Categorization of free-text problem lists: an effective method of capturing clinical data

Area of Science:

• Medical Informatics
• Health Information Systems
• Clinical Data Management

Background:

• Problem lists are crucial for organizing patient data in electronic health records.
• Effective categorization of problems is necessary for utilizing problem lists in billing, research, and decision support.
• Current systems require enhanced methods for structured problem data entry.

Purpose of the Study:

• To describe and evaluate a problem list component within the On-line Medical Record (OMR) system.
• To assess the effectiveness of a combined free-text entry and categorization scheme.
• To analyze the categorization rates and potential for improvement.

Main Methods:

• The OMR system utilizes a free-text entry mechanism combined with a categorization scheme based on an 846-term dictionary.
• Analysis of 118,040 problems entered over six years by 477 clinicians across 22,311 patient records.
• Lexical analysis of problem names to assess mapping to existing medical terminologies (Meta 1.4, Snomed 3.0, Read 3.0).

Main Results:

• A total of 118,040 problems were analyzed, with an average of 5.1 problems per patient record.
• 66% of all problems were successfully categorized using the problem dictionary.
• Lexical analysis indicated that 80% of problem names could be mapped to standard terminologies, suggesting room for dictionary expansion.

Conclusions:

• A problem list entry scheme combining free-text entry and optional dictionary-based categorization achieves a high proportion of categorized problems.
• System improvements, including dictionary refinement and expansion, can further increase categorization rates.
• The OMR system demonstrates a viable approach to structured problem list management in electronic health records.