Automated Disease Coding for Large-scale Patient Records Share

Large volumes of health-related/patient data in the form of medical records, diagnostic reports, discharge summaries, clinical notes, and medical scans, are continuously generated in modern hospitals, due to extensive digitization efforts.  ICD9 disease coding is an important task in a hospital, as part of which a trained medical coder with domain knowledge assigns disease-specific, standardized codes called ICD9 codes to a patient's admission record. As hospital billing and insurance claims are based on the assigned ICD9 codes, the coding task requires high precision, but is often prone to human error, which has resulted in an annual spending of $25 billion in the US to improve coding efficacy. Hence, automated disease coding approaches using Medical Natural Language Modeling are proposed as a significant solution to this problem.

Principal Investigator : Dr. Sowmya Kamath S | Contact : sowmyakamath@nitk.edu.in

Project Status :

Funded By :

Estimated Budget : 500000