Text

Background: Case-based learning (CBL) is a well-established approach to foster clinical competences [1]. However, it is crucial to understand what exactly makes a case difficult, as learning outcomes depend on case demands such as case-inherent complexity [2]. Accordingly, this study aims to examine how complexity influences case difficulty for learners, assuming that an increase in complexity leads to an increased difficulty.

Methods: Utilizing an innovative standardized evaluation system (CASE) [3], we operationalized medical case complexity as the summed CASE score, quantifying the amount of information provided across predefined categories (e.g., imaging, laboratory). Across three prior studies (N=342; 45 cases), we used an explanatory item response model with random intercepts for cases and participants to link CASE-category information to case difficulty (diagnostic accuracy; one dataset dichotomized at ≥0.5), to estimate how a one-point increase in category-specific complexity changes the odds of a correct diagnosis.

Results: Between-case variability exceeded between-person variability (item variance ≈1.57; person variance ≈0.35), indicating substantial differences in case difficulty beyond individual ability. Across datasets, imaging information was consistently associated with lower diagnostic accuracy (β≈-.19, p<.001; OR≈0.82), whereas laboratory (β≈.034, p≈.009; OR≈1.03-1.04) and physical examination information (β≈.043, p≈.031; OR≈1.04) showed small positive associations; medical history showed no robust effect.

Discussion: As expected, our findings indicate that case features systematically shape case difficulty as reflected in diagnostic accuracy, which may partly explain why effects of CBL vary across studies. Diagnostic performance appears at least partly case-dependent, highlighting the relevance of case composition when assembling case sets for different learner levels. Moreover, CASE-based complexity was not uniformly difficulty-increasing: imaging may act as a challenge driver, warranting additional instructional support (e.g., interpretation frameworks, prompts, preparatory instruction), whereas laboratory and physical-examination information support correct diagnoses. Overall, difficulty may not be inferred from information quantity alone, supporting a more fine-grained view of case features.

Take home messages:

• More information alone does not make cases harder; difficulty depends on the type of information, as captured by the CASE categories.
• CASE may enable transparent, comparable descriptions of case materials across studies and teaching settings.
• CASE-based profiling could help identify cases that may benefit from additional instructional support. This could further enable the construction of cases that are precisely tailored to different skill levels of medical students.
• Future research could examine how case complexity can be aligned with learners’ prerequisites.

Literatur

[1] Kassirer JP. Teaching clinical reasoning: case-based and coached. Acad Med. 2010;85(7):1118-1124. DOI: 10.1097/ACM.0b013e3181d5dd0d
[2] Braun LT, Lenzer B, Fischer MR, Schmidmaier R. Complexity of clinical cases in simulated learning environments: proposal for a scoring system. GMS J Med Educ. 2019;36(6):Doc80. DOI: 10.3205/zma001288
[3] Öhler KS, Hilger E, Stadler M, Hege I, Papa F, Schmidmaier R, Fischer MR, Weidenbusch M, Zottmann J. Assessing the complexity of clinical cases. In: Gemeinsame Jahrestagung der Gesellschaft für Medizinische Ausbildung (GMA) und des Arbeitskreises zur Weiterentwicklung der Lehre in der Zahnmedizin (AKWLZ). Halle (Saale), 15.-17.09.2022. Düsseldorf: German Medical Science GMS Publishing House; 2022. DocV-10-05. DOI: 10.3205/22gma063