A single-center, open-label, randomized, parallel, non-inferiority clinical trial was conducted at a general hospital. The study received ethical committee approval from our institution and was registered on clinicaltrials.org with identifier: NCT06189625.

The study included patients diagnosed with mild ankle sprain treated in the pediatric emergency service (PES) of our center. A mild or grade I sprain was defined as presenting with mild edema, involvement of a single ligament (detected by palpation), and no joint instability (assessed using the anterior drawer and talar tilt tests) [7]. Inclusion criteria were: age between 5 and 16 years, diagnosis of grade I sprain, and informed consent signed by parents. Exclusion criteria included: chronic foot and/or ankle injury, pre-existing conditions with an increased risk of fractures, developmental disorders, presence of a penetrating injury or other associated musculoskeletal injuries, such as fractures, apophyseal lesions, etc., injury occurring more than 72 h prior, inability for telephone contact, or language barrier.

The randomization sequence was created using Epidat 4.2 software, ensuring allocation concealment, with a constant 1:1 allocation ratio.

Upon evaluation by a pediatrician and diagnosis of mild ankle sprain, patients and their guardians were invited to participate in the study and provided with all necessary information as recommended by the Helsinki Declaration. Upon agreement to participate, informed consent was obtained. Then, the treatment group assignment was revealed to both the attending physician and the patient. Blinding of the treatment at the time of care and during clinical reviews or data analysis was not possible due to the nature of the intervention.

During history taking and physical examination, age, sex, relevant personal history, and previous ankle sprains on the same side were documented. Injury mechanism, time until arrival at the PES, prior analgesia administration, and pain level using a numerical scale (0–10) were also recorded. Physical examination documented pain location, inability to take more than two steps at the consultation, and presence of edema, bruising, or joint instability.

Adequate analgesia was ensured during the emergency consultation.

Once the informed consent was signed, each patient was randomly assigned to either the restraint group, where a functional bandage was applied for 5 days, or the non-restraint group, where only common recommendations were implemented.

Due to financial constraints and the need to use a standardized type of restraint during emergency care to minimize heterogeneous results in the restraint group, the following functional bandage was chosen (Fig. 1). It consisted of a standardized bandage using 6 cm silk tape and 7.5 cm Tensoplast. Emergency nurses underwent dual training, consisting of an instructional video and a half-hour practical session. Reminder posters were placed in the procedure room where such interventions are typically performed.

Functional bandage technique

All patients, from both restraint and non-restraint groups, were provided with common recommendations upon discharge. They were advised to take ibuprofen for 2–3 days and subsequently if in pain, as well as to apply local cold for a maximum of 10 min several times a day. Relative rest was recommended, encouraging weight-bearing as tolerated, and using crutches if necessary. All patients were advised to refrain from sports for at least a week or more if needed and resume athletic activity, gradually, once pain-free during ambulation. The use of an elastic ankle brace only during sports activity was also recommended for the following weeks or months after the injury to prevent re-injury.

Four researchers conducted follow-up calls to patients at 5 , 14, and 30 days post-injury using a predefined questionnaire. If no response was received on the first attempt, a second call was made the same day, and in case of a repeated lack of response, another attempt was made the following day, with a non-response recorded as a loss. For patients who did not respond to two or more follow-ups, a review of the electronic medical record was conducted in the month following their initial emergency visit via a centralized database, which records any visit to a public healthcare facility, whether primary care or hospital, in the province where the study was conducted.

Study data were collected and managed using REDCap electronic data capture tools hosted at our institution.

The primary outcome assessed was the score on “The Oxford Ankle Foot Questionnaire for Children” (OXAFQ-C) (https://acortar.link/oFt4Ne) at 5, 14, and 30 days, a validated scale for the pediatric population aged 5 to 16 years to assess the functionality of the injured ankle, comprising 14 items across three domains: physical, social/daily activities, and emotional. Each item is scored on a frequency scale: never (4), rarely (3), sometimes (2), often (1), always (0). Higher scores indicate better functionality. This scale can also be converted into a percentage scale. A difference of more than 10 points in the OXAFQ-C percentage score was considered clinically significant. Licenses for academic use were obtained prior to study commencement.

As secondary variables, both groups were assessed for the presence of pain at the time of review (PES and 5, 14, and 30 days follow-up) and, if pain was present, its severity on a numerical scale (0–10). Days requiring oral analgesia and days requiring crutches use were also recorded. It was also documented, for the restraint group, adverse effects of the bandage and tolerance to it using an ordinal scale (very poor, poor, fair, good, very good). In the non-restraint group, the use of an elastic ankle brace or any other form of restraint despite recommendations was recorded. Finally, patient satisfaction regarding the care received in the PES, the assigned treatment, and their recovery was recorded using a numerical scale (0–10).

In the absence of previous studies evaluating the use of non-restraint approaches in patients with ankle sprains, as well as research investigating the outcomes of the OXAFQ-C questionnaire during their recovery process, the initial sample size calculation assumed a non-inferiority margin of 10 points on the percentage scale of the OXAFQ-C at 5 days and a standard deviation (SD) of 25 for this calculation. Thus, comparing two independent means with a two-sided test and balanced random allocation (Group 1 size/Group 2 size ratio = 1), with an alpha risk of 5% and a power of 80%, the calculated sample size was 156 patients (78 in each group). Assuming a dropout rate of 10%, the established sample size was 172 (86 in each group). Due to the lack of prior data, an interim analysis was planned after the inclusion of the first 50 patients to recalculate the sample size. With the preliminary data obtained, the final estimated sample size was 100 (50 in each group) for a non-inferiority margin of 10 points on the OXAFQ-C percentage scale at 5 days, an SD of 15, an alpha risk of 5%, a power of 90%, and a dropout rate of 30%.

An intention-to-treat analysis and a per-protocol analysis were planned. Given an estimated loss to follow-up greater than 5% but less than 40%, the best–worst case analysis proposed by Jakobsen et al. [13] was used, where missing data from the restraint group were replaced with the mean + 1 SD of that group (beneficial outcome) and those from the non-restraint group with the mean − 1 SD (harmful outcome).

Categorical variables were described using percentages, and continuous variables as means and standard deviations (SD) if normally distributed (Kolmogorov–Smirnov and/or Shapiro–Wilk) or medians and interquartile ranges otherwise. The bivariate analysis of categorical variables was conducted using the Chi-square test or Fisher's exact test, and that of continuous variables with the Student’s t-test or its non-parametric equivalent (the Mann–Whitney U test).

All statistical tests were considered significant at a 5% level. The analysis was performed using JASP software version 0.18.