Abstract
Introduction: Foot morphology in the general population has been shown to change with age, and active older adults have reported a need for wide-fitting footwear. Methods: This study recruited 374 women active in racket sports and team sports in the UK who had their feet scanned while 50% weight bearing. Participants were grouped into 10-year age bands ranging from 18–29 years to 70–79 years. Data analysis was performed on the widths, heights, and circumferences of participants’ right feet normalised to foot length, as well as an assessment of hallux valgus angle and deformity. Results: The 18–29-year group had significantly smaller measures of foot width, ball of foot circumference and short heel circumference (p < 0.05, η2 = 0.042, η2 = 0.056) compared to the older groups. The foot dorsum height and circumference at 50% foot length were significantly less in the oldest age groups compared to the middle age groups (p = 0.0001, η2 = 0.055 and p = 0.0007, η2 = 0.044, respectively). There was some evidence of increased hallux valgus deformity with age. Conclusion: Designers and manufacturers of athletic footwear should be aware of the changes in foot morphology with age in order to provide more inclusive footwear.
Introduction
Foot morphology in the general population has been reported to change with age. In particular, there is often an increase in forefoot width with age, which may be related to swollen feet because of venous deficiencies, reduced medial longitudinal arch height, and hallux valgus and other foot and toe deformities including those associated with rheumatoid arthritis [1‒4]. In Caucasian women between 20 and 80 years old from the general population, an increase in ball width, ball circumference, circumference around the heel, and circumference around the instep has been reported per decade [4]. In women from Hong Kong, a moderately greater foot width was reported in older women (mean ± SD age 73 ± 7 years) compared to younger women (24 ± 4 years), while there was little to no effect of age on the length between heel and either first or fifth metatarsal head [5]. Toe deformities are typically more prevalent in older women compared to older men [6‒9]. Women also have a heightened susceptibility to musculoskeletal injuries as they age compared to males [10]. Changes in foot dimensions with age have clear implications for shoe design, as obtaining a proper shoe fit can be achieved if shoe design is based on the individual’s foot shape or the average of the target population [11‒14]. The importance of athletic footwear being available in suitable wide-fitting options has been raised by active older adults [15, 16]; however, to our knowledge, no data exist on the foot dimension of active older women.
Shoe design should be based on the feet of the target customers [17] but also be appropriate for the demands of the activity for which they are intended to be worn. Traditionally, lasts for athletic shoes for women have been based on scaled-down versions of men’s lasts [18‒20], despite well-documented sex differences in foot morphology across the lifespan [1, 17, 18, 21‒24]. In participants under 60 years of age, women’s feet are reportedly shorter and narrower than men’s feet, with a lesser instep height and circumference [18, 21, 22, 24]. In those over 60 years old, ball width, ball circumference, instep height and circumference are also smaller in women than men, with additionally greater pronation and increased hallux valgus observed in women than men [1, 23]. Shoes for sports which involve changes of direction (e.g., court sports) will have particular requirements such as sufficient stability to avoid excessive supination [25], but the players themselves may also have characteristic foot shapes as foot structure can be influenced by habitual loading and physical activity [26]. Shoes for court sports are of particular relevance to older adults as racket sports and modified team sports are popular sports in later life [27‒32]. Therefore, to make footwear suitable for older women engaging in court sports, it is important to have data on the foot shape of this group from which to design shoes.
Three-dimension foot scanning is a straightforward means of obtaining foot dimensions from a large sample in a short period of time [14]. Conversely, measuring foot dimensions manually can be tedious and dependent on the measurer and their instruments [13]. Using a 3D foot scanner has been recommended over digital calipers and digital or ink-based footprint methods for its superior accuracy and precision [33]. Taking a 3D foot scan when the foot is loaded has real-world relevance in which to inform the last design and can enable a good match between foot and shoe [34]. From half-weight bearing to full-weight bearing, foot length, ball width, and heel width can increase [35, 36]. Medial longitudinal arch height will also decrease with increasing load [35‒37]. However, changes in foot dimensions from half-body weight to full-body weight are more modest than from unloaded to loaded [35‒37]. As it is quick and easy to obtain 3D foot scans, it is possible to obtain scans both 50% and 100% weight bearing in the same session. Arguably, with greater load, the more the foot morphology would be representative of a dynamic game situation.
The specific variables used to characterise foot morphology depend on the application of the data. Foot length, width at the forefoot and heel, instep height, and girth measurements are informative for shoe design to ensure a proper fit [13, 17, 19, 34, 38]. While 3D foot scanning allows for a sophisticated exploration of foot form and shape with the use of statistical shape-function models [39‒41], it is difficult to incorporate foot shape in the last design, so linear and girth measurements are more relevant to industrial applications than shape information [13]. Absolute foot dimensions have value for the grading system of shoe design by size [17, 22, 42]; however, relative measurements expressed as a percentage of foot length are informative for comparing groups like sex, gender, or weight groups across sizes [1, 18, 43]. Additionally, quantifying the presence of a hallux valgus deformity and the hallux valgus angle could be informative to shoe design as older adults report that athletic shoes reinforced around the bunion are painful [15, 16]. Information on the prevalence and severity of hallux valgus is currently lacking in the 3D foot scan literature [44].
The aim of this study was to provide a novel data set characterising foot morphology of adult women active in court sports across different ages by quantifying foot dimensions and hallux valgus deformity using 3D foot scans at 50% and 100% weight bearing. Our primary hypothesis was that foot width, ball width and ball circumference would increase with age and arch height would decrease with age. Second, we hypothesised that the hallux valgus angle and the presence of a hallux valgus deformity would increase with age. Additionally, we hypothesised modest differences between 50% and 100% weight bearing, in particular greater foot width, ball width, and ball circumference and lesser arch height in 100% weight bearing compared to 50% weight bearing. The intention is that the data from this study will quantify age-related differences in the foot morphology of active women and inform athletic footwear design, allowing athletic footwear designs to become more inclusive.
Methods
Participants
A total of 374 women were recruited and scanned at sports tournaments and training sessions/matches around the UK and from local sports clubs and personal contacts at the university. The sample size per age group was comparable to other literature on the effect of age on foot morphology [5, 45]. Participants were eligible if they were a woman over the age of 18 years of age and played a team or racket sport at any level and frequency. Ethical approval was granted from the university (519475) and each participant provided either written or digital informed consent.
Scanning Procedure
The 3D foot scans were recorded barefoot using a Tiger Scanner (Materialise, Belgium). The scanner uses laser technology and nine cameras with a manufacturer-reported accuracy of 0.5 mm and a scan time of 5–15 s. A single scan of each foot was taken first 50% weight bearing, with the foot of interest in the scanner volume and the contralateral foot stood on a platform to the side. Participants were instructed to look straight ahead and stand in a relaxed standing position with weight evenly distributed between their feet (to approximate 50% weight bearing). The procedure was then repeated 100% weight bearing, in which one foot was in the scanner volume and the other raised in the air. Participants were permitted to use the handrail to assist with balance if necessary; however, they were asked not to put their weight on the handrail. The scanner was wiped with an antibacterial wipe between participants.
Measurements
The manufacturer software (Materialise Footscan version 9.9.1) was used to obtain automated foot length, foot width, arch height, and arch length dimensions (Table 1) and an estimated UK shoe size (allowing 15 mm toe space). Measurements obtained from the scanner software were manually entered into Microsoft Excel. Additional analysis was performed in MATLAB (R2023b, Mathworks Inc.) using an adapted version of a programme written by J. Anderson (PhD). Outcome variables and their definitions are included in Table 1, which each corresponds to one of the multiple definitions provided in the Institute of Electrical and Electronics Engineers (IEEE) Standards Association White Paper on foot measurement technology [46]. The MATLAB analysis aligns the foot so that the most inferior proximal and distal points of the foot sit on x = 0 and the proximal centre of the foot and distal centre of the foot sit on z = 0. This method is analogous to “Method 2” described by the British Standards Institute [47]. The analysis requires some user input to define landmarks and the inflection point on the dorsal surface of the foot which was performed by authors J.R. and R.B. A negative hallux valgus angle was interpreted as measurement error and converted to zero. A subjective assessment of the presence of hallux valgus deformity was performed on a two-dimensional cross-section of the three-dimensional data according to the Manchester Scale, again by authors J.R. and R.B., who are non-clinicians [48]. The Manchester scale was developed as a means of non-invasively classifying hallux valgus deformity from grade 1 (no deformity) to grade 4 (severe deformity) by comparison to standardised photographs; the scale has previously been validated against x-ray measurements [49]. Due to some data collection occurring at grassed venues, in some cases, there were small amounts of grass and debris included in the scan which could affect measurements. In these cases, the scans were first imported into MeshLab [50] to remove unwanted vertices. The edited scans were then reimported into the Materialise database for a recalculation of foot length and foot width and the edited scan was used for the MATLAB analysis. When time allowed, ethnicity was obtained as part of a marketing questionnaire administered by our industry partners.
Statistical Analysis
Reliability was assessed using subsets of data for within scan (the same scan analysed twice), between scans (the same foot recorded 50% weight bearing three times) and between raters (the same scans analysed by both J.R. and R.B.). For intra-scan reliability 20 scans, each from a different participant (10 left feet and 10 right feet) were analysed. The scans selected for intra-scan reliability were from participants between 43 and 78 years of age (mean ± SD: 61 ± 12 years old). For inter-scan and inter-rater reliability, analysis was performed on both feet of 13 younger women between 19 and 22 years of age (mean ± SD: 20 ± 1 years old) and 8 older women between 51 and 72 years of age (mean ± SD: 57 ± 7 years old). The reliability statistics calculated were the typical error (SEM [SDΔ/√2]), minimal detectable difference (MDD), intraclass correlation (ICC; 3,1), and coefficient of variation (CV). The ICC (3,1) and SEM were calculated using an open source Microsoft Excel spreadsheet for consecutive pairwise analysis [51]. The MDD was calculated as 1.5*SEM [52]. The CV was calculated as (SEM/grand mean) × 100 [52, 53]. A weighted kappa statistic [54] was calculated to determine the level of agreement between authors when grading hallux valgus deformity. The strength of agreement was interpreted as: <0 is poor; 0.00–0.20 is slight; 0.21–0.40 is fair; 0.41–0.60 is moderate; 0.61–0.80 is substantial; and 0.81–1.00 is almost perfect, according to thresholds suggested by Landis and Koch [54].
For absolute values and the differences between weight-bearing conditions, outliers were determined to be values more than three standard deviations from the mean. Outliers were corrected if due to an error in data entry or data analysis; otherwise, outliers were retained in the data set. Data were assessed for normality by visual inspection of histograms and using a threshold of an absolute skewness ≥2 or an absolute kurtosis value ≥7 for substantial non-normality [55]. To assess the effect of increasing percentage of weight bearing, allowing for persisting outliers, the difference between sessions was calculated as 100% weight bearing – 50% weight bearing and compared using paired t tests and Wilcoxon tests for data that were non-normally distributed. Effect sizes were calculated as Cohen’s d and r for parametric and non-parametric data, respectively.
For assessing the effect of age, where appropriate, variables were normalised to foot length (measures were expressed as a percentage of foot length, foot length x). Participants were grouped by age groups of 18–29 years, 30–39 years, 40–49 years, 50–59 years, 60–69 years, and 70–79 years. Further statistical analysis was performed on the right foot only as per previous work to satisfy the assumption of data independence [1, 43, 56]. A one-way analysis of variance (ANOVA) was performed with age group as a factor. Groups were checked for homogeneity of variance using Bartlett’s test. In the event of substantial departures from normality, or unequal variances, a Kruskal-Wallis test was performed. Effect sizes were calculated as eta squared (η2) with threshold values interpreted as small (0.01), medium (0.06) and large (≥0.14) [57]. A post hoc analysis was performed using a Bonferroni test when there was a significant effect of age group following an ANOVA and Dunn’s test [58] with a Bonferroni correction when there was a significant effect of age group following a Kruskal-Wallis test. Level of significance was set at α = 0.05. Statistical analysis was performed in MATLAB using the Statistics and Machine Learning Toolbox, SPSS Version 29 for Windows (IBM Corp, Armonk, NY) and Microsoft Excel. A checklist for reporting 3D scanning studies [44] is included in the online supplementary material (for all online suppl. material, see https://doi.org/10.1159/000541732).
Results
Participant Characteristics
The number of participants in each group is presented in Table 2. Approximately 90% of the participants were Caucasian. The distribution of UK women’s shoe sizes by age group is represented in Figure 1.
Participant numbers per age group
. | 18–29 . | 30–39 . | 40–49 . | 50–59 . | 60–69 . | 70–79 . |
---|---|---|---|---|---|---|
n | 50 | 41 | 48 | 97 | 98 | 40 |
. | 18–29 . | 30–39 . | 40–49 . | 50–59 . | 60–69 . | 70–79 . |
---|---|---|---|---|---|---|
n | 50 | 41 | 48 | 97 | 98 | 40 |
Reliability
Reliability results are presented in Tables 3, 4. Reliability was generally better within the same scan or between scans than between assessors. For intra-scan variability, ICC values were 0.91 or greater, CV was between 0.12% and 2.74% and MDD was between 0.2 mm and 4.1 mm. Between scan reliability, ICCs were between 0.80 (arch height from the scanner) and 0.99, CV was generally around 2%, except for arch height from the scanner (7.14%) and hallux valgus angle (32.73%). MDD was between 0.6 mm and 3.7 mm. For inter-rater reliability, ICC ranged from 0.71 (heel to 1st metatarsophalangeal joint [MTPJ1]) to >0.99, CV was between 0.18% and 2.27%, except for hallux valgus angle (19.96%). The MDD from the inter-rater reliability varied from 0.25 mm to 6.02 mm and was used for interpretation of results in the context of measurement error as it was greater than the other measures of reliability. The MDD was most influenced by selecting the MTPJ joint location as MDD was highest in heel to MTPJ1 and heel to 5th metatarsophalangeal joint (MTPJ5). There was a fair agreement between assessors in hallux valgus deformity grading (κ = 0.362, p < 0.001). The arch height from the scanner software had a CV of 7.14% between scans, which was more than twice all other values obtained from the scanner software and it was observed that the scanner would select a different point within the arch area to measure the height between scans and an arch height below around 10 mm could not be detected. Therefore, this measure was not used for further analysis; instead, arch height was inferred from the normalised height at 50% of foot length which has been to define the arch height index in the literature [59] (described as instep height elsewhere [1]).
Intra-scan and between-scan reliability
Variable . | Typical error . | Lower CI . | Upper CI . | MDD . | ICC . | Lower CI . | Upper CI . | CV, % . |
---|---|---|---|---|---|---|---|---|
Intra-scan (same scan analysed twice for variables requiring user input) | ||||||||
Heel to MTPJ1 | 0.2 mm | 0.17 | 0.32 | 0.3 mm | >0.99 | >0.99 | >0.99 | 0.12 |
Heel to MTPJ5 | 1.5 mm | 1.14 | 2.16 | 2.2 mm | 0.98 | 0.94 | 0.99 | 0.92 |
Ball of foot circumference | 2.8 mm | 2.11 | 4.01 | 4.1 mm | 0.95 | 0.88 | 0.97 | 1.17 |
Ball width diagonal | 1.5 mm | 1.17 | 2.21 | 2.3 mm | 0.91 | 0.79 | 0.96 | 1.55 |
Max height over MTPJ | 0.3 mm | 0.23 | 0.43 | 0.5 mm | 0.99 | 0.98 | >0.99 | 0.76 |
Short heel circumference | 1.7 mm | 1.27 | 2.4 | 2.5 mm | 0.99 | 0.98 | >0.99 | 0.52 |
Hallux length | 0.1 mm | 0.1 | 0.17 | 0.2 mm | >0.99 | >0.99 | >0.99 | 0.19 |
Hallux valgus angle | 0.4° | 0.27 | 0.51 | 0.5° | >0.99 | >0.99 | >0.99 | 2.74 |
Between scan outputs from scanner (3 scans) | ||||||||
Foot length | 1.0 mm | 0.83 | 1.12 | 1.4 mm | 0.99 | 0.98 | 0.99 | 0.39 |
Foot width | 1.0 mm | 0.88 | 1.18 | 1.5 mm | 0.93 | 0.90 | 0.96 | 1.06 |
Arch height | 1.1 mm | 0.91 | 1.29 | 1.6 mm | 0.80 | 0.69 | 0.88 | 7.14 |
Arch length | 4.3 mm | 3.77 | 5.06 | 6.4 mm | 0.85 | 0.77 | 0.90 | 2.48 |
Shoe size | 0.2 | 0.14 | 0.19 | 0.2 | 0.98 | 0.96 | 0.99 | 2.79 |
Between scan outputs from MATLAB analysis (3 scans) | ||||||||
Foot length x | 0.9 mm | 0.77 | 1.09 | 1.3 mm | 0.99 | 0.99 | >0.99 | 0.37 |
Heel to MTPJ1 | 2.5 mm | 2.13 | 3.02 | 3.7 mm | 0.88 | 0.81 | 0.93 | 1.40 |
Heel to MTPJ5 | 2.2 mm | 1.90 | 2.69 | 3.3 mm | 0.91 | 0.84 | 0.95 | 1.38 |
Ball width diagonal | 0.8 mm | 0.68 | 0.96 | 1.2 mm | 0.97 | 0.95 | 0.98 | 0.84 |
Ball width straight | 0.5 mm | 0.45 | 0.63 | 0.8 mm | 0.98 | 0.97 | 0.99 | 0.57 |
Max height over MTPJ | 0.7 mm | 0.63 | 0.89 | 1.1 mm | 0.96 | 0.93 | 0.98 | 1.94 |
Height at 50% of foot length | 0.7 mm | 0.57 | 0.81 | 1.0 mm | 0.98 | 0.97 | 0.99 | 1.11 |
Circumference at 50% of foot length | 0.8 mm | 0.64 | 0.91 | 1.1 mm | 0.99 | 0.99 | >0.99 | 0.34 |
Ball of foot circumference | 1.5 mm | 1.25 | 1.78 | 2.2 mm | 0.98 | 0.96 | 0.99 | 0.65 |
Short heel circumference | 1.4 mm | 1.17 | 1.66 | 2.1 mm | 0.99 | 0.98 | 0.99 | 0.45 |
Heel breadth | 0.4 mm | 0.37 | 0.52 | 0.6 mm | 0.98 | 0.97 | 0.99 | 0.71 |
Hallux valgus angle | 3.3° | 2.79 | 3.96 | 4.9° | 0.90 | 0.84 | 0.94 | 32.73 |
Variable . | Typical error . | Lower CI . | Upper CI . | MDD . | ICC . | Lower CI . | Upper CI . | CV, % . |
---|---|---|---|---|---|---|---|---|
Intra-scan (same scan analysed twice for variables requiring user input) | ||||||||
Heel to MTPJ1 | 0.2 mm | 0.17 | 0.32 | 0.3 mm | >0.99 | >0.99 | >0.99 | 0.12 |
Heel to MTPJ5 | 1.5 mm | 1.14 | 2.16 | 2.2 mm | 0.98 | 0.94 | 0.99 | 0.92 |
Ball of foot circumference | 2.8 mm | 2.11 | 4.01 | 4.1 mm | 0.95 | 0.88 | 0.97 | 1.17 |
Ball width diagonal | 1.5 mm | 1.17 | 2.21 | 2.3 mm | 0.91 | 0.79 | 0.96 | 1.55 |
Max height over MTPJ | 0.3 mm | 0.23 | 0.43 | 0.5 mm | 0.99 | 0.98 | >0.99 | 0.76 |
Short heel circumference | 1.7 mm | 1.27 | 2.4 | 2.5 mm | 0.99 | 0.98 | >0.99 | 0.52 |
Hallux length | 0.1 mm | 0.1 | 0.17 | 0.2 mm | >0.99 | >0.99 | >0.99 | 0.19 |
Hallux valgus angle | 0.4° | 0.27 | 0.51 | 0.5° | >0.99 | >0.99 | >0.99 | 2.74 |
Between scan outputs from scanner (3 scans) | ||||||||
Foot length | 1.0 mm | 0.83 | 1.12 | 1.4 mm | 0.99 | 0.98 | 0.99 | 0.39 |
Foot width | 1.0 mm | 0.88 | 1.18 | 1.5 mm | 0.93 | 0.90 | 0.96 | 1.06 |
Arch height | 1.1 mm | 0.91 | 1.29 | 1.6 mm | 0.80 | 0.69 | 0.88 | 7.14 |
Arch length | 4.3 mm | 3.77 | 5.06 | 6.4 mm | 0.85 | 0.77 | 0.90 | 2.48 |
Shoe size | 0.2 | 0.14 | 0.19 | 0.2 | 0.98 | 0.96 | 0.99 | 2.79 |
Between scan outputs from MATLAB analysis (3 scans) | ||||||||
Foot length x | 0.9 mm | 0.77 | 1.09 | 1.3 mm | 0.99 | 0.99 | >0.99 | 0.37 |
Heel to MTPJ1 | 2.5 mm | 2.13 | 3.02 | 3.7 mm | 0.88 | 0.81 | 0.93 | 1.40 |
Heel to MTPJ5 | 2.2 mm | 1.90 | 2.69 | 3.3 mm | 0.91 | 0.84 | 0.95 | 1.38 |
Ball width diagonal | 0.8 mm | 0.68 | 0.96 | 1.2 mm | 0.97 | 0.95 | 0.98 | 0.84 |
Ball width straight | 0.5 mm | 0.45 | 0.63 | 0.8 mm | 0.98 | 0.97 | 0.99 | 0.57 |
Max height over MTPJ | 0.7 mm | 0.63 | 0.89 | 1.1 mm | 0.96 | 0.93 | 0.98 | 1.94 |
Height at 50% of foot length | 0.7 mm | 0.57 | 0.81 | 1.0 mm | 0.98 | 0.97 | 0.99 | 1.11 |
Circumference at 50% of foot length | 0.8 mm | 0.64 | 0.91 | 1.1 mm | 0.99 | 0.99 | >0.99 | 0.34 |
Ball of foot circumference | 1.5 mm | 1.25 | 1.78 | 2.2 mm | 0.98 | 0.96 | 0.99 | 0.65 |
Short heel circumference | 1.4 mm | 1.17 | 1.66 | 2.1 mm | 0.99 | 0.98 | 0.99 | 0.45 |
Heel breadth | 0.4 mm | 0.37 | 0.52 | 0.6 mm | 0.98 | 0.97 | 0.99 | 0.71 |
Hallux valgus angle | 3.3° | 2.79 | 3.96 | 4.9° | 0.90 | 0.84 | 0.94 | 32.73 |
MTPJ, metatarsophalangeal joint; MTPJ1, 1st metatarsophalangeal joint; MTPJ5, 5th metatarsophalangeal joint; CI, confidence interval; ICC, intraclass correlation; MDD, minimal detectable difference; CV, coefficient of variation.
Inter-rater reliability
Variable . | Typical error . | Lower CI . | Upper CI . | MDD . | ICC . | Lower CI . | Upper CI . | CV, % . |
---|---|---|---|---|---|---|---|---|
Inter-rater reliability from MATLAB analysis | ||||||||
Foot length x | 0.4 mm | 0.4 | 0.6 | 0.65 mm | >0.99 | >0.99 | >0.99 | 0.18 |
Ball foot width diagonal | 0.7 mm | 0.6 | 0.9 | 1.06 mm (0.45%) | 0.97 | 0.95 | 0.99 | 0.75 |
Ball foot width straight | 0.2 mm | 0.1 | 0.2 | 0.25 mm (0.47%) | >0.99 | >0.99 | >0.99 | 0.18 |
Heel breadth | 0.6 mm | 0.5 | 0.8 | 0.94 mm (0.40%) | 0.96 | 0.93 | 0.98 | 1.04 |
Heel to MTPJ1 | 4.0 mm | 3.3 | 5.1 | 6.02 mm (1.80%) | 0.71 | 0.52 | 0.83 | 2.27 |
Heel to MTPJ5 | 2.7 mm | 2.2 | 3.4 | 4.03 mm (1.73%) | 0.86 | 0.75 | 0.92 | 1.66 |
Max height over MTPJ | 0.7 mm | 0.6 | 0.9 | 1.03 mm (0.42%) | 0.96 | 0.93 | 0.98 | 1.82 |
Height at 50% of foot length | 1.0 mm | 0.8 | 1.3 | 1.51 mm (0.59%) | 0.97 | 0.94 | 0.98 | 1.67 |
Circumference at 50% of foot length | 0.8 mm | 0.7 | 1 | 1.21 mm (0.45%) | 0.99 | 0.99 | >0.99 | 0.36 |
Ball of foot circumference | 1.5 mm | 1.3 | 1.9 | 2.27 mm (0.97%) | 0.98 | 0.96 | 0.99 | 0.67 |
Short heel circumference | 1.2 mm | 1 | 1.5 | 1.77 mm (0.72%) | 0.99 | 0.99 | >0.99 | 0.38 |
Hallux valgus angle | 2.2° | 1.8 | 2.9 | 3.4° | 0.94 | 0.9 | 0.97 | 19.96 |
Variable . | Typical error . | Lower CI . | Upper CI . | MDD . | ICC . | Lower CI . | Upper CI . | CV, % . |
---|---|---|---|---|---|---|---|---|
Inter-rater reliability from MATLAB analysis | ||||||||
Foot length x | 0.4 mm | 0.4 | 0.6 | 0.65 mm | >0.99 | >0.99 | >0.99 | 0.18 |
Ball foot width diagonal | 0.7 mm | 0.6 | 0.9 | 1.06 mm (0.45%) | 0.97 | 0.95 | 0.99 | 0.75 |
Ball foot width straight | 0.2 mm | 0.1 | 0.2 | 0.25 mm (0.47%) | >0.99 | >0.99 | >0.99 | 0.18 |
Heel breadth | 0.6 mm | 0.5 | 0.8 | 0.94 mm (0.40%) | 0.96 | 0.93 | 0.98 | 1.04 |
Heel to MTPJ1 | 4.0 mm | 3.3 | 5.1 | 6.02 mm (1.80%) | 0.71 | 0.52 | 0.83 | 2.27 |
Heel to MTPJ5 | 2.7 mm | 2.2 | 3.4 | 4.03 mm (1.73%) | 0.86 | 0.75 | 0.92 | 1.66 |
Max height over MTPJ | 0.7 mm | 0.6 | 0.9 | 1.03 mm (0.42%) | 0.96 | 0.93 | 0.98 | 1.82 |
Height at 50% of foot length | 1.0 mm | 0.8 | 1.3 | 1.51 mm (0.59%) | 0.97 | 0.94 | 0.98 | 1.67 |
Circumference at 50% of foot length | 0.8 mm | 0.7 | 1 | 1.21 mm (0.45%) | 0.99 | 0.99 | >0.99 | 0.36 |
Ball of foot circumference | 1.5 mm | 1.3 | 1.9 | 2.27 mm (0.97%) | 0.98 | 0.96 | 0.99 | 0.67 |
Short heel circumference | 1.2 mm | 1 | 1.5 | 1.77 mm (0.72%) | 0.99 | 0.99 | >0.99 | 0.38 |
Hallux valgus angle | 2.2° | 1.8 | 2.9 | 3.4° | 0.94 | 0.9 | 0.97 | 19.96 |
MTPJ, metatarsophalangeal joint; MTPJ1, 1st metatarsophalangeal joint; MTPJ5, 5th metatarsophalangeal joint; CI, confidence interval; ICC, intraclass correlation; MDD, minimal detectable difference (normalised); CV, coefficient of variation.
Effect of Weight Bearing Condition
The only differences between weight-bearing condition (Table 5) that were greater than the MDD were foot width straight and foot width diagonal, which were around 1 mm greater in the 100% weight-bearing condition than 50% weight-bearing condition and short heel circumference which was 2.8 mm greater in the 100% weight-bearing condition than 50% weight-bearing condition. Relative to the 50% weight-bearing condition, the 100% weight-bearing condition consisted of more missing values and was at times of a lower scan quality due to the greater demand on balance for the participants and more movement during the scan. Therefore, only the 50% weight-bearing scans were used for further analysis. The 100% weight-bearing scan data are included in the repository for completeness in case of interest as reference values for footwear manufacturers.
Differences between 50% and 100% weight bearing
. | Mean 50% weight bearing . | SD 50% weight bearing . | Mean 100% weight bearing . | SD 100% weight bearing . | Average difference . | CIs/(IQR) . | p value . | Effect size (d, CIs)/r . |
---|---|---|---|---|---|---|---|---|
Foot length, mm (n = 347) | 246 | 11 | 246 | 11 | 0 | (−0.34 to 0.11) | 0.34 | −0.05 (−0.16 to 0.05) |
Foot width, mm (n = 347) | 97 | 5 | 98 | 6 | 1 | (1.09 to 1.46) | <0.001 | 0.73 (0.62 to 0.85) |
Arch height, mm (n = 275) | 16 | 3 | 16 | 3 | 0 | (−0.70 to 0.25) | 0.35 | −0.06 (−0.18 to 0.06) |
Arch length, mm (n = 344) | 175 | 8 | 175 | 9 | 0 | (−0.79 to 0.32) | 0.42 | −0.04 (−0.15 to 0.06) |
Foot length x, mm (n = 347) | 246.1 | 10.7 | 245.9 | 10.8 | −0.2 | (−0.38 to −0.07) | 0.004 | −0.16 (−0.26 to −0.05) |
Heel to MTPJ1, mm (n = 347) | 179.4 | 8.3 | 179.5 | 8.3 | 0.1 | (−0.25 to 0.46) | 0.55 | 0.03 (−0.07 to 0.14) |
Heel to MTPJ5, mm (n = 347) | 161.1 | 7.6 | 161.0 | 7.5 | −0.1 | (−0.47 to 0.31) | 0.68 | 0.02 (−0.13 to −0.08) |
Ball foot width diagonal, mm (n = 347) | 98.7 | 5.6 | 99.8 | 5.6 | 1.1 | (0.94 to 1.22) | <0.001 | 0.82 (0.69–0.94) |
Ball foot width straight, mm (n = 347) | 95.5 | 5.3 | 96.5 | 5.4 | 1.1 | (0.97 to 1.2) | <0.001 | 1.1 (0.97 to 1.2) |
Max height over MTPJ, mm (n = 347) | 40.4 | 3.3 | 40.3 | 3.4 | −0.1 | (−0.26 to 0.01) | 0.08 | −0.1 (−0.2 to 0.01) |
Height at 50% of foot length, mm (n = 347) | 64.1 | 5.0 | 64.1 | 4.4 | 0.1 | |||
Median (IQR) | 63.9 | 6.5 | 64.2 | 5.8 | 0.3 | (1.8) | <0.001 | 0.20 |
Circumference at 50% of foot length, mm (n = 347) | 234.1 | 11.7 | 234.4 | 11.5 | 0.3 | |||
Median (IQR) | 233.6 | 16.4 | 233.7 | 16.7 | 0.5 | (1.9) | <0.001 | 0.3 |
Ball of foot circumference, mm (n = 346) | 238.4 | 12.5 | 240.1 | 12.6 | 1.7 | (1.39 to 1.97) | <0.001 | 0.61 (0.50 to 0.73) |
Short heel circumference, mm (n = 347) | 317.7 | 14.7 | 320.6 | 14.7 | 3.0 | (2.69 to 3.22) | <0.001 | 1.18 (1.0 to 1.3) |
Heel breadth, mm (n = 347) | 62.7 | 4.0 | 63.1 | 3.8 | 2.8 | (0.27 to 0.63) | <0.001 | 0.27 (0.16 to 0.38) |
Hallux valgus angle, ° (n = 347) | 9.7 | 8.6 | 9.1 | 8.7 | −0.5 | |||
Median (IQR) | 9.6 | 14.8 | 9.1 | 14 | 0 | (1.5) | <0.001 | 0.2 |
. | Mean 50% weight bearing . | SD 50% weight bearing . | Mean 100% weight bearing . | SD 100% weight bearing . | Average difference . | CIs/(IQR) . | p value . | Effect size (d, CIs)/r . |
---|---|---|---|---|---|---|---|---|
Foot length, mm (n = 347) | 246 | 11 | 246 | 11 | 0 | (−0.34 to 0.11) | 0.34 | −0.05 (−0.16 to 0.05) |
Foot width, mm (n = 347) | 97 | 5 | 98 | 6 | 1 | (1.09 to 1.46) | <0.001 | 0.73 (0.62 to 0.85) |
Arch height, mm (n = 275) | 16 | 3 | 16 | 3 | 0 | (−0.70 to 0.25) | 0.35 | −0.06 (−0.18 to 0.06) |
Arch length, mm (n = 344) | 175 | 8 | 175 | 9 | 0 | (−0.79 to 0.32) | 0.42 | −0.04 (−0.15 to 0.06) |
Foot length x, mm (n = 347) | 246.1 | 10.7 | 245.9 | 10.8 | −0.2 | (−0.38 to −0.07) | 0.004 | −0.16 (−0.26 to −0.05) |
Heel to MTPJ1, mm (n = 347) | 179.4 | 8.3 | 179.5 | 8.3 | 0.1 | (−0.25 to 0.46) | 0.55 | 0.03 (−0.07 to 0.14) |
Heel to MTPJ5, mm (n = 347) | 161.1 | 7.6 | 161.0 | 7.5 | −0.1 | (−0.47 to 0.31) | 0.68 | 0.02 (−0.13 to −0.08) |
Ball foot width diagonal, mm (n = 347) | 98.7 | 5.6 | 99.8 | 5.6 | 1.1 | (0.94 to 1.22) | <0.001 | 0.82 (0.69–0.94) |
Ball foot width straight, mm (n = 347) | 95.5 | 5.3 | 96.5 | 5.4 | 1.1 | (0.97 to 1.2) | <0.001 | 1.1 (0.97 to 1.2) |
Max height over MTPJ, mm (n = 347) | 40.4 | 3.3 | 40.3 | 3.4 | −0.1 | (−0.26 to 0.01) | 0.08 | −0.1 (−0.2 to 0.01) |
Height at 50% of foot length, mm (n = 347) | 64.1 | 5.0 | 64.1 | 4.4 | 0.1 | |||
Median (IQR) | 63.9 | 6.5 | 64.2 | 5.8 | 0.3 | (1.8) | <0.001 | 0.20 |
Circumference at 50% of foot length, mm (n = 347) | 234.1 | 11.7 | 234.4 | 11.5 | 0.3 | |||
Median (IQR) | 233.6 | 16.4 | 233.7 | 16.7 | 0.5 | (1.9) | <0.001 | 0.3 |
Ball of foot circumference, mm (n = 346) | 238.4 | 12.5 | 240.1 | 12.6 | 1.7 | (1.39 to 1.97) | <0.001 | 0.61 (0.50 to 0.73) |
Short heel circumference, mm (n = 347) | 317.7 | 14.7 | 320.6 | 14.7 | 3.0 | (2.69 to 3.22) | <0.001 | 1.18 (1.0 to 1.3) |
Heel breadth, mm (n = 347) | 62.7 | 4.0 | 63.1 | 3.8 | 2.8 | (0.27 to 0.63) | <0.001 | 0.27 (0.16 to 0.38) |
Hallux valgus angle, ° (n = 347) | 9.7 | 8.6 | 9.1 | 8.7 | −0.5 | |||
Median (IQR) | 9.6 | 14.8 | 9.1 | 14 | 0 | (1.5) | <0.001 | 0.2 |
Values in bold represent difference between conditions greater than MDD. Differences were calculated as 100% weight bearing minus 50% weight bearing. Mean/median, standard deviation, and interquartile range values are reported to the number of decimal places of the measurement output.
MTPJ, metatarsophalangeal joint; MTPJ1, 1st metatarsophalangeal joint; MTPJ5, 5th metatarsophalangeal joint; IQR, interquartile range.
Effect of Age on Anatomical Measures (50% Weight Bearing)
There was a significant effect of age on normalised foot width from the scanner output (p = 0.007, η2 = 0.042 [0.004–0.077]), ball width diagonal (p = 0.002, η2 = 0.051 [0.008–0.089]), foot width straight (p = 0.001, η2 = 0.052 [0.009–0.091]), and heel breadth (p = 0.0007, η2 = 0.056 [0.011–0.095]). The greater mean widths of the groups 40 years and over compared to the 18–29-year group were greater than the MDD. There was also a significant effect of age on ball of foot circumference and short heel circumference (p = 0.005, η2 = 0.045 [0.005–0.081] and p = 0.003, η2 = 0.047 [0.006–0.084], respectively). Again, the greater values of groups 40 years and over compared to the 18–29-year group were greater than the MDD. There was a significant effect of age on height at 50% of foot length (p = 0.0001, η2 = 0.055) and circumference at 50% of foot length (p = 0.0007, η2 = 0.044), calculated with the Kruskal-Wallis test due to unequal variances. For these measures, the normalised values of the middle two age groups were greater than the oldest two age groups, with a difference greater than the MDD. There was no effect of age on heel-to-MTPJ1 distance (p = 0.606, η2 = 0.010 (0.001–0.024) or heel-to-MTPJ5 distance (p = 0.246, η2 = 0.018 (<0.001–0.040). Mean anatomical variables normalised to foot length are displayed in Figure 2a, b, denoting significant pairwise comparisons. As there is not a proportional increase in foot width and foot length [17], absolute foot length by ball width diagonal and foot width straight are represented in Figure 3a, b.
a, b Mean anatomical measures normalised to foot length for each age group. * represents significant difference from 18 to 29 years group (p < 0.05); $ represents significant difference from 60 to 69 years group (p < 0.05); and # represents significant difference from 70 to 79 years group (p < 0.05).
a, b Mean anatomical measures normalised to foot length for each age group. * represents significant difference from 18 to 29 years group (p < 0.05); $ represents significant difference from 60 to 69 years group (p < 0.05); and # represents significant difference from 70 to 79 years group (p < 0.05).
Absolute foot length versus right ball width diagonal (a) and right ball width straight (b). Grey circles represent individual data points for all data.
Absolute foot length versus right ball width diagonal (a) and right ball width straight (b). Grey circles represent individual data points for all data.
Hallux valgus angle per age group is presented in Figure 4. There was a significant effect of age group on hallux valgus angle (p < 0.0001), identified using the Kruskal-Wallis test. The median of the 60–69 year-group (12.6°) was significantly greater than the median of the 18–29-year group (8.9°, p = 0.025, η2 = 0.17), the 30–39-year group (7.2°, p < 0.001, η2 = 0.24) and the 40–49-year group (0°, p < 0.001, η2 = 0.28). The median of the 70–79-year group (10.8°) was significantly greater than the median of the 30–39-year group (7.2°, p = 0.039, η2 = 0.21) and the median of the 40–49-year group (0°, p = 0.004, η2 = 0.24). The median hallux valgus deformity grading was “no deformity” for each group.
Discussion
This study is the first to provide data on the foot dimensions and hallux valgus deformity in adult women active in court sports across different ages. In partial support of our primary hypothesis, measurements of foot width and ball of foot circumference were greater in the older age groups compared to the youngest age group. While we are limited in the inferences that can be made from this cross-sectional study, the findings are in agreement with previous cross-sectional work from the Netherlands and Hong Kong, which found increases in foot width and girth of women’s feet with increased age [4, 5]. In contrast, one study reported greater absolute foot width in younger Japanese adults, than middle aged and older adults; however, this could be attributed to the younger group having larger feet (greater mean foot length) [45]. An increase in heel breadth and short heel circumference with age was found in the present study similar to an increase in absolute short heel circumference with age also reported previously [4]. The lack of a difference in foot width and circumference between the youngest group and oldest group was unexpected. This may be contributed to by the relatively low sample size in the 70–79-year group (n = 40), which was the lowest of the groups, resulting in insufficient statistical power to detect significant effects. The existence of fewer over 70-year-olds active in court sports compared to those over 50 s and 60 s [60] may also have contributed to this observation since those recruited are likely to represent particularly healthy/resilient individuals who have been able to keep playing sports longer than their peers. It is also possible that there is a real lack of difference in these variables between the oldest and youngest groups, as has been reported previously [4].
There was some evidence to support our hypothesis that arch height would decrease with age. While the arch height measured from the scanner was unreliable, we used the dorsum height at 50% of foot length as a surrogate measure, previously used to reliably quantify arch height index [56]. This measure also defines instep height, and older adults with flat feet have been shown to have significantly lower instep height than older adults without flat feet [1]. The observations in the present study of lower foot height and circumference at 50% foot length in the 60–69-year group and 70–79-year group compared to the 40–49-year group and lower height, and circumference at 50% of foot length in the 70–79-year group compared to the 50–59-year group, are in line with previous reports of a gradual lowering of the medial longitudinal arch from middle age [61]. It has been suggested that a gradual lowering of the medial longitudinal arch may be due to the weakening and lengthening of the tibialis posterior with age [3]. In addition, changes in the function of intrinsic foot muscles which also contribute to dynamic support of the medial longitudinal arch [37] may contribute to the observed lower arch. This lower arch may have implications for foot pain and injuries associated with flatter feet [7]. The observation of lower arch, together with evidence that ability for footwear to accommodate an orthotic device has been identified as important to 44% of older adults engaged in court sports [16], suggests that manufacturers of athletic shoes for this population would be prudent to consider incorporation of a removable insole. This would allow the possible addition of arch support in shoes, should the wearer choose to add external support in combination with or instead of performing strengthening exercises.
The current study has provided partial support for an increase in the hallux valgus deformity with age. There was a significant effect of age on the hallux valgus angle; however, a substantial number of measurements had to be converted from negative to zero and reliability was lower than other measures suggesting the measurement method was sub-optimal and results should, therefore, be interpreted with caution. Future work could consider calculating the hallux valgus angle from a 2D image, for instance, taken from beneath the sole of the foot using the scanner from this study. Although there was no difference in hallux valgus deformity grading with age, the reliability between assessors was only categorised as “fair.” This grading scale with stronger reliability was developed with podiatrists with between 9 and 19 years of experience [48], whereas the authors within this study are not trained podiatrists. Nonetheless, from Figure 4, it can be seen that more participants in the groups 50 and over than those under 50 had a hallux valgus angle of over 20°, which would constitute a moderate deformity according to some definitions [62], even allowing for differences in sample sizes between the groups. Not only does the presence of hallux valgus affect foot width but it can also mean that footwear is uncomfortable, particularly when there is stitching or reinforcement around the bunion area [15, 16], which footwear designers should also be mindful of.
A further purpose of this study was to provide reference data which can inform future athletic shoe design. These data, provided in the repository as both normalised values and absolute values for shoe design, will be beneficial for use by footwear companies, enhancing inclusive footwear design. Given there is not a proportional increase in foot width and foot length [17], normalising measures to foot length is an imperfect means of comparing between age groups. From Figure 1, it can be seen that not all age groups consisted of participants with shoe sizes at the extremes. However, from Figure 3, it can be seen that the trendline for absolute foot width versus length represents lower foot widths for the 18–29-year group than the other age groups, which is in line with our normalised results. Also from the individual data points in Figure 3, a large spread in values of foot length versus widths can be seen, supporting the argument by Jurca and colleagues [17] that providing shoes in at least three different widths per size would be necessary to account for the variability in the population, which is rarely the case in the current athletic shoe market. The repository also contains both the 50% and 100% weight-bearing conditions, although we cannot guarantee exactly 50% of the load as we were not able to record ground reaction force. The increase in foot width with load was expected, where the increase was of a lesser magnitude than one study [35] and a greater magnitude than another [36]. The greater short heel circumference observed in the 100% weight-bearing condition than in the 50% weight-bearing condition may be a consequence of postural change as participants may have been leaning forward more in the 100% weight-bearing condition than in 50% weight-bearing condition to assist with balance. A comparison of age groups in the 100% weight-bearing condition was inhibited by the lower data quality and missing data in that condition.
The observation that there was no effect of age on distance from the heel to MTPJ1 or MTPJ5 suggests that age does not affect the location of the forefoot bending axis, an important characteristic utilised in footwear design. While no difference in these measures was expected with age and is consistent with previous literature [5], these were also some of the least reliable measures. If measurement of MTPJ locations is required in future work, the placement of markers on the feet prior to scanning as per some previous studies [1, 4] could be used to facilitate identification of the joint axis. Similarly, a marker on the navicular could allow for the calculation of navicular height [1] as another measure of medial longitudinal arch height. The use of skin markers in this way would be beneficial for the determination of anatomical locations but would introduce additional data collection time, where priority was given to recruitment of participant numbers in data collection for much of the current study, taking place at competitive tournaments.
A further limitation of this study was that in the interest of time, body mass was not recorded which can affect variables such as foot width [43]; however, our results support the findings of previous work in the general population accounting for body mass [4]. Future studies would benefit from accounting for body mass index and other participant characteristics such as foot dominance if possible. Our findings are from a heterogeneous, female, largely Caucasian sample, which may not be generalisable to all.
In conclusion, this study provides novel data for older adults who play court sports, by quantifying differences in foot width, foot circumference, hallux valgus deformity and arch height index/instep height with age. Greater values for measures of foot width, ball of foot circumference and short heel circumference were observed for the older groups compared with the younger. In addition, lower foot dorsum height and circumference at 50% foot length were detected for the oldest age groups compared to the middle age groups. There was also some evidence of increased hallux valgus deformity with age. It is suggested that the increased awareness of age-related differences in foot morphology and accompanying data for this population can inform athletic footwear design so that more inclusive athletic footwear designs can be developed.
Acknowledgements
The authors would like to thank Jenny Anderson (PhD) for sharing MATLAB code which was adapted for the analysis of the 3D foot scans in this study. We would like to thank our industry partners IDA Sports for their assistance in data collection and access to survey data.
Statement of Ethics
This study protocol was reviewed and approved by the Faculty of Health and Life Sciences, Sport and Health Sciences Ethics Committee, Approval No. 519475. Written informed consent was obtained from participants to participate in the study.
Conflict of Interest Statement
The authors have no conflicts of interest to declare.
Funding Sources
This study was performed as part of a project funded by UKRI Innovate UK (10025932). The funder had no role in the design, data collection, data analysis, and reporting of this study.
Author Contributions
J.R. contributed to the design of the study, collected and analysed data, contributed to interpretation of data, and wrote the draft of the manuscript. S.D. contributed to the design of the study; collection, analysis, and interpretation of the data, and writing the draft of the manuscript. R.B. contributed to collection, analysis, and interpretation of the data and critically reviewed the manuscript.
Data Availability Statement
The data that support the findings of this study are openly available in ReShare (https://reshare.ukdataservice.ac.uk/). Further enquiries can be directed to the corresponding author.