An explainable CNN approach for medical codes prediction from clinical text

Abstract Background Clinical notes are unstructured text documents generated by clinicians during patient encounters, generally are annotated with International Classification of Diseases (ICD) codes, which give formatted information about the diagnosis and treatment. ICD code has shown its potentia...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Shuyuan Hu, Fei Teng, Lufei Huang, Jun Yan, Haibo Zhang
Formato:	article
Lenguaje:	EN
Publicado:	BMC 2021
Materias:	ICD coding Machine learning Attention mechanism Convolutional neural network Computer applications to medicine. Medical informatics R858-859.7
Acceso en línea:	https://doaj.org/article/3fa31aee624647a9993899b1ea411720
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:3fa31aee624647a9993899b1ea411720
record_format	dspace
spelling	oai:doaj.org-article:3fa31aee624647a9993899b1ea4117202021-11-21T12:28:50ZAn explainable CNN approach for medical codes prediction from clinical text10.1186/s12911-021-01615-61472-6947https://doaj.org/article/3fa31aee624647a9993899b1ea4117202021-11-01T00:00:00Zhttps://doi.org/10.1186/s12911-021-01615-6https://doaj.org/toc/1472-6947Abstract Background Clinical notes are unstructured text documents generated by clinicians during patient encounters, generally are annotated with International Classification of Diseases (ICD) codes, which give formatted information about the diagnosis and treatment. ICD code has shown its potentials in many fields, but manual coding is labor-intensive and error-prone, lead to researches of automatic coding. Two specific challenges of this task are (1) given an annotated clinical notes, the reasons behind specific diagnoses and treatments are implicit; (2) explainability is important for practical automatic coding method, the method should not only explain its prediction output but also have explainable internal mechanics. This study aims to develop an explainable CNN approach to address these two challenges. Method Our key idea is that for the automatic ICD coding task, the presence of informative snippets in the clinical text that correlated with each code plays an important role in the prediction of codes, and an informative snippet can be considered as a local and low-level feature. We infer that there exists a correspondence between a convolution filter and a local and low-level feature. Base on the inference, we come up with the Shallow and Wide Attention convolutional Mechanism (SWAM) to improve the CNN-based models’ ability to learn local and low-level features for each label. Results We evaluate our approach on MIMIC-III, an open-access dataset of ICU medical records. Our approach substantially outperforms previous results on top-50 medical code prediction on MIMIC-III dataset, the precision of the worst-performing 10% labels in previous works is increased from 0% to 53% on average. We attribute this improvement to SWAM, by which the wide architecture with attention mechanism gives the model ability to more extensively learn the unique features of different codes, and we prove it by an ablation experiment. Besides, we perform manual analysis of the performance imbalance between different codes, and preliminary conclude the characteristics that determine the difficulty of learning specific codes. Conclusions Our main contributions can be summarized into the following three: (1) We present local and low-level features, a.k.a. informative snippets play an important role in the automatic ICD coding task, and the informative snippets extracted from the clinical text provide explanations for each code. (2) We propose that there exists a correspondence between a convolution filter and a local and low-level feature. A combination of wide and shallow convolutional layer and attention layer can help the CNN-based models better learn local and low-level features. (3) We improved the precision of the worst-performing 10% labels from 0 to 53% on average.Shuyuan HuFei TengLufei HuangJun YanHaibo ZhangBMCarticleICD codingMachine learningAttention mechanismConvolutional neural networkComputer applications to medicine. Medical informaticsR858-859.7ENBMC Medical Informatics and Decision Making, Vol 21, Iss S9, Pp 1-12 (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	ICD coding Machine learning Attention mechanism Convolutional neural network Computer applications to medicine. Medical informatics R858-859.7
spellingShingle	ICD coding Machine learning Attention mechanism Convolutional neural network Computer applications to medicine. Medical informatics R858-859.7 Shuyuan Hu Fei Teng Lufei Huang Jun Yan Haibo Zhang An explainable CNN approach for medical codes prediction from clinical text
description	Abstract Background Clinical notes are unstructured text documents generated by clinicians during patient encounters, generally are annotated with International Classification of Diseases (ICD) codes, which give formatted information about the diagnosis and treatment. ICD code has shown its potentials in many fields, but manual coding is labor-intensive and error-prone, lead to researches of automatic coding. Two specific challenges of this task are (1) given an annotated clinical notes, the reasons behind specific diagnoses and treatments are implicit; (2) explainability is important for practical automatic coding method, the method should not only explain its prediction output but also have explainable internal mechanics. This study aims to develop an explainable CNN approach to address these two challenges. Method Our key idea is that for the automatic ICD coding task, the presence of informative snippets in the clinical text that correlated with each code plays an important role in the prediction of codes, and an informative snippet can be considered as a local and low-level feature. We infer that there exists a correspondence between a convolution filter and a local and low-level feature. Base on the inference, we come up with the Shallow and Wide Attention convolutional Mechanism (SWAM) to improve the CNN-based models’ ability to learn local and low-level features for each label. Results We evaluate our approach on MIMIC-III, an open-access dataset of ICU medical records. Our approach substantially outperforms previous results on top-50 medical code prediction on MIMIC-III dataset, the precision of the worst-performing 10% labels in previous works is increased from 0% to 53% on average. We attribute this improvement to SWAM, by which the wide architecture with attention mechanism gives the model ability to more extensively learn the unique features of different codes, and we prove it by an ablation experiment. Besides, we perform manual analysis of the performance imbalance between different codes, and preliminary conclude the characteristics that determine the difficulty of learning specific codes. Conclusions Our main contributions can be summarized into the following three: (1) We present local and low-level features, a.k.a. informative snippets play an important role in the automatic ICD coding task, and the informative snippets extracted from the clinical text provide explanations for each code. (2) We propose that there exists a correspondence between a convolution filter and a local and low-level feature. A combination of wide and shallow convolutional layer and attention layer can help the CNN-based models better learn local and low-level features. (3) We improved the precision of the worst-performing 10% labels from 0 to 53% on average.
format	article
author	Shuyuan Hu Fei Teng Lufei Huang Jun Yan Haibo Zhang
author_facet	Shuyuan Hu Fei Teng Lufei Huang Jun Yan Haibo Zhang
author_sort	Shuyuan Hu
title	An explainable CNN approach for medical codes prediction from clinical text
title_short	An explainable CNN approach for medical codes prediction from clinical text
title_full	An explainable CNN approach for medical codes prediction from clinical text
title_fullStr	An explainable CNN approach for medical codes prediction from clinical text
title_full_unstemmed	An explainable CNN approach for medical codes prediction from clinical text
title_sort	explainable cnn approach for medical codes prediction from clinical text
publisher	BMC
publishDate	2021
url	https://doaj.org/article/3fa31aee624647a9993899b1ea411720
work_keys_str_mv	AT shuyuanhu anexplainablecnnapproachformedicalcodespredictionfromclinicaltext AT feiteng anexplainablecnnapproachformedicalcodespredictionfromclinicaltext AT lufeihuang anexplainablecnnapproachformedicalcodespredictionfromclinicaltext AT junyan anexplainablecnnapproachformedicalcodespredictionfromclinicaltext AT haibozhang anexplainablecnnapproachformedicalcodespredictionfromclinicaltext AT shuyuanhu explainablecnnapproachformedicalcodespredictionfromclinicaltext AT feiteng explainablecnnapproachformedicalcodespredictionfromclinicaltext AT lufeihuang explainablecnnapproachformedicalcodespredictionfromclinicaltext AT junyan explainablecnnapproachformedicalcodespredictionfromclinicaltext AT haibozhang explainablecnnapproachformedicalcodespredictionfromclinicaltext
_version_	1718419001794625536

An explainable CNN approach for medical codes prediction from clinical text

Ejemplares similares