A free-text processing system to capture physical findings: Canonical Phrase Identification System (CAPIS)

Area of Science:

• Medical Informatics
• Clinical Research Data Management

Background:

• Free-text medical records pose challenges for clinical research data extraction.
• Detailed patient information is crucial for advancing medical studies.

Purpose of the Study:

• To develop and evaluate a prototype system for extracting physical examination findings from dictated admission summaries.
• To enhance an existing clinical database with structured physical examination data.

Main Methods:

• A concept-based free-text processing algorithm was employed to identify target physical examination findings.
• The system's extraction performance was compared against an independent investigator's manual extraction.
• Evaluation metrics included sensitivity (recall) and positive predictive value (precision).

Main Results:

• The prototype system demonstrated high performance in extracting physical examination findings.
• Sensitivity of 92% indicated the system recalled most relevant findings.
• Precision of 96% confirmed the high accuracy of the extracted data.

Conclusions:

• The developed prototype system effectively extracts physical examination findings from free-text medical records.
• This automated approach can significantly improve the efficiency and accuracy of clinical research data enrichment.
• The system shows promise for overcoming barriers in utilizing unstructured clinical data for research.