Formatting and searching a massive, multi-parameter clinical information database
Author(s)
Kyaw, Tin Htet
DownloadFull printable version (10.39Mb)
Other Contributors
Massachusetts Institute of Technology. Dept. of Electrical Engineering and Computer Science.
Advisor
Roger G. Mark.
Terms of use
Metadata
Show full item recordAbstract
Formatting data and executing time-oriented queries on a massive, multi-parameter clinical information database poses significant computational challenges. The challenges encountered in converting high-resolution waveform and trend signals in the MIMIC II (Multi-parameter Intelligent Monitoring for Intensive Care II) database from an error-prone proprietary format to a stable open-source WFDB (Waveform Database) format is presented in the first half of this thesis. The design and implementation of a search engine that is capable of executing time-series queries on clinical information in the MIMIC II database such as lab results, medications, and nurse-verified values from bedside monitors is presented in the second half of this thesis. The search engine employs simple algorithms with little storage overhead to identify time periods in patient records that satisfy time series criteria based on thresholds and gradients of unevenly-sampled measurements. Results from queries executed on the search engine to detect physiological events of clinical interest were presented. Case studies on patient records returned as hits for queries were performed to review the strengths and limitations of the search engine.
Description
Thesis (M. Eng.)--Massachusetts Institute of Technology, Dept. of Electrical Engineering and Computer Science, 2005. Includes bibliographical references (p. 117-120).
Date issued
2005Department
Massachusetts Institute of Technology. Department of Electrical Engineering and Computer SciencePublisher
Massachusetts Institute of Technology
Keywords
Electrical Engineering and Computer Science.