Comparison-Bot: an Automated Preliminary-Final Report Comparison System

Area of Science:

• Radiology Education
• Medical Informatics
• Health Systems Science

Background:

• Substantial discrepancies exist between preliminary and final radiology reports, hindering resident and fellow education.
• Barriers to learning from these discrepancies include high study volume, remote finalization, and difficulty accessing prior reports.

Purpose of the Study:

• To develop and evaluate a system for automatically summarizing and presenting differences between preliminary and final radiology reports.
• To enhance the educational experience of radiology trainees by facilitating efficient review of report changes.

Main Methods:

• Developed an automated system to compile and email weekly summaries of report differences.
• Implemented a trainee-accessible dashboard with custom reporting and direct links to studies in Picture Archiving and Communication Systems (PACS).
• Highlighted reports with significant changes, particularly in the impression, for focused review.

Main Results:

• Departmental surveys indicated the system is easy to understand and improves the educational experience.
• The system provides descriptive statistics on report changes by trainee level, attending, and exam type.
• The system is designed for easy portability to other departments with Health Level 7 (HL7) data access.

Conclusions:

• The automated report difference summary system effectively addresses barriers to learning from preliminary-to-final report discrepancies.
• This tool enhances radiology trainee education by providing accessible, highlighted feedback and direct image access.
• The system offers valuable data for departmental quality improvement and can be broadly implemented.