Grundlagen der Mustererkennung in der radiologischen Diagnostik - Entwicklung und Evaluation eines studentischen Lernkurses

Pattern recognition builds the foundation for radiological reports. There are several radiological patterns, which are based on a certain number of archetypes. The basis for the current study is the consideration, that humans have been able to program a computer to recognize certain patterns and categorize them by the implementation of computer aided diagnosis (CAD). This advancement lead to the hypothesis, that we should also have the ability to impart the basic principles of pattern recognition to radiology students. Therefore, the online learning course “Pattern recognition in diagnostic radiology 2011“ has been de-signed. The novel learning course was implemented in the summer semester in 2011 in the setting of the radiology course “Imaging methods, radiation therapy and radiation protection“ for medical students in their first clinical year. Throughout the course, we evaluated to which extent the online course lead to an increase in test results in the small group seminars. In addition, we evaluated the students’ acceptance regarding the modified course concept and novel teaching method of pattern recognition: studying pattern recognition through classifications and transferring certain archetypes onto different structures, instead of memorizing different diagnoses. The online learning course includes 98 pages, 89 images, 16 questions and 3 tables. By means of especially developed classification organigrams for patterns in parenchymatous organs, tubular structures and visceral cavities, examples of different patterns were illustrated and explained. Evaluation of the learning course included its implementation in the small group seminars within the modified course concept. All of the students had four seminars with the prerequisite to work through Power-Point-Files beforehand. These files contained sample cases of typical radiological patterns. Within the seminars the students had to answer questions regarding these patterns using a TED-system. The small groups were divided into an intervention and a control group. After the first two seminars only the intervention group had access to the online learning course. The last two seminars included the same questions and served as a follow-up and evaluation of the students learning success. At the end of the radiology course the online learning course and the new didactic approach were evaluated by an online questionnaire. The test results of the intervention and control group, user-tracking data of the learning course and the online questionnaire were analysed with descriptive statistics. A total of 21 students have finished the online learning course. The comparison of groups did not show a significant difference in test results of the seminars. Furthermore, no correlation between the duration of working with the online course and test results could be observed. 59 of 61 students took part in the online questionnaire. About 50% of the students evaluated the learning course as “good“, one third as “satisfactory“. Almost 80% of the learning course content was rated as distinct and comprehensive. Previous knowledge in pattern recognition was rated as “inadequate“ in 79%. The new didactic approach of imparting pattern recognition appeared beneficial to most students. The question in dispute is, to what extent the test results reflect the actual learning success of the learning course. Analysis of the test results may imply that the learning course did not lead to a better understanding of pattern recognition. In reference to the different learning theories, e.g. behaviourism and cognitivism, there are several factors that influence learning. Some aspects of these learning theories are incorporated when learning new patterns. The fact that the actual test results did not meet our expectations does thus not mean that no transfer has taken place. Due to the fact, that most students favour the new didactic approach and rated the learning course as good, limiting factors such as procedural flaws and a small study population have to be considered when interpreting the test results. Hence, this study has to be valued as exploratory.