Evaluating Open-Source Full-Text Search Engines for Matching ICD-10 Codes.

Area of Science:

• Medical Informatics
• Computer Science

Background:

• Accurate retrieval of diagnostic codes is crucial for healthcare data analysis and billing.
• International Classification of Diseases, 10th Revision (ICD-10) codes are widely used but can be challenging to search efficiently.
• Free, open-source search engines offer potential solutions for improving diagnostic code matching.

Purpose of the Study:

• To evaluate the performance of various free, open-source full-text search engines in matching ICD-10 diagnostic codes.
• To identify the minimum number of keywords required for accurate ICD-10 code retrieval through full-text search.
• To compare the effectiveness of different types of search engines (Java-based, JavaScript, relational databases) for this task.

Main Methods:

• Extracted keywords from the text of each ICD-10 code.
• Constructed full-text search queries using combinations of these keywords.
• Executed queries against the entire ICD-10 code database.
• Identified the code with the highest relative score as the match.
• Determined the minimum keyword set for accurate matching.

Main Results:

• The study successfully identified the minimum keyword combinations for accurate ICD-10 code matching across different search engines.
• Performance varied among the evaluated engines, highlighting differences in their full-text search capabilities.
• The research provides insights into the effectiveness of various open-source solutions for diagnostic code retrieval.

Conclusions:

• Free, open-source search engines can be effectively utilized for matching ICD-10 diagnostic codes.
• Understanding keyword requirements is essential for optimizing full-text search accuracy in medical coding.
• The findings support the use of these engines for improving the efficiency and accuracy of healthcare data management.