Methods Inf Med 2008; 47(06): 560-568
DOI: 10.3414/ME9125
Original Article
Schattauer GmbH

High Speed Clinical Data Retrieval System with Event Time Sequence Feature

With 10 Years of Clinical Data of Hamamatsu University Hospital CPOE
M. Kimura
1   Hamamatsu University, School of Medicine, Hamamatsu, Japan
,
S. Tani
1   Hamamatsu University, School of Medicine, Hamamatsu, Japan
,
H. Watanabe
1   Hamamatsu University, School of Medicine, Hamamatsu, Japan
,
Y. Naito
1   Hamamatsu University, School of Medicine, Hamamatsu, Japan
,
T. Sakusabe
2   Fujita Health University, School of Health Sciences, Toyoake, Japan
,
H. Watanabe
3   The University of Tokyo, Dept. of Planning, Information and Management, Tokyo, Japan
,
J. Nakaya
4   Tokyo Medical and Dental University, Information Center for Medical Sciences, Tokyo, Japan
,
F. Sasaki
5   NEC Corporation, Tokyo, Japan
,
T. Numano
6   NTT Data Tokai Corporation, Nagoya, Japan
,
T. Furuta
6   NTT Data Tokai Corporation, Nagoya, Japan
,
T. Furuta
1   Hamamatsu University, School of Medicine, Hamamatsu, Japan
› Author Affiliations
Further Information

Publication History

Publication Date:
18 January 2018 (online)

Preview

Summary

Objectives: This paper illustrates a high speed clinical data retrieving system, from 10 years of data of operating hospital information system for the purposes of research, evidence creation, patient safety, etc., even incorporating time sequence of causal relations.

Methods: Total of 73,709,298 records of 10 years at Hamamatsu University Hospital (as of June 2008) are sent from HIS to retrieval system in HL7 v2.5 format. Hierarchical variable length database is used to install them.

Results: A search for “listing patients who were prescribed Pravastatin (Mevalotin and generic drugs, any titer)” took 1.92 seconds. “Pravastatin (any) prescribed and recorded AST >150 within two weeks” took 112.22 seconds. Searching conditions can be set to be more complex, connected by Boolean operator and/or. This system called D*D is in operation at Hamamatsu University Hospital since August 2002. It is used for 48,518 times (monthly average of 703 searches). Neither searching, nor background export of data from HIS caused delay of routine operating CPOE.

Conclusions: Search database outside of routine operating CPOE, with daily export of order data in HL7 v2.5 format, is proved to provide excellent search environment without causing trouble. Hierarchical representation gives high-speed search response, especially with time sequence of events.