Automatic Matching of ICD-10 codes to Diagnoses in Discharge Letters

This paper presents an approach for automatic mapping of International Classification of Diseases 10th revision (ICD-10) codes to diagnoses extracted from discharge letters. The proposed algorithm is designed for processing free text documents in Bulgarian language. Diseases are often described in the medical patient records as free text using terminology, phrases and paraphrases which differ significantly from those used in ICD-10 classification. In this way the task of diseases recognition (which practically means e.g. assigning standardized ICD codes to diseases’ names) is an important natural language processing (NLP) challenge. The approach is based on multiclass Support Vector Machines method, where each ICD-10 4 character classification code is considered as single class. The problem is reduced to multiple binary classifiers and classification is done by a max-wins voting strategy.