Computer-based Virtual Environment Simulations for Differential Diagnosis in Medical Education

Abstract

A major problem in medical education is that the clinical knowledge and clinical reasoning skills acquired through university teaching do not transfer well to clinical practice. Despite acquiring massive amounts of content knowledge about the functioning of the human body, medical students struggle to transfer that knowledge to one of the core disciplinary practices - differential diagnosis. The lack to transfer may stem from current instruction methods which is focused primarily on imparting massive amounts of basic content knowledge without adequate attention to situate this knowledge in disciplinary practice. A possible solution to this problem is to expose and link the learning of medical students to the practice of differential diagnosis. This approach is supported by theories of situated cognition. Whilst I acknowledge that there are several options to integrate situated learning, I aimed to explore the use of medical computer-based virtual environment (CVE) simulations. In empirical research it has yet to be shown how, when, and why CVE simulations are effective in medical education and enhancing transfer. Accordingly, the first major goal of this thesis was to overcome the issue of transfer in differential diagnosis by using a situated learning approach. The second goal was to explore how, when, and why CVE simulations are effective and they enhance transfer. Inter alia, the ability to transfer is important because medical students cannot be confronted with all possible situations in their medical studies they will face later in their professional career as doctors. In my studies, I focused on clinical knowledge and the conceptual aspects of differential diagnosis and not on the manual execution of procedures. Due to Corona-Virus-19 issues, most study activities in this thesis were conducted remotely where participants worked individually from home. In the first study of this thesis, I used a problem-solving prior to direct instruction versus instruction first experimental design to examine the effect of problem-solving in CVEs on isomorphic testing and transfer outcomes assessing declarative and conceptual clinical knowledge and clinical reasoning skills, along with evoked learning mechanisms. I found neither of the learning activity sequences to be superior to the other. However, when looking at the two learning activities individually, I found that problem-solving in CVEs as well as direct instruction are equally effective for learning declarative and conceptual clinical knowledge. On the other hand, problem-solving in CVE with formative feedback is more effective for learning clinical reasoning skills than mere instruction. In the second cohort-based study, I explored the influence of semester-long exposure to problem-solving in CVE prior to introductory lectures versus interactive group discussions preceding the lectures in an introductory course on performance in a successive advanced course in the medical trajectory, both related to differential diagnosis. Using a double transfer experimental design, I compared the effects of the two teaching approaches on acquisition of clinical knowledge, transfer of clinical reasoning skills, and students’ satisfaction and self-confidence in learning. I found both group discussion and CVE problem-solving preceding an instructional lecture to prepare students for the acquisition of declarative and conceptual clinical knowledge in an advanced differential diagnosis course. However, the group discussions preceding the lecture seem to be more effective than CVE problem-solving preceding the lecture. Furthermore, I found interactive group discussions preceding the lecture in the basic differential diagnosis course to be more effective in enhancing clinical reasoning skills transfer, but only when the transfer problem featured specific configurations. However, I did not find a difference in satisfaction and self-confidence in learning with simulations between cohorts nor any mediating effect of simulation design features on learning outcomes. However, CVE problem-solving preceding the instructional lecture was equally effective than group discussions in (a) enhancing transfer of clinical reasoning skills when specific testing configurations were met and (b) preparing students for the acquisition of declarative and conceptual clinical knowledge. Consequently, from a practical side of view CVE problem-solving seems to be a plausible alternative to group discussions as preparatory activity for following instruction. In the third study I extended the research on the problem-solving prior to instruction approach and put the findings of study 1 into practice to further enhance the acquisition and transfer of clinical knowledge and clinical reasoning skills trough problem-solving in CVE prior to instruction. I did so by altering the timing of provided formative feedback during the CVE problem-solving phase. This was represented by instant, delayed, and no feedback provision. Again, I evaluated triggered learning mechanism. The findings of this third study indicated that all three timings of investigated feedback are eligible because none of them had detrimental effects on learning. However, if and at what point in time feedback should be provided to optimally enhance learning, heavily depends on the targeted learning outcome, which might be clinical declarative or conceptual knowledge, clinical reasoning skills, and respective acquisition or transfer. With my thesis I (a) filled the gap in the learning sciences research of how, when, and why CVE simulations are effective in medical education, (b) examined the extent to which medical CVE simulations can enhance isomorphic testing outcomes and transfer of clinical knowledge and clinical reasoning skills, and (c) derived principles for when and how to best implement such CVE patient scenarios for differential diagnosis into a medical curriculum. The findings of my first study indicate that the problem-solving learning activity in CVEs seems to be more effective for learning clinical reasoning skills than direct instruction. When combined, the sequence of the two learning activities seems to play a minor role for learning outcomes as there were no significant differences in learning outcomes. The findings of my second study indicate that students might benefit better from interactive group discussion as preparatory activity for an introductory lecture than from individual CVE problem-solving for enhancing transfer of clinical knowledge and clinical reasoning skills. However, CVE problem-solving as preparatory activity for following instruction seems to be a plausible alternative. Furthermore, the second study revealed some supplementary details to the first study. It indicated that collaborative activities prior to instruction might be more effective for learning than individual problem-solving preceding the instruction, especially for learning clinical reasoning skills. The findings of my third study indicate that whether and when feedback during or after CVE problem-solving prior to instruction should be implemented, primarily depends on whether clinical knowledge or clinical reasoning skills, respectively their acquisition or transfer is the targeted learning outcome. The findings of the present thesis have a high level of ecological validity because all studies took place in settings where (a) data were derived from real teaching settings or (b) students had to perform learning and testing tasks autonomously from home due to Corona-Virus-19 issues.