Foot orthotics plays an integral part in the management of various musculoskeletal conditions affecting the lower extremities (Choo et al., 2020; Collins et al., 2007). Foot orthotics, or orthotic insoles, typically aim to redistribute plantar pressure and support the arch of the foot. Orthotic insoles can be fabricated either traditionally using Plaster of Paris (PoP) or digitally using three-dimensional (3D) scanning and additive manufacturing (Rogati et al., 2021; Powers et al., 2022; Lee et al., 2014; Farhan et al., 2021; Barrios-Muriel et al., 2020). The fabrication process begins with careful examination and assessment of the patient, their lower limb mechanics, and foot condition to ensure the capture of accurate geometric information for a well-fitted orthotic insole (Landorf and Keenan, 2000). This assessment includes marking key anatomical landmarks in the foot to measure the foot geometry for guiding the orthotic fabrication process (Telfer et al., 2012).

Landmark identification is central to both traditional and digital fabrication methods. In traditional fabrication, the anatomical landmarks are identified by experienced clinicians through manual palpation to help measure the foot using callipers and measuring tape (Mohammed and Albadwi, 2024; Telfer and Woodburn, 2010). Whereas the digital manufacturing approach streamlines the traditional process by using 3D surface scanning as a reliable, efficient, and precise method to capture foot and ankle morphology (Rogati et al., 2021; Powers et al., 2022; Lee et al., 2014; Farhan et al., 2021; Barrios-Muriel et al., 2020). This technology replicates the foot data as 3D volumes/meshes which are used to design customized orthotic insoles. 3D meshes of the foot can be measured by orthotic designers easily in a digital environment using digital landmarks. The selection of these landmarks is based on their distinctiveness, consistency across different foot models, and the algorithms of the orthotic design software program. Few anatomical landmarks, such as navicular tuberosity, and medial and lateral malleoli, can be identified more easily than the others such as the head of metatarsals and the centre of the heel, which may require training to locate correctly on a digital scan. Collectively, these landmarks are used to guide decision-making during the fabrication process (Lee et al., 2014; Kouchi and Mochimaru, 2011; Kouchi et al., 1999). The key digital landmarks needed to design an orthotic insole include the head of the first and fifth metatarsal and centre of the heel on the plantar region and the highest point of the medial longitudinal arch (Telfer et al., 2012; Telfer and Woodburn, 2010).

Anatomical landmarks in the foot are used as fixed reference points for registration to assist accurate alignment for generating a measurement axis, to enhance the reliability and accuracy of foot measurements (Telfer and Woodburn, 2010; Wu et al., 2018; Witana et al., 2006; Barisch-Fritz et al., 2014; Goonetilleke et al., 2009). Once the model is registered, the user calculates distances and circumferences with the help of these predefined landmarks. These measurements form the foundation for modifying the custom orthotic insole tailored to the specific needs of individuals (Lee et al., 2014; Farhan et al., 2021; Telfer and Woodburn, 2010). Landmark identification in 3D models can follow one of three primary approaches:

Palpation-guided approach- The anatomical landmarks can be marked manually through palpation by clinicians before the 3D scanning process for later identification in the digital designing process. These visible landmarks can then be selected in the 3D scans for accurate alignment with the world coordinate system (Wu et al., 2018; Witana et al., 2006; Xiong et al., 2009; Luo et al., 2009).

Studies have demonstrated that changes in foot morphology can be measured using landmarks, whether palpated (Wu et al., 2018; Witana et al., 2006; Xiong et al., 2009; Luo et al., 2009), manually scan-derived (Pantazi and Vasilescu, 2025; Van den Herrewegen et al., 2014) or Artificial intelligence (AI)-based (Rogati et al., 2021; Wu et al., 2018; Wang et al., 2018). Recent advances in commercial 3D scanning systems, including tablet-based scanners, increasingly incorporate basic automated or AI assisted landmarking algorithms. While AI-based approaches are promising for their potential to automate landmark identification, they require extensive validation to ensure their reliability and clinical applicability, especially in the foot orthotic industry, where the foot morphology could be affected by the pathology (Yahara et al., 2005). Moreover, many of the current AI and computational techniques rely on advanced imaging modalities such as radiographs (Shelton et al., 2019; Noh et al., 2024; Ryu et al., 2022) or CT scans (Subburaj et al., 2009), and require complex software capable of geometric analysis (Rogati et al., 2021), statistical shape modelling (Schuster et al., 2021), or deep learning algorithms (Ryu et al., 2022). These approaches limit their feasibility for routine use in podiatry clinics and orthotic manufacturing settings where time, budget, and technical expertise are limited. Consequently, despite increasing adoption of automated tools, manual landmark identification, either through palpation or clinician-guided, remains the current gold standard, particularly in complex or pathological cases. Therefore, there remains a gap between research advancements and their translation into clinical and commercial orthotic practice (Choo and Chang, 2023; Williams et al., 2016). Consequently, this study will disregard the automated landmark detection technique and consider the manual identification method as the scan-derived approach.

The use of palpation-guided and scan-derived anatomical landmarks is common in both research and clinical settings. Studies have testified to the use and importance of anatomical landmarks to generate accurate measurements (Lee et al., 2014; Witana et al., 2006; Barisch-Fritz et al., 2014; Goonetilleke et al., 2009; Zhao et al., 2008; Liu et al., 1999; Bunch, 1988; Tsung et al., 2003). Clarkson et al. (Clarkson et al., 2016) reported in their study that the palpation-guided identification process is time-consuming but most reliable and yields the greatest accuracy. Similar results were observed by Lee et al. (2014) in their study, reporting that the palpation-guided approach performed better when compared to other traditional methods (digital callipers, and digital and ink footprint) of generating foot measurements. A review study conducted by Telfer and Woodburn (2010) summarized the use of 3D scanning technologies in custom orthotic development and reported that bony landmarks identified on the foot using markers that can be seen in the 3D scans appear to be the most reliable method to obtain girth measurements. However, this accuracy is not without limitations. Errors may arise from both the practitioner’s expertise in locating landmarks manually and the designer’s accuracy in selecting those points digitally, potentially compounding the overall measurement error (Lee et al., 2014; Kouchi and Mochimaru, 2011; Kouchi et al., 1999; Clarkson et al., 2016).

While scan-derived methods are time-efficient and support remote assessments, their accuracy remains uncertain in comparison to manual-palpated methods for orthotic designing. Consequently, the implication of using palpation-guided or scan-derived landmarks in orthotic design remains unclear. It is crucial to address this as orthotic devices are most effective when customized precisely to an individual’s anatomy (Telfer et al., 2012). Moreover, understanding the impact of different weight-bearing positions on anatomical landmark identification is crucial, as foot scans in clinical practice are often captured in varying weight-bearing conditions depending on the individual’s needs or clinician’s preference such as non-weight-bearing (NWB), partial weight-bearing (PWB), half weight bearing (HWB), full-weight bearing (FWB), each reflecting different levels of load on the foot. These positional differences can influence the visibility and accuracy of landmark placement, whether palpation-guided or scan-derived, which may in turn affect foot measurements and the precision of orthotic designs (Chhikara et al., 2025).

With the growing integration of digital technologies in orthotic practice, podiatrists are now starting to perform 3D foot scans and transmit the digital models to orthotic fabrication laboratories. Equipped with additive manufacturing technologies, these facilities rely on in-house orthotic designers to develop customized orthotic insoles based on the scanned data and clinical specifications provided by the podiatrist. Although this division of responsibilities enhances workflow efficiency, it may potentially introduce challenges in translating clinical information between podiatrists and designers. Accuracy in identifying anatomical landmarks may vary between podiatrists (possessing critical anatomical and biomechanical knowledge) and designers (possessing expertise in 3D modelling and digital design software). This disparity in expertise highlights the need to critically examine how effectively anatomical information is communicated from clinics to fabrication laboratories.

While previous studies have examined scanner accuracy, foot measurement reliability, or weight-bearing effects independently, none have evaluated how landmarking method (palpation-guided, scan-derived), scan condition (NWB, WB), and user background (clinicians, designers) collectively influence the reliability of anatomical measurements and the downstream consistency of orthotic designs. Although the analysis of foot morphology is common in the literature, the corresponding orthotic designs are rarely examined. This study provides the first integrated assessment of variability across the digital orthotic workflow, from landmark placement to orthotic designing, thereby identifying how discrepancies propagate into final device geometry. By quantifying the sources of variability, the study contributes new evidence to support the development of standardised digital orthotic design protocols that minimise user-dependent differences and improve design consistency.

This study was conducted in two phases. The primary study focused on reliability of digital foot measurement and the secondary preliminary study focussed on analysing the consistency of digital foot orthotic designs.

This multi-user, multi-designer workflow has not previously been evaluated in orthotic research and enables direct comparison of measurement reliability and resulting design variability under controlled changes in landmarking method and weight-bearing condition.

Unlike previous studies that focused solely on scanner accuracy or foot measurement reliability in isolation, this study provides the first integrated assessment of how landmarking method, user expertise, and weight-bearing condition collectively influence foot measurements and orthotic design consistency. With advancement in technology, the orthotic industry is inclining towards switching the traditional tools and techniques with the digital ones for improved efficiency and scalability (Barrios-Muriel et al., 2020; Binedell et al., 2020; Kogler and Hovorka, 2021; Yuen et al., 2023; Senbekov et al., 2020). However, this transition has introduced newfound challenges, particularly in the standardisation of anatomical landmarking which forms the foundation for foot measurements and orthotic designs. With this shift, anatomical expertise resides with podiatrists while digital expertise lies with designers, creating a mismatch that challenges consistent landmarking and measurements. This study investigated the impact of palpation-guided versus scan-derived landmarking on the reliability of foot measurements (phase 1) and consistency of orthotic designs (phase 2). Across NWB and HWB conditions, intra-user reliability remained excellent for both methods, whereas inter-user reliability dropped substantially for scan-derived landmarks in clinically important measures such as arch height, MT1-to-heel, MT5-to-heel, and heel width. The findings from this study assist in clarifying whether design inconsistencies stem from user skill level, variability in landmarking methods, or phase-specific changes in foot morphology under different weight-bearing conditions. Notably, the observed low inter-user ICC values for plantar measurements highlight a critical limitation of relying on unguided digital landmark estimation in clinical and commercial orthotic workflows. The findings from this study are valuable for clinical and orthotic fabrication settings aiming to optimize their digital workflows, train staff effectively, and enhance outcome consistency across varying level of user expertise.

The reliability of foot measurements was assessed by three different users on 24 foot scans, scanned in NWB and HWB condition. The findings demonstrate that palpation-guided landmarks improve both the intra- and inter-user reliability when making digital foot measurements, especially in key measurements such as arch height, MT1 to hell, MT5 to heel and heel width. Intra-user reliability remained excellent for both approaches (>0.9), indicating that each user was consistent within their own repeated measurements. However, the inter-user ICC values dropped markedly for the scan-derived approach, particularly for MT1-heel, MT5-heel, and heel width, where the ICC decreased from 0.87-0.99 (with landmarks) to 0.31–0.55 (without landmarks). This marked reduction to poor inter-user agreement demonstrates that unguided, scan-derived landmark estimation (“digital guessing”) is inherently unreliable between operators, particularly for plantar measurements that lack clear visual reference points. Notably, arch height in NWB also showed reduced agreement (0.85 vs. 0.55), whereas in HWB, the ICC was slightly higher without landmarks (0.75 vs. 0.63), likely due to the more uniform foot contour under loading. These findings highlight that landmark visibility critically affects inter-user consistency for key clinical measures. In both weight bearing conditions, palpation-guided landmarks indicated higher ICC values, representing better reliability across same and different users. These results align with prior studies (Kouchi and Mochimaru, 2011; Aynechi et al., 2011; Dong et al., 2021; Gibelli et al., 2022; Weinberg et al., 2004), such as one conducted by Powers et al. (2022), which reported higher consistency for measurements generated from manual palpation. Similar trends have been observed in landmark placement across facial scans by Aynechi et al. (Gibelli et al., 2022), highlighting broader relevance of this issue.

For foot scans with scan-derived landmarks, where the user had to estimate the point in the digital scan, reliability was relatively low. The inclusion of user C, with a lower bound of anatomical expertise, highlights the dependency of scan-derived landmarking reliability on anatomical training rather than solely on software proficiency. Especially for users with limited anatomical knowledge, several key measurements such as arch point and inter-metatarsal distance demonstrated reduced reliability, emphasizing that accuracy of digital landmarks does not solely depend on the software but is also significantly influenced by the user’s clinical understanding. The lower reliability scores for the user with no anatomical training suggest that scan-derived landmarking may not be suitable for novice users lacking expertise, also reported by previous studies (Kouchi and Mochimaru, 2011; Kouchi et al., 1999). Regardless of 3D design experience, the translation of patient-specific anatomical data from clinical to industrial manufacturing settings is only effective when industrial orthotic designers are trained in anatomy and identification of bony landmarks. In addition, scan-derived landmarks may not fully substitute for anatomical palpation in certain pathologies, particularly in cases of foot swelling or in presence of lesions (Vega-et al., 2023; Rodriguez et al., 2023). This study was performed with healthy participants as a baseline. As such, the findings may not fully reflect the reduced landmarking reliability likely to occur in pathological feet (e.g., oedema or Charcot deformity), warranting further investigation in clinical populations In addition, the orthotic design consistency analysis was based on a single subject; therefore, the findings represent a preliminary assessment and should be interpreted with caution. Weight-bearing conditions affected measurement reliability differently across parameters. Flattening of the arch in HWB scans made the arch contour more uniform, slightly improving inter-user reliability compared to NWB, whereas NWB conditions showed higher variability, particularly in the plantar and hindfoot regions. These findings suggest that weight-bearing must be carefully considered when designing orthotics or assessing plantar geometry.

For a preliminary evaluation of the orthotic designs produced for palpation-guided and scan-derived landmarking methods, the consistency was assessed by measuring orthotic shape conformity within a tolerance region of ±1 mm (Redaelli et al., 2021). The consistency between orthotic designs based on palpated landmarks by podiatrists when compared to designs based on scan-derived landmarks demonstrated highest conformity for the PWB position compared to the NWB position. This marked increase in conformity under PWB conditions suggest that partial loading of the foot stabilises the soft tissue and the medial longitudinal arch,thereby reducing ambiguity in arch interpretation during the orthotic design process. These findings indicate that PWB scans may provide a more reproducible representation of foot geometry for digital orthotic design, particularly in the arch region where design discretion is greatest. The variation in NWB condition could be observed particularly in the arch height (up to 4 mm), which could be significant for the orthotic user, this effect was less visible (up to 2 mm) in the PWB condition. An error of up to 4 mm in arch height may substantially alter plantar load distribution, particularly in the MLA region where orthotic correction is most targeted. This error exceeds the acceptable tolerance value in the orthotic design industry (±1 mm) (Redaelli et al., 2021) and may contribute to discomfort, reduced compliance, or device rejection. These findings underscore the importance of both anatomical landmarking strategy and weight-bearing protocol when designing digital orthoses, as seemingly small geometric errors can translate into clinically meaningful outcomes.

In contrast, forefoot and hindfoot regions in the orthotic designs showed higher conformity regardless of weight-bearing condition. However, it is important to consider that the orthotic designs in the study were made by designers with several years of experience working with podiatrists. A similar trend was observed in inter-designer consistency, where providing all designers with identical palpation-guided landmarks (same podiatrist’s landmarks given to all designers) resulted in lower consistency than the intra-designer consistency for non-identical palpation-guided landmarks (landmarks placed by different podiatrists). These findings indicate that while differences in podiatrist landmark placement can introduce variability, experienced designers may reduce this impact through their ability to adapt to varied inputs. Additional findings from the consistency analysis are presented in detail in the Supplementary Material. It should be noted that the orthotic designs were not assessed through physical or experimental trials, representing a limitation of the study.

Both study phases indicate that anatomical knowledge plays a critical role in interpreting foot morphology. While palpation ensures reliable anatomical landmark identification, designing the corresponding arch height in the orthotic designs requires additional biomechanical expertise and a standardised design approach, especially in clinically critical region such as medial longitudinal arch. Further studies involving designers with varying experience levels (novices in training vs. experts) and multiple participants with varying foot shapes and pathologies is essential.

Orthotic designing has traditionally been a practical and experience-based approach, relying heavily on the expertise of Allied Health Professionals (Holtkamp et al., 2024). The digital era is reforming this process by enhancing reliability and consistency by reducing the variability associated with different users and their level of experience. This study shows that the digital design process still requires manual input from the designers and is still affected by human error. There is variation between orthotic designs created by different designers and there is variation in the placement of palpated landmarks by different podiatrists. While digital tools enhance workflow efficiency and scalability, tasks like landmark identification still require the clinical expertise of trained professionals (Wickramasinghe et al., 2022). If palpated landmarks are used, podiatrists and designers would benefit from more collaborative training to align their approaches and reduce variability. Also, emerging technologies such as ultrasound-based landmark detection, motion capture systems and AI-driven automated landmarking could enhance landmarking accuracy thus supporting more standardised orthotic design workflows (Rogati et al., 2021; Wu et al., 2018; Wang et al., 2018). However, their integration in the actual clinical practice currently remains limited due to the high cost, complexity and workflow compatibility, thus making them less viable in practical settings (Choo and Chang, 2023; Williams et al., 2016).

This study provides the first integrated evaluation of how palpation-guided versus scan-derived landmarking, user expertise, and foot weight-bearing conditions collectively influence foot measurement reliability and orthotic design consistency. A common question while adopting digital technology for the purpose of orthotic designing is whether only to rely on manual palpation or entrust digital identification of anatomical landmarks in the 3D foot scans. This uncertainty leads to concerns about which approach ensures greater reliability and higher consistency. This study aimed to address this issue by analysing the reliability of foot measurement and consistency in orthotic designs with palpation-guided and scan-derived landmarks. The findings suggest that palpation-guided landmarks demonstrate enhanced reliability for general foot measurements and relative consistency of orthotic designs. However, the detailed consistency outcome showed mixed results due to variability introduced by the orthotic designer’s discretion. Variations in foot loading conditions influenced both measurement reliability and orthotic design consistency. Overall, palpation remains a valuable aid in this controlled study and is likely even more crucial in less controlled clinical settings. However, if palpation needs to be integrated with the digital designing methods, collaborative training between podiatrists and designers will ensure that clinical insight and technical expertise are effectively merged to improve the accuracy of arch height in orthotic designs.