This PhD thesis explores the enhancement of training and assessment methods for osteosynthesis of
proximal femoral fractures through simulation-based approaches. The research is presented in four
interconnected studies, each contributing to a framework for improving education in this surgical area.
The first study identifies essential technical procedures in orthopedic surgery and traumatology that
should be included in a simulation-based training curriculum. Through expert consensus, this study
evaluates numerous procedures and prioritizes 33 of them, with osteosynthesis of proximal femoral
fractures ranked second. This systematic evaluation ensures that the curriculum addresses the most
relevant areas, providing a solid foundation for developing targeted educational programs.
The second study examines the training of orthopedic surgeons in hip-fracture osteosynthesis,
focusing on understanding learning curves and establishing proficiency standards. Utilizing a virtual
reality (VR) simulator, the study tracks the performance of both novice and experienced surgeons in
procedures involving cannulated screws, Hansson pins, and sliding hip screws. The findings indicate
that training time or the number of training iterations are not effective parameters for predicting
when trainees reach proficiency. This study supports the shift from time-based to proficiency-based
learning, recommending benchmarks that ensure trainees achieve the necessary competence before
progressing to supervised clinical practice.
Building on previous research, the third study evaluates short antegrade femoral nail osteosynthesis
across multiple centers to establish proficiency standards. This study uses the same VR simulator
platform as in the previous study but introduces new software and a developed test for the specific
procedure. The results showed that trainees who trained to a proficiency standard performed much
better than those who trained without a set goal. This emphasizes the importance of setting clear
proficiency thresholds to enhance skill development and ensure that trainees are proficient before
performing real-life surgery under supervision. By proposing proficiency thresholds and expanding the
understanding of their necessity, the study contributes to the broader goal of standardizing simulationbased
surgical education.
The final study explores an innovative approach to performance assessment using deep learning
techniques. A convolutional neural network (CNN) model is developed to analyze end-procedure
images from simulation-based training in proximal femoral fracture osteosynthesis. The model's
predictions are compared with traditional simulator metrics and clinical evaluations, showing strong
correlations, and indicating the potential for automated, objective assessment tools. This study aims to
address the limitations of current assessment methods by offering a scalable and consistent evaluation
framework, potentially enhancing the effectiveness of simulation-based medical education.
These studies collectively seek to advance simulation-based education for osteosynthesis of proximal
femoral fractures by identifying key training procedures, proposing proficiency standards, and
exploring technological innovations for performance assessment. The thesis advocates for a structured
and evidence-based approach to surgical training, ensuring that trainees develop the skills necessary
to perform complex procedures competently before transitioning to supervised clinical practice. By
incorporating expert consensus, proficiency-based learning, multicenter evaluations, and technological
advancements, this work aims to contribute to a more effective framework for orthopedic surgical
education, ultimately improving the quality of training in the field of proximal femoral fracture
osteosynthesis.
|