| Author: | Zhang, Liying |
| Title: | Ergonomic design of diabetic insoles and analysis of plantar pressure based on machine learning method |
| Advisors: | Yick, Kit-lun (SFT) Yip, Joanne (SFT) |
| Degree: | Ph.D. |
| Year: | 2024 |
| Subject: | Footwear -- Design Pressure -- Measurement Foot -- Ulcers Diabetes -- Complications -- Prevention Hong Kong Polytechnic University -- Dissertations |
| Department: | School of Fashion and Textiles |
| Pages: | xxii, 189 pages : color illustrations |
| Language: | English |
| Abstract: | Diabetic foot ulceration is a major complication of diabetes mellitus, and it occurs in approximately 15% of diabetic patients and 1% end up with an amputation if not well managed. These will significantly affect the life quality of diabetics and result in public health burden. Footwear/insole, when properly designed, is a cost-effective way of reducing high plantar pressure which would then redistributing the abnormal plantar pressure to prevent foot ulcers. The shape of the foot provides important information for footwear/insole fit and wear comfort. Therefore, an accurate and reliable method to obtain the foot contours is paramount for orthopaedic footwear/insole design to provide optimal fit and foot protection. However, scientific knowledge on the surface deformation of foot shape at different stances especially in dynamic situations that enable precise support and foot protection, as well as redistribution of plantar pressure for diabetic patients is minimal. The primary goal of this study is to design, with reference to foot scanning experiments and foot biomechanical analysis based on machine learning method, an insole with enhanced fit and offloading performance for diabetics to prevent the development of diabetic foot ulcers. To achieve the research objective, a foot care programme including face-to-face questionnaire, foot assessment and 3D foot scanning was firstly conducted to understand the features and needs of footwear/insole as well as foot morphologies of a total of 48 diabetic subjects with Type 1 or Type 2 diabetes mellitus. The results suggest that the diabetics prefer sports shoe with a wider toe box. Most of the diabetics show low adherence to orthopaedic insole, more than half of them use flat insole with no specific functions in their daily activity, which may further cause foot health problems like arch collapse, and foot ulceration caused by abnormal plantar pressure. To investigate the relationship between offloading performance and insole material properties, 4 different insole materials were used. The performance of the designed 3D insoles on the whole and regional plantar pressure offloading effect during walking and subjective perception of wear comfort with 4 different insole materials were examined through a laboratory wear trial of 40 diabetic patients by using in-shoe Pedar® system. In addition, the relationship between offloading performance and insole material property was also qualitatively analysed. The results suggest that the pressure of whole plantar was reduced, especially the insoles designed with PORON® Medical 4708 and Nora Lunalastik EVA. 3D structure of the insole increases the contact area of the foot-insole interface with arch support and transfer the loads from high pressure regions to the support region. Therefore, the pressure of MTHs (metatarsal heads) and heel regions was reduced, while the pressure of midfoot was increased. Amongst the 4 insole materials, PORON® Medical 4708 and Nora Lunalastik EVA show better pressure redistribution capacity because they are softer with better energy absorption and perceived comfort. Insole materials should be chosen carefully to achieve optimal plantar pressure redistribution and wear comfort. To obtain the plantar pressure at different situations more efficiently, in view of the qualitative association found between ink footprint images and plantar pressure distribution, a novel approach was proposed to use the footprints to predict the dynamic plantar pressure in different foot/insole conditions. Multilayer perceptron neural network model was introduced to extract the ink density information from the footprint images. By integrating the location information of each sensor on the Pedar insole with its corresponding footprint image block, the quantitative pressure value of each sensor in different foot scenarios can be predicted. A total of 52 diabetics were invited to collect the data for model training through a laboratory wear trial experiment. The results show that the mean absolute error of the prediction models for the barefoot and 4 insole conditions (PORON® Medical 4708, Pe-Lite, Nora Lunalight A fresh, and Nora Lunalastik EVA) is 5.51% (33.1 kPa), 3.99% (23.9 kPa), 4.85% (29.1 kPa), 4.25% (25.5 kPa) and 3.57% (21.4 kPa) respectively. The plantar of the foot is divided into 5 regions to provide more precise insole material recommendations for each plantar region of the foot. The findings help clinicians confidently and precisely prescribe orthotic insoles for each diabetic patient, modify and replace the insole material for each foot region, which greatly reduces the operating time and cost to evaluate insole offloading performance. To further improve the insole fit during motion, a novel 4D foot scanning system was introduced to capture the foot geometry of 19 diabetic patients in half weight bearing standing and self-selected walking speed conditions (Zhang et al., 2023). To speed up the scanning image processing, point cloud registration for image reorientation and algorithms to automatically extract key points and foot measurements related to footwear/insole design are developed. Thirteen key foot anthropometric measurements of five representative frames of scanned foot images were extracted for an in-depth analysis of the foot deformations during the complete roll-over process. The results show that maximum deformation of length and girth dimensions are found at first toe contact. Width dimensions have maximum deformation at heel take off during the foot roll-over process. The findings provide detailed practical recommendations on design features and tolerance for improving diabetic footwear/insole fit and wear comfort during motion. The insole prototype was modified using silicone mould technology. By applying metatarsal pad, arch support and heel cup on a porous structure, the insole fit and wear comfort can be improved during motion. With selecting suitable material for different foot regions determined from AI prediction models, the risk for diabetic foot ulcers development caused by abnormal plantar pressure can be reduced. A laboratory wear trial of 23 diabetic patients was carried out to confirm the offloading effect of these modifications on insole prototype by using the in-shoe Pedar® system. Experimental results confirmed that as compared to the commercially available insole and insole prototype, the proposed diabetic insole can better redistribute the plantar pressure during dynamic walking with satisfactory subjective wear comfort. This study enhanced our knowledge on the foot geometry and changes of foot anthropometric measurements of the diabetics at static and dynamic situations, as well as their needs and preferences of footwear/insole. These findings on foot shape deformations provide important information for diabetic footwear/insole geometry design to improve the fit and wear comfort for the diabetics at different situations. The AI prediction models help to predict the plantar pressure distribution at 4 different insole situations, which can provide a good idea for material selection and evaluation for clinicians during the customised insole design process with involving more insole structure and materials in the future research. The AI prediction model provides reference for clinicians to recommend insole material from these 4 materials for each foot region with enhanced offloading effect, thus reducing the risk for development of diabetic foot ulcerations. Compared to traditional insole fabrication process, this study not only provides more foot geometrical information for diabetic insole design, but also proposes a more efficient and reliable approach for insole performance evaluation. Additionally, silicone mould method also reduces the cost for insole fabrication compared to traditional insole fabrication methods. |
| Rights: | All rights reserved |
| Access: | open access |
Copyright Undertaking
As a bona fide Library user, I declare that:
- I will abide by the rules and legal ordinances governing copyright regarding the use of the Database.
- I will use the Database for the purpose of my research or private study only and not for circulation or further reproduction or any other purpose.
- I agree to indemnify and hold the University harmless from and against any loss, damage, cost, liability or expenses arising from copyright infringement or unauthorized usage.
By downloading any item(s) listed above, you acknowledge that you have read and understood the copyright undertaking as stated above, and agree to be bound by all of its terms.
Please use this identifier to cite or link to this item:
https://theses.lib.polyu.edu.hk/handle/200/13057