Abstract
For people with diabetes, a good-fitting shoe is essential for reducing the risk of ulcers and eventual amputation. However, there is a lack of 3D data explaining differences in foot shape between people with and without diabetes and how individual factors might influence these differences; these data are vital for adequate shoe design and prevention of diabetic foot ulcers. This study quantifies the differences in external foot shapes of people with and without diabetes and peripheral neuropathy and examines which factors might affect these variations. One-hundred thirty-six foot scans of older adults with and without diabetes and peripheral neuropathy were used to create and assess foot shape models against demographic and health factors. Principal component analysis (PCA) showed that the feet of people with diabetes and with neuropathy are not necessarily clustered into a particular foot shape but have more pronounced features in specific foot variations (e.g., ankle width, arch height, hallux abduction, and edema vs. atrophic feet) compared to people without diabetes and neuropathy. The mean pairwise distance (intrascore spread) in PC1 and PC2 for individuals with diabetes and neuropathy was 43% larger than for those without diabetes and neuropathy and 24% larger than for those with neuropathy but no diabetes. Partial least squares regression (PLSR) showed potential predicting the presence of diabetes and neuropathy; however, additional data are required to support the trend. Analyses, such as PCA and PLSR, could be useful for determining how to quantify these changes to design more appropriate footwear for these populations.