Learning performance with interactive simulations in medical education: Lessons learned from results of learning complex physiological models with the HAEMOdynamics SIMulator Andreas Holzinger * , Michael D. Kickmeier-Rust, Sigi Wassertheurer, Michael Hessinger Head of Research Unit HCI4MED, Institute of Medical Informatics, Statistics and Documentation (IMI), Medical University Graz (MUG), Auenbruggerplatz 2/V, A-8036 Graz, Austria article info Article history: Received 6 March 2008 Received in revised form 11 August 2008 Accepted 22 August 2008 Keywords: Medical simulation Learning performance Cognitive load Multimedia Educational software development abstract Objective: Since simulations are often accepted uncritically, with excessive emphasis being placed on technological sophistication at the expense of underlying psychological and educational theories, we evaluated the learning performance of simulation software, in order to gain insight into the proper use of simulations for application in medical education. Design: The authors designed and evaluated a software packet, following of user-centered development, which they call Haemodynamics Simulator (HAEMOSIM), for the simulation of complex physiological models, e.g., the modeling of arterial blood flow dependent on the pressure gradient, radius and bifurca- tions; shear–stress and blood flow profiles depending on viscosity and radius. Measurements: In a quasi-experimental real-life setup, the authors compared the learning performance of 96 medical students for three conditions: (1) conventional text-based lesson; (2) HAEMOSIM alone and (3) HAEMOSIM with a combination of additional material and support, found necessary during user-centered development. The individual student’s learning time was unvarying in all three conditions. Results: While the first two settings produced equivalent results, the combination of additional support and HAEMOSIM yielded a significantly higher learning performance. These results are discussed regard- ing Mayer’s multimedia learning theory, Sweller’s cognitive load theory, and claims of prior research on utilizing interactive simulations for learning. Conclusion: The results showed that simulations can be beneficial for learning complex concepts, how- ever, interacting with sophisticated simulations strain the limitation of cognitive processes; therefore successful application of simulations require careful additional guidance from medical professionals and a certain amount of previous knowledge on the part of the learners. The inclusion of pedagogical and psychological expertise into the design and development of educational software is essential. Ó 2008 Elsevier Ltd. All rights reserved. 1. Introduction and motivation for research Dynamic, interactive computer simulations, designed to teach complex processes and concepts, have become very popular in all do- mains of science education, for example physics, chemistry and biology, as demonstrated in the high number of sales. The nature of such simulation ranges from compelling visualizations (Chittaro, 2001; Johnson, MacLeod, Parker, & Weinstein, 2004) to educational computer games (Ebner & Holzinger, 2007; Kickmeier-Rust et al., 2007). Simulations as learning tools are often engaging and, when it is assured that such tools are validated for accuracy and for learning outcome, can be valuable for medical education (Bergin & Fors, 2003; Dev et al., 2002; Hmelo & Day, 1999; Holzinger, Emberger, Wassertheurer, & Neal, 2008). Basically, a major advantage of learning with interactive simula- tions can be seen in the highly constructivist nature of such exploratory learning processes. However, we can observe in practice that, be- cause they are technologically and graphically appealing, there is a tendency to apply interactive simulations uncritically and enthusiastically. Despite the huge enthusiasm for educational simulations generally, many questions still remain as to their effective de- sign. Since developers are now able to include a wide range of visual and auditory elements, the complexity of design decisions is steadily increasing. This immediately raises the importance of gaining insight into the behavior of the end-users, consequently raising the appropriateness of a user-centered development (Edwards & Holland, 1992; Holzinger, 2003; Norman & Draper, 1986). Exploratory, self-directed, and goal-oriented learning with interactive and dynamic media, such as simulations, is highly demanding from the 0360-1315/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.compedu.2008.08.008 * Corresponding author. Tel.: +43 316 385 3883; fax: 43 316 385 3590. E-mail address: andreas.holzinger@meduni-graz.at (A. Holzinger). Computers & Education 52 (2009) 292–301 Contents lists available at ScienceDirect Computers & Education journal homepage: www.elsevier.com/locate/compedu