A Cross-Sectional Study on Whether Comprehensively Gathering Information From Medical Records Is Useful for the...

Area of Science:

• Medical Informatics
• Clinical Decision Support Systems
• Bayes' Theorem Application

Background:

• Developing effective clinical decision support systems (CDSS) is crucial for improving healthcare.
• Bayes' theorem is a statistical method that can be applied to diagnostic reasoning.
• The utility of comprehensive medical record data extraction for CDSS development requires investigation.

Purpose of the Study:

• To evaluate the feasibility and effectiveness of using comprehensive medical record data extraction with Bayes' theorem for building CDSS.
• To determine if this method can overcome existing challenges in medical informatics.

Main Methods:

• A retrospective cross-sectional study analyzed 270 patient records from a single emergency center.
• Manual extraction of diagnoses, unique keywords, and annotated keywords (positive/negative) was performed.
• Cross-tables were generated to calculate likelihood ratios, and probability functions for new keywords were modeled.

Main Results:

• 623 unique keywords, 26 diagnoses, and 6,904 annotated keywords were extracted.
• Likelihood ratios were calculable for only 1.70% of combinations, with 0.15% showing significant differences.
• A power function with a constant best fit new unique keyword appearance; comprehensive data gathering incurs high human costs.

Conclusions:

• Comprehensive information gathering from medical records using Bayes' theorem is impractical and costly for CDSS development.
• This method does not fundamentally address medical informatics issues in building effective CDSS.
• Alternative approaches are recommended for developing robust clinical decision support systems.