By Topic

Context-Sensitive Correlation of Implicitly Related Data: An Episode Creation Methodology

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$33 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

4 Author(s)
Roderick Y. Son ; Med. Imaging Inf. Group, Univ. of California, Los Angeles, CA ; Ricky K. Taira ; Hooshang Kangarloo ; Alfonso F. CÁrdenas

Episode creation is the task of classifying medical events and related clinical data to high-level concepts, such as diseases. Challenges in episode creation result in part because of data, in the patient record, only implicitly being associated with their respective episodes. Furthermore, traditional approaches have been limited to using feature-poor claims records to generate episodes. The accurate correlation of data to their episodes is valuable in health outcomes research to discern resource utilization with respect to medical conditions. This paper describes a combinatorial optimization approach for constructing episodes, which supports the incorporation of heterogeneous data types. Aspects of this approach include an episode model for characterizing the generation of data elements as a result of a process, a methodology for identifying the relationships between implicit processes and the data elements generated by the processes, a measure for evaluating candidate episode configurations, and an energy-minimization methodology for addressing episode creation. An implementation of this work, called Episode Creation Version 2 (EC2), has been applied on patient records with various episode types, which present with knee pain. EC2 demonstrated data element classification precision and recall scores of 78% and 82%, respectively. Significant improvements in precision and recall were observed over a traditional healthcare services approach.

Published in:

IEEE Transactions on Information Technology in Biomedicine  (Volume:12 ,  Issue: 5 )