Evaluating the educational impact of tomography-based high-resolution 3D-printed distal radius fracture models Authors: Ramazan Kurul, Busra Inal, Murat Diramali, Hamza Ozer, Kevser Ezgi Erdogan Published: 2025 ABSTRACT AbstractBackgroundDistal radius fractures (DRFs) are among the most common orthopedic injuries and are often challenging for students to fully understand when taught using only 2D radiographs or healthy models. This study aimed to evaluate the educational impact of tomography-based 3DP DRF models on medical students’ knowledge acquisition and anatomical understanding compared with conventional teaching methods.MethodsThe study was designed as a parallel-group, single-blinded prospective randomized controlled trial conducted over two days of structured education. Ninety-four second year physiotherapy students (19.62 ± 1.2 years; 45 F, 49 M) were enrolled and randomly assigned into intervention (n = 47) and control (n = 47) groups. The intervention group received tomography-based fracture 3DP training, while the control group received a healthy 3D-printed model. Primary outcomes included theoretical knowledge (MCQs) and practical case-solving. Secondary outcomes were Test-Taking Motivation Questionnaire (TTMQ); Test Anxiety Inventory (TAI), puzzle assembly time, satisfaction, and willingness-to-recommend ratings. Assessments were conducted pre- and post-training, with knowledge retention re-assessed at one month after education. Statistical analyses were performed using independent samples t-tests and chi-square tests for between-group comparisons, paired t-tests for within-group changes.ResultsThe intervention group included 47 students (19.51 ± 1.1 years; 36 F, 11 M), while the control group consisted of 47 students (19.74 ± 1.3 years; 33 F, 14 M). Demographic characteristics were similar between the groups. In the 3D group, theoretical knowledge scores improved (50.68 [45.34, 55.93], p < .001), TTMQ scores increased (28.06 [24.22, 31.91], p < .001), and TAI scores decreased (–10.89 [–14.33, − 7.46], p < .001). Similarly, in the control group, theoretical knowledge scores improved (49.46 [44.86, 54.03], p < .001), TTMQ scores increased (21.60 [17.13, 26.06], p < .001), and TAI scores decreased (–9.70 [–14.35, − 5.05], p < .001). Post-training comparisons showed that the intervention group achieved significantly higher scores in case-solving (–18.55 [–23.87, − 13.23], p < .001), MCQ performance (4.17 [0.15, 8.18], p = .042), puzzle assembly time (11.34 [4.24, 18.44], p = .002), TTMQ (8.04 [3.20, 12.87], p = .030), and TAI (–4.42 [–7.54, − 1.30], p = .030). No significant between-group differences were observed for color recognition (p = .301), weight identification (p = .161), detail recognition (p = .669), or knowledge retention (p = .160).ConclusionsWithin the limits of this study, tomography-based fracture-specific 3D printed models led to greater improvements in knowledge, motivation, reduced anxiety, and better practical performance compared with conventional teaching methods. These models showed educational benefits; further research is needed to confirm long-term and clinical impacts.Trial registrationclinicaltrials.gov NCT06061003 24/09/2023.Supplementary InformationThe online version contains supplementary material available at 10.1186/s12909-025-08164-w. Introduction Distal radius fracture (DRF) is the most common upper extremity fracture in adults, involving a complex multiarticular system and presenting with diverse fracture patterns [1]. Beyond their high incidence, DRFs are linked to substantial morbidity and increased mortality in older populations, underscoring the need for effective educational strategies in this field [2]. The choice of treatment depends largely on the specific type of fracture [2]. DRFs are often associated with significant wrist deformities and complications, which can occur in 6% to 80% of cases, depending on the severity of the injury and the treatment method used. These complications may result in altered wrist biomechanics, pain, and long-term disability [1]. In medical education, learning about DRF is therefore crucial. Conventional instruction relies on textbooks, radiographic and ultrasound images, cadavers, and 2D illustrations. Although cadavers are still considered the most appropriate source of tissue for manipulation in clinical practice, there are many religious, ethnic, legal, cultural, social, and financial barriers to their use [3]. These limitations highlight the need for alternative educational tools. The development of technology has made 3D printers more accessible and cost-effective, enabling their use in schools and even at home. 3D printing (3DP) has been commonly used in medical education, particularly in neuroanatomy, cardiac, and abdominal anatomy, where it has demonstrated superiority over traditional methods regarding test scores and student satisfaction [4–6]. More broadly, 3DP has been used for student and residency education [4, 7], patient training [8], and pre-surgical planning [9], with applications across multiple anatomical regions including the spine, skull, pelvis, and extremities. Recent advances have also extended its use to fracture models such as the tibial plateau, acetabulum, and distal radius [8, 10, 11]. Alongside augmented and virtual reality (VR), high-fidelity 3D-printed models represent a shift toward immersive and interactive learning tools. These technologies improve spatial understanding and increase student motivation [12, 13]. Furthermore, the integration of serious game principles within these virtual platforms has been shown to increase student motivation and engagement [14]. The use of 3D-printed models in anatomy and clinical education can also be conceptualized within the framework of established learning theories such as Cognitive Load Theory and Kolb’s Learning Theory. Printed models may reduce cognitive load by allowing learners to easily manipulate and visualize complex anatomical structures [15, 16]. According to Kolb’s theory, active participation and learning through hands-on engagement is more effective than passive approaches, and 3D-printed models can support this process [17]. Despite these advantages, research specifically addressing fractures remains scarce in undergraduate education [10, 18]. Although 3D models have demonstrated clinical utility for patient care, their role in medical education is less established, particularly for tomography-based DRF models. A recent randomized clinical trial reported the usefulness of 3D models for intra-articular DRF education but also emphasized that evidence in undergraduate teaching remains limited [11]. Tomography-based DRF models offer additional benefits, including improved visualization, diagnostic accuracy, and support for surgical planning, with some studies reporting reductions in fluoroscopy use and operative time [19]. However, their production requires advanced imaging and printing infrastructure, which increases cost and preparation time, and long-term clinical benefits remain unclear [20, 21]. To date, no study has evaluated whether tomography-based, fracture-specific 3D printed distal radius models improve undergraduate students’ anatomical comprehension and clinical reasoning compared with healthy 3D models in physiotherapy education. We hypothesized that exposure to fracture-specific 3D printed DRF models would improve anatomical comprehension and learning motivation in comparison with healthy 3D models. Materials and methods Study design and participants This study was a parallel-group, prospective, assessor-blinded, randomized controlled trial conducted at Bolu Abant Izzet Baysal University, Department of Physiotherapy and Rehabilitation, Turkey. The study was carried out between September 2024 and December 2024. (clinicaltrials.gov NCT06061003 24/09/2023). The study design is shown in Fig. 1. This study was approved by the Ethics Committee of Bolu Abant Izzet Baysal University (2023/295–410) and all participants signed informed consent. The inclusion criteria were willingness to participate and successful completion of prior anatomy training. Participants were excluded if they had previously failed the anatomy course (score < 30/100), had attended clinical courses on wrist and elbow fracture rehabilitation, had a history of fractures in this region, or did not complete the questionnaires and practical exam. Flow Diagram Illustrating the Study Design and Components of the Educational Intervention MCQs: Multiple choices questionnaires, TTMQ: Test Taking Motivation Questionnaire, TAI: Test Anxiety Inventory A total of 108 students were enrolled; 14 were excluded due to incomplete assessments, leaving 94 participants randomized to either the intervention group (n = 47) or the control group with healthy DRF 3DP models (n = 47). Stratification was based on sex, Purdue Spatial Visualization Test – Rotations (PSVT-R) scores [22], hand laterality, and general upper extremity anatomy exam scores to ensure baseline equivalence across critical cognitive and demographic variables. A computer-generated randomization sequence was created by an independent researcher (K.E.E.) not involved in outcome assessments. Group assignments were concealed using sequentially numbered, opaque, sealed envelopes. Upon enrollment, envelopes were opened in sequence to allocate participants. Because groups used different models, participant blinding was not feasible. However, all outcome assessments were performed by two investigators (R.K. and M.D.) who were blinded to group allocation throughout data collection and analysis. Students’ anatomical knowledge of the upper extremity was assessed using 30-item multiple-choice questionnaires (MCQs) prior to the study, in order to eliminate knowledge retention bias. Sample size calculation was performed using G*Power, based on prior studies of 3D-printed skeletal models that used theoretical exam scores, with an estimated effect size of 0.56 [23]. It was determined that a total of 82 participants were required to achieve 80% power with a significance level of p <.05. To account for a potential dropout rate of 10%, 90 students were targeted for recruitment. 3D model Preparation and printing Computed Tomography (CT) scans of a healthy distal upper extremity, with a section thickness of 0.625 mm, were extracted as DICOM files. The 3DSlicer 4 [24] software was used to segment the DICOM files and generate Standard Triangle Language (STL) files. The bone models were reviewed for anatomical accuracy by both an anatomist and a radiologist. The radiologist then retrospectively examined the CT images to identify the most common fracture lines for each type of fracture. These fracture lines were marked on the layered STL images using Blender (Blender Foundation, Amsterdam, the Netherlands), creating a model representing the most common fracture line for each scenario. Fracture types were categorized into eight models based on the Frykman classification, along with Smith, Colles, Chauffeur, Barton, and Hutchinson fracture models. These fractures, which are common in the region and require different rehabilitation programs, were developed as case-based learning models (Fig. 2). 3D-Printed Models Showing Different Colored Representations of a Healthy Radius Bone (A) Beige Resin, (B) Vanilla Resin, (C) White Resin, (D) Grey Resin Chitubox 1.9 software (Chitubox Inc, Guangdong, China) was used to prepare the STL file for printing, with a layer height of 0.05 mm and 30% hollowing of the model to approximate the real bone weight. A resin printer (Elegoo Saturn 3, Elegoo Inc, HK, China) was used, with an exposure time of 2.5 s per layer. The models were washed with 95% isopropyl alcohol and cured with ultraviolet light for 10 min. Colors similar to those used in standard plastic teaching models were selected as references for printing. Printed models were rated by students with visual analog scale for color, anatomical accuracy, level of detail, and texture. The students were asked to rate the clarity and distinguishability of anatomical details based on the color of the printed models, 7 students (7.45%) preferred gray, 21 (22.34%) preferred beige, 40 (42.55%) selected vanilla, and 26 (27.66%) chose white. Based on student preferences, the vanilla-colored resin, which was rated highest for anatomical clarity and visual appeal, was selected for model printing (Fig. 3). Representative 3D-Printed Models of Various Radius and Ulna Fracture Types (A) Ulnar Shaft Fracture, (B) Ulnar Styloid Fracture, (C) Ulnar Comminuted Fracture, (D) Radius Comminuted Fracture, (E) Smith or Colles Fracture depending on the displacementdirection, (F) Barton Fracture, (G) Radius Intraarticular Fracture, (H) Radius Styloid Fracture Educational program All students attended an extracurricular 60-minutes lecture on wrist fractures, accompanied by a slideshow presentation. Then, students in the intervention group used fracture-specific 3DP models. Each student was allotted 30 min to manipulate the models. Students in the control group received healthy 3DP models of the radius and ulna. The following day, all students participated in a 2-hour lecture followed by a 1-hour practical, case-based decision-making session focused on rehabilitation principles for wrist fractures. Outcome assessments All students received practical and theoretical training over the course of two days. Following a three-hour break after the training sessions on the second day, all students were asked to complete the primary outcomes (MCQ-based theoretical knowledge test and case-solving practical exam) as well as the secondary outcomes (TTMQ, TAI, knowledge retention, puzzle assembly task, and satisfaction/recommendation ratings). To assess knowledge retention, students’ MCQ scores were reassessed one month after the post-tests. The MCQ was developed using exam questions from the lecturer’s question pool over the past five years. A pool of 182 MCQs was assessed through item analysis, and those with exceeding 0.5 were from item analysis marked for selection. Then, an expert committee of two members with over 10 years of experience in the subject chose from these questions. If unanimity was not achieved among the committee members, the final decision was made through discussion and consensus. Thirty of these items were selected according to Bloom’s taxonomy levels, and the test was administered to fourth-year students, as they had already completed this subject, for the purpose of factor analysis. Then, 25 test items were selected from question pools that had undergone thorough item analysis, including evaluation of lower and upper group scores. Practical exam questions were designed as 16 unique cases covering all fracture types, with subsections addressing basic regional anatomical knowledge, indications/contraindications, assessment capabilities, and clinical decision-making in accordance with the domains of the revised Bloom’s taxonomy [25]. TTMQ was used to measure students’ motivation levels. This questionnaire consists of 22 items across 7 subsections: effort, expectations, importance, interest, utility, and test anxiety. All items were rated on a 4-point Likert scale, ranging from 1 (strongly disagree) to 4 (strongly agree). Total scores ranged from 22 to 88, and the questionnaire has demonstrated high validity and reliability among university students [26, 27]. To assess changes in students’ anxiety levels during exams, a short version of the 5-item TAI was used. This inventory includes subsections that focus on worry and emotionality domains (ease of concentration, comfort with learning materials, and the alignment of exam questions) with course content. All items are rated on a 4-point Likert scale, ranging from 1 (rarely), 2 (Sometimes), 3 (Often) to 4 (Always). Scores on the test range from 20 to 80 [28]. Acceptable levels of discriminant validity and internal reliability were observed for the test’s worry, emotionality, and total score components [29]. Participants completed a feedback questionnaire on the 3D-printed learning models, using a visual analog scale (0 = not at all, 100 = extremely) to rate their satisfaction and likelihood of recommending the models to peers. In addition, all students evaluated the weight and anatomical detail of the 3D-printed models in comparison with cadaveric radius and ulna bones, also using a visual analog scale. Participants were asked to match each fractured bone part correctly, and the total time required to complete the entire assembly was recorded in seconds. Statistical analyses The chi-square test was used to compare the distribution of sex and model colors. The Shapiro-Wilk test was used to assess the normality of continuous variables. Age, general anatomy scores, PSVT-R, TTMQ, TAI and hand laterality scores were found to follow a normal distribution. Paired-sample t tests were used for within-group pre–post comparisons, and independent-sample t tests for between-group differences. The interrater reliability of the Bloom’s taxonomy classifications was tested with Krippendorff’s alpha due to the nominal (dichotomous) nature of the coding. The content validity of the MCQ test was evaluated by an expert committee and its internal consistency reliability was examined using Cronbach’s alpha. The suitability of the data for factor analysis was assessed with the Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy and Bartlett’s test of sphericity. A p-value of less than 0.05 and Cronbach’s α ≥ 0.70 were considered statistically acceptable. All statistical analyses were performed using the SPSS statistical package for Windows, version 24 (IBM Inc., Armonk, NY, USA). Results A total of 108 students were enrolled; but 14 of these students did not complete the assessment process and they were excluded from the study. 94 students were included in the final analysis (Fig. 4). Flowchart of participant enrollment, allocation, follow-up, and analysis The demographic and baseline characteristics of the participants in both the 3D printed and control groups are shown in Table 1. There were no significant differences between the 3D-printed and control groups in age (p =.361), spatial visualization (PVST-R, p =.317), hand laterality (p =.258), general anatomy knowledge (p =.321), MCQ (p =.321), GPA (p =.069) sex distribution (p =.484), TTMQ (p =.544), or TAI (p =.489), confirming baseline homogeneity. Normality of all outcome variables was confirmed by the Shapiro–Wilk test (MCQ, p =.357; TTMQ, p =.110; TAI, p =.617; case solving, p =.191; knowledge retention, p =.209; puzzle time, p =.210; satisfaction, p =.059; recommendation, p =.060). Table 1Demographic and baseline information of studentsIntervention (n = 47)Control (n = 47)X ± SDX ± SDtpAge (years)19.51 ± 1.1019.74 ± 1.350.9170.361PVST-R20.46 ± 5.7921.57 ± 4.821.0060.317Hand Laterality29.27 ± 8.8431.36 ± 8.901.1390.258General Anatomy Knowledge Scores57.26 ± 18.0160.65 ± 14.720.9980.321Grade point average3.13 ± 0.263.02 ± 0.35−1.8400.069Sexn (%)n (%)χ²pMale11 (23.4)14 (29.8)0.4900.484Female36 (76.6)33 (70.2)PSVT-R Purdue Spatial Visualization Test Rotations, t independent samples t test, χ² chi-square test, p <.05 Demographic and baseline information of students PSVT-R Purdue Spatial Visualization Test Rotations, t independent samples t test, χ² chi-square test, p <.05 Table 2 shows the pretest and posttest comparison of MCQs, TTMQ, and TAI scores within groups. In the control group, MCQ scores improved significantly (mean difference = 49.46, 95% CI [44.86–54.03], p <.001), TTMQ scores increased (21.60, 95% CI [17.13–26.06], p <.001), and TAI scores decreased (–9.70, 95% CI [–14.35 to − 5.05], p <.001). In the 3D-printed group, MCQ scores also improved significantly (50.68, 95% CI [45.34–55.93], p <.001), TTMQ scores increased (28.06, 95% CI [24.22–31.91], p <.001), and TAI scores decreased (–10.89, 95% CI [–14.33 to − 7.46], p <.001). Table 2Results of Pre- and Post-Intervention scores within groupsGroupPre-testX ± SDPost-testX ± SDMean Difference95% CI [Lower, Upper]pMCQsIntervention (n = 47)31.14 ± 13.7681.95 ± 19.3950.63[45.34, 55.93]< 0.001Control (n = 47)28.42 ± 13.0777.78 ± 10.2049.44[44.86, 53.03]< 0.001TTMQIntervention (n = 47)30.85 ± 13.7058.91 ± 10.9428.06[24.21, 31.91]< 0.001Control (n = 47)29.27 ± 11.4950.87 ± 12.6021.59[17.12, 26.06]< 0.001TAIIntervention (n = 47)38.51 ± 9.4727.61 ± 6.38−10.89[−14.32, −7.46]< 0.001Control (n = 47)41.74 ± 12.3732.04 ± 8.65−9.70[−14.35, −5.05]< 0.001MCQs Multiple-Choice questionnaires, TTMQ Test-Taking Motivation Questionnaire, TAI Test Anxiety Inventory, Paired sample t-test, p <.05 Results of Pre- and Post-Intervention scores within groups MCQs Multiple-Choice questionnaires, TTMQ Test-Taking Motivation Questionnaire, TAI Test Anxiety Inventory, Paired sample t-test, p <.05 The comparison of knowledge retention between the post-test and the one-month follow-up showed scores decreased significantly at one month in both groups: in the control group by 4.77 points (95% CI [3.17–6.30], p <.001), and in the intervention group by 3.02 points (95% CI [1.13–4.91], p =.002). Post-intervention comparisons are presented in Table 3. The intervention group outperformed the control group in several outcomes. MCQ scores were higher in the intervention group (4.17, [0.15–8.18], p =.042). Similarly, TTMQ scores favored the intervention group 8.04, [3.20–12.87], p =.001). TAI scores were lower, indicating reduced anxiety, in the intervention group compared to controls (–4.42 [–7.54 to − 1.30], p =.006). Case-solving performance was also higher in the intervention group (–18.83 [–23.8 to − 13.2], p <.001). Table 3Comparison of outcome measurements between the groupsIntervention (n = 47)Control (n = 47)X ± SDX ± SDMean Difference 95% CI [Lower, Upper]tpEffect sizeMCQs-post81.95 ± 9.3977.78 ± 10.204.17 [0.15, 8.18]−2.0610.0420.425TTMQ58.91 ± 10.9450.87 ± 12.608.04 [3.20, 12.87]−3.3020.0010.681TAI27.61 ± 6.3832.04 ± 8.65−4.42 [−7.54, −1.30)2.8200.0060.582Case solving79.89 ± 14.3561.06 ± 11.92–18.83 [–23.8, − 13.2]−6.917< 0.0011.427Knowledge Retention−3.02 ± 6.44−4.85 ± 5.461.83 [−0.69, 4.18]−1.4710.1600.306Satisfaction VAS (cm)7.82 ± 1.185.11 ± 1.63−2.71 [−3.29, −2.12]−9.233< 0.0011.905Recommendation VAS (cm)7.63 ± 1.374.93 ± 1.57−2.70 [−3.31, −2.09]−8.862< 0.0011.839Puzzle Time (sec)98.59 ± 12.29109.93 ± 21.2011.34 [4.23, 17.44]3.1710.0020.654Weight VAS (cm)8.34 ± 1.078.65 ± 1.060.31 [−0.12, 0.74]1.4130.1610.290Detail VAS (cm)8.14 ± 1.448.27 ± 1.480.13 [−0.47, 0.73]0.4290.6690.080MCQs Multiple-choice questionnaires, TTMQ Test Taking Motivation Questionnaire, TAI Test Anxiety Inventory, independent sample t-test, p <.05 Comparison of outcome measurements between the groups MCQs Multiple-choice questionnaires, TTMQ Test Taking Motivation Questionnaire, TAI Test Anxiety Inventory, independent sample t-test, p <.05 In terms of practical performance, puzzle assembly time was significantly faster in the intervention group (11.34 [4.23–17.44], p =.002). Satisfaction levels were markedly higher in the intervention group (–2.71 [–3.29 to − 2.12], p <.001), as was willingness to recommend (–2.70 [–3.31 to − 2.09], p <.001). No significant between-group differences were observed for knowledge retention (1.83 [–0.69 to 4.18], p =.160), weight perception (0.31 [–0.12 to 0.74], p =.161), or detail recognition (0.13 [–0.47 to 0.73], p =.669). Discussion This study shows that the use of fracture-specific 3D-printed models enhanced students’ learning outcomes across cognitive (knowledge), affective (motivation, anxiety), and psychomotor (case-solving) domains. While both groups improved post-intervention, the 3D model group demonstrated significantly greater gains, suggesting added value from pathology-specific, hands-on anatomical tools. The intervention group students had higher theoretical and motivation questionnaire scores compared to the control, although both groups improved after the education. Similar to our study, previous studies have reported improved test scores with 3DP education compared with 2D images [6], cadavers [30], textbooks [31], CT [32], and anatomical atlases [33]. Also, one study showed that students receiving 3DP education achieved higher fracture knowledge scores than those trained with traditional methods [34]. The higher scores in the intervention group may be attributed to the more realistic nature of the 3D-printed distal fracture model compared to healthy 3DP. In addition to demonstrating the positive effects of fracture-specific 3DP models on students’ anatomical learning, we believe that providing a hands-on experience of distal radius fracture such as the fracture structure, fracture lines, and anatomical location offers significant value from a physiotherapy and rehabilitation student’s perspective. We found that case-solving scores were higher in the intervention group, which is consistent with the literature. Jiang et al., found that integrating 3DP with problem-based learning had enhanced clinical reasoning skills [35]. Similarly, another study reported that training with 3DP models improved clinical planning and reasoning skills in dental students [36]. These findings may be interpreted through Cognitive Load Theory and Kolb’s Experiential Learning Theory. By providing tangible, three-dimensional representations of fractures, the models help reduce extraneous cognitive load and facilitate mental reconstruction of fracture patterns, thereby supporting rehabilitation planning. For the intervention group, we observed higher motivation scores, lower anxiety levels, and greater satisfaction, which aligns with previous studies [37, 38]. Case-based learning has been shown to increase confidence, satisfaction [39] improve interaction [40], and more enjoyable learning experiences [18]. The ability to directly observe fracture patterns and receive tactile as well as visual feedback may explain these benefits by deepening understanding, supporting realistic planning of rehabilitation strategies, and fostering mastery of theoretical knowledge and basic clinical skills. Collectively, these processes may strengthen confidence, enhance motivation, and reduce anxiety. However, it’s important to note that high satisfaction scores might be partially driven by the novelty of using 3D technology rather than by its intrinsic educational superiority. Future studies should aim to isolate the novelty effect by incorporating longitudinal designs. Although immediate post-training outcomes favored the intervention group, knowledge retention assessed one month later was similar in both groups, suggesting that both teaching methods were similarly effective in long-term recall of anatomical information. However, both groups exhibited a similar reduction at the one-month follow-up. The absence of a significant difference in the knowledge retention test may be explained either by the strong initial learning gains in both groups or by the possibility that the elapsed time was insufficient to capture knowledge decay. Similar to our findings, one study found no difference in knowledge retention between 3DP heart models and conventional methods [41] while another studies showed better knowledge retention in craniosynostosis anatomy and 3DP fracture education [42, 43]. One possible explanation for this discrepancy is that, unlike previous studies which compared 3DP models with conventional education, our study compared two different types of 3DP-based interventions. This may suggest that once key concepts are internalized, the format of learning plays a less prominent role in long-term memory. Additionally, the retention assessment time points across studies were heterogeneous, which may have influenced the outcomes. We showed that satisfaction and enjoyment further favored the intervention group. Consistent with prior reports, students expressed higher satisfaction with fracture-specific 3DP education compared to conventional methods [44, 45]. Neijhoft et al. specifically noted that students in 3DP group expressed greater enjoyment compared with traditional methods [46]. These results may be explained by the multisensory experience provided by fracture-based 3DP models, which more realistically integrate tactile, visual, and spatial modalities compared to healthy 3DP models. In terms of practical performance, the intervention group completed puzzle assembly tasks approximately 10% faster, which may reflect both improved case understanding and reduced anxiety. This aligns with Zeng et al., who reported that combining 3DP models with case-based learning improved theoretical and practical performance and fostered greater confidence [40]. Our study also highlights technical and practical considerations. With decreasing printer costs and improved efficiency, orthopedic and anatomical education increasingly integrates 3DP. However, important limitations remain, including model durability, filament layer thickness, rough surface texture, and anatomical accuracy issues [47, 48]. According to Zhou et al., the primary limitation of filament-printed models was the lack of anatomical detail and the inability to replicate the biomechanical properties of real tissue and they could not replicate tactile properties of real bone [49]. To address these, we employed resin-based stereolithography (SLA), producing anatomically detailed models comparable to real bones. Resin-based SLA models offered detailed structures; while not equivalent to cadaveric bone, they provided high anatomical fidelity for teaching purposes. A notable strength of this study is the use of standardized STL files for both fracture and healthy models, ensuring structural consistency. Additionally, the models were highly cost-effective, averaging $1.80 per pair compared with commercial models priced at $5.40–$24.99. This underscores the scalability and accessibility of in-house SLA printing for educational purposes. This substantial cost disparity highlights the economic advantage of in-house 3D printing for educational applications. The ability to produce anatomically accurate, weight-adjusted models at a fraction of the commercial cost reinforces the practicality, accessibility, and scalability of this approach, particularly in physiotherapy and medical education settings. Limitations This study has several limitations. The one-month follow-up may not capture longer-term retention or clinical transfer of knowledge. Additionally, the possible influence of student learning styles or the novelty effect of 3D printing on satisfaction and motivation outcomes should be acknowledged. It was conducted at a single institution, which may limit generalizability. A key direction for future research is the long-term impact of different educational modalities on clinical skill application [50]. Although our study assessed knowledge retention, longitudinal studies are needed to examine this effect in greater detail. In addition, evaluating technology adoption and collaborative learning processes will provide a more comprehensive understanding of their pedagogical contributions. Directly comparing 3D-printed models with fully immersive VR simulations could further highlight the respective advantages of physical and virtual haptic experiences and clarify how these approaches can complement one another in anatomical education. The only practical limitation encountered during the study was the technical challenge of 3D printing under high-humidity conditions, which extended the model preparation time but did not impact the study’s execution or outcomes. Conclusion In conclusion, our study found tomography-based 3DP distal radius fracture models provided short-term benefits in anatomical knowledge, motivation, and clinical decision-making skills compared with healthy 3DP models. These findings suggest that integrating 3D-printed models derived from actual fractures into the educational curriculum can improve both cognitive and affective learning outcomes in physiotherapy and rehabilitation training. This approach may also help future clinicians better understand fracture morphology and rehabilitation planning, potentially leading to more accurate assessments, safer interventions, and improved patient care. This approach may provide a cost-effective and more advantageous alternative to conventional anatomical teaching tools. Future research should assess the persistence of benefits beyond novelty effects, and evaluate multi-institutional generalizability. Supplementary Information Supplementary material 1. Supplementary material 2. Supplementary material 3. Abbreviations 3D Printing Augmented reality Computed tomography Digital imaging and communications in medicine Distal radius fracture Grade point average Multiple-choice questionnaires Test anxiety inventory Test taking motivation questionnaire Purdue spatial visualization test – rotations Randomized controlled trial Standard triangle language Visual analog scale Virtual reality Authors’ contributions The concept and manuscript design were led by R.K., who also provided overall supervision and critical revision. B.I contributed to manuscript preparation, designed the figures and tables, and performed the statistical analysis. Randomization carried out by K.E.E. and data extraction and model preparation were performed by M.D. and H.O. R.K. and M.D. conducted the case based exams. All authors reviewed the manuscript. Funding The authors declare that no funding was received for this study. Data availability The datasets used and analyzed during the current study and STL files are available from the corresponding author on request. Declarations Ethics approval and consent to participate The study was approved by the Ethics Committee of Bolu Abant Izzet Baysal University (Clinical Research Ethics Committee No: 2023/295–410), and written informed consent was obtained from all participants in accordance with the principles of the Helsinki Declaration. Consent for publication All authors consent to the publication of the manuscript. Competing interests The authors declare no competing interests. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary material 1. Supplementary material 2. Supplementary material 3. Data Availability Statement The datasets used and analyzed during the current study and STL files are available from the corresponding author on request.