Background Medical education has shifted from memory-based practice to evidence-based decisions. The question arises: how can we ensure that all students get correct and systematic information? Visually based, computerised diagnostic decision support system (VCDDSS, VisualDx) may just fit our needs. A pilot study was conducted to investigate its role in medical education and clinical practice.

Methods This was a prospective study, including one consultant dermatologist, 51 medical students and 13 dermatology residents, conducted in the dermatology teaching clinic at China Medical University Hospital from 30 December 2014 to 21 April 2015. Clinical diagnoses of 13 patients were made before and after using VCDDSS. Questionnaires were filled out at the end. The consultant dermatologist's diagnosis was defined as the standard answer; the Sign test was used to analyse diagnostic accuracy and the Fisher exact test to analyse questionnaires.

Results There was an 18.75% increase in diagnostic accuracy after use of VCDDSS (62.5–81.25%; p value <0.01). Significant associations were found in diagnostic assistance in terms of user factors such as accessibility, interface satisfaction, quality of imaging, textual description, and a Chinese language interface option (p value<0.01).

Conclusions This study demonstrated that VCDDSS increases diagnostic accuracy by 18.75%, which means we can avoid possible misdiagnosis, provide better treatment, and avoid waste of medical resources. The user satisfaction is high. We expect wider application of this kind of decision support system in clinical practice, medical education, residency training, and patient education in the future. Further large-scale studies should be planned to confirm its application.