Background
Footwear plays an important role in protecting the foot from extremes of temperature, moisture and mechanical trauma. However, since the development and widespread popularity of fashion footwear in the 17
th century, the functional aspect of footwear has largely been supplanted by the requirements of fashion. It has been suggested that shoe selection may be primarily based on aesthetic considerations, many of which are incompatible with the optimal function of the foot [
1]. This is of particular concern in older people, as studies have shown that between 26 and 50% of older people wear shoes that are either too short or too narrow [
2‐
5]. Several factors may be responsible for this, including fashion influences (particularly in older women [
6,
7]), not measuring foot dimensions when purchasing shoes [
8], or the limited availability of commercially-available, low-cost footwear that adequately caters for the broader older foot [
9,
10].
There is growing evidence of an association between wearing shoes that are too small and the development of foot problems in older people. In a sample of 176 retirement village residents aged 62 to 96 years, Menz and Morris [
11] found that wearing shoes substantially narrower than the foot was associated with corns on the toes, hallux valgus deformity and foot pain, whereas wearing shoes shorter than the foot was associated with lesser toe deformity. Similarly, in an ambulatory population of 213 people aged 60 to 80 years, Chaiwanichsiri et al. [
3] reported that those who wore shoes that were too narrow were twice as likely to report foot pain. Most recently, a survey of 399 community-dwelling people aged over 60 years revealed that 61% of women and 30% of men reported foot pain when wearing shoes (most commonly in the forefoot and toes), and that women with foot pain exhibited a broader forefoot than those without pain [
12].
Given the high prevalence of ill-fitting footwear in older people and its relationship to foot problems, accurate fitting of shoes in this age-group is an important consideration. Therefore, as part of a larger randomised controlled trial assessing the effectiveness of off-the-shelf footwear in reducing foot pain in Australian Department of Veterans’ Affairs (DVA) recipients not eligible for medical grade footwear [
13], this study was undertaken to evaluate whether the fitting protocol employed in the trial resulted in the appropriate selection of shoe sizes. To do this, we compared the last dimensions of the shoes allocated to each participant based on an assessment using the Brannock device® [
14] to corresponding foot dimensions obtained with a high resolution 3D foot scanner, and documented participants’ perceptions of the fit and comfort of the shoes.
Discussion
The aim of this study was to evaluate the accuracy of shoe fitting in older people who were allocated to the intervention group of a randomised controlled trial assessing the effectiveness of off-the-shelf footwear in reducing foot pain [
13]. To do this, we compared the last dimensions of the shoes allocated to each participant to foot dimensions obtained with a high resolution 3D foot scanner. The findings suggest that the fitting protocol used in the study resulted in the selection of appropriately fitting shoes, as evidenced by the shoes being slightly longer than the foot and having equivalent ball width and ball girth measurements. Participants’ overall perceptions of shoe fit and comfort were also very high, providing subjective confirmation of the objective measurements.
Our fitting procedure used a combination of standardised measurement using the Brannock device® and participant feedback to determine the final shoe size allocation. Interestingly, just over half of the sample (55%) was allocated a shoe size that differed to the Brannock device®-determined size. The most common reason for this discrepancy was having a difference between right and left foot sizes, necessitating the selection of a shoe to fit the larger foot. This is consistent with a recent study which reported that just under half of the population has a difference of at least half a US shoe size between left and right feet [
4]. However, several other factors contributed to the allocation of a different sized shoe, such as participant preference, accommodation for orthoses and expected foot swelling, suggesting that foot dimensions alone cannot be used to select the optimum shoe size. Indeed, Nácher et al. [
27] have shown that a detailed statistical model incorporating 14 anthropometric foot variables was only 66% accurate in identifying the preferred shoe size. Taken together, these findings support the view that fitting shoes is both an art and a science [
28] and that optimum shoe size selection needs to take into account a range of factors specific to the individual in addition to accurate measurement of foot dimensions.
On average, the allocated shoes were 23.6 mm longer than the corresponding foot length measurements obtained with the 3D scanner. Although it is widely accepted that shoes need to be slightly longer than the foot to allow for elongation when standing and walking, the recommended distance for the gap between the longest toe and the end of the shoe is essentially arbitrary, varying between 10 to 20 mm (or a ‘thumb’s width’) in the literature [
9,
22,
24,
29‐
32]. The 23.6 mm gap in our study is slightly larger than the upper limit of these recommendations, which suggests that participants may have been allocated shoes that were relatively long. There are three main reasons for this. First, our fitting criteria required that the larger size be allocated to participants with discrepant foot lengths, as no split-sizes were possible within the practical constraints of the study. Second, all participants needed to accommodate some form of insole or orthosis with a contoured heel cup. Third, we allowed for some degree of participant preference in shoe allocation, and five participants requested a shoe size longer than the size determined by the Brannock device®. We consider it unlikely that the slightly longer fitting of the allocated shoes in this study would be detrimental, as shoe discomfort [
33] and footwear-related foot disorders [
11] are most often related to wearing shoes that are too short rather than too long, and participants reported the shoes to be well fitting and comfortable. However, the 16 week follow-up planned for this study will help to identify whether the fit of the shoes is optimal after longer periods of wear.
There were no significant differences between foot dimensions and allocated shoe dimensions in relation to ball width or ball girth, although there was a trend towards the shoes having slightly larger ball width (mean difference 1.4 mm,
p = 0.066). This finding needs to be considered in the context of foot deformation when standing and walking, and diurnal variation in foot volume. It has previously been shown that ball width increases by approximately 4% from non-weight bearing to full weight bearing [
34‐
36], and by an additional 3% when walking compared to standing [
37]. Furthermore, in healthy older adults, it has been shown that overall foot volume increases by 1.4% after ten minutes of walking [
38] and it is likely that older people with venous insufficiency would demonstrate even greater increases in foot volume during the course of a day due to accumulation of oedema [
10]. These findings suggest that optimum fitting of shoes requires some allowance for soft tissue expansion, particularly in the forefoot. In our study, the 3D foot scans were taken in a fully weight bearing position, so any soft tissue expansion that occurred from sitting to standing was accounted for. Therefore, only the dimensional changes related to foot loading during gait and oedema would need to be accommodated by the deformation of the upper of the shoe. Given that the shoes used in our trial have a highly pliable Lycra® (elastane) upper, we are confident that the close fit between foot and shoe dimensions in the forefoot represents appropriate fitting. However, this may not have been the case had we selected shoes with less pliable upper materials such as leather.
The findings of this study need to be considered in the context of several limitations. First, we were limited to three variables to characterise the accuracy of shoe fitting. Although several other dimensions can be extracted from 3D foot scans [
39], we were only able to obtain length, ball width and ball girth measurements of the shoe lasts from the manufacturer for comparison. These are important parameters, however we acknowledge that a wider array of measurements, such as heel width, instep height, ankle girth and relative lengths of the toes would have provided a more detailed insight into the fit of the shoes [
10,
40,
41]. Second, our assessment of participants’ perceptions of fit and comfort related to the shoe as a whole, and did not delineate the relative fit and comfort of the rearfoot, midfoot and forefoot. Although perceived fit at different foot locations are significantly correlated to each other [
32], they are likely to have different implications for the optimum shape and dimensions of the shoe. Third, participants’ reporting of the fit and comfort of the shoes may have been influenced by Rosenthal effects. Fourth, participants in this study were recruited from a veterans’ affairs database, and were required to be ineligible for medical grade footwear to be included. It is therefore possible that older veterans with more pronounced foot deformity were excluded, skewing our sample towards those with more ‘normal’ feet. However, the foot dimensions of participants in our study were very similar to those reported in 668 older people (mean age 64 years) by Chantelau and Gede [
9] and 312 community-dwelling older people (mean age 71 years) reported by Mickle et al. [
10]. Furthermore, the prevalence of foot disorders in our sample was similar to previous population-based studies of older people [
42‐
44]. This suggests that our findings may be broadly generalisable to the older population, with the exception of those with marked foot deformity who are unable to be accommodated in regular off-the-shelf footwear. Finally, these findings may not be generalisable to other footwear brands, as (i) we used a Brannock device® marked with sizings specific to the Dr Comfort® range, and (ii) the Brannock device® only measures linear ball
width, so it cannot be assumed that the appropriate fitting of ball
girth obtained in this study would necessarily translate to other shoes, as the relationship between ball width and girth is not standardised.
In conclusion, this study has shown that shoe size selection in older people using the Brannock device® combined with participant feedback resulted in the allocation of Dr Comfort® shoes with last dimensions that were well matched to the dimensions of the foot determined by a high resolution 3D foot scanner. There are two main implications of these findings. Firstly, in the context of the randomised controlled trial, we can be confident that the protocol used resulted in the provision of appropriately-fitting shoes to the intervention group. However, the longer term follow-up of these participants will assist in determining whether this approach is effective at reducing foot pain. Secondly, in the broader context of clinical practice, our findings suggest that the Brannock device® is a useful clinical tool, but optimum shoe size selection in this age-group may need to take into account a range of factors specific to the individual in addition to accurate measurement of foot dimensions.
Competing interests
HBM and SEM are Editor-in-Chief and Deputy Editor-in-Chief, respectively, of Journal of Foot and Ankle Research. It is journal policy that editors are removed from the peer review and editorial decision-making processes for papers they have co-authored. The remaining authors declare that they have no competing interests.
Authors’ contributions
HBM and NF conceived the idea and obtained funding for the study. HBM, NF and SEM designed the study protocol. MA and SR collected and entered the data. HBM conducted the statistical analysis and drafted the manuscript. All authors assisted with the writing of the manuscript, and read and approved the final manuscript.