The International Classification of Diseases, Tenth Revision (ICD-10) is a medical classification system developed by the World Health Organization (WHO). It serves as a global standard for diagnosing and classifying diseases, symptoms, abnormal findings, complaints, social circumstances, and external causes of injury or diseases. ICD-10 allows for the systematic recording, analysis, interpretation, and comparison of mortality and morbidity data across different countries and time periods.
Introduced in 1990 to replace the ninth revision (ICD-9), ICD-10 provides a significantly more detailed classification system, which includes codes for diseases, signs and symptoms, abnormal findings, complaints, and external causes of injury or diseases. This detail allows for a more precise description of diagnoses and procedures, facilitating the collection of global health information for statistical purposes and epidemiological studies.
The system is organized into chapters based on the type of disease or condition, and each entry is assigned a unique code. These codes are used worldwide in health care and related industries, aiding in the management of health care, decision-making, funding allocations, and research.
ICD-10 has been updated periodically through revisions to reflect advances in health and medical science. The most current version at any time is essential for ensuring accurate and up-to-date classification of health conditions.
As for this project, we know that the LLMs could help with identifying the ICD codes from the doctors' notes, then we will try to compare how different models work differentlly in similar tasks(classificaiton)
This project is inspired by Mimic-IV-ICD: A new benchmark for eXtreme MultiLabel Classificatio, and beyond that, we would like to dig further on using different models for ICD classification.
Kefeng - Meditron 7B
Victor - Falcon 1B
John - Biomistral 7B
Yepeng - Mistral 7B
Nithin - Gemma