Auto-mapping Clinical Documents to ICD-10 using SNOMED-CT.

Area of Science:

• Medical Informatics
• Clinical Documentation Improvement
• Natural Language Processing in Healthcare

Background:

• Physicians face significant burdens from manual medical record review and ICD-10 coding for financial claims.
• Thai healthcare systems generate large volumes of discharge summaries and ICD-10 codes annually, contributing to administrative overhead.

Purpose of the Study:

• To develop and present an automated approach for analyzing discharge summary notes.
• To reduce the manual workload of healthcare professionals in mapping diagnoses to ICD-10 codes.

Main Methods:

• A novel approach combining SNOMED-CT and natural language processing (NLP) was developed.
• The process involved three key steps: data cleaning, keyword extraction from discharge summaries, and keyword-to-ICD-10 matching.
• Approximate matching techniques were employed for enhanced accuracy in the mapping process.

Main Results:

• The proposed automated system demonstrates potential for accurately mapping clinical diagnoses to ICD-10 codes.
• The integration of SNOMED-CT facilitates a more robust and comprehensive coding process.
• The approach shows promise in significantly decreasing the manual paperwork burden on physicians.

Conclusions:

• Automating the ICD-10 mapping process using NLP and SNOMED-CT is feasible and beneficial.
• This method offers a scalable solution to improve efficiency in clinical documentation and healthcare administration.
• Further development and implementation can lead to substantial time and resource savings in healthcare systems.