TY - GEN
T1 - Granite
T2 - 5th IEEE International Conference on Healthcare Informatics, ICHI 2017
AU - Henderson, Jette
AU - Ho, Joyce C.
AU - Kho, Abel N.
AU - Denny, Joshua C.
AU - Malin, Bradley A.
AU - Sun, Jimeng
AU - Ghosh, Joydeep
N1 - Funding Information:
The authors would like to thank Suriya Gunasekar for her input on inducing sparsity. This work was supported by NSF grant 1418504.
Publisher Copyright:
© 2017 IEEE.
PY - 2017/9/8
Y1 - 2017/9/8
N2 - One of the most formidable challenges electronic health records (EHRs) pose for traditional analytics is the inability to map directly (or reliably) to medical concepts or phenotypes. Among other things, EHR-based phenotyping can help identify and target patients for interventions and improve real-time clinical decisions. Existing phenotyping approaches often require labor-intensive supervision from medical experts or do not focus on generating concise and diverse phenotypes. Sparsity in phenotypes is key to making them interpretable and useful to clinicians, while diversity allows clinicians to grasp the main features of a patient population quickly.In this paper, we introduce Granite, a diversified, sparse nonnegative tensor factorization method to derive phenotypes with limited human supervision. Compared to existing high-throughput phenotyping techniques, Granite yields phenotypes with much more distinct (non-overlapping) elements that can, as an artifact, capture rare phenotypes. Moreover, the resulting concise phenotypes retain predictive powers comparable to or surpassing existing dimensionality reduction techniques. We evaluate Granite by comparing its resulting phenotypes with those generated using state-of-the-art, high-throughput methods on simulated as well as real EHR data. Our algorithm offers a promising and novel data-driven solution to rapidly characterize, predict, and manage a wide range of diseases.
AB - One of the most formidable challenges electronic health records (EHRs) pose for traditional analytics is the inability to map directly (or reliably) to medical concepts or phenotypes. Among other things, EHR-based phenotyping can help identify and target patients for interventions and improve real-time clinical decisions. Existing phenotyping approaches often require labor-intensive supervision from medical experts or do not focus on generating concise and diverse phenotypes. Sparsity in phenotypes is key to making them interpretable and useful to clinicians, while diversity allows clinicians to grasp the main features of a patient population quickly.In this paper, we introduce Granite, a diversified, sparse nonnegative tensor factorization method to derive phenotypes with limited human supervision. Compared to existing high-throughput phenotyping techniques, Granite yields phenotypes with much more distinct (non-overlapping) elements that can, as an artifact, capture rare phenotypes. Moreover, the resulting concise phenotypes retain predictive powers comparable to or surpassing existing dimensionality reduction techniques. We evaluate Granite by comparing its resulting phenotypes with those generated using state-of-the-art, high-throughput methods on simulated as well as real EHR data. Our algorithm offers a promising and novel data-driven solution to rapidly characterize, predict, and manage a wide range of diseases.
KW - Computational phenotyping
KW - Data mining
KW - Electronic health records
KW - Feature extraction
KW - Health information management
KW - Tensor factorization
UR - http://www.scopus.com/inward/record.url?scp=85032389331&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85032389331&partnerID=8YFLogxK
U2 - 10.1109/ICHI.2017.61
DO - 10.1109/ICHI.2017.61
M3 - Conference contribution
AN - SCOPUS:85032389331
T3 - Proceedings - 2017 IEEE International Conference on Healthcare Informatics, ICHI 2017
SP - 214
EP - 223
BT - Proceedings - 2017 IEEE International Conference on Healthcare Informatics, ICHI 2017
A2 - Cummins, Mollie
A2 - Facelli, Julio
A2 - Meixner, Gerrit
A2 - Giraud-Carrier, Christophe
A2 - Nakajima, Hiroshi
PB - Institute of Electrical and Electronics Engineers Inc.
Y2 - 23 August 2017 through 26 August 2017
ER -