TY - GEN
T1 - Simultaneous prognosis and exploratory analysis of multiple chronic conditions using clinical notes
AU - Joshi, Shalmali
AU - Koyejo, Oluwasanmi
AU - Resurreccion, Kristine
AU - Ghosh, Joydeep
N1 - Publisher Copyright:
© 2015 IEEE.
PY - 2015/12/8
Y1 - 2015/12/8
N2 - We propose a novel approach for simultaneous exploratory analysis and prognosis of multiple chronic conditions that can enable healthcare professionals to harness the vast amounts of medical data available in electronic health records (EHRs) in a comprehensive manner for early treatment, the discovery of previously unknown symptoms and exploration of new associations in the context of individual chronic conditions. In particular, this manuscript focuses on the use of early clinical notes available in EHRs to diagnose multiple chronic conditions. While most methods for EHR analysis have focused on single disease predictions, multiple conditions called co morbidities tend to occur simultaneously with the primary chronic condition in many patients. To this end, we model each chronic condition as a latent variable in a novel supervised topic model which results in the prediction of multiple illnesses associated with each patient. In addition, the supervised topic model also provides an interpretable graphical visualization of each chronic condition for detailed exploratory analysis. We compare the proposed model to relevant state-of-the art methods, demonstrating both its quantitative and qualitative merits.
AB - We propose a novel approach for simultaneous exploratory analysis and prognosis of multiple chronic conditions that can enable healthcare professionals to harness the vast amounts of medical data available in electronic health records (EHRs) in a comprehensive manner for early treatment, the discovery of previously unknown symptoms and exploration of new associations in the context of individual chronic conditions. In particular, this manuscript focuses on the use of early clinical notes available in EHRs to diagnose multiple chronic conditions. While most methods for EHR analysis have focused on single disease predictions, multiple conditions called co morbidities tend to occur simultaneously with the primary chronic condition in many patients. To this end, we model each chronic condition as a latent variable in a novel supervised topic model which results in the prediction of multiple illnesses associated with each patient. In addition, the supervised topic model also provides an interpretable graphical visualization of each chronic condition for detailed exploratory analysis. We compare the proposed model to relevant state-of-the art methods, demonstrating both its quantitative and qualitative merits.
UR - http://www.scopus.com/inward/record.url?scp=84966297363&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=84966297363&partnerID=8YFLogxK
U2 - 10.1109/ICHI.2015.36
DO - 10.1109/ICHI.2015.36
M3 - Conference contribution
AN - SCOPUS:84966297363
T3 - Proceedings - 2015 IEEE International Conference on Healthcare Informatics, ICHI 2015
SP - 243
EP - 252
BT - Proceedings - 2015 IEEE International Conference on Healthcare Informatics, ICHI 2015
A2 - Fu, Wai-Tat
A2 - Balakrishnan, Prabhakaran
A2 - Harabagiu, Sanda
A2 - Wang, Fei
A2 - Srivatsava, Jaideep
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 3rd IEEE International Conference on Healthcare Informatics, ICHI 2015
Y2 - 21 October 2015 through 23 October 2015
ER -