This study was a single-centre, parallel randomised controlled trial which was conducted among patients with clubfoot selected from the orthopaedic clinic at the University Teaching Hospital (UTH) in Lusaka, Zambia. The study involved two groups who were required to start the intervention at the same time. Group A was subjected to both AFO and FAB (experiment) while Group B was subjected to the FAB only (control group).

This study was conducted at the orthopaedic clinic of the UTH in Lusaka, Zambia. University Teaching Hospital, a third-level hospital located along Nationalist Road, serves as a referral centre and training facility for healthcare providers. The Consolidated Standards of Reporting Trials (CONSORT) requirements were followed in this study.

The study population comprised all male and female children receiving Ponseti clubfoot management at UTH. The inclusion criteria were children of both sexes undergoing Ponseti clubfoot management with a minimum age of 1 year and a maximum age of 3 years, with idiopathic clubfoot who had completed the correction phase and were in the maintenance phase on a night bracing regimen. They were required to be able to stand and walk independently or with minimal support. The exclusion criteria were patients with syndromic clubfoot and patients who underwent a surgical operation rather than the Ponseti method.

Sample size determination was conducted using G-Power 3.1.9.7 software. To achieve 80% power with 5% type I error rate, a minimum of 13 participants per group was required. This calculation was based on relapse rates reported by Janicki et al. (2011), where recurrence occurred in 83% of cases using AFO and 31% with FAB following correction with the Ponseti method. The sample size estimation utilised the z-test family to compare two independent proportions. To account for an anticipated 20% dropout rate, the required sample size was adjusted using the formula:

A total of 40 children with clubfoot were assessed for eligibility, of whom 6 were excluded for not meeting the inclusion criteria. Three participants dropped out during data collection, 1 from the experiment group and 2 from the control group. Intention-to-treat was used to include all the participants in the analysis (Figure 1).

Thirty-four participants were randomly assigned to either the AFO+FAB or FAB group with a 1:1 ratio using simple randomisation in Stata 17, utilising computer-generated numbers. An independent researcher conducted the randomisation, and group assignments were concealed to prevent selection bias. To ensure methodological rigour, outcome assessors who are Ponseti-trained physiotherapists were blinded to group allocation, minimising assessment bias. Opaque sealed envelopes were used during the allocation. Participants’ enrolment numbers were used in place of names.

Based on the study objectives, three (n = 3) assessment instruments were used to collect data. Ankle range of motion (ROM) was measured using a goniometer, assessing dorsiflexion, eversion and inversion. This ensured objective evaluation of mobility improvements over time. Goniometry involves the art and science of measuring the range of motion in each joint plane, and the goniometer is a widely used tool (Gandbhir et al. 2020). The scale has two possible ranges: from 0 degree to 180 degrees and vice versa. Studies have shown that a universal goniometer has been found to have good to excellent reliability (Gogia et al. 1987; Hanks & Myers 2023). Clubfoot severity and relapse were evaluated using the Pirani scoring system, a validated tool for assessing the severity of clubfoot deformity and identifying potential relapses. Relapse was defined as a deterioration in foot alignment necessitating additional intervention. The Pirani scoring system has three points used as its basis: 0 for no abnormality, 0.5 for moderate abnormality and 1 for severe abnormality. The Pirani scoring system has been proven to be both valid and reliable (Pirani, Hodges & Sekeramayi 2008). Functional activity was assessed using the Ponseti functional score, which evaluates a child’s ability to perform weight-bearing and ambulatory activities, providing insight into mobility improvements. It is a sum of a 100-point assessment scale, of which 30 comes from the practitioner, while 70 comes from the patient. This has been proven reliable in assessing functional outcome (Laaveg & Ponseti 1980).

The intervention involved incorporating an AFO alongside the standard FAB to enhance adherence and reduce clubfoot relapses. Designed to provide additional support and improve comfort, the AFO aimed to increase compliance during daytime activities, potentially lowering relapse rates. Participants were randomly assigned to two groups: Group A received standard treatment with the FAB worn only during sleep, while Group B used both the FAB (during sleep) and the AFO (throughout the day, except during bathing and passive stretching). The AFOs used were custom-made at the hospital using thermoplastic orthotic material. The parents were shown how to fit the AFO on the child and were instructed to fit it during the daytime while the child was awake and playing. The FABs were readily available at the hospital. They were fitted according to the foot size, adjusted for abduction and dorsi flexion using a bar which is designed between the two braces. Parents were instructed to wear during nap time and sleeping time. They were also educated on the importance of compliance. The intervention was implemented for a period of 6 months at the UTH by two Ponseti-trained physiotherapists, each with over 5 years of experience in clubfoot management. During reviews, monitoring was carried out by inspecting the device for indications of usage, such as worn-out straps, and parents being able to point out if the child had outgrown the size of the device.

Data were collected at three time points: baseline, midline and endline over a 6-month period. Standardised assessment tools, including the Pirani scoring system, Ponseti functional score and the goniometer, were used to ensure consistency across all data collection stages. The Pirani score is a widely used clinical tool for assessing the severity of clubfoot, based on six deformity indicators in the midfoot and hindfoot (Pirani et al. 1999). To enhance data reliability, trained assessors conducted measurements using these instruments. Additionally, assessors remained blinded to group details to minimise assessment bias. This structured approach strengthened the rigour and validity of the study’s outcome evaluations.

The study assessed three key parameters: ankle range of motion, clubfoot severity and relapse and functional activity. The co-primary endpoints were improvements in ankle ROM and functional activity from baseline. The secondary outcome was the occurrence of relapse. All parameters were measured at baseline, midline and endline to track progress over the study period.

Data were analysed using Stata version 17.0 for Windows. The choice of statistical tests was guided by data distribution and the nature of comparisons. Independent t tests were performed for normally distributed demographic variables, while Wilcoxon rank-sum tests were used for non-normally distributed variables when comparing groups. We used the Chi-square test to compare categorical variables; Fisher’s exact test was used when assumptions for a Chi-square test were violated.

The effect of treatment on ROM, functional activity outcome, and relapse was analysed using mixed-effect linear regression models to account for the repeated measurements within participants over time. The primary fixed effects were treatment group (treatment vs. control), time period and their interaction. The interaction term assessed whether the change in outcomes over time differed between treatment groups. All models were adjusted for potential confounding, including age, number of casts, sex of the child, bilateral involvement (bilateral vs. unilateral), affected side and tenotomy status. Model estimates were reported as regression coefficients (β) with corresponding 95% confidence interval (CI). The significance level for the statistical tests was set at 5% (p < 0.05), with results below this threshold considered statistically significant. An intentional-to-treat approach was applied to the analysis. All randomised participants were included in the groups to which they were originally assigned, regardless of whether they completed all follow-up assessments. This approach preserved the benefits of randomisation and provided an unbiased estimate of treatment effectiveness under the real-world conditions (McCoy 2017). The incomplete repeated measures were accounted for by the use of mixed-effect models (Brown 2021).

Ethical clearance to conduct this study was obtained from the University of Zambia Health Sciences Research Ethics Committee (202301270106). Certification by the National Health Research Authority (NHRA) was granted under registration number NHRAR-R 572/21/06/2023. The study was also registered with the Pan African Clinical Trial Registry (PACTR; https://pactr.samrc.ac.za) under registration number PACTR202502554262049. Institutional permission was secured before the study commenced. Informed consent was obtained from all participants’ legal guardians, with assent forms signed. Risks and potential harm were minimised by ensuring that only certified research assistants who are also Ponseti-trained physiotherapists attend to the study participants. Enrolment numbers were used in place of participants’ names. The principal investigator kept all participant records in a password-protected file.

As shown in Table 1, baseline characteristics were comparable between the two groups in terms of age (months), number of casts, affected side and tenotomy history (p > 0.05). However, a significant difference was observed in sex distribution, with a higher proportion of female children in the intervention group compared to the control group (71.4% vs. 35.0%, p = 0.037).

The mixed-effects model results in Table 2 show that participants in the treatment group (AFO+FAB) had significantly higher dorsiflexion ROM than the control group (FAB) at baseline (β = 7.16, p = 0.036). The ROM declined significantly over time across all groups (β = –4.47, p < 0.001). A significant interaction between treatment and time (β = 2.96, p = 0.007) for dorsiflexion revealed that the AFO+FAB group declined less in dorsiflexion ROM than the control group. No significant treatment interactions were observed for eversion or inversion. This is visualised in Figure 2, where the predicted means from the model demonstrate the diverging trajectories between the treatment and control groups.

Table 3 shows the change in patients’ functional activity from baseline to the end of the study in the two groups (AFO+FAB and the FAB only). The study shows a reduction (β = –15.62, p < 0.001) in Ponseti functional score in the control group over time (FAB only). A significant interaction between treatment and time (β = 19.01, p < 0.001) for the Ponseti functional score revealed that the AFO+FAB group had a higher functional activity score at endline compared to the control group. Figure 3 visually represents the model’s predicted means, which show how the Ponseti functional scores differ between treatment and control groups.

Table 4 shows that the mean Pirani score significantly increased in the FAB intervention group at the end of the study (β = 0.76, p < 0.001). There was a significant interaction between treatment and time (β = –0.51, p = 0.018) for the Pirani score, indicating that the AFO+FAB group had a lower Pirani score at endline compared to the control group. A visual representation of how the model’s predicted Pirani means differ between treatment and control groups is shown in Figure 4.

The study had some limitations in that although randomisation was used in allocating participants to treatment groups, significant differences were observed at baseline for some variables, such as sex, which could have introduced biases and affected the results of the study. The small number of children in each group limits the generalisability of the findings. Furthermore, treatment compliance was partially assessed. Although adherence to orthotics was occasionally noted during clinic visits (evidenced by whether the child wore the brace and its condition), no systematic methods, such as structured caregiver interviews or adherence logs, were employed to monitor consistent usage at home. This informal assessment may not fully capture variations in daily use. Future studies should consider adopting standardised and objective methods to more accurately monitor compliance, like the wear time diary (Griffiths et al. 2022).

Recommendation for further clinical trials in future should be on a large scale involving a larger sample size to improve generalisability. It may involve multiple clubfoot centres and spread across a long period in order to evaluate the sustained effects of the AFO+FAB intervention on relapse rates. Incorporation of an AFO in the maintenance phase of Ponseti management should be considered because of the potential to prevent relapses.