the scientific validity of the study protocol, assessment of study quality and conduct, as well as for the scientific quality of the final study report.
Composition of the data monitoring committee, its role and reporting structure {21a}
The authors understand the composition for a standard data monitoring committee (DMC) for this trial is challenging and might be impossible. Lack of funding is the main barrier. Unlike a clinical trial where a DMC and interim analysis must be formulated, this non-interventional trial, with relatively small numbers, does not investigate a pharmacological or medicinal product and does not expose patients to significant harm. Both treatment methods are already part of routine practice. As such, the trial team will closely monitor any adverse event or harm that might arise and act accordingly on a daily basis (see below).
Adverse event reporting and harms {22}
Collected case report forms will be checked daily by the research team before being stored in the trial locker; any adverse events or harm will be communicated with the study team.
Frequency and plans for auditing trial conduct {23}
The trial conduct is continuously audited in the departmental audit meeting (three monthly).
Plans for communicating important protocol amendments to relevant parties (e.g., trial participants, ethical committees) {25}
Any change to the trial protocol will be communicated with the ethical committees and trial registry.
Dissemination plans {31a}
The results of this trial will be published in one of the peer-reviewed medical journals.