Automated ICD-10 code assignment of nonstandard diagnoses via a two-stage framework

作者：

Highlights：

• We propose a two-stage code assignment framework based on the hierarchical structure of the ICD-10 code, which can improve the performance of the automated code assignment.

• We propose a text-matching model called DNMM, which can produce a satisfying matching result to enhance the assignment of the ICD-10 codes.

• We propose an approach to overcome the training data sparsity issue by introducing more supervised information.

• The evaluation results show that our proposed framework achieves good performance.

摘要

•We propose a two-stage code assignment framework based on the hierarchical structure of the ICD-10 code, which can improve the performance of the automated code assignment.•We propose a text-matching model called DNMM, which can produce a satisfying matching result to enhance the assignment of the ICD-10 codes.•We propose an approach to overcome the training data sparsity issue by introducing more supervised information.•The evaluation results show that our proposed framework achieves good performance.

论文关键词：Automated ICD-10 code assignment,Two-stage framework,Text-matching

论文评审过程：Received 6 April 2020, Revised 18 July 2020, Accepted 7 August 2020, Available online 15 August 2020, Version of Record 29 August 2020.

论文官网地址：https://doi.org/10.1016/j.artmed.2020.101939