Improving Clinical Reasoning Skills With Machine Learning K-Means Algorithm

The enhancement of clinical reasoning is crucial in health sciences education for producing skilled practitioners. This study explores whether machine learning, particularly the K-means clustering algorithm, can detect technical and conceptual errors occurring while students are engaged in problem-solving. The study's main questions ask to what extent machine learning provides opportunities for a personalized approach towards educational interventions aimed at certain types of reasoning deficits.

A new method was proposed to classify students on clinical reasoning skills by integrating K-means clustering with Bloom's taxonomy. The approach gathered learners in clusters at different levels of cognition, starting from very basic cognitive processes of recalling factual knowledge to fully advanced clinical problematization. It was these reverse-engineered clusters that allowed the design of pedagogy that targeted the specific cognitive needs of the groups.

Clustering using the K-means method provides valuable insights into performance patterns in student behaviour that extend beyond the limitations of conventional assessments. By placing students on a continuum of reasoning abilities, educators were able to take action to respond to individual learning paths. Such interventions could be applied in real time at the scale necessary for effective targeted instruction, which is essential for closing reasoning gaps.

The combination of machine learning, especially K-means clustering, and educational theory, such as Bloom's taxonomy, results in electronic-high-scale, multi-evidence, personalized clinical training. This is another theorem on how machine learning enables teaching and individual learning by a student in various cognitive domains.