User-oriented evaluation of a medical image retrieval system for radiologists

Area of Science:

• Medical Informatics
• Radiology Information Systems
• Human-Computer Interaction

Background:

• Medical image retrieval systems are crucial for accessing visual data in the biomedical literature.
• Evaluating system usability and effectiveness is essential for clinical adoption.
• User-centered design is key to developing efficient radiology information retrieval tools.

Purpose of the Study:

• To conduct a user-oriented evaluation of a text- and content-based medical image retrieval system.
• To assess the usability and identify areas for improvement in the system and its interface.
• To investigate the added value of the system for radiology information retrieval and gather insights for future system development.

Main Methods:

• Iterative user testing with 16 radiologists performing information-seeking tasks.
• Monitoring user interaction, measuring usability, recording retrieval success rates, and collecting feedback via surveys.
• Refining the system and study design based on iterative feedback.

Main Results:

• High average success rates: 87% for image retrieval and 78% for case retrieval.
• Successful searches completed in under 3 minutes; users adapted quickly to new tools.
• Positive user satisfaction, with feedback identifying specific areas for enhancement, such as modality filtering.

Conclusions:

• Iterative evaluation effectively gathered detailed user feedback on medical image retrieval systems.
• Radiologists rapidly adopted system functionalities but desired additional features.
• The methodology and findings offer valuable insights for refining current and developing future user-oriented medical image retrieval systems.