Automated medical coding is an area in Clinical Natural Language Processing to assign diagnosis or procedure medical codes to free-text clinical notes. The domain is a sub-field of document classification and information extraction.
Below is a curation of papers (mostly peer-reviewed) and datasets in this field, mainly since the application of deep learning to this field (around 2017). Given the many new papers and datasets published, I may have lost some of them.
The repository will not be actively updated given the many publications in this domain. But hope it provides a good starting point for your research on this topic!
-Automated Clinical Coding: What, Why, and Where We Are? - a perspective paper about automated clinical coding, its current states and technical challenges, in npj Digital Medicine, 2022
-A Unified Review of Deep Learning for Automated Medical Coding - a focus on deep learning and its main building blocks specific for clinical coding as a healthcare AI application, in ACM Computing Surveys, 2024
-A review on deep neural networks for ICD coding - technical summary, deep learning, summary of public datasets, in IEEE TKDE, 2022
-AI-based ICD coding and classification approaches using discharge summaries: A systematic literature review, (arXiv version) - review of work between 2010-2021, in Expert Systems with Applications, 2022
-Computer-assisted clinical coding: A narrative review of the literature on its benefits, limitations, implementation and impact on clinical coding professionals - an application-oriented review of computer-assisted clinical coding, in Health Information Management Journal, 2020
-A systematic literature review of automated clinical coding and classification systems - one of the earliest reviews in automated clinical coding, in JAMIA, 2010
We tried to use tags to provide simple and shallow sub-topics for easy retrieval. These are by no means accurate or exhaustive.
The tags so far include: resource, knowledge (knowledge-augmented with ontologies/hierarchies/descriptions), explainability, human-in-the-loop, few/zero-shot, analysis (analysis-focused), multimodal, PLM (pre-trained language models), LLM (large language models), CNN (Convolutional Neural Networks), NER+L (named entity recognition and linking), GNN (Graph Neural Networks).
-ClinicalMamba: A Generative Clinical Language Model on Longitudinal Clinical Notes - LLM, few/zero-shot
-Can GPT-3.5 Generate and Code Discharge Summaries? - LLM, analysis, few/zero-shot
-Surpassing GPT-4 Medical Coding with a Two-Stage Approach - LLM, analysis
-Automated clinical coding using off-the-shelf large language models - NeurIPS 2023 Deep Generative Models for Health workshop LLM, PLM, few/zero-shot, knowledge
-Towards Automatic ICD Coding via Knowledge Enhanced Multi-Task Learning - in CIKM 2023 knowledge, NER+L, GNN
-MDACE: MIMIC Documents Annotated with Code Evidence - in ACL 2023 - official repository resource, explainability
-Automated Medical Coding on MIMIC-III and MIMIC-IV: A Critical Review and Replicability Study - in SIGIR 2023 - official implementation resource
-Multi-label Few-shot ICD Coding as Autoregressive Generation with Prompt - in AAAI 2023 PLM
-AnEMIC: A Framework for Benchmarking ICD Coding Models - in EMNLP 2022 demo
-Can Current Explainability Help Provide References in Clinical Notes to Support Humans Annotate Medical Codes? - in LOUHI@EMNLP 2022 explainability
-This Patient Looks Like That Patient: Prototypical Networks for Interpretable Diagnosis Prediction from Clinical Text - in AACL-IJCNLP 2022 explainability
-Automatic ICD Coding Exploiting Discourse Structure and Reconciled Code Embeddings - in COLING 2022
-TreeMAN: Tree-enhanced Multimodal Attention Network for ICD Coding - in COLING 2022 multimodal
-Knowledge Injected Prompt Based Fine-tuning for Multi-label Few-shot ICD Coding - in Findings of EMNLP 2022 knowledge PLM few/zero-shot
-Concatenating BioMed-Transformers to Tackle Long Medical Documents and to Improve the Prediction of Tail-End Labels - in ICANN 2022 PLM few/zero-shot
-Entity Anchored ICD Coding - in AMIA 2022 knowledge NER+L
-An exploratory data analysis: the performance differences of a medical code prediction system on different demographic groups - in ClinicalNLP@NAACL 2022 analysis
-PLM-ICD: Automatic ICD Coding with Pretrained Language Models - in ClinicalNLP@NAACL 2022 - official implementation PLM
-Horses to Zebras: Ontology-Guided Data Augmentation and Synthesis for ICD-9 Coding - in BioNLP@ACL 2022 knowledge NER+L
-Model Distillation for Faithful Explanations of Medical Code Predictions - in BioNLP@ACL 2022 explainability
-ICDBigBird: A Contextual Embedding Model for ICD Code Classification - in BioNLP@ACL 2022 PLM
-Code Synonyms Do Matter: Multiple Synonyms Matching Network for Automatic ICD Coding - in ACL 2022 - official implementation knowledge
-A Novel Framework Based on Medical Concept Driven Attention for Explainable Medical Code Prediction via External Knowledge - in Findings of the ACL 2022 knowledge, explainability
-Active learning for medical code assignment - ACM CHIL 2021 workshop human-in-the-loop
-Multitask Recalibrated Aggregation Network for Medical Code Prediction - multi-task learning - in ECML-PKDD 2021 knowledge
-Effective Convolutional Attention Network for Multi-label Clinical Document Classification - in EMNLP 2021 CNN
-Meta-LMTC: Meta-Learning for Large-Scale Multi-Label Text Classification - Meta-learning for few- or zero-shot multi-label classification - in EMNLP 2021
-CoPHE: A Count-Preserving Hierarchical Evaluation Metric in Large-Scale Multi-Label Text Classification A novel metric for hierarchical multi-label classification, applied to MIMIC-III ICD coding - in EMNLP 2021 knowledge
-Description-based Label Attention Classifier for Explainable ICD-9 Classification - with Longformer and label descriptions - in W-NUT@EMNLP 2021 knowledge
-Read, Attend, and Code: Pushing the Limits of Medical Codes Prediction from Clinical Notes by Machines - Attention-based model, human-level coding results - in MLHC 2021 - leaderboard on paper with code - video knowledge
-Does the Magic of BERT Apply to Medical Code Assignment? A Quantitative Study Evaluation of BERT on MIMIC-III ICD coding. in Computers in Biology and Medicine, 2021 PLM
-Counterfactual Supporting Facts Extraction for Explainable Medical Record Based Diagnosis with Graph Network - in NAACL 2021 explainability
-Automatic ICD Coding via Interactive Shared Representation Networks with Self-distillation Mechanism in ACL 2021
-Analyzing Code Embeddings for Coding Clinical Narratives - in Findings of the ACL 2021 knowledge
-JLAN: medical code prediction via joint learning attention networks and denoising mechanism in BMC Bioinformatics, 2021
-Explainable Automated Coding of Clinical Notes using Hierarchical Label-Wise Attention Networks and Label Embedding Initialisation in JBI, 2021. explainability
-Multi-label Few/Zero-shot Learning with Knowledge Aggregated from Multiple Label Graphs. In EMNLP 2020 knowledge, few/zero-shot
-An Empirical Study on Large-Scale Multi-Label Text Classification Including Few and Zero-Shot Labels - (i) Improvement on zero-shot learning and (ii) the idea of Graph-aware Annotation Proximity (GAP), an graph-based look into the coding process, and (iii) BERTs' underpreformance on MIMIC-III. In EMNLP 2020 few/zero-shot
-BERT-XML: Large Scale Automated ICD Coding Using BERT Pretraining - in ClinicalNLP workshop at EMNLP 2020 PLM
-A Label Attention Model for ICD Coding from Clinical Text - In IJCAI 2020.
-Generalized Zero-Shot Text Classification for ICD Coding - Generalised Zero-shot learning with Generative adversial training, the ICD hierarchy with descriptions, and Graph Recurrent Neural Networks. In IJCAI 2020. few/zero-shot
-Towards Interpretable Clinical Diagnosis with Bayesian Network Ensembles Stacked on Entity-Aware CNNs in ACL 2020. explainability
-HyperCore: Hyperbolic and Co-graph Representation for Automatic ICD Coding - Hyperbolic embedding + Graph Convolutional Networks. In ACL 2020. knowledge
-Clinical-Coder: Assigning Interpretable ICD-10 Codes to Chinese Clinical Notes in System Demonstrations, ACL 2020. explainability
-Experimental Evaluation and Development of a Silver-Standard for the MIMIC-III Clinical Coding Dataset in BioNLP at ACL 2020. knowledge, analysis, NER+L
-Dynamically Extracting Outcome-Specific Problem Lists from Clinical Notes with Guided Multi-Headed Attention in PMLR 2020. - official code. explainability
-EHR Coding with Multi-scale Feature Attention and Structured Knowledge Graph Propagation - in CIKM 2019 knowledge
-ICD Coding from Clinical Text Using Multi-Filter Residual Convolutional Neural Network - in AAAI 2019 CNN
-Multimodal Machine Learning for Automated ICD Coding Ensembling models from unstructured text, semi-structured text and structured tabular data for ICD coding. (Keyang Xu, Mike Lam, Jingzhi Pang, Xin Gao, Charlotte Band, Piyush Mathur, Frank Papay, Ashish K. Khanna, Jacek B. Cywinski, Kamal Maheshwari, Pengtao Xie, Eric P. Xing ; Proceedings of the 4th Machine Learning for Healthcare Conference, PMLR 106:197-215, 2019.) multimodal
-Ontological attention ensembles for capturing semantic concepts in ICD code prediction from clinical text - Multi-view convolution + multi-task learning. In LOUHI 2019 at EMNLP. knowledge
-Clinical Concept Extraction for Document-Level Coding - In BioNLP@ACL 2019. knowledge NER+L
-Few-Shot and Zero-Shot Multi-Label Learning for Structured Label Spaces - Few-shot and zero-shot learning with Graph Convolutional Neural Networks and the ICD hierarchy with descriptions. In EMNLP 2018. few/zero-shot
-Explainable Prediction of Medical Codes from Clinical Text - CNN with labelwise attention and the benchmark MIMIC preprocessed datasets. In NAACL-HLT 2018. explainability, resource
-Towards automated clinical coding - International Journal of Medical Informatics, 2018
-Towards Automated ICD Coding Using Deep Learning
-Automatic Diagnosis Coding of Radiology Reports: A Comparison of Deep Learning and Conventional Classification Methods in BioNLP 2017
-MIMIC-IV and MIMIC-IV-Note
