Plantar pressure assessment is commonly used to provide insight into the plantar loading characteristics of the paediatric foot. Due to their common use in both a clinical and research setting, it is necessary to ensure that the protocols for plantar pressure assessment in children can reproduce plantar pressure measures of dynamic foot function on different occasions. To date no study has also considered the reliability of protocols for the assessment of plantar pressure data in typically developing children and therefore, the purpose of this study was to determine the reliability of plantar pressure variables in a group of children aged 7 to 11 years, during barefoot level walking.
The results demonstrated that the collection of reliable plantar pressure variables is possible in children for all foot segments except the lesser toes. The within-session ICCs for the seven analysed regions of the foot ranged from 0.50 to 0.92 for peak pressure, 0.47 to 0.93 for peak force, 0.46 to 0.91 for pressure-time integrals and 0.17 to 0.92 for force-time integrals. All variables, except at the lesser toes recorded consistently moderate-to-good levels of reliability, whereas the lesser toe region reported poor levels of reliability, particularly for the force-time integral (0.17). However, the ICC is a unitless value and does not provide an indication of absolute variability and therefore further analysis using CoVs (%) was conducted. This analysis showed a similar pattern to the ICCs in that the reliable variables demonstrated smaller CoV percentages (10.22 to 19.84%) for all foot segments in comparison to the lesser toes (27.15 to 56.08%). These results indicate that within a single testing session, repeated measurements at the lesser toes are associated with reduced reliability and increased variability during gait, in comparison to the other six regions of the foot.
Assessment of systematic differences between sessions indicated that peak pressure, peak force, pressure-time integral and force-time integral at the lesser toe region exhibited a significant mean difference between sessions (p < 0.05). The remaining six regions, across all four variables, did not display any systematic differences between sessions, captured one week apart. Relative reliability between sessions was consistently good, with all regions of the foot, except the lesser toes reporting ICC values greater than 0.79. The lesser toe region again reported lower reliability (0.58-0.68) across the four variables. The between-session CoVs for the seven analysed regions of the foot ranged from 6.81 to 29.64% for peak pressure, 22.83 to 51.83% for peak force, 10.13% to 31.75% for pressure-time integrals and 21.85 to 56.61% for force-time integrals. Due to the absence of an agreed upon criteria for the assessment of CoV values, it is difficult to comment upon the acceptability of the values derived from this study. However it is important to note that the CoV's reported in this study are typically higher for the four measured variables across the seven regions of the foot, in comparison to those reported in the investigation into the reliability of repeated plantar pressure measurements in adults [20, 23] and may indicate an increased variability in the foot loading patterns of children during gait.
The midfoot and lesser toe region displayed the greatest percentage difference for all four variables, highlighting the greater variability within different regions of the foot. The findings at the midfoot and lesser toe regions in this study are consistent with previous reports in adults [20, 23] and indicate that these regions of the foot may be subject to inherent variability during gait. The reduced reliability between sessions at the lesser toes reported in this study are also consistent with those reported by Gurney et al. . Findings from the adult study were that reliability was reduced in areas of the foot where loading was less typical, in this instance it was the lateral lesser toes (defined as the 3rd-5th toes). Also of interest, the authors reported reduced reliability at the medial mid-foot (ICC < 0.8) which was not noted in the present study. Although these results are of interest direct comparison between studies is difficult due to differences in the participants, methodologies adopted and in the sensor technology used. The study by Gurney et al.  assessed reliability using the Novel EMED® plantar pressure platform, in comparison to the TekScan MatScan® system used in this study. The Novel EMED® platform used by Gurney et al.  has a slightly higher spatial resolution of 2 sensors/cm2 in comparison to 1.4 sensors/cm2 of the TekScan MatScan® system; it also has a slightly higher sampling frequency of 50 Hz in comparison to 40 Hz respectively. Consequently, although the collection of reliable plantar pressure data appears possible in children, the lower spatial resolution of the TekScan MatScan® system may limit the validity of this system in accurately being able to isolate small regions of the foot, such as the lesser toes and highlights the importance of the resolution of a system when assessing the plantar pressures in children with small foot sizes.
The work by Gurney et al.  highlighted that the definition of foot segments is not uniform or consistent across different studies. Gurney et al.  divided the mid-foot into lateral and medial midfoot whereas the present study defined the mid-foot as one segment. The definition of the foot segments between studies limits direct comparison. Furthermore, the methods used for segmental division of the foot vary between systems and can be subject to error resulting from the sensor resolution of the platform and anatomical knowledge of the observer's involved . Recently, Deschamps et al.  conducted a study looking at the reliability of manually determined masks across the forefoot in plantar pressure footprints and concluded that that the masking of small segments of the foot be conducted with caution. This is in agreement with the work of Latour et al.  and Urry and Wearing  whom both commented that the lower reliability at the lesser toes may be due to limitations with sensor technology in isolating this small region of the foot, particularly in young children.
Normative pressure data have been published by Alvarez et al.  who identified age-related differences in plantar pressure profiles in a sample of 146 children ranging from 1.6 years-14.9 years. In their study the foot was divided into five segments however the lesser toes were not considered separately, rather as part of the lateral forefoot. They reported that comparable foot pressure profiles could be identified across the age groups which were: (i) children under the age of two years; (ii) children aged two-five years and (iii) children older than five years. This work is interesting because it suggests that whilst there is debate in the literature regarding the development of a mature gait , the loading characteristics of children over five years are consistent and can yield reliable data. However, with the limitation of the lesser toes being incorporated as part of the lateral forefoot, further work is required to explore the reliability of this specific region of the foot in children.
Plantar pressure assessment is complex and challenging in young children and it is important to acknowledge that the capture of reliable data is dependent on a number of different factors, such as the participants (age, pathology and developmental status), instrumentation (sensor technology and validity of instrumentation) and the adopted protocols (abbreviated vs. midgait protocols, gait velocity). Factors such as walking speed have been reported to affect pressure variables  and where appropriate should be taken into consideration. To the authors' knowledge, there are no studies which have attempted to standardise or control cadence during plantar pressure studies in children and thus would suggest that standardisation may not be appropriate in the child. However, this is an area that warrants further investigation.
There are several limitations to this present study that need to be considered when interpreting the findings. First, only typically developing children were recruited, so the reliability of these measurements cannot be generalised to a clinical population. Confounding variables such as pain or asymmetrical gait commonly reported in symptomatic populations may have a significant impact upon the reproducibility of plantar measurements taken one week apart. Secondly, Zammit et al.  has previously discussed limitations with the use of the TekScan MatScan® including factors such as the manual masking procedure for the determination of plantar pressure outputs for each individual as well as the relatively low sampling frequency (40 Hz) and spatial resolution (1.4 sensors/cm2). All of which may affect the validity and reliability of repeated measures and in accurately isolating small regions of the foot as seen in children. Finally, the results from this study can only apply to the age group under investigation. Further work is required to investigate the reliability of plantar pressure measurements in a younger group of children, due to the potential influence of the continuing development of their foot structure. Additional research would also be required to confirm these results in an older cohort of children due to any influence postural changes have upon musculoskeletal development during puberty and adolescence.
The present study has demonstrated that reliable plantar pressure data can be collected in children aged 7-11 years, however it must be acknowledged that the results presented can only apply to the sample under investigation and to the instrumentation used in this study. Further work is required to explore the reliability of plantar pressure data collection across children of varying ages and postural development to determine the factors which influence the reliability of plantar pressure data capture. There is also a need to explore the reliability of plantar pressure assessment where deformity is present.