TEACHING VETERINARY RADIOLOGY: DOES COMPARISON HELP?

Introduction/Purpose
Comparison learning is an approach for learning complex visual tasks. As described in human medicine, by comparing radiographs with pathology and without pathology, veterinary students could learn to discriminate relevant disease-related information to recognize the disease. We hypothesized that exposure to a training by side-by-side comparison of normal to abnormal radiographs would lead to higher visual diagnostic accuracy and possibly to a higher ability to describe the features of a known disease on canine thoracic radiographs.
Material and methods
Twenty veterinary students were randomly assigned to either a group that compared radiographs showing thoracic disease with normal images (group 1) or to a group that only studied abnormal radiographs (group 2). All students had their theoretical radiology teaching and 13 of them had also received the practical teaching of the study curriculum, evenly distributed between the 2 groups. Twenty-nine radiographs of small animal thorax were used. The procedure consisted in three experimenter-supervised phases: 1. training, 2. visual recognition test, 3. feature description test. In the training phase, each screen showed two radiographs with the name of the disease present in each. In the group 1 (pathology/normal condition), a radiograph of a patient and a normal image were shown next to each other. In the group 2 (pathology/pathology condition), two radiographs of patients with the same disease were shown next to each other. The 9 screens were presented twice, with the diseases in a different order for the first and second runs. A Mann-Whitney test was used to compare the success rates of groups of students.
Results
On visual recognition test, there was no statistical difference in visual diagnostic accuracy between groups. When students gave the wrong answer, they often diagnosed the item as another disease of similar distribution (diffuse or focal). Students who received the practical teaching and students of group 1 had a higher accuracy for normal thoracic radiographs. On feature description test, no significant effect of comparison learning was found, but focal diseases were better described than diffuse diseases with a significant difference between these.
Discussion/Conclusions
Results show that comparison with normal images did not help in recognizing or describing thoracic pathologies but helps to recognize normal images.