Abstract
[Purpose] According to the Orthotics (4th ed), positive models fabricated using the conventional impression foam-based modeling method require multiple modifications, including reductions at the waist girth, lateral aspect of the calcaneus, medial longitudinal arch, lateral longitudinal arch, and metatarsal arch as well as additions to account for soft tissue deformation around the dorsum of the foot and the calcaneus. Achieving accurate fabrication therefore requires considerable working time and technical proficiency. [Participants and Methods] In this study, we developed the arch support molding method (ASMM), which enables acquisition of a post-modification model by having the participants step on a dedicated arch support during modeling, and evaluated the usefulness of this method. Both feet of 10 healthy participants (5 men and 5 women) were modeled using both the ASMM and the conventional method, and the time required for each procedure and the resulting shapes were compared. [Results] The results showed that ASMM required less working time compared with the conventional method, and the obtained shapes were closer to those of the post-modified positive models. [Conclusion] These findings indicate the high utility of ASMM.
Key words: Foot orthosis, Modeling, 3D CAD
INTRODUCTION
A wide variety of modeling methods are used in prosthetics and orthotics. Although the most widely used method is plaster bandage-based modeling, various other materials and techniques have also been employed. Even among plaster bandage-based methods, multiple variations exist depending on the modeling site and the type of devices to be fabricated. For example, in transtibial prostheses, plaster bandage methods include circumferential wrapping as used in the patellar tendon bearing socket, as well as a two-stage modeling method such as that used for total surface bearing sockets, in which a plaster splint is first applied followed by circumferential wrapping. The advantages of two-stage modeling include accurate capture of the anterior bony contour of the residual limb and the ability to control limb volume by applying pressure to the residual limb with elastic bandages. However, because this method compresses and displaces soft tissues, it requires an understanding of bone morphology, soft tissue, and skin conditions1). Accordingly, these methods require considerable experience and expertise. Other methods include weight-bearing modeling using dedicated modeling platforms3), including the Ipso Cast system developed by Ottobock2). Weight-bearing modeling has the advantage of capturing a shape closer to that encountered during actual use by applying the load to the residual limb, although it requires initial investment for a dedicated modeling platform. In addition, pressure-based modeling using systems such as Icecast by Össur4) are also available. Go et al. reported that pressure-based modeling has the advantage of enabling modeling based on the hydrostatic theory5), which involves uniform compression of the entire residual limb6). Furthermore, methods that do not use plaster bandages, such as three-dimensional (3D) scanning of the residual limb, have also been reported. Compared with plaster bandage-based methods, 3D scanning requires less time and avoids contamination of the surrounding area with plaster7). Although plaster bandages remain the mainstream for modeling ankle-foot orthoses and knee-ankle-foot orthoses, alternative methods include fabrication based on traced outlines of the lower limb and measurement values8), as well as modeling using plastic casts as a substitute for plaster bandages9).
In the case of foot orthoses, the Orthotics (4th ed) describes the conventional impression foam-based modeling as the standard approach. This method has the advantage of easily reproducing the shape of the sole of a living person’s foot and not causing fear to the participant because it does not involve cutting with a blade. However, on the other hand, the positive model created by taking a mold multiple modifications, including reductions at the waist girth, lateral aspect of the calcaneus, medial longitudinal arch, lateral longitudinal arch, and metatarsal arch, as well as additions to account for soft tissue deformation around the dorsum of the foot and the calcaneus10). Therefore, achieving accurate fabrication requires substantial working time and technical proficiency. Regarding the fabrication of foot orthoses, Uno reported that “the acquisition of knowledge and technical skills requires many years of experience”11). With the aim of reducing working time, we developed a novel modeling method, referred to as the Arch Support Molding Method (ASMM), inspired by weight-bearing modeling used for prostheses. This method enables acquisition of a post-modification positive model by having the participant step on a dedicated arch support (Fig. 1) during modeling.
Fig. 1.
Dedicated arch support.
ASMM has the advantage of allowing the acquisition of a certain model shape even with low skill in shoe orthotic fabrication, and because a modified model can be obtained, the modification process and time can be shortened. However, on the other hand, it requires specialized equipment, and because there are only two sizes, it may not be suitable if the foot length of the person is too large or too small.
The objectives of this study were to evaluate the working time and shapes obtained using ASMM compared with the conventional method and to demonstrate its usefulness.
PARTICIPANTS AND METHODS
Two sizes of arch supports, large and small, were selected based on the foot length of the participant, and a mold was taken. The dimensions of the dedicated arch support are as follows.
The large size measures 180 mm in length, 80 mm in width, 10 mm in height, with an arch support section of 125 mm and an arch support section width of 55 mm.
For the small size, the overall length was 165 mm, the width was 75 mm, the height was 10 mm, the arch support section length was 105 mm, and the arch support section width was 45 mm. The ASMM modeling procedure was performed as follows. A sheet of tracing paper was placed under the foot, and an outline of the foot was traced. The following anatomical landmarks were transferred onto the tracing paper: the tip of the second toe, the prominence of the calcaneus, the head of the first metatarsal, and the head of the fifth metatarsal. A line connecting the first and fifth metatarsal heads (MP line) was drawn, and a dedicated arch support was positioned with its apex at the MP line and fixed in place using double-sided tape or similar material (Fig. 2). A wedge-shaped toe spring was attached to the toes and similarly fixed with double-sided tape. The foot was then wrapped from above the ankle to the tips of the toes with plastic wrap to prevent direct contact between the skin and the plaster bandage. A cutting string was placed over the wrap along the dorsum of the foot. A non-elastic plaster bandage was applied by wrapping from slightly above the ankle to the toes. The foot was then placed onto the dedicated arch support, referencing the traced foot outline. Subsequently, manual pressure was applied from the dorsum of the foot to reproduce the shape of the arch support (Fig. 3). After the plaster bandage had hardened, the dorsum of the foot was incised using a cutting tool, and the negative model was removed from the foot.
Fig. 2.
Preparation for modeling.
Fig. 3.
Modeling procedure.
Participants were 10 undergraduate students from Niigata University of Health and Welfare (5 men and 5 women, mean age 19.7 ± 0.9 years, mean height 166.7 ± 8.4 cm, mean body weight 60.0 ± 9.2 kg). Both feet of each participant were modeled, for a total of 20 feet.
For modeling, plaster bandages (Plasrun Gyps; Alcare Co., Ltd., Tokyo, Japan) and impression foam (Keiai Orthopedic Appliance Co., Ltd., Tokyo, Japan) were used.
Working time was measured using a stopwatch. For shape comparison, 3D data of the positive models were acquired using a 3D scanner (EinScan H2, SHINING 3D, Hangzhou, China) and analyzed using CAD inspection software (ZEISS INSPECT Optical 3D, ZEISS, Oberkochen, Germany).
One researcher who had been trained in modeling procedures using both the conventional method and ASMM, as well as the modification of positive models, performed the modeling using the two aforementioned methods. The time required for modeling and for modification of the fabricated positive models was measured. A second researcher acquired 3D data of the positive models before and after modification, using 3D scanning and measured the dimensions using 3D CAD measurement software. The time required for each process—modeling, subtractive modification, and additive modification—was measured. The procedures were performed under the following conditions:[During modeling]
- All necessary materials and tools were prepared before starting modeling.
- The order in which the two methods were performed was randomized.
- Participants performed only the instructed movements.[During subtractive modification]
- The procedure was started with all necessary tools prepared.
- The procedure was started after removing irregularities in the plaster wrapping of the positive model.[During additive modification]
- The procedure was started with all necessary tools prepared.
- An equal amount of plaster was prepared before starting the procedure.
- No additional toe allowance was applied.
For 3D scanning of the positive models, three types of models were analyzed: the unmodified positive model created using the conventional method (“Conventional”), the unmodified positive model created using ASMM (“ASMM”), and the modified positive model created using the conventional method, which served as the reference (“Target”). Six parameters were measured for these three models: heel girth, ball girth, MP line width, lateral longitudinal arch height, medial longitudinal arch height, and metatarsal arch height. Specifically, the circumferences of the heel girth and ball girth, the width of the MP line, and the heights of the lateral longitudinal arch, medial longitudinal arch, and metatarsal arch from the plantar surface were measured. To clearly define the measurement locations of the arches, the positions of each arch were defined with reference to anatomical models and textbooks12). The definitions were as follows: the lateral longitudinal arch was defined at 25% of the foot length from the heel, the medial longitudinal arch at 35%, and the metatarsal arch at 55%. Cross-sectional views at these positions were created on the 3D models, and measurements were performed. The lateral and medial longitudinal arches were measured according to the procedures shown in Table 1, and the metatarsal arch was measured according to the procedures shown in Table 2.
Table 1. Measurement procedure for the medial and lateral longitudinal arches.
Table 2. Measurement procedure for the metatarsal arch.
To compare working time and the obtained shapes, the distribution of the data was visually assessed using frequency distribution tables. Because many of the results did not follow a normal distribution, nonparametric tests were used for subsequent comparisons13). For the comparison of working time, all time measurements were converted to seconds, and the Wilcoxon signed-rank test was performed for each task using data from all 20 feet. For the comparison of the obtained shapes, all measurements were converted to millimeters, and comparisons among the three groups were performed using Bonferroni-adjusted multiple comparisons. In comparing both working time and shape, we referred to the previous study by Sato et al. and set the significance level at 1%9).
Statistical analyses were performed using JSTAT (Ver. 22.1 J).
This study was approved by the Ethics Review Committee of Niigata University of Health and Welfare (Approval No.: 19605-250714). All participants received written and verbal explanations regarding the study objectives, methods, and protection of personal information, and provided written informed consent prior to participation.
RESULTS
The comparison of working time showed that the time required for modeling, subtractive modification, and additive modification was significantly shorter with ASMM than with the conventional method (Table 3).
Table 3. Comparison of working time (s).
| Evaluation item | ASMM | Conventional method | p-value |
| Modeling | 827 (761, 869.5) | 1069 (949.25, 1148.75) | 0.0001 |
| Subtractive modification | 299.5 (268.75, 335.25) | 549.5 (491.25, 607.25) | 0.0001 |
| Additive modification | 116.5 (86.75, 135) | 201 (180.5, 207.25) | 0.0003 |
Median (Q1, Q3). ASMM: arch support molding method.
Comparison of the obtained shapes revealed that heel girth in the ASMM models was significantly smaller than that in the Conventional and Target models. For ball girth, values in the ASMM model were significantly smaller than those in the Conventional models and showed little difference from the Target models. Similarly, the MP line width in the ASMM models was significantly smaller than that in the Conventional models and showed minimal difference from the Target models. For the medial longitudinal arch, values in the ASMM model were significantly higher than those in the Conventional models and showed little difference from the Target models. For the lateral longitudinal arch, values in the ASMM and Conventional models were significantly lower than that in the Target models, although the values in the ASMM models were closer to those in the Target models. For the metatarsal arch, significant differences were observed among all three groups (ASMM, Conventional, and Target), although the values in the ASMM models were closer to those in the Target models (Table 4).
Table 4. Comparison of obtained shapes (mm).
| Evaluation item | ASMM | Conventional method | Target value |
| Heel girth*† | 302.24 (289.90, 315.74) | 310.85 (300.31, 324.04) | 312.15 (306.62, 324.70) |
| Ball girth* | 230.56 (219.64, 244.84) | 237.68 (223.89, 249.93) | 230.48 (223.43, 249.50) |
| MP line width*‡ | 94.63 (89.99, 98.96) | 98.02 (90.93, 103.21) | 94.32 (90.03, 101.49) |
| Lateral longitudinal arch†‡ | 4.91 (4.53, 5.16) | 4.33 (3.99, 4.66) | 5.68 (5.15, 6.32) |
| Medial longitudinal arch*‡ | 9.91 (9.40,10.50) | 8.20 (7.65, 9.14) | 10.08 (9.59, 10.67) |
| Metatarsal arch*†‡ | 7.74 (6.56, 8.41) | 1.00 (0.53, 1.88) | 6.16 (5.91, 6.56) |
*: ASMM vs. conventional method p<0.01. †: ASMM vs. target value comparison p<0.01. ‡: comparison of conventional method vs. target value p<0.01. Median (Q1, Q3). ASMM: arch support molding method.
DISCUSSION
In the comparison of working time for the three processes, ASMM reduced the time required by 242 s for modeling, 250 s for subtractive modification, and 84.5 s for additive modification compared with the conventional method. These reductions are likely attributable to the absence of the impression foam step during modeling and the smaller number of modification areas required, resulting in an overall reduction in working time to approximately two-thirds. In the comparison of the obtained shapes, heel girth in the ASMM models was smaller than that in the Conventional and Target models. This may be because, in ASMM, the load was applied from the dorsum of the foot to emphasize the shape of the dedicated arch support, resulting in a reduced circumference. This could potentially affect the shoe’s design, as it could result in a smaller opening for the shoe or a shorter strap length during manufacturing. For ball girth, values in the ASMM models were significantly smaller than those in the Conventional models and showed little difference from the Target models. This can be explained by the fact that, similar to the weight-bearing modeling3), stepping onto the dedicated arch support and applying load from the dorsum of the foot reproduces the shape of the arch support, elevates the metatarsal arch, and results in a shape closer to that of the post-modification model. Regarding the appropriate use of the dedicated arch support, it is considered necessary, as in total surface bearing modeling, to understand soft tissue displacement and bone morphology1). Similarly, the MP line width in the ASMM models was smaller than that in the Conventional models and showed minimal difference from the Target models. This is likely due to the same mechanism as for ball girth, whereby stepping onto the dedicated arch support and applying the load onto the support reproduces its shape and elevates the metatarsal arch, resulting in a shape closer to that of the post-modification model. For the medial longitudinal arch, values in the ASMM models were higher than those in the Conventional models and showed little difference from the Target models. This may be because ASMM does not use impression foam during modeling, enabling the medial longitudinal arch to be more easily reproduced with a manual technique. For the lateral longitudinal arch, both ASMM and the conventional method produced values lower than the target value, although ASMM was closer to the target value. This suggests that, although simply stepping on the dedicated arch support is insufficient to elevate the lateral longitudinal arch to be closer to the post-modification model, ASMM still provides a certain degree of improvement, given that the values were closer to the target value. In contrast, the conventional method may not allow sufficient manipulation of the lateral longitudinal arch due to the use of impression foam. For the metatarsal arch, significant differences were observed among the ASMM, Conventional, and Target models, although the values in the ASMM models were closer to those in the Target models. This may be because ASMM elevates the metatarsal arch by having the participant step on the dedicated arch support, thereby reproducing the shape of the post-modification model. The arch height was set at 10 mm, which may have been higher than the target value obtained through modification. In contrast, in the conventional method, the metatarsal arch was lower because no subtractive modification was applied to this region10).
As a limitation of this study, the researcher who performed modeling and model modification was an undergraduate student, which may have resulted in longer working times compared with those of a professional prosthetist/orthotist who routinely fabricates foot orthoses. However, since the areas to be modified are the same even when used by prosthetists and orthotists, a reduction of about 30% in working time is expected, suggesting high practicality in clinical settings. In addition, in the comparison of dimensions, the modified positive model produced using the conventional method was used as the target value. However, because the modifications were performed by a student, the accuracy of modification may differ from that performed by a professional prosthetist/orthotist. Furthermore, even among prosthetists and orthotists, the correction values are not consistent, which can lead to systematic errors. In contrast, it is believed that systematic errors can be reduced by providing correction values. Furthermore, since the participants in this study were healthy individuals, their foot condition differs from that of participants requiring orthotic shoes due to foot deformities, which may affect the results. Similarly, because the arch support height is fixed at 10 mm, it may not be suitable for participants with excessively high or low arches.
Despite these limitations, this study quantitatively evaluated the working time and the accuracy of reproducing post-modification model shapes across different modeling methods, thereby demonstrating the usefulness of the ASMM modeling method. The findings of this study are expected to inform the selection of appropriate modeling methods in clinical practice according to patient characteristics and environmental conditions.
The remaining challenge based on the present results is that, with conventional dedicated arch support, the lateral longitudinal arch tends to be lower than that of the post-modification model. Therefore, the support height in the lateral longitudinal arch region should be increased. In contrast, the metatarsal arch tends to be higher than that of the post-modification model, indicating that the support height in the metatarsal arch region should be reduced.
Funding and Conflict of interest
None.
REFERENCES
- 1.Tazawa E, editor.: Manual for prosthetic fabrication, 2nd ed. Tokyo: Ishiyaku Publishers, 2022, pp 69–183. [Google Scholar]
- 2.IpsoCast: Prosthetic modeling technique. ifu_743G15_Ipso_Cast_20210416_jp.pdf (Accessed Apr. 15, 2026)
- 3.743G15 IpsoCast instruction manual. https://assets.ctfassets.net/8ks1shyq5m87/6gx07hq064AvReIMesPyQw/273fa3de9e819b450ef8ba3c1ef07d7a/ifu_743G15_Ipso_Cast_20210416_jp.pdf (Accessed Apr. 15, 2026)
- 4.Direct Socket toolkit. https://www.ossur.com/ja-jp/%E7%BE%A9%E8%82%A2/%E3%82%BD%E3%82%B1%E3%83%83%E3%83%88/direct-socket-tool-kit (Accessed Apr. 15, 2026)
- 5.Kristinsson O: The ICEROSS concept: a discussion of a philosophy. Prosthet Orthot Int, 1993, 17: 49–55. [DOI] [PubMed] [Google Scholar]
- 6.Go T, Katsuhira J, Toe Y, et al. : Shape analysis of transtibial residual limbs and sockets using a 3D scanner: relationship between compression value and residual limb length. J Jpn Soc Prosthet Orthot, 2022, 38: 226–233. [Google Scholar]
- 7.Kodama Y: Application of 3D digital technology in prosthetics and orthotics in France. J Jpn Soc Prosthet Orthot, 2020, 36: 194–197. [Google Scholar]
- 8.Hiroshige K, Endo M, eds.: Introduction to prosthetics and orthotics. In: Textbook of physical therapy: prosthetics and orthotics. Tokyo: Medical View Co., Ltd., 2023, pp 1–31. [Google Scholar]
- 9.Sato S, Ishikawa K, Yajima K, et al. : Verification of the usefulness of shape acquisition using plastic casts during fabrication of a metal strut ankle-foot orthosis. J Phys Ther Sci, 2024, 36: 430–434. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Tobimatsu Y, Takashima T, editors.: Orthopedic shoes and foot orthoses. In: Orthotics, 4th ed. Tokyo: Ishiyaku Publishers, 2022, pp 17–46. [Google Scholar]
- 11.Uno A: Advances in therapeutic shoes for children: background and current status of prosthetists/orthotists. J Jpn Soc Med Study Footwear, 2023: 123–130. [Google Scholar]
- 12.Sakai T, Matsumura G, eds.: Lower limb. In: PROMETHEUS Atlas of Anatomy: General Anatomy and Musculoskeletal System, 3rd ed. Tokyo: Igaku-Shoin, 2021, pp 410–574. [Google Scholar]
- 13.Yamamoto S, Ishii S, Ehara Y, eds.: Nonparametric tests. In: Practical rehabilitation statistics (with analysis software), 2nd ed. Tokyo: Nankodo, 2024, pp 71–89. [Google Scholar]