Automatic ICD-10 coding algorithm using an improved longest common subsequence based on semantic similarity.

Area of Science:

• Medical Informatics
• Natural Language Processing
• Health Information Management

Background:

• Clinical diagnoses in medical records often use free text that differs from standardized International Classification of Diseases 10th revision (ICD-10) terminology.
• Accurate mapping between clinical free-text diagnoses and ICD-10 codes is crucial for data standardization, research, and healthcare management.
• Existing methods may struggle with the nuances and variations in Chinese medical terminology.

Purpose of the Study:

• To develop and evaluate an improved approach for the automatic mapping of Chinese clinical diagnoses to ICD-10 disease names.
• To enhance the accuracy of processing and standardizing Chinese medical diagnoses using natural language processing techniques.

Main Methods:

• The study proposes an improved method combining Longest Common Subsequence (LCS) algorithm with semantic similarity.
• The approach focuses on mapping free-text Chinese disease names to standardized ICD-10 disease names.
• The improved LCS algorithm is designed to handle variations in Chinese medical terms more effectively.

Main Results:

• The proposed method demonstrates increased accuracy in processing Chinese disease names for mapping to ICD-10.
• The integration of semantic similarity alongside an improved LCS enhances the robustness of the mapping process.
• The findings suggest a more reliable automated system for standardizing Chinese clinical diagnoses.

Conclusions:

• The developed approach offers a significant improvement for automatic Chinese diagnosis mapping to ICD-10.
• This method can enhance the quality and consistency of health data by improving the accuracy of disease classification.
• The study contributes to the field of medical informatics by providing a more effective tool for handling clinical terminology variations.