Artificial intelligence (AI)-based methods have emerged as powerful tools to transform medical care. Although machine learning classifiers (MLCs) have already demonstrated strong performance in image-based diagnoses, analysis of diverse and massive electronic health record (EHR) data remains challenging. Here, we show that MLCs can query EHRs in a manner similar to the hypothetico-deductive reasoning used by physicians and unearth associations that previous statistical methods have not found. Our model applies an automated natural language processing system using deep learning techniques to extract clinically relevant information from EHRs. In total, 101.6 million data points from 1,362,559 pediatric patient visits presenting to a major referral center were analyzed to train and validate the framework. Our model demonstrates high diagnostic accuracy across multiple organ systems and is comparable to experienced pediatricians in diagnosing common childhood diseases. Our study provides a proof of concept for implementing an AI-based system as a means to aid physicians in tackling large amounts of data, augmenting diagnostic evaluations, and to provide clinical decision support in cases of diagnostic uncertainty or complexity. Although this impact may be most evident in areas where healthcare providers are in relative shortage, the benefits of such an AI system are likely to be universal.
Description
Evaluation and accurate diagnoses of pediatric diseases using artificial intelligence | Nature Medicine
%0 Journal Article
%1 liang2019evaluation
%A Liang, Huiying
%A Tsui, Brian Y.
%A Ni, Hao
%A Valentim, Carolina C. S.
%A Baxter, Sally L.
%A Liu, Guangjian
%A Cai, Wenjia
%A Kermany, Daniel S.
%A Sun, Xin
%A Chen, Jiancong
%A He, Liya
%A Zhu, Jie
%A Tian, Pin
%A Shao, Hua
%A Zheng, Lianghong
%A Hou, Rui
%A Hewett, Sierra
%A Li, Gen
%A Liang, Ping
%A Zang, Xuan
%A Zhang, Zhiqi
%A Pan, Liyan
%A Cai, Huimin
%A Ling, Rujuan
%A Li, Shuhua
%A Cui, Yongwang
%A Tang, Shusheng
%A Ye, Hong
%A Huang, Xiaoyan
%A He, Waner
%A Liang, Wenqing
%A Zhang, Qing
%A Jiang, Jianmin
%A Yu, Wei
%A Gao, Jianqun
%A Ou, Wanxing
%A Deng, Yingmin
%A Hou, Qiaozhen
%A Wang, Bei
%A Yao, Cuichan
%A Liang, Yan
%A Zhang, Shu
%A Duan, Yaou
%A Zhang, Runze
%A Gibson, Sarah
%A Zhang, Charlotte L.
%A Li, Oulan
%A Zhang, Edward D.
%A Karin, Gabriel
%A Nguyen, Nathan
%A Wu, Xiaokang
%A Wen, Cindy
%A Xu, Jie
%A Xu, Wenqin
%A Wang, Bochu
%A Wang, Winston
%A Li, Jing
%A Pizzato, Bianca
%A Bao, Caroline
%A Xiang, Daoman
%A He, Wanting
%A He, Suiqin
%A Zhou, Yugui
%A Haw, Weldon
%A Goldbaum, Michael
%A Tremoulet, Adriana
%A Hsu, Chun-Nan
%A Carter, Hannah
%A Zhu, Long
%A Zhang, Kang
%A Xia, Huimin
%D 2019
%J Nature Medicine
%K accurate ai deep diagnoses evaluation learning
%N 3
%P 433--438
%R 10.1038/s41591-018-0335-9
%T Evaluation and accurate diagnoses of pediatric diseases using artificial intelligence
%U https://doi.org/10.1038/s41591-018-0335-9
%V 25
%X Artificial intelligence (AI)-based methods have emerged as powerful tools to transform medical care. Although machine learning classifiers (MLCs) have already demonstrated strong performance in image-based diagnoses, analysis of diverse and massive electronic health record (EHR) data remains challenging. Here, we show that MLCs can query EHRs in a manner similar to the hypothetico-deductive reasoning used by physicians and unearth associations that previous statistical methods have not found. Our model applies an automated natural language processing system using deep learning techniques to extract clinically relevant information from EHRs. In total, 101.6 million data points from 1,362,559 pediatric patient visits presenting to a major referral center were analyzed to train and validate the framework. Our model demonstrates high diagnostic accuracy across multiple organ systems and is comparable to experienced pediatricians in diagnosing common childhood diseases. Our study provides a proof of concept for implementing an AI-based system as a means to aid physicians in tackling large amounts of data, augmenting diagnostic evaluations, and to provide clinical decision support in cases of diagnostic uncertainty or complexity. Although this impact may be most evident in areas where healthcare providers are in relative shortage, the benefits of such an AI system are likely to be universal.
@article{liang2019evaluation,
abstract = {Artificial intelligence (AI)-based methods have emerged as powerful tools to transform medical care. Although machine learning classifiers (MLCs) have already demonstrated strong performance in image-based diagnoses, analysis of diverse and massive electronic health record (EHR) data remains challenging. Here, we show that MLCs can query EHRs in a manner similar to the hypothetico-deductive reasoning used by physicians and unearth associations that previous statistical methods have not found. Our model applies an automated natural language processing system using deep learning techniques to extract clinically relevant information from EHRs. In total, 101.6 million data points from 1,362,559 pediatric patient visits presenting to a major referral center were analyzed to train and validate the framework. Our model demonstrates high diagnostic accuracy across multiple organ systems and is comparable to experienced pediatricians in diagnosing common childhood diseases. Our study provides a proof of concept for implementing an AI-based system as a means to aid physicians in tackling large amounts of data, augmenting diagnostic evaluations, and to provide clinical decision support in cases of diagnostic uncertainty or complexity. Although this impact may be most evident in areas where healthcare providers are in relative shortage, the benefits of such an AI system are likely to be universal.},
added-at = {2019-06-04T11:24:17.000+0200},
author = {Liang, Huiying and Tsui, Brian Y. and Ni, Hao and Valentim, Carolina C. S. and Baxter, Sally L. and Liu, Guangjian and Cai, Wenjia and Kermany, Daniel S. and Sun, Xin and Chen, Jiancong and He, Liya and Zhu, Jie and Tian, Pin and Shao, Hua and Zheng, Lianghong and Hou, Rui and Hewett, Sierra and Li, Gen and Liang, Ping and Zang, Xuan and Zhang, Zhiqi and Pan, Liyan and Cai, Huimin and Ling, Rujuan and Li, Shuhua and Cui, Yongwang and Tang, Shusheng and Ye, Hong and Huang, Xiaoyan and He, Waner and Liang, Wenqing and Zhang, Qing and Jiang, Jianmin and Yu, Wei and Gao, Jianqun and Ou, Wanxing and Deng, Yingmin and Hou, Qiaozhen and Wang, Bei and Yao, Cuichan and Liang, Yan and Zhang, Shu and Duan, Yaou and Zhang, Runze and Gibson, Sarah and Zhang, Charlotte L. and Li, Oulan and Zhang, Edward D. and Karin, Gabriel and Nguyen, Nathan and Wu, Xiaokang and Wen, Cindy and Xu, Jie and Xu, Wenqin and Wang, Bochu and Wang, Winston and Li, Jing and Pizzato, Bianca and Bao, Caroline and Xiang, Daoman and He, Wanting and He, Suiqin and Zhou, Yugui and Haw, Weldon and Goldbaum, Michael and Tremoulet, Adriana and Hsu, Chun-Nan and Carter, Hannah and Zhu, Long and Zhang, Kang and Xia, Huimin},
biburl = {https://www.bibsonomy.org/bibtex/233f283cd00d1483d1dc716cfa152ddcb/nosebrain},
description = {Evaluation and accurate diagnoses of pediatric diseases using artificial intelligence | Nature Medicine},
doi = {10.1038/s41591-018-0335-9},
interhash = {a240eb0358fd832029d953102c110ab9},
intrahash = {33f283cd00d1483d1dc716cfa152ddcb},
issn = {1546170X},
journal = {Nature Medicine},
keywords = {accurate ai deep diagnoses evaluation learning},
number = 3,
pages = {433--438},
refid = {Liang2019},
timestamp = {2019-06-04T11:24:17.000+0200},
title = {Evaluation and accurate diagnoses of pediatric diseases using artificial intelligence},
url = {https://doi.org/10.1038/s41591-018-0335-9},
volume = 25,
year = 2019
}