CodeMagic: Semi-Automatic Assignment of ICD-10-AM Codes to Patient Records

There is huge amount of data generated in modern hospitals, such as medical records, lab reports and doctor notes. Health care systems employ a large number of categorization and classification systems to assist data management for a variety of tasks such as patient care, record storage and retrieval, statistical analysis, insurance and billing. One of these systems is the International Classification of Diseases (ICD), which is the official system of assigning codes to diagnoses and procedures associated with hospital utilization. After a patient’s clinical treatment, ICD codes are assigned to that patient’s health records and insurance companies take these labels into account for reimbursement. On the other hand, label assignment by human annotators (clinic coders) may not be reliable. For example, in a Turkish hospital, two human experts examined 491 pre-labeled patient records and indicated that more than half of the records are assigned wrong ICD codes. An ICD-10-AM code indicates a classification of a disease, symptom, procedure or injury. Codes are organized hierarchically, where top level entries are general groupings and bottom level codes indicate specific symptoms or diseases and their location [9]. Each specific, low-level code consists of 4 or 5 digits, with a decimal after the third. Overall, there are thousands of codes that cover a broad range of medical conditions [9]. Only trained experts can properly perform the task, making the process of coding documents both expensive and unreliable since a coder must