Inferring multimodal latent topics from electronic health records
© 2020, The Author(s). Electronic health records (EHR) are rich heterogeneous collections of patient health information, whose broad adoption provides clinicians and researchers unprecedented opportunities for health informatics, disease-risk prediction, actionable clinical recommendations, and prec...
| Main Authors: | , , , , , , , , , , , , , , , , , , |
|---|---|
| Other Authors: | |
| Format: | Article |
| Language: | English |
| Published: |
Springer Science and Business Media LLC
2021
|
| Online Access: | https://hdl.handle.net/1721.1/136021 |
| _version_ | 1811088567273783296 |
|---|---|
| author | Li, Yue Nair, Pratheeksha Lu, Xing Han Wen, Zhi Wang, Yuening Dehaghi, Amir Ardalan Kalantari Miao, Yan Liu, Weiqi Ordog, Tamas Biernacka, Joanna M Ryu, Euijung Olson, Janet E Frye, Mark A Liu, Aihua Guo, Liming Marelli, Ariane Ahuja, Yuri Davila-Velderrain, Jose Kellis, Manolis |
| author2 | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory |
| author_facet | Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory Li, Yue Nair, Pratheeksha Lu, Xing Han Wen, Zhi Wang, Yuening Dehaghi, Amir Ardalan Kalantari Miao, Yan Liu, Weiqi Ordog, Tamas Biernacka, Joanna M Ryu, Euijung Olson, Janet E Frye, Mark A Liu, Aihua Guo, Liming Marelli, Ariane Ahuja, Yuri Davila-Velderrain, Jose Kellis, Manolis |
| author_sort | Li, Yue |
| collection | MIT |
| description | © 2020, The Author(s). Electronic health records (EHR) are rich heterogeneous collections of patient health information, whose broad adoption provides clinicians and researchers unprecedented opportunities for health informatics, disease-risk prediction, actionable clinical recommendations, and precision medicine. However, EHRs present several modeling challenges, including highly sparse data matrices, noisy irregular clinical notes, arbitrary biases in billing code assignment, diagnosis-driven lab tests, and heterogeneous data types. To address these challenges, we present MixEHR, a multi-view Bayesian topic model. We demonstrate MixEHR on MIMIC-III, Mayo Clinic Bipolar Disorder, and Quebec Congenital Heart Disease EHR datasets. Qualitatively, MixEHR disease topics reveal meaningful combinations of clinical features across heterogeneous data types. Quantitatively, we observe superior prediction accuracy of diagnostic codes and lab test imputations compared to the state-of-art methods. We leverage the inferred patient topic mixtures to classify target diseases and predict mortality of patients in critical conditions. In all comparison, MixEHR confers competitive performance and reveals meaningful disease-related topics. |
| first_indexed | 2024-09-23T14:04:04Z |
| format | Article |
| id | mit-1721.1/136021 |
| institution | Massachusetts Institute of Technology |
| language | English |
| last_indexed | 2024-09-23T14:04:04Z |
| publishDate | 2021 |
| publisher | Springer Science and Business Media LLC |
| record_format | dspace |
| spelling | mit-1721.1/1360212023-12-22T18:46:25Z Inferring multimodal latent topics from electronic health records Li, Yue Nair, Pratheeksha Lu, Xing Han Wen, Zhi Wang, Yuening Dehaghi, Amir Ardalan Kalantari Miao, Yan Liu, Weiqi Ordog, Tamas Biernacka, Joanna M Ryu, Euijung Olson, Janet E Frye, Mark A Liu, Aihua Guo, Liming Marelli, Ariane Ahuja, Yuri Davila-Velderrain, Jose Kellis, Manolis Massachusetts Institute of Technology. Computer Science and Artificial Intelligence Laboratory © 2020, The Author(s). Electronic health records (EHR) are rich heterogeneous collections of patient health information, whose broad adoption provides clinicians and researchers unprecedented opportunities for health informatics, disease-risk prediction, actionable clinical recommendations, and precision medicine. However, EHRs present several modeling challenges, including highly sparse data matrices, noisy irregular clinical notes, arbitrary biases in billing code assignment, diagnosis-driven lab tests, and heterogeneous data types. To address these challenges, we present MixEHR, a multi-view Bayesian topic model. We demonstrate MixEHR on MIMIC-III, Mayo Clinic Bipolar Disorder, and Quebec Congenital Heart Disease EHR datasets. Qualitatively, MixEHR disease topics reveal meaningful combinations of clinical features across heterogeneous data types. Quantitatively, we observe superior prediction accuracy of diagnostic codes and lab test imputations compared to the state-of-art methods. We leverage the inferred patient topic mixtures to classify target diseases and predict mortality of patients in critical conditions. In all comparison, MixEHR confers competitive performance and reveals meaningful disease-related topics. 2021-10-27T20:30:26Z 2021-10-27T20:30:26Z 2020 2021-01-05T19:26:58Z Article http://purl.org/eprint/type/JournalArticle https://hdl.handle.net/1721.1/136021 en 10.1038/S41467-020-16378-3 Nature Communications Creative Commons Attribution 4.0 International license https://creativecommons.org/licenses/by/4.0/ application/pdf Springer Science and Business Media LLC Nature |
| spellingShingle | Li, Yue Nair, Pratheeksha Lu, Xing Han Wen, Zhi Wang, Yuening Dehaghi, Amir Ardalan Kalantari Miao, Yan Liu, Weiqi Ordog, Tamas Biernacka, Joanna M Ryu, Euijung Olson, Janet E Frye, Mark A Liu, Aihua Guo, Liming Marelli, Ariane Ahuja, Yuri Davila-Velderrain, Jose Kellis, Manolis Inferring multimodal latent topics from electronic health records |
| title | Inferring multimodal latent topics from electronic health records |
| title_full | Inferring multimodal latent topics from electronic health records |
| title_fullStr | Inferring multimodal latent topics from electronic health records |
| title_full_unstemmed | Inferring multimodal latent topics from electronic health records |
| title_short | Inferring multimodal latent topics from electronic health records |
| title_sort | inferring multimodal latent topics from electronic health records |
| url | https://hdl.handle.net/1721.1/136021 |
| work_keys_str_mv | AT liyue inferringmultimodallatenttopicsfromelectronichealthrecords AT nairpratheeksha inferringmultimodallatenttopicsfromelectronichealthrecords AT luxinghan inferringmultimodallatenttopicsfromelectronichealthrecords AT wenzhi inferringmultimodallatenttopicsfromelectronichealthrecords AT wangyuening inferringmultimodallatenttopicsfromelectronichealthrecords AT dehaghiamirardalankalantari inferringmultimodallatenttopicsfromelectronichealthrecords AT miaoyan inferringmultimodallatenttopicsfromelectronichealthrecords AT liuweiqi inferringmultimodallatenttopicsfromelectronichealthrecords AT ordogtamas inferringmultimodallatenttopicsfromelectronichealthrecords AT biernackajoannam inferringmultimodallatenttopicsfromelectronichealthrecords AT ryueuijung inferringmultimodallatenttopicsfromelectronichealthrecords AT olsonjanete inferringmultimodallatenttopicsfromelectronichealthrecords AT fryemarka inferringmultimodallatenttopicsfromelectronichealthrecords AT liuaihua inferringmultimodallatenttopicsfromelectronichealthrecords AT guoliming inferringmultimodallatenttopicsfromelectronichealthrecords AT marelliariane inferringmultimodallatenttopicsfromelectronichealthrecords AT ahujayuri inferringmultimodallatenttopicsfromelectronichealthrecords AT davilavelderrainjose inferringmultimodallatenttopicsfromelectronichealthrecords AT kellismanolis inferringmultimodallatenttopicsfromelectronichealthrecords |