The Wicked Problem of Selection and Competency-Based Education: Applying TRIZ Inversion to the Graduate Medical...

Graduate medical education (GME) selection processes may pose significant barriers to implementing competency-based health professions education (CBHPE). Current selection metrics are largely nonpredictive of future success, and the removal of clerkship grades has given rise to a shadow economy of effort focused on publications, research hours, and other activities that detract from clinical training. This misalignment between selection practices and CBHPE principles constitutes a wicked problem that requires innovative approaches to develop new solutions.

We conducted a TRIZ (Theory of Inventive Problem Solving) exercise, a Liberating Structure that challenges assumptions by asking what would make a process fail spectacularly. Eight participants representing three countries, two continents, and four medical specialties convened via video conference in October 2025. Participants served as proxies for key stakeholder groups (students, program directors, medical school leadership, and patients/society), identified elements of the worst-possible selection system, compared them to current practices, and proposed solutions.

Participants identified numerous elements characterizing suboptimal selection systems, including inequitable and biased processes, arbitrary decision-making, gaming of metrics, resource-intensive processes that fail to differentiate candidates, misinformation from training sites, and prioritization of self-interest over patient needs. Many of these elements were recognized as present in current GME selection systems worldwide.

Recommendations to align selection with CBHPE included embedding social accountability throughout selection, leveraging coproduced competency frameworks, prioritizing workforce diversity, ensuring equity in assessment and selection, incentivizing a growth mindset through a transparent learner handover, and implementing structural innovations, including preference signaling or potentially lottery systems, to better align CBHPE and selection.