This study was a case control design and approved by Monash Health Human Research Ethics Committee (HREC number: 15405A). Ethics approval for The 1000 Norms Project was provided by National Health and Medical Research Council of Australia Centre for Research Excellence in Neuromuscular Disorders (NHMRC 1031893). All parents/carers of participants provided written informed consent, and child participants assented.

Participants were recruited from private practice clinics, public health outpatient and community clinics. Participants were eligible if they were between the aged 4–16 years, visually demonstrated a toe walking gait, and were diagnosed with idiopathic toe walking gait by a multidisciplinary clinic with medical and allied health team. If potential participants were toe walking and had not seen the multidisciplinary clinic, they were screened with a validated exclusionary tool [3]. The screener was clinician researcher who had > 8 years working in public health community-based paediatric gait screening clinics.

Participants were excluded if they had lower limb pain at the time of initial screening, had previous ITW treatment with ankle foot or full length orthotics, recent prescription of stretches or strengthening program targeted at the lower limb that the child was adhering to at a dosage that was deemed potentially having clinical impact, or serial casting or Botulinum Toxin-A as part of their ITW treatment within the past 12 months.

Normative raw data was sourced from two data sets. Strength and active range of motion data was collected from the 1000 Norms Project. The 1000 Norms Project is an observational study investigating outcome measures of self-reported health and physical function in 1000 healthy individuals aged 3 to 101 years [15]. A secondary normative data source was used for comparing the weight bearing lunge test in a straight leg position between children who have ITW with their age matched peers [7].

The following demographic data was collected from the parent: child’s age (years), sex, weight (kg), height (m), parent-reported duration of toe walking (years), parental estimate of percentage of time awake that the child toe walks (% of walking).

The primary outcome of interest was the weight bearing ankle joint range of motion. Additional lower limb range measures, and strength measures were also assessed (Table 1).

Weight bearing ankle joint range of motion was assessed with a calibrated digital inclinometer. Active joint ranges of motion were assessed with a universal goniometer. The starting positions, limb positions and assessment task/movement were performed as per the 1000 Norms Project protocol, a summary of this is described in Table 1.

Lower limb muscle strength testing was undertaken with the Citec handheld dynamometer (Citec dynamometer CT 3001, CIT technics, Groningen, the Netherlands). Each participant was assessed using the “make” technique to measure strength and were directed to exert a maximal force against the hand-held dynamometer [16]. The starting position, and limb position and movement assessed are described in Table 1. The universal goniometer, and digital inclinometer (Laser Depot, Adelaide, Australia) and hand-held dynamotor have all demonstrated high reliability when used according to the set measurement protocol [16].

The principal researcher (AC) had experience in utilising these measurement techniques (AC), however prior to data collection, had peer support to match technique with measurement protocols with an experienced physiotherapist and podiatrist. The principal researcher was responsible for all participant testing. For the range of motion measures, each participant was asked to perform the movement to their end range and hold while the tester recorded the active or weightbearing range of motion. For the strength measures, each participant had a practice trial at submaximal effort, then were instructed to perform three maximal voluntary contractions lasting three to 5 seconds each. Given the age of participants, rapport with the participant was obtained prior to testing. Instructions and encouragement were individualised to the personality and patients’ age to account for difference cognitive abilities. Participants were given a resting period of 10 seconds in between each contraction. All data were entered into an online spreadsheet. It was pre-planned that where a participant was unable to perform the measure, no data were recorded for that item and treated as missing data.

Data were analysed with the use of Stata 13 (Stata Corp, College Station Texas). Descriptive synthesis of demographic data were completed. Anthropometric measures were described in means (Standard Deviations = SD) or frequencies (%) after confirming normal distribution of data. As ITW is only diagnosed when the child toe walks symmetrically, only right leg measures were used. This has been found to satisfy assumptions of data independence where there is likelihood of high correlation between two limbs [15].

We originally explored any differences between groups using logistic regression. The data from the ITW group and normative group were originally compared using univariate logistic regression analysis to determine any group differences in each individual measure. Backwards step multivariable linear regression analysis was then conducted for each individual range of motion or strength variable taking into account other variables identified as impacting range of motion or strength. Where there were variables that were highly intercorrelated (r > 0.7), for example, height and weight, only one variable was included to avoid multicollinearity. The preliminary multivariable model for each measure were built with variables identified in univariate analysis that revealed a value of p ≤ 0.2. During the analysis, variables were then removed based on least significant fit in a backward stepwise fashion. This backward step removal continued until all remaining model variables had a p-value < 0.05 [17]. Regression coefficients (Coef) were reported to understand direction of the relationship, and 95% confidence intervals (CI) were also reported.

A sample size of 26 participants was calculated to achieve 80% power, and to detect an effect size of 0.81 as a result of differences in ankle range of motion (primary outcome) between the ITW cohort and their non-toe walking peers using an α criterion of 0.05 [7].