Novice and expert self-regulated learning phase transitions in medical diagnosis: Implications for adaptive and intelligent systems

Expertise plays a significant role in shaping self-regulated learning (SRL) by influencing how individuals set goals, monitor progress, employ strategies, and reflect on their learning process. However, comprehensive data on this link is sparse in medical contexts. This paper investigates the transitions of SRL phases during clinical-reasoning tasks with a multimedia system, CresME, designed to elicit clinical-reasoning processes using illness scripts. We investigate whether experts utilize more frequent and diverse SRL phase transitions and have better diagnostic performance than novices. Thirty-four participants from a North American Medical School were trained to think-aloud and solved five clinical cases related to the common cough with CResME. Verbalizations were transcribed and coded for SRL phases based on Zimmerman and Moylan’s socio-cognitive model of SRL. Sequential pattern mining revealed that experts exhibited less frequent but more diverse SRL phase transitions than novices, yet these relations did not always result in better diagnostic performance. Instead, the relations between expertise, SRL, and diagnostic performance were dependent on the case. These insights hold implications for assessing SRL phases during clinical reasoning activities to guide just-in-time and personalized support with multimedia systems in medical education.