Structured radiology reporting: A 4 year case study of 160,000 reports

This paper analyzes the use of structured radiology reports, including the ability to produce natural English sentences from structured input.
natural language generation
rule-based systems
Published

November 25, 2001

Citation

Gerald D. Berman, Richard N. Gray, David Liu, and James J. Tyhurst. 2001. “Structured radiology reporting: A 4 year case study of 160,000 reports”. Presented at the Integrating the Healthcare Enterprise (IHE) Symposium of the Radiological Society of North America (RSNA) 2001 Annual Meeting, November 25 - 30, 2001.

References

PDF document: structurad-rsna-2001.pdf

Abstract

This paper analyzes the use of structured radiology reports since 1997 at Midway Hospital Medical Center, which is a full-modality 150-bed community hospital located in Los Angeles, California. We describe the transition from a conventional transcription-based reporting system to one in which radiologists create reports for twelve radiology modalities directly through a structured reporting system. More than 160,000 structured radiology reports have been produced with the structured system.

Compared to transcription, the structured reporting system has:

  1. Improved the quality and consistency of reports.
  2. Enabled instantaneous dissemination of reports to improve patient care and to decrease hospital stays.
  3. Eliminated transcription costs.
  4. Automated abstraction of reports with CPT and ICD-9 codes.

The success of this system depends crucially on the completeness of structured report templates, as well as the ability to produce natural English sentences from structured input.