Evaluation of a 3D-Printed Model for Simulating Closed Reduction and Percutaneous Pinning of Unstable Distal Radius Fractures

Abstract

Introduction:

We evaluated construct validity of a 3D-printed displaced, unstable distal radius fracture (DRF) model for closed reduction and percutaneous pinning (CRPP). We hypothesized that clinical DRF experience levels would correlate with performance metrics during model CRPP.

Methods:

Thirteen volunteers across a spectrum of DRF experience participated, including orthopaedic surgery residents and hand surgery specialists (fellows and attendings). Participants completed a survey assessing confidence and experience with DRF CRPP; watched a video demonstrating model CRPP and evaluation procedures; independently performed model CRPP with a single, power-driven Kirschner wire; and provided anonymous feedback on the model. An independent observer recorded each participant's Kirschner wire locations, degrees of final sagittal-plane DRF model tilt on fluoroscopy, number of fluoroscopic images, and procedural time. We grouped participants based on experience level for analyses.

Results:

Group 1 consisted of six nonspecialized orthopaedic surgery residents and group 2 consisted of seven hand surgery specialists (five fellows; two attendings). Group 1 had less baseline confidence and fewer prior DRF CRPPs. Group 2 performed markedly better than group 1 in all metrics, including Kirschner wire entry point and final trajectory; final sagittal plane tilt (7° vs. 20° dorsal, P < 0.05); number of fluoroscopic images (19 vs. 43, P < 0.05); and procedure time (2:37 vs. 6:01, P < 0.05). All participants recommended the model for educational use.

Discussion:

Hand surgery attendings and fellows performed better than residents in all metrics for CRPP with our displaced, unstable DRF model, indicating appropriate construct validity. All participants recommended the model for procedural education.

Distal radius fractures (DRFs) are one of the most common musculoskeletal injuries, comprising 15% to 25% of all adult and pediatric fractures.¹ DRF closed reduction and percutaneous pinning (CRPP) is an important basic skill to acquire and can be used clinically for definitive management of some fractures and for provisional fixation during open reduction and internal fixation and other procedures.²

Simulations for other clinical scenarios have been successfully developed to augment surgical education.³ Artificial simulations can be performed repetitively in low-risk, low-stress environments with lower costs and greater efficiency than standard clinical- or cadaver-based teaching methods.⁴ There is a paucity of literature and surgical simulation tools specific to the upper extremity and DRF. In prior work, we confirmed construct validity and educational utility for orthopaedic surgery residents of a novel, low-cost, high-fidelity, 3D-printed displaced, unstable DRF model for closed reduction and splint application.⁵ Other authors have also investigated simulation of nonsurgical DRF management with artificial models.^4,6

Our primary goal for the current study was to evaluate the construct validity of our same DRF model for CRPP (Figure 1). We hypothesized that (1) the model would have construct validity, in that metrics of performance would positively correlate with greater clinical experience levels⁷ and (2) participants would rate the model as educationally beneficial.

Figure 1.

(Top) photographs and (bottom) fluoroscopic images showing the model with a Kirschner wire in place.

Materials and Methods

Study Design

This was a single-center simulation-based study approved by our institutional review board. We obtained written informed consent from all participants before their enrollment. Enrollment and continued participation were completely and explicitly voluntary throughout the study period.

We recruited participants through convenience sampling across our center's orthopaedic surgery residents, orthopaedic- and plastic-background hand surgery fellows, and hand surgery attendings. Each participant first independently filled an anonymized survey regarding their confidence and experience with DRF CRPP. We 3D-printed polymer models from computer-assisted design software that allows for creation of a displaced, unstable DRF fragment that rests in ∼25° of dorsal angulation. The fragment was reducible to ∼neutral tilt and redisplaced if not held in position. Please refer to our prior publication for more details on the model.⁸ Participants watched an instructional video created by a nonparticipating attending hand surgeon study member outlining how to use the DRF model for CRPP and what the evaluation rubric would entail. Each participant then performed DRF model CRPP with a single 2.0-mm Kirschner wire while using mini C-arm fluoroscopy. Finally, each participant voluntarily completed an anonymous survey to collect feedback on the model.

Study Outcomes

Preparticipation Survey

We collected, through an anonymous survey, study participant experience levels and their Accreditation Council for Graduate Medical Education log values or attending-level estimates for the number of DRF reduction and pinning procedures they had performed clinically. We assessed clinical comfort with DRF CRPP on a 5-item Likert scale.

Distal Radius Fracture Model Closed Reduction and Percutaneous Pinning Evaluations

A single independent orthopaedic-background hand surgery fellow directly assessed each participant throughout the DRF model CRPP procedures. This observer recorded participant metrics for the following: time to DRF model CRPP completion, number of fluoroscopic images taken during the procedure, degrees of DRF fragment tilt on final lateral view fluoroscopy, Kirschner wire entry site location (defined as ideal at the styloid tip, acceptable between the styloid tip and the fracture, or unacceptable at or proximal to the fracture level), and Kirschner wire trajectory (defined as crossing the fracture with ideal bicortical purchase, crossing the fracture with marginal bicortical purchase, or unacceptable as not crossing the fracture with any degree of bicortical purchase). With this iteration of the model, the maximum possible correction of the baseline approximately 25° of fracture fragment dorsal angulation was to an approximately neutral position.

Postparticipation Survey

We gathered postparticipation feedback with an anonymous survey on the use of the model for simulation of DRF CRPP. This included asking if they would recommend the model for learning and improving DRF CRPP skills. They also rated the model's educational value on a 5-item Likert scale. We also included a free response section for any additional qualitative feedback.

Statistical Analyses

For comparative analyses, we divided participants into two groups based on experience level. Group 1 was composed of orthopaedic surgery residents and group 2 was composed of hand surgery fellows and attendings. We used Shapiro-Wilk tests to assess the normality of the data distributions. We compared continuous data between the groups using two-sided independent t-tests and compared categorical data using Fisher exact tests. We determined statistical significance as anything less than an alpha value set at 0.05.

Results

Group 1 had six orthopaedic surgery residents and group 2 had seven higher-level participants (five hand surgery fellows and two fellowship-trained hand surgery attendings). The breakdown of postgraduate years (PGYs) for participating residents was one PGY-5, two PGY-3s, two PGY-2s, and one PGY-1. All fellows had completed at least 6 months of clinical fellowship before participation. All attendings had been in clinical practice for more than 3 years after their fellowship. Overall, group 1 ranked lower baseline confidence than group 2 regarding DRF CRPP (Likert scale range 1 to 3 vs. 4 to 5 out of five, respectively). Half of the group 1 participants had no prior clinical DRF CRPP experience, whereas all participants in group 2 had performed DRF CRPP clinically.

Procedural Outcomes

We confirmed normality of the data distributions for appropriate selection of descriptive and comparative analyses. The average time for completion of DRF model CRPP for group 1 was markedly longer than for group 2 (6:01 vs. 2:37, respectively, P < 0.05; Figure 2). Group 1 had markedly worse average residual degree of DRF model dorsal angulation compared with group 2 (20° vs. 7° of dorsal angulation, respectively, P < 0.05). Group 1 took markedly more total fluoroscopic images during DRF model CRPP than group 2 (43 vs. 19, P < 0.05). Group 1 had variable Kirschner wire starting points with most being between the radial styloid tip and the fracture. All group 2 participants used the ideal Kirschner wire starting point at the radial styloid tip (Figure 3, A and B). Final Kirschner wire trajectory rating was markedly worse for group 1 than for group 2 (one ideal, four marginal, one unacceptable versus six ideal, one marginal, respectively, P < 0.05).

Figure 2.

Graph showing procedure completion times: residents versus fellows/attendings.

Figure 3.

Radiographs showing (A) pin entry points and trajectories for a group 1 resident and (B) group 2 hand surgery attending.

Qualitative Feedback

All participants recommended the model for use with future learners of DRF CRPP. All participants rated the model as “useful” or “very useful” other than one who rated it as “neutral.” Comments largely focused on the advantages of allowing practice using surgical power tools in a more controlled, safer, and lower stress environment than standard clinical teaching. Many also noted the ability to feel realistic aspects of surgery, such as the tactile feedback of appropriate bicortical pinning, as strong features of the model. Free response feedback included: “great model to learn … can feel the bicortical purchase” and “excellent model for understanding the feel, maneuvering, and spatial awareness of pinning.”

Discussion

We demonstrated substantial construct validity for CRPP using our displaced, unstable DRF 3D-printed model, with the more experienced group of participants performing better in all metrics. All participants recommended the model for use with surgical learners, highlighting its potential to enhance surgical education. These findings support our hypotheses. We propose this model as appropriate for the consideration of further use in surgical education.

A next validation step with this model could be to determine if education with the model leads to measurable improvements in retention of skills through retesting at a later date. This could also include establishment of a more efficient learning timeline to decrease cumulative risks of fluoroscopy for surgical providers.⁹

Wu et al¹⁰ also developed a 3D-printed hand and wrist model for pinning, which was found to be effective for teaching per five orthopaedic surgeons who tested the model. Investigators using supracondylar humerus fracture models have also demonstrated their utility for residency education and improving surgical pinning performance.^11-13 Although other upper extremity models such as these are available, we are unaware of any that include a high-fidelity, displaced, unstable DRF option for CRPP. We also performed more formal construct validation of our model with comparative assessments across multiple experience levels. In DRF, CRPP may be used for definitive or provisional fixation. Displaced DRFs are often reduced and stabilized with a single Kirschner wire for pediatric cases or for temporary intraoperative use for adult cases. Although displaced DRFs are otherwise commonly further stabilized, we designed our study to minimize variables and facilitate standardization to initially assess the model's construct validity for potential educational and competency assessment purposes. This approach still allowed us to detect substantial and statistically significant differences in technical abilities between residents and more experienced surgeons. For example, the residents averaged 20° of residual dorsal tilt after reduction and pinning, compared with fellows and attendings who averaged 7°. Our findings highlight the importance of an appropriately placed initial Kirschner wire to maintain reduction of unstable fragments. We think that these findings further support the construct validity of this model. As such, it is an important skill for new learners to acquire. Beyond this, being able to practice fracture reductions and stabilizations with models, power tools, and fluoroscopy in a controlled fashion can reinforce skills used across many procedures. This highlights another tool for teaching, particularly in the modern era of duty hour restrictions.¹⁴ Furthermore, the usefulness of this model, including for CRPP exposure, could be likely extrapolated to other procedural fields and for medical student education and exploring career interest, possible areas of future evaluation.

Our study had several limitations. The evaluations, although capturing largely objective metrics, were unblinded and performed by a single independent observer. Furthermore, our sample size was small, and we performed the study at a single center, limiting the generalizability of our findings. The effect sizes were large enough to already detect statistically significant differences between the groups even with these sample sizes. Future research could include a greater number of participants, a wider range of experience levels, multiple sites, and potential evaluator blinding with deidentified video recordings to reduce bias and confirm our findings. To further assess construct validity, it would be useful to study whether the model could discern between experience levels within the groups studied (eg, a PGY-1 versus a PGY-3) with a larger sample size.

In conclusion, our 3D-printed displaced, unstable DRF model has construct validity for CRPP, with more experienced participants doing markedly better on all performance metrics. All participants recommended the model to learn and improve skills related to DRF care. Future model investigations could determine the ability for learners to efficiently acquire and retain adequate technical abilities and comfort with CRPP. Further validation could also assess translational clinical outcomes.