Skip to Main content Skip to Navigation

Using Case-Based Reasoning and Argumentation to Assist Medical Coding

Michael Schnell 1, 2
1 K Team - Data Science, Knowledge, Reasoning and Engineering
LORIA - NLPKD - Department of Natural Language Processing & Knowledge Discovery
Abstract : The aim of the National Cancer Registry (NCR) in Luxembourg is to collect data about cancer and the quality of cancer treatment. To obtain high quality data that can be compared with other registries or countries, the NCR follows international coding standards and rules, such as the International Classification of Diseases for Oncology (ICD-O). These standards are extensive and complex, which complicates the data collection process. The operators, i.e. the people in charge of this process, are often confronted with situations where data is missing or contradictory, preventing the application of the provided guidelines. To assist in their effort, the coding experts of the NCR answer coding questions asked by operators. This time consuming for experts. To help reduce this burden on experts and to facilitate the operators’ task, this project aims at implementing a coding assistant that would answer coding questions. From a scientific point of view, this thesis tackles the problem of extracting the information from a set of data sources under a given set of rules and guidelines. Case-based reasoning has been chosen as the method for solving this problem given its similarity with the reasoning process of the coding experts. The method designed to solve this problem relies on arguments provided by coding experts in the context of previously solved problems. This document presents how these arguments are used to identify similar problems and to explain the computed solution to both operators and coding experts. A preliminary evaluation has assessed the designed method and has highlighted key areas to improve. While this work focused on cancer registries and medical coding, this method could be generalized to other domains.
Document type :
Complete list of metadata
Contributor : Michael Schnell Connect in order to contact the contributor
Submitted on : Sunday, January 17, 2021 - 11:08:46 AM
Last modification on : Saturday, October 16, 2021 - 11:26:10 AM
Long-term archiving on: : Sunday, April 18, 2021 - 6:08:08 PM


Files produced by the author(s)


  • HAL Id : tel-03112664, version 1


Michael Schnell. Using Case-Based Reasoning and Argumentation to Assist Medical Coding. Artificial Intelligence [cs.AI]. Université de Lorraine, 2020. English. ⟨NNT : 2020LORR0168⟩. ⟨tel-03112664⟩



Record views


Files downloads