Medical Code Prediction

Context: Prediction of medical codes from clinical notes is both a practical and essential need for every healthcare delivery organization within current medical systems. Automating annotation will save significant time and excessive effort by human coders today. A new milestone will mark a meaningful step toward fully Autonomous Medical Coding in machines reaching parity with human coders' performance in medical code prediction.

Question: What exactly is the medical code prediction problem?

Answer: Clinical notes contain much information about what precisely happened during the patient's entire stay. And those clinical notes (e.g., discharge summary) is typically long, loosely structured, consists of medical domain language, and sometimes riddled with spelling errors. So, it's a highly multi-label classification problem, and the forthcoming ICD-11 standard will add more complexity to the problem! The medical code prediction problem is to annotate this clinical note with multiple codes subset from nearly 70K total codes (in the current ICD-10 system, for example).

Papers

Recently Added Most Hyped Most Active Needs Verification Most Verified

Showing 1–27 of 27 papers

Title	Date	Tasks	Status	Hype
Uncertainty-aware abstention in medical diagnosis based on medical texts	Feb 25, 2025	Anxiety DetectionMedical Code Prediction	—Unverified	0
An Unsupervised Approach to Achieve Supervised-Level Explainability in Healthcare Records	Jun 13, 2024	Adversarial RobustnessExplainable Artificial Intelligence (XAI)	CodeCode Available	2
Effective Medical Code Prediction via Label Internal Alignment	May 9, 2023	Medical Code PredictionPrediction	—Unverified	0
Automated Medical Coding on MIMIC-III and MIMIC-IV: A Critical Review and Replicability Study	Apr 21, 2023	Medical Code Prediction	CodeCode Available	1
Can Current Explainability Help Provide References in Clinical Notes to Support Humans Annotate Medical Codes?	Oct 28, 2022	Knowledge DistillationMedical Code Prediction	—Unverified	0
Knowledge Injected Prompt Based Fine-tuning for Multi-label Few-shot ICD Coding	Oct 7, 2022	Contrastive LearningMedical Code Prediction	CodeCode Available	1
Automatic ICD Coding Exploiting Discourse Structure and Reconciled Code Embeddings	Oct 1, 2022	EpidemiologyMedical Code Prediction	CodeCode Available	0
HiCu: Leveraging Hierarchy for Curriculum Learning in Automated ICD Coding	Aug 3, 2022	Medical Code PredictionMulti-Label Classification	CodeCode Available	1
An exploratory data analysis: the performance differences of a medical code prediction system on different demographic groups	Jul 1, 2022	Medical Code Prediction	—Unverified	0
A Novel Framework Based on Medical Concept Driven Attention for Explainable Medical Code Prediction via External Knowledge	May 1, 2022	Medical Code PredictionPrediction	—Unverified	0
Code Synonyms Do Matter: Multiple Synonyms Matching Network for Automatic ICD Coding	Mar 3, 2022	Medical Code PredictionRepresentation Learning	CodeCode Available	1
A Novel Framework Based on Medical Concept Driven Attention for Explainable Medical Code Prediction via External Knowledge	Nov 16, 2021	Medical Code PredictionPrediction	—Unverified	0
Effective Convolutional Attention Network for Multi-label Clinical Document Classification	Nov 1, 2021	ClassificationDocument Classification	—Unverified	0
Multitask Balanced and Recalibrated Network for Medical Code Prediction	Sep 6, 2021	Medical Code PredictionMulti-Task Learning	CodeCode Available	1
Read, Attend, and Code: Pushing the Limits of Medical Codes Prediction from Clinical Notes by Machines	Jul 10, 2021	Medical Code PredictionMulti-Label Classification Of Biomedical Texts	—Unverified	0
Parameter Selection: Why We Should Pay More Attention to It	Jul 8, 2021	Medical Code PredictionMulti-Label Classification	—Unverified	0
Modeling Diagnostic Label Correlation for Automatic ICD Coding	Jun 24, 2021	DiagnosticMedical Code Prediction	CodeCode Available	0
Medical Code Prediction from Discharge Summary: Document to Sequence BERT using Sequence Attention	Jun 15, 2021	Medical Code PredictionMedical Diagnosis	CodeCode Available	0
Multitask Recalibrated Aggregation Network for Medical Code Prediction	Apr 2, 2021	Medical Code PredictionPrediction	CodeCode Available	1
Does the Magic of BERT Apply to Medical Code Assignment? A Quantitative Study	Mar 11, 2021	Medical Code PredictionTransfer Learning	—Unverified	0
An Explainable CNN Approach for Medical Codes Prediction from Clinical Text	Jan 14, 2021	AttributeDocument Classification	CodeCode Available	0
Explainable Automated Coding of Clinical Notes using Hierarchical Label-wise Attention Networks and Label Embedding Initialisation	Oct 29, 2020	DiagnosticMedical Code Prediction	CodeCode Available	1
A Label Attention Model for ICD Coding from Clinical Text	Jul 13, 2020	Medical Code Prediction	CodeCode Available	1
ICD Coding from Clinical Text Using Multi-Filter Residual Convolutional Neural Network	Nov 25, 2019	Medical Code Prediction	CodeCode Available	1
Medical code prediction with multi-view convolution and description-regularized label-dependent attention	Nov 5, 2018	Medical Code Prediction	—Unverified	0
Explainable Prediction of Medical Codes from Clinical Text	Feb 15, 2018	Medical Code PredictionPrediction	CodeCode Available	1
MIMIC-III, a freely accessible critical care database	May 24, 2016	Blood pressure estimationData Integration	CodeCode Available	0

Show:10 25 50

All datasets MIMIC-III MIMIC-IV ICD-10 MIMIC-IV ICD-9 MIMIC-IV-ICD-10-full MIMIC-IV-ICD10-top50 MIMIC-IV-ICD9-full MIMIC-IV-ICD9-top50

Benchmark Results

#	Model	Metric	Claimed	Verified	Status
1	GKI-ICD	Micro-F1	61.2	—	Unverified
2	PLM-CA	Micro-F1	60	—	Unverified
3	MSMN+KEPTLongformer	Micro-F1	59.9	—	Unverified
4	EffectiveCAN	Micro-F1	58.9	—	Unverified
5	Discnet+RE	Micro-F1	58.8	—	Unverified
6	RAC	Micro-F1	58.6	—	Unverified
7	MSMN	Micro-F1	58.4	—	Unverified
8	LAAT	Micro-F1	57.5	—	Unverified
9	JointLAAT	Micro-F1	57.5	—	Unverified
10	MSATT-KG	Micro-F1	55.3	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	PLM-ICD	Precision@8	69.9	—	Unverified
2	LAAT	Precision@8	68.9	—	Unverified
3	MultiResCNN	Precision@8	67.8	—	Unverified
4	CAML	Precision@8	66.8	—	Unverified
5	Bi-GRU	Precision@8	62.6	—	Unverified
6	CNN	Precision@8	60.3	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	PLM-ICD	AUC Macro	97.2	—	Unverified
2	LAAT	AUC Macro	96	—	Unverified
3	MultiResCNN	AUC Macro	95.1	—	Unverified
4	Bi-GRU	AUC Macro	93.8	—	Unverified
5	CAML	AUC Macro	90.7	—	Unverified
6	CNN	AUC Macro	89.4	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	MSMN	Macro-AUC	97.07	—	Unverified
2	Joint LAAT	Macro-AUC	93.64	—	Unverified
3	LAAT	Macro-AUC	92.96	—	Unverified
4	PLM	Macro-AUC	91.85	—	Unverified
5	CAML	Macro-AUC	89.91	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	MSMN	F1 (micro)	74.15	—	Unverified
2	PLM-ICD	F1 (micro)	73.27	—	Unverified
3	Joint LAAT	F1 (micro)	72.85	—	Unverified
4	LAAT	F1 (micro)	72.56	—	Unverified
5	CAML	F1 (micro)	67.56	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	MSMN	Macro AUC	96.79	—	Unverified
2	PLM-ICD	Macro AUC	96.61	—	Unverified
3	Joint LAAT	Macro AUC	95.57	—	Unverified
4	LAAT	Macro AUC	95.18	—	Unverified
5	CAML	Macro AUC	93.45	—	Unverified

#	Model	Metric	Claimed	Verified	Status
1	MSMN	AUC Macro	95.13	—	Unverified
2	PLM-ICD	AUC Macro	94.97	—	Unverified
3	Joint LAAT	AUC Macro	94.92	—	Unverified
4	LAAT	AUC Macro	94.88	—	Unverified
5	CAML	AUC Macro	93.07	—	Unverified