Improving Skin Cancer Diagnostics Through a Mobile App With a Large Interactive Image Repository: Randomized...

Area of Science:

• Medical Education
• Dermatology
• Artificial Intelligence in Healthcare

Background:

• Skin cancer diagnosis is complex and requires extensive physician training.
• Mastery in diagnosing skin conditions necessitates prolonged, dedicated practice.

Purpose of the Study:

• To evaluate if self-paced pattern recognition training using a digital image repository improves primary care physicians' (PCPs') diagnostic skills and confidence in skin cancer detection.
• To assess the impact of training with clinical and dermoscopic images of skin lesions via a large-scale interactive image repository (LIIR).

Main Methods:

• 115 PCPs were randomized into intervention and control groups.
• The intervention group used a quiz-based smartphone app with an LIIR for 8 days.
• Diagnostic skills were assessed via multiple-choice questionnaires pre- and post-intervention.

Main Results:

• The intervention group showed a statistically significant improvement of 14.3 percentage points in diagnostic accuracy (P<.001).
• Participants' ability to recognize malignant lesions improved by 6.6 percentage points, and correct diagnosis setting improved by 10.5 percentage points.
• Diagnostic confidence increased by 32.9% in the intervention group, with no significant change in the control group.

Conclusions:

• Self-paced, app-based training using a digital LIIR significantly enhances PCPs' diagnostic accuracy for skin cancer.
• This digital learning approach offers an effective method to improve dermatological diagnostic skills in primary care.