Medical Code Prediction
15 papers with code • 7 benchmarks • 7 datasets
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).
LibrariesUse these libraries to find Medical Code Prediction models and implementations
Our method aggregates information across the document using a convolutional neural network, and uses an attention mechanism to select the most relevant segments for each of the thousands of possible codes.
The innovations of our model are two-folds: it utilizes a multi-filter convolutional layer to capture various text patterns with different lengths and a residual convolutional layer to enlarge the receptive field.
MIMIC-III (‘Medical Information Mart for Intensive Care’) is a large, single-center database comprising information relating to patients admitted to critical care units at a large tertiary care hospital.
In this paper, we propose a new label attention model for automatic ICD coding, which can handle both the various lengths and the interdependence of the ICD code related text fragments.
Explainable Automated Coding of Clinical Notes using Hierarchical Label-wise Attention Networks and Label Embedding Initialisation
LE initialisation consistently boosted most deep learning models for automated medical coding.
Our approach substantially outperforms previous results on top-50 medical code prediction on MIMIC-III dataset.
Nevertheless, automated medical coding is still challenging because of the imbalanced class problem, complex code association, and noise in lengthy documents.
Medical coding translates professionally written medical reports into standardized codes, which is an essential part of medical information systems and health insurance reimbursement.
So we propose a model based on bidirectional encoder representations from transformers (BERT) using the sequence attention method for automatic ICD code assignment.
To address this problem, we propose a two-stage framework to improve automatic ICD coding by capturing the label correlation.