A low cost silicone mandibular model for suturing skills training among undergraduate medical students: development and comparative evaluation

Abstract

Background

Simulation based suturing training allows novice learners to practise technical skills before clinical application. Conventional flat suturing pads are useful for early repetitive practice but provide limited anatomical context. This study aimed to develop an affordable silicone mandibular model for suturing skills training among undergraduate medical students and to compare students’ perceptions of this model with those of a conventional flat silicone suturing pad.

Methods

A dry human mandible from an anatomy education inventory was scanned using a three dimensional scanner. The digital model was modified in Blender and a two component negative mould was designed and printed. A platinum based RTV-2 silicone was used to produce the mandibular suturing model. Third year undergraduate medical students practised simple interrupted sutures on both the mandibular model, labelled as Model A and a conventional flat suturing pad, labelled as Model B. Participants evaluated both models using the same anonymous structured questionnaire. Likert scale responses were compared using the Wilcoxon signed rank test. Internal consistency was assessed using Cronbach’s alpha. Holm correction was applied across the 17 item level comparisons.

Results

A total of 107 students completed the questionnaire. The mean age was 21.85 ± 1.39 years; 52 participants were female and 55 were male. Cronbach’s alpha was 0.982 for Model A and 0.983 for Model B. The total questionnaire score was slightly higher for Model B than for Model A, but the difference was not statistically significant (median 69 [63–79] vs. 74 [63–83], p = 0.060). Unadjusted item level comparisons showed higher ratings for the flat pad in four domains. After Holm correction for 17 item level comparisons, only realistic needle passage remained statistically significant in favour of the flat pad (Holm adjusted p = 0.014). No item showed a statistically significant advantage in favour of the mandibular model. The estimated material cost of each mandibular model was approximately 2 USD, excluding the reusable upper mould.

Conclusions

The silicone mandibular model did not demonstrate superiority over the conventional flat silicone suturing pad in students’ evaluations. After correction for multiple comparisons, only realistic needle passage was rated significantly higher for the flat pad. However, the mandibular model was affordable, reproducible and openly shareable and received learner ratings of similar magnitude for anatomical realism, spatial awareness, difficult access practice, durability and overall satisfaction. These perception based findings suggest that the model may have potential as a complementary tool for anatomically contextualized suturing practice rather than as a replacement for conventional flat pads. Future studies using objective performance measures are needed to evaluate its effect on suturing competence and skill transfer.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12909-026-10030-2.

Background

Acquiring basic surgical skills is an important component of undergraduate medical education. Suturing, in particular, requires not only cognitive understanding of procedural steps but also repeated psychomotor practice, appropriate instrument handling, visuospatial orientation and tissue manipulation. Simulation based training provides a safe environment in which novice learners can practise technical skills before performing procedures on patients, while reducing patient related risk and allowing structured feedback [1]. This approach is consistent with deliberate practice principles, which emphasize well defined learning objectives, focused repetition, rigorous assessment and specific feedback for progressive skill development [1, 2].

Traditional suturing education often relies on flat silicone pads, animal tissues, cadaveric materials or commercially available task trainers. Although these models provide opportunities for repetitive practice, each has limitations. Biological models may present ethical, logistical, cost related and reproducibility concerns, whereas simple synthetic models may lack anatomical accuracy and realistic haptic feedback [3]. In addition, flat suturing pads do not reproduce the restricted access, curved surfaces and three dimensional anatomical context encountered in clinical practice. Previous work in dental and surgical education has emphasized that anatomically accurate models may better support learners in understanding spatial constraints and preparing for clinical procedures [4, 5].

Recent advances in three dimensional (3D) imaging, computer aided design and additive manufacturing have enabled the development of patient specific or anatomy based physical models for procedural training. In dental education, individualized 3D printed models based on real patient situations have been used for surgical procedures such as root tip resection and have been compared with conventional typodont models [6]. Similarly, multi material 3D printed models have been evaluated for bronchoscopy simulation, demonstrating the potential of anatomically accurate, patient specific models for procedural training [7]. More recently, 3D printed simulators have been used in oral surgery and cleft lip education, with studies reporting high educational value, improved learner experience and in some cases, better operational performance compared with traditional teaching materials [3, 8].

Recent studies have extended this evidence across several levels of dental and surgical education. In undergraduate conservative dentistry, 3D printed models have been compared with commercial training models [9]. A low cost 3D printed model for gingival flap surgery was reported as practical, realistic and useful for hands on training [10]. In a randomized cohort study of suturing education, 3D printed models were evaluated using procedural time and structured performance scores, illustrating the value of combining perception based outcomes with objective performance measures [11]. Patient specific mandibular and maxillary models have also been evaluated among postgraduate periodontics and implantology students, with favourable learner perceptions but persistent concerns regarding tactile realism [12].

Despite these developments, producing realistic and accessible soft tissue simulators remains challenging. Studies evaluating 3D printed incision and suturing simulators have reported high ratings for visual realism and educational usefulness, but tactile feedback, tissue resistance, adhesion, detachment and suture resistance remain areas requiring improvement [3]. These limitations are particularly relevant for suturing education, where needle passage, tissue elasticity, tearing resistance and knot placement are central to the learning experience. Therefore, there is still a need for low cost, reproducible and anatomically meaningful models that can provide learners with a more realistic context than flat suturing pads while remaining feasible for routine undergraduate teaching. Silicone based and hybrid soft tissue simulators have also been developed for cleft and craniofacial surgical training using cast silicone components, three dimensionally printed moulds or printed anatomical supports [13–15]. These studies demonstrate the feasibility of combining compliant soft tissue components with rigid, anatomically shaped structures, while also identifying soft tissue realism as an area requiring further refinement. Mechanical testing of elastomers proposed for oral mucosa simulation has demonstrated substantial differences in tensile, compressive, hardness and wettability characteristics, supporting task specific material selection rather than the assumption that a single elastomer is optimal for all simulation purposes [16].

Accessibility, reproducibility and affordability are key determinants of whether simulation models can be integrated into routine undergraduate training. Cadaveric materials and commercial high fidelity simulators may offer valuable realism, but their use is often constrained by cost, availability, maintenance requirements and legal, ethical, cultural or logistical factors. Three dimensional printed and anatomy based models have therefore been proposed as scalable educational tools that can support anatomy and procedural skills training without replacing all aspects of cadaver based learning [7, 17, 18].

In the present study, the digital model and fabrication workflow were designed to be reproducible with commonly available three dimensional printing and silicone casting materials. By making the digital model file available online, the study aims to enable other institutions to reproduce the model and adapt it to their own educational settings. Open source hybrid simulators combining three dimensionally printed components with cast silicone have been proposed as a means of reducing dependence on expensive commercial surgical trainers [14]. A recent scoping review of three dimensionally printed models in oral surgery simulation highlighted substantial heterogeneity in model design, participant groups and evaluation methods, as well as the need for further validation [19]. Such open and low cost approaches may contribute to more equitable access to anatomically contextualized procedural skills training, particularly in institutions with limited access to cadaveric materials or commercial simulators.

The mandible provides a clinically relevant anatomical framework for suturing practice because of its curved morphology, spatial constraints and association with oral and maxillofacial procedures [4, 6]. The mandible was selected because its curved alveolar arch and three dimensional contours provide a clinically meaningful, non flat substrate for anatomically contextualized suturing practice. A silicone layer formed over the alveolar region can approximate the spatial configuration of gingival soft tissues and provide margins and contours for practising needle entry, tissue approximation and knot placement. In contrast, coating a generic long bone surface with silicone would not reproduce a comparable oral soft tissue geometry or training context. Recent work has demonstrated that a scanned mandibular model and a digitally generated negative model can be used to create a controlled space for simulated oral mucosa [20]. Patient specific oral surgical models combining printed hard tissue structures with simulated soft tissues have likewise been developed for intraoral procedural training [21]. However, commonly used suturing models in undergraduate training are often non anatomical, while anatomically contextualized models have primarily been described in dental or specialty specific surgical education. There is limited evidence regarding learner perceptions and the educational use of a low cost silicone mandibular model developed from an anatomical bone specimen for suturing skills training among undergraduate medical students. Such a model may help bridge the gap between basic suturing practice and anatomically contextualized procedural learning.

This study aimed to develop a low cost silicone mandibular model for suturing skills training and to compare undergraduate medical students’ perceptions of this model with those of a conventional flat silicone suturing pad. The study evaluated learners’ perceptions of anatomical realism, tissue feel, technical suitability, ease of use, perceived educational contribution and overall acceptability. The null hypothesis was that there would be no significant difference between the silicone mandibular model and the conventional flat silicone suturing pad in the total questionnaire score or individual item ratings.

Methods

Study design and ethical approval

This study was designed as a comparative educational evaluation study conducted among undergraduate medical students. Ethical approval was obtained from the Non Interventional Clinical Research Ethics Committee of Kütahya Health Sciences University, Türkiye (decision no: 2026/06–17; date: 04 May 2026). Participation was voluntary and anonymous data were collected from all participants. No patient data or identifiable personal information were used in the study.

Participants

The study included 107 third year undergraduate medical students at Kütahya Health Sciences University Faculty of Medicine. All participants attended the suturing skills training session and voluntarily completed the evaluation form. No identifying information such as name, student number or contact details was collected.

Development of the silicone mandibular model

An anonymized dry human mandible from the educational inventory of the Department of Anatomy was used as the anatomical basis for model development. The mandible was scanned using a Creality CR-Scan Raptor three dimensional scanner (Shenzhen Creality 3D Technology Co., Ltd., Shenzhen, China). During scanning, three dimensionally printed reflector pyramids and reflective markers were used to improve tracking and surface acquisition (Fig. 1A). The scanning process generated a digital three dimensional representation of the mandible, which was visualized using Creality Scan version 4.17 during acquisition (Fig. 1B).

Fig. 1.

Three dimensional scanning of the dry mandible. A dry human mandible from the anatomy education inventory was scanned using a Creality CR-Scan Raptor three dimensional scanner. Reflective markers and three dimensionally printed reflector pyramids were used to improve tracking during scanning. A shows the physical scanning setup and B shows real time point cloud acquisition in the scanning software

The scanned mandibular model was imported into Blender version 4.4 for digital post processing. The model was cleaned, and the alveolar margins were digitally modified and thickened to make the structure suitable for repeated suturing practice. A negative mould was then designed in Blender based on the processed mandibular model (Fig. 2). The mould was divided into upper and lower components. The upper component was designed as a reusable mould part for repeated silicone casting. The lower component was designed to remain attached to the cured silicone model, thereby providing a stable polylactic acid (PLA) base for each silicone mandibular model. The lower component incorporated concealed PLA retention bridges within the casting space to mechanically retain the cured silicone on the lower base. This design allowed subsequent models to be produced by reusing the upper mould and printing only a new lower base component when required.

Fig. 2.

Digital design of the mandibular model and mould components. The scanned mandibular model was processed in Blender. The alveolar margins were digitally modified to make the model suitable for suturing practice and a two component negative mould was designed. The upper mould was designed for repeated use whereas the lower component served as a printable base attached to each silicone model

The print files were prepared and sliced using Bambu Studio version 2.2.1. The upper and lower negative mould components were printed using Creality Hyper PLA RFID filament (1.75 mm, Peach Fuzz; Shenzhen Creality 3D Technology Co., Ltd., Shenzhen, China) with a Bambu Lab X1E three dimensional printer (Bambu Lab, Shenzhen, China). No photopolymer resin was used at any stage of the fabrication process. The mould components and lower bases were produced by fused filament fabrication using Hyper PLA, whereas the soft suturing model was produced by casting platinum based RTV-2 silicone. Hyper PLA was selected because it was readily available, relatively inexpensive and suitable for the rapid fabrication of reusable casting components. A similar open source hybrid simulator used PLA printed mould components for silicone casting [14]. No comparative assessment of alternative filament materials was performed in the present study. During the study, three upper mould components and 20 lower mould components were produced. Each lower mould component required approximately 2 h to print, and each upper mould component required approximately 6 h. The lower mould component required approximately 72 g of filament, whereas the upper mould component required approximately 223 g.

Before silicone casting, the upper and lower mould components were aligned, and Dalin spray baby oil (Kopaş Kozmetik Pazarlama Sanayi A.Ş., Istanbul, Türkiye) was applied to the silicone contacting surfaces of both mould components as a release agent to facilitate demoulding. An AKSİL platinum based RTV-2 silicone with a hardness of 10 Shore A (AKSİL, Konya, Türkiye) was used for casting. Silicone was selected because cast silicone can be used to create compliant soft tissue components within three dimensionally printed moulds and has previously been applied in craniofacial surgical simulators [13, 14]. Mechanical testing of candidate oral mucosa elastomers has shown that hardness, tensile and compressive behaviour and wettability differ substantially among materials [16]. In the present proof of concept study, the 10 Shore A platinum based RTV-2 silicone was selected pragmatically because it was available, readily castable and sufficiently soft to permit needle penetration. It was not selected through a formal comparative material screening study and should not be interpreted as an optimal substitute for human oral soft tissue.

AKSİL RTV-2 red silicone colour paste (PMS 186 C; AKSİL, Konya, Türkiye) was added to the silicone mixture. Approximately a pinhead sized amount of pigment was sufficient for 70 g of silicone prepared by combining equal amounts of components A and B. After the silicone mixture was poured into the mould, it flowed around and encapsulated the concealed retention bridges incorporated into the lower PLA component. Elastic bands were used to secure the mould components and prevent leakage or accidental separation during curing. After curing for approximately 4–6 h at room temperature, the reusable upper mould was removed, whereas the cured silicone remained mechanically secured to the printed lower base through encapsulation of the retention bridges (Fig. 3A).

Fig. 3.

Suturing models used in the study. A The silicone mandibular model used as Model A in the questionnaire, shown next to the original dry mandible used for three dimensional scanning. The mandibular model was produced using platinum based RTV-2 silicone with a hardness of 10 Shore A and remained attached to a printed polylactic acid base after curing. B The commercially available flat silicone suturing pad used as Model B in the questionnaire

The digital mandibular model files and a brief fabrication guide were deposited in Zenodo [29]. These openly available resources support reproduction and adaptation of the model.

Educational procedure and comparator model

The suturing training session was delivered jointly by three faculty members from the departments of general surgery, cardiovascular surgery and orthopaedics and traumatology. None of the instructors were included as investigators in the present study. Before the training session, the instructors were informed about the study design and were introduced to both training models.

Two suturing models were used during the session: the newly developed silicone mandibular model, which was labelled as Model A in the questionnaire (Fig. 3A) and a conventional flat silicone suturing pad, which was labelled as Model B in the questionnaire (Fig. 3B). The conventional model was a commercially available flat silicone suturing pad designed for basic suturing practice. The mandibular model and the flat pad were both used for teaching and practising simple interrupted sutures.

Before the 60 min hands on practice period, the faculty instructors demonstrated the simple interrupted suture technique on both training models. The demonstration period was separate from the subsequent practice periods. Each participant was provided with an individual unbranded suturing set comprising a needle holder, forceps and suture material. These sets had been purchased before the study, and manufacturer, city and country information was unavailable. Participants used the same instruments and suture material while practising on both models. Only the simple interrupted suture technique was practised during the session. The flat silicone pad contained existing linear incisions before the training session. No artificial incision was created in the mandibular model. Instead the empty alveolar sockets and adjacent anatomical grooves provided gaps and opposing silicone margins that served as suturing sites. The flat pad was a commercially available unbranded product that had been purchased before the study. Therefore information regarding its manufacturer, city and country of manufacture was unavailable.

Before the students entered the classroom, the mandibular models and flat pads were placed alternately at adjacent stools. Students selected their seats freely and were not assigned to a particular starting model. Consequently, some participants began with the mandibular model, whereas others began with the flat pad (Fig. 4). Each participant practised suturing on the model initially placed in front of them for 30 min. The models were then passed to the participant seated one place to the left, ensuring that each participant switched to the other model and practised for an additional 30 min. Thus, the hands on practice period lasted 60 min, with equal practice time allocated to each model. During both practice periods, the instructors circulated in the classroom and provided assistance when requested or when a student appeared to experience difficulty. This feedback was individualized and was not delivered using a formal standardized checklist or script. To reduce expectation and novelty bias, students were not informed before or during the evaluation that one of the models had been developed specifically for the study. Until completion of the questionnaire, both models were presented to students as two separate suturing training models. During the questionnaire, students were asked only to evaluate the two models; no information was provided regarding the origin, production method or commercial status of the mandibular model.

Fig. 4.

Suturing practice session. Undergraduate medical students practised simple interrupted sutures using both the conventional flat silicone suturing pad and the silicone mandibular model during the same training session. The students’ faces were blurred to protect privacy

Data collection instrument

After completing suturing practice on both models, participants completed an anonymous structured questionnaire. No previously validated instrument specifically designed for this type of model comparison was identified among the relevant literature reviewed during questionnaire development. Therefore, the questionnaire was developed by the author based on the educational objectives of the training session and previously published studies evaluating three dimensional anatomical and surgical simulation models [3, 4, 6, 21, 22]. These studies informed the inclusion of domains related to anatomical realism, three dimensional structure, visual appearance, tactile properties, needle passage, technical usability, perceived clinical relevance, educational contribution and overall satisfaction.

Before administration, the draft questionnaire was reviewed for clarity and comprehensibility by research assistants, graduate students and medical specialists within the department. The wording and organization of the items were revised based on their feedback. This review was intended to improve item clarity and comprehensibility and did not constitute formal psychometric validation. During questionnaire administration, participants were informed both in writing and verbally that they could ask the researcher for clarification if required. No participant reported an item as unclear or requested clarification regarding the meaning of any questionnaire item. The questionnaire also included demographic questions, previous suturing experience, previous experience with a model or simulation trainer, photographs of both models and one open ended question. Model specific items were rated on a five point Likert scale ranging from 1 = strongly disagree to 5 = strongly agree.

The questionnaire was administered in Turkish using Google Forms. The original Turkish Google Forms questionnaire is provided as Additional file 1 and the English translation of the questionnaire is provided as Additional file 2. The questionnaire referred to the models only as Model A and Model B and did not disclose the origin, production method or commercial status of either model.

Outcome measures

The quantitative outcomes were the total questionnaire score for each model, ranging from 17 to 85, and the paired item level ratings across the 17 model specific questionnaire items. Demographic variables, previous suturing experience and previous experience with a model or simulation trainer were treated as descriptive participant characteristics. Open ended comments were reviewed descriptively.

Statistical analysis

Data were analysed using IBM SPSS Statistics version 18.0. Categorical variables were presented as frequencies and percentages. Likert scale responses were summarized using median and interquartile range. Because each participant evaluated both models, paired comparisons between Model A and Model B were performed using the Wilcoxon signed rank test. Item level comparisons were exploratory. To control the family wise error rate across the 17 item level Wilcoxon signed rank tests, Holm’s sequential Bonferroni procedure was applied. Both unadjusted and Holm adjusted p values were reported. The total score comparison was analysed separately and was not included in the Holm adjustment. A p value of < 0.05 was considered statistically significant for the total score comparison, whereas statistical significance for the item level comparisons was determined using Holm adjusted p values. The internal consistency of the 17 Likert type questionnaire items was assessed using Cronbach’s alpha separately for the mandibular model and flat pad. Effect sizes for the Wilcoxon signed rank comparisons were calculated as r= |Z|/√N, where Z was the standardized Wilcoxon test statistic and N was the number of non-zero paired differences. Effect size estimates were interpreted descriptively.

Open ended responses were reviewed descriptively and grouped according to recurring themes related to model realism, usability, tissue properties and suggestions for improvement.

Results

Participant characteristics

A total of 107 third year undergraduate medical students participated in the study and completed the questionnaire. The mean age of the participants was 21.85 ± 1.39 years, with a median age of 22 years and an age range of 20–27 years. Of the participants, 52 were female (48.6%) and 55 were male (51.4%). Twenty two participants (20.6%) reported previous suturing experience, whereas 85 participants (79.4%) had no previous suturing experience. Thirty participants (28.0%) reported previous experience with suturing on a model or simulation trainer, whereas 77 participants (72.0%) had no such experience (Table 1).

Table 1.

Participant characteristics

Variable	Value
Number of participants	107
Age, mean +/- SD	21.85 +/- 1.39
Age, median	22
Age range	20–27
Female, n (%)	52 (48.6%)
Male, n (%)	55 (51.4%)
Previous suturing experience, n (%)	22 (20.6%)
No previous suturing experience, n (%)	85 (79.4%)
Previous suturing practice on a model / simulation trainer, n (%)	30 (28.0%)
No previous suturing practice on a model / simulation trainer, n (%)	77 (72.0%)

Internal consistency of the questionnaire

The internal consistency of the Likert type questionnaire items was high for both models. Cronbach’s alpha was 0.982 for Model A and 0.983 for Model B (Table 2).

Table 2.

Total questionnaire scores and internal consistency

Measure	Mandibular model	Flat pad	p value	Effect size (r)
Cronbach’s alpha	0.982	0.983	-
Total score, mean +/- SD	66.55 +/- 17.62	68.39 +/- 18.25	-	-
Total score, median [IQR]	69 [63–79]	74 [63–83]	0.060	0.21
Total score, range	17–85	17–85	-

Effect size was calculated as r= |Z|/√ N, where N represents the number of non-zero paired differences

Comparison of total questionnaire scores

The total questionnaire score was slightly higher for Model B than for Model A; however, this difference did not reach statistical significance. The mean total score was 66.55 ± 17.62 for Model A and 68.39 ± 18.25 for Model B. The median total score was 69 [63–79] for Model A and 74 [63–83] for Model B (Wilcoxon signed-rank test, p = 0.060). The effect size for the total score comparison was r = 0.21 (Table 2).

Item level comparison of the mandibular model and flat pad

Unadjusted item level paired comparisons showed higher ratings for the flat pad than for the mandibular model for perceived similarity to a real clinical situation (p = 0.010), realistic needle passage (p = 0.001), suitability for practising different suturing techniques (p = 0.004) and contribution to preparation for clinical suturing (p = 0.037). After Holm correction for the 17 item level comparisons, only realistic needle passage remained statistically significant in favour of the flat pad (Holm adjusted p = 0.014). The differences in perceived similarity to a real clinical situation (Holm adjusted p = 0.146), suitability for practising different suturing techniques (Holm adjusted p = 0.063) and contribution to preparation for clinical suturing (Holm adjusted p = 0.522) did not remain statistically significant. No questionnaire item showed a statistically significant advantage in favour of the mandibular model. Effect sizes for all item level comparisons are presented in Table 3. The largest observed effect was for realistic needle passage (r = 0.52), favouring the flat pad.

Table 3.

Item level comparison of the mandibular model and flat pad

Item	Mandibular model median [IQR]	Mandibular model mean ± SD	Flat pad median [IQR]	Flat pad mean ± SD	Unadjusted p value	Holm adjusted p value	Effect size (r)
Anatomical structure was realistic	4 [3–5]	3.94 ± 1.10	4 [3–5]	3.87 ± 1.19	0.678	1.000	0.06
Three dimensional structure helped understanding suturing practice	4 [3–5]	3.90 ± 1.20	4 [4–5]	4.07 ± 1.16	0.071	0.782	0.26
Provided an experience similar to a real clinical situation	4 [3–5]	3.68 ± 1.19	4 [3–5]	3.91 ± 1.18	0.010	0.146	0.39
Helped gain spatial awareness for suture placement	4 [4–5]	3.83 ± 1.18	4 [4–5]	3.94 ± 1.26	0.126	0.958	0.25
Appearance was similar to real tissue	4 [3–5]	3.86 ± 1.06	4 [4–5]	3.96 ± 1.16	0.114	0.958	0.24
Tactile feel was satisfactory	4 [3.5-5]	3.83 ± 1.07	4 [3–5]	3.91 ± 1.19	0.254	1.000	0.18
Needle passage felt realistic	4 [3–5]	3.71 ± 1.21	4 [3–5]	3.96 ± 1.15	0.001	0.014	0.52
Working with instruments was comfortable and controlled	4 [4–5]	3.91 ± 1.21	4 [4–5]	4.03 ± 1.15	0.131	0.958	0.25
Knot tying practice was possible	4 [4–5]	4.07 ± 1.24	5 [4–5]	4.21 ± 1.13	0.065	0.782	0.31
Suturing was technically feasible	4 [4–5]	4.00 ± 1.22	5 [4–5]	4.12 ± 1.25	0.096	0.958	0.26
Different suturing techniques could be practised	4 [3–5]	3.88 ± 1.22	4 [4–5]	4.12 ± 1.20	0.004	0.063	0.45
Suturing from different angles was possible	4 [4–5]	3.98 ± 1.15	4 [4–5]	4.09 ± 1.19	0.140	0.958	0.23
Difficult-access suturing practice was possible	4 [3.5-5]	3.90 ± 1.26	4 [3–5]	3.87 ± 1.28	0.857	1.000	0.03
Model did not tear or deform easily	4 [4–5]	4.04 ± 1.14	4 [4–5]	3.92 ± 1.35	0.376	1.000	0.15
Helped improve basic suturing skills	4 [4–5]	4.03 ± 1.20	5 [4–5]	4.15 ± 1.26	0.101	0.958	0.29
Contributed to preparation for clinical suturing	4 [4–5]	3.97 ± 1.20	5 [4–5]	4.10 ± 1.27	0.037	0.522	0.35
Overall satisfaction	4 [4–5]	4.03 ± 1.11	4 [4–5]	4.16 ± 1.15	0.056	0.722	0.34

Unadjusted p values were obtained using paired Wilcoxon signed rank tests. Holm adjusted p values were calculated across the 17 item level comparisons to control the family wise error rate. Statistical significance was determined using Holm adjusted p values. Bold type indicates statistical significance after adjustment. Effect sizes were calculated as r= |Z|/√N using the standardized Wilcoxon test statistic and the number of non-zero paired differences. Effect sizes are presented as absolute values; the direction of the difference should be interpreted from the model ratings

Although no item significantly favoured the mandibular model, it received numerically higher mean ratings for anatomical realism, difficult access suturing practice and resistance to tearing or deformation. The differences in these domains were not statistically significant. The mandibular model also received ratings of similar magnitude to those of the flat pad for spatial awareness and overall satisfaction (Table 3).

Open ended responses

Open ended comments were provided by 16 participants. Because the responses were brief and were not collected for formal qualitative analysis, they were reviewed descriptively. Representative comments, translated from Turkish, included: “The mandibular model was more realistic, whereas the other model was more suitable for practice and trying different techniques”; “Both models were suitable for practice and were very good. I do not think either had an overwhelming advantage over the other”; and “A more realistic training experience could be achieved with a model that could be fixed more securely to the surface.” Other comments referred to the training session and included requests for longer practice time and additional suturing materials.

Fabrication output and estimated material cost

During model fabrication, three reusable upper mould components and 20 lower base components were produced. Each silicone mandibular model required approximately 70 g of platinum based RTV-2 silicone and approximately 72 g of polylactic acid filament for the lower base. The estimated material cost per model was about 100 Turkish lira when the reusable upper mould cost was excluded. Costs were calculated according to local material prices in May 2026. The estimated material cost of each model was approximately 2 USD. Mineral oil, used as a release agent and red silicone pigment each cost approximately 1 USD in total; however, less than 5% of each material was used during the study and their contribution to the per model cost was therefore negligible.

Discussion

Principal findings

This study developed an affordable silicone mandibular model for suturing skills training among undergraduate medical students and compared students’ perceptions of the mandibular model with those of a conventional flat silicone suturing pad. The mandibular model was generally well accepted and received ratings similar to those of the flat pad in several domains, including anatomical realism, spatial awareness for suture placement, difficult access suturing practice, durability during practice and overall satisfaction. In the unadjusted item level analyses, the flat pad received higher ratings for realistic needle passage, suitability for practising different suturing techniques, perceived similarity to a real clinical situation and contribution to preparation for clinical suturing. However, after Holm correction for multiple comparisons, only the difference in realistic needle passage remained statistically significant. The total questionnaire score was also slightly higher for the flat pad, although this difference did not reach statistical significance. These findings represent learner reported perceptions obtained immediately after the training session and should not be interpreted as evidence of superior skill acquisition or procedural competence. No objective performance assessment was conducted to determine whether either model improved suturing performance.

Based on the observed learner ratings, the mandibular model should not be interpreted as superior to or as a replacement for conventional flat suturing pads. Although its proposed use as a complementary, anatomically contextualized tool after basic flat pad practice was not directly tested in the present study, this potential role is consistent with evidence that greater simulator fidelity does not necessarily result in superior learning outcomes [23].

Interpretation of findings

Flat silicone pads provide a broad, stable and uniform surface, which may make needle passage, knot placement and repeated practice easier for novice learners. In contrast, the mandibular model introduces a curved anatomical surface and spatial constraints. Although these features were intended to provide a more anatomically meaningful context, they may also increase technical difficulty for beginners. This interpretation is consistent with evidence that task and environmental complexity in simulation based education should be sufficient to promote learning, but not so high that it compromises performance or increases cognitive load [24]. One possible explanation for the lower ratings of the mandibular model in some domains is the greater task complexity introduced by its curved anatomical form, in addition to limitations related to the model material. However, task complexity and cognitive load were not measured in the present study. Therefore this explanation should be regarded as hypothesis generating rather than as an observed finding.

The present study did not evaluate or compare different educational sequences. Nevertheless, based on the relative technical simplicity of flat pads and the anatomical constraints of the mandibular model, a staged progression from flat pad practice to anatomically contextualized practice may warrant prospective evaluation [2, 24]. The finding that no questionnaire item significantly favoured the mandibular model is important. It indicates that anatomical resemblance alone is not sufficient to make a training model educationally superior. For suturing education, haptic properties such as tissue resistance, elasticity, needle passage and stability during knot tying are central to the learner experience. Similar issues have been reported in previous studies evaluating three dimensional or synthetic simulators, where visual realism and educational value were often rated positively, while tactile feedback and tissue behaviour remained challenging aspects requiring further refinement [3].

Comparison with previous studies

The present findings are consistent with the broader simulation literature suggesting that greater anatomical or physical realism does not automatically translate into superior learner ratings or educational outcomes. Norman et al. reviewed studies comparing high fidelity and low fidelity simulation and reported that the relationship between simulator fidelity and transfer of learning was limited rather than linear [25]. Similarly, Hamstra et al. argued that simulator fidelity should not be understood only as the extent to which a simulator looks, feels or acts like a human patient; instead, simulation design should focus on the functional features required for the intended learning task [26]. This perspective helps explain why the anatomically contextualized mandibular model in the present study did not outperform the flat pad, despite being derived from a real human mandible.

The present results are also compatible with previous evidence on task complexity in simulation based education. Haji et al. showed that task complexity can influence novice learners’ cognitive load and simulation based learning outcomes [27]. In our study, the mandibular model introduced a curved and spatially constrained working surface, whereas the flat pad offered a broad, stable and technically permissive surface. Therefore, the higher ratings of the flat pad in some technical domains may reflect the suitability of simpler models for early skill acquisition, rather than a lack of educational value in the mandibular model.

To provide a structured comparison with the existing literature, Table 4 summarizes selected studies that are particularly relevant to the present work in terms of simulator characteristics, participant populations, outcome measures and principal findings [3, 4, 6, 11, 15, 21]. Considerable methodological heterogeneity is evident across these studies. Hanisch et al., Smail et al. and Gholami et al. relied primarily on learner or expert evaluations, although Gholami et al. also assessed the dimensional accuracy of the printed model [3, 6, 21]. In contrast, Zhu et al. and Wright et al. incorporated objective measures of procedural performance, including completion time, structured performance scores, knowledge assessment and Objective Structured Assessment of Technical Skills scores [11, 15]. Across the oral and dental simulation studies, anatomical or visual realism was generally evaluated favourably, whereas soft tissue behaviour and haptic fidelity remained recurring limitations [3, 6, 21]. The present study adds a paired, perception based comparison of an anatomy based mandibular model and a conventional flat suturing pad but does not provide evidence of objective skill acquisition.

Table 4.

Comparison of selected simulation studies relevant to suturing and anatomically contextualized surgical training

Study	Simulator characteristics and training context	Participants	Outcome measures	Principal findings
Macluskey et al. [4]	Four models: tabletop silicone dental model, restricted access tabletop model, phantom head mounted model and Thiel cadaver; suturing and mucoperiosteal flap training	67 fourth year dental students; 46 provided feedback after the first session and 28 after the second	Primarily subjective preparedness, confidence and model preference ratings; a subsequent formative checklist assessment was not designed as a direct objective comparison of the four models	The Thiel cadaver was rated as the best model for suturing and flap training. Confidence regarding mucoperiosteal flap procedures remained relatively limited.
Hanisch et al. [6]	Patient specific multimaterial 3D printed apicoectomy model with a silicone gingival mask, compared with a commercial typodont	35 students evaluated the typodont and 33 evaluated the 3D printed model	Subjective visual analogue scale ratings; no objective procedural performance assessment	The 3D printed model was not rated as inferior to the typodont. Tearing and limitations of the gingival mask were reported.
Smail et al. [3]	Multimaterial PolyJet simulator for gingival incision, flap elevation and suturing	69 participants: 27 undergraduate students, 19 postgraduate learners and 23 expert oral surgeons	Eleven item five point Likert questionnaire and free text feedback; no objective performance assessment	Visual realism and perceived educational usefulness were rated positively. Tissue adhesion, detachment, thickness and suture resistance required refinement.
Zhu et al. [11]	3D printed intestinal anastomosis models compared with a basic suturing pad in a randomized training design	16 sixth year medical students	Objective procedural time and structured performance scores, together with post training questionnaire findings	The 3D model group demonstrated improved performance and outperformed the pad group in the later training stage.
Wright et al. [15]	Multimaterial craniofacial simulator for fronto orbital advancement osteotomy	Five junior and five senior plastic surgery residents	Knowledge Assessment, OSATS, Global Rating Scale and subjective simulation experience scale	Junior residents demonstrated significant improvements in knowledge, OSATS and global performance scores after training. Soft tissue realism remained an area for development.
Gholami et al. [21]	Patient specific maxillary model reproducing printed hard tissue and simulated soft tissue for gingival flap surgery	50 final year dental students and 18 periodontists	Objective cloud to cloud dimensional accuracy assessment; subjective VAS ratings and free text feedback; no objective procedural performance assessment	The model showed high dimensional accuracy. Students rated it more favourably than periodontists, while experts identified limitations in soft tissue tactile feedback.
Present study	Anatomy based silicone mandibular model derived from a dry human mandible and compared with a conventional flat silicone suturing pad	107 third year medical students	Paired learner reported Likert ratings, open ended feedback and Holm adjusted item level comparisons; no objective procedural performance assessment	Total scores did not differ significantly. After Holm correction, only realistic needle passage was rated significantly higher for the flat pad. The mandibular model was affordable, reproducible and openly shareable.

Abbreviations: 3D three dimensional, VAS Visual analogue scale, OSATS Objective Structured Assessment of Technical Skills. The table presents selected studies for contextual comparison and does not constitute a systematic review

The model had an estimated material cost of approximately 2 USD, excluding the reusable upper mould, and its digital files were deposited openly to support reproduction and adaptation. These practical advantages do not establish educational effectiveness and should be considered alongside haptic suitability and objectively measured learning outcomes [9, 14, 20, 28].

Overall, the findings support a task oriented interpretation: simulation tools should be selected according to the learning objective, learner experience and task complexity rather than physical realism alone [1, 24–28].

Strengths and limitations

This study has several strengths. First, the mandibular model was developed from a real dry human mandible, providing an anatomy based framework rather than a purely generic or flat training surface. Second, all participants evaluated both models using the same questionnaire items, allowing paired comparison of learner perceptions. Third, students were not informed that Model A had been developed by the author for this study before completing the questionnaire, which may have reduced expectation and novelty bias. Fourth, the model was designed to be affordable and reproducible and the digital files and fabrication guide were made openly available through Zenodo to support external use and adaptation.

The study also has limitations. It was conducted at a single institution and included only third year medical students, which limits the generalizability of the findings to other learner groups, disciplines and educational settings. Accordingly, the findings should not be extrapolated to dental students, residents, experienced practitioners or clinical performance without direct evaluation in those populations. The questionnaire was developed by the author with reference to previous simulation model evaluation studies and was reviewed by research assistants, graduate students and medical specialists within the department for clarity and comprehensibility. The wording and organization of the items were revised based on their feedback. However, this process did not constitute formal psychometric validation, and the questionnaire was not subjected to formal content validity assessment, construct validity testing or pilot testing in an independent learner sample. Although internal consistency was high, Cronbach’s alpha alone does not establish questionnaire validity, and the very high alpha values may partly reflect conceptual overlap among some items. In addition, the study did not include structured evaluations from the instructors who delivered the training. Therefore, the findings reflect only the students’ perceptions and do not capture faculty perspectives on the models’ instructional usability, technical limitations or suitability for teaching.

The outcomes were based on learner perceptions collected immediately after the training session rather than on objective assessments of suturing performance, skill acquisition or long term retention. No procedural checklist, instructor rated performance score, time to completion measure, suture placement accuracy assessment or practical examination result was available for analysis. Therefore, the study cannot determine whether practice on either model improved procedural competence or resulted in superior skill acquisition. The findings should be interpreted as a comparative evaluation of learner perceptions rather than as evidence of educational effectiveness.

The study also did not evaluate transfer of learning to clinical wound repair or performance on real tissues. The mechanical properties of the silicone material, such as elasticity, tear resistance and needle penetration force, were not objectively measured.

Although the alternating classroom arrangement generated both possible model sequences, allocation to the starting model was not randomized because students selected their seats freely, and the individual starting sequence was not recorded as a participant level variable. The crossover procedure may have reduced systematic order effects; however, it could not eliminate potential influences of practice, increasing familiarity with the task, adaptation to the instruments or fatigue. Because participant level sequence data were unavailable, these potential effects could not be examined statistically, which limits the internal validity of the comparison.

Finally, although Holm correction was applied to control the family wise error rate, the item level analyses were exploratory and the study was not specifically powered for individual questionnaire item comparisons. These findings should therefore be interpreted cautiously.

Future research

Future studies should evaluate the mandibular model using objective performance based outcomes in addition to learner perceptions and should include structured feedback from instructors and experienced practitioners to assess the models from teaching and professional perspectives. An Objective Structured Assessment of Technical Skills (OSATS) based checklist, instructor rated suturing performance score, time to task completion, suture placement accuracy, knot security and model deformation during practice could provide more direct evidence of educational effectiveness. Objective evaluation frameworks have previously been applied in three dimensionally printed surgical simulation using OSATS, global rating scales and knowledge assessments [15], while randomized suturing research has incorporated procedural time and structured performance scores [11]. Future comparative studies should use a formally randomized and counterbalanced crossover design, record each participant’s starting sequence and include sequence or order effects in the statistical analysis.

Further refinement should focus on haptic properties and model fixation, including the comparison of silicones with different Shore hardness values or multilayer material designs [16]. Future research should prospectively test whether staged use of the model after basic flat pad training provides any educational advantage. The possible usefulness of the mandibular model for dental students, dentists, oral and maxillofacial surgery residents and other surgical trainees remains a hypothesis and was not evaluated in the present study. Future studies should directly assess these populations to determine whether greater clinical experience or familiarity with anatomically constrained oral procedures influences their perceptions or performance. Multicentre studies involving different institutions, disciplines and learner groups would further clarify the generalizability and educational role of the model.

Conclusions

This study developed an affordable and reproducible silicone mandibular model for suturing skills training among undergraduate medical students. The mandibular model did not demonstrate superiority over the conventional flat silicone suturing pad in students’ evaluations. After correction for multiple comparisons, the flat pad received a significantly higher learner rating only for realistic needle passage. Nevertheless, the mandibular model received learner ratings of similar magnitude for anatomical realism, spatial awareness, difficult access practice, durability and overall satisfaction. These perception based findings suggest that the mandibular model may have potential as a complementary, anatomically contextualized training resource rather than as a replacement for conventional flat pads. Further studies using objective performance measures and formally randomized, counterbalanced crossover designs, with participant level sequence recording and evaluations from instructors and experienced practitioners, are needed to determine its effect on suturing competence and skill transfer.

Author contributions

Ömer Can KIZILAY, conceived and designed the study, developed the silicone mandibular model, collected and analysed the data, interpreted the findings and wrote and approved the final manuscript.

Funding

This research received no specific grant from any funding agency in the public, commercial or not for profit sectors.

Data availability

The digital mandibular model files are openly available in Zenodo at https://doi.org/10.5281/zenodo.20135650 .

Declarations

Ethics approval and consent to participate

This study was approved by the Non Interventional Clinical Research Ethics Committee of Kütahya Health Sciences University, Türkiye (decision no: 2026/06–17; date: 04 May 2026). All methods involving human participants were carried out in accordance with relevant guidelines and regulations, including the Declaration of Helsinki and its subsequent amendments. Informed consent was obtained from all participants before participation. Participation was voluntary and anonymous data were collected.

Consent for publication

Written consent for publication was obtained from the students shown in Fig. 4. The students’ faces were blurred to protect privacy.

Competing interests

The authors declare no competing interests.