Design of an ECG Database Management System

Glenn Guillemette, Robert Antinoro, Floyd Fulton, Sophia Zhou

Cardiology Division, Philips Medical Systems, Oxnard, CA, U.S.A.

Correspondence: Sophia Zhou, Ph.D., Advanced Algorithm Research Center, Philips Medical Systems, 1201 N. Rice Ave.,
Oxnard, CA 93030, U.S.A. E-mail: sophia.zhou@philips.com, phone 1 805 988 7424, fax 1 805 604 0493

1.  Introduction

Medical research and education requires ECG databases [Norman, 1998], and patient care requires ECG database management systems [Nose, 1987; Wang, 1996; Gamon, 2002]. Design, development and maintenance of a sophisticated ECG database in an era of rapid evolution of clinical and information technology has become not only complex, but also expensive. This paper reports on the design approach of Philips ECG Management System and its openness in data access.

2.  Methods

Relational Database. SQL Server is the core in database management. The Philips EMS database is composed of normalized tables containing patient demographics, digital ECG waveform and ECG interpretation. Essentially, all ECG data are exposed via individual columns in relational tables. This design enables use of simple to advanced SQL queries for information retrieval.

User Access and Workflow. Secure web-based applications for ECG viewing and printing from the Philips EMS are supported over the intranet or internet. As part of a well designed data management system, the workflow layer is configurable to meet the needs of a wide variety of practices, including automatic pre-processing such as diagnostic serial comparison, routing of incoming ECGs for review, notification via email, automatic in‑box assignment, additional special processing on confirmation of ECGs, including exporting the ECG in a variety of formats, such as XML, and Health Level Seven (HL7) to a variety of external systems, including hospital information system (HIS) billing, order management, and enterprise clinical information systems. The system supports full auditing and tracking of ECG access and the operations and actions performed on individual ECGs.

System Communications. A rich set of communications protocols is supported, including TCP/IP over local area network (LAN) and wide area network (WAN), and transmission control protocol/point-to-point protocol (TCP/PPP) over modem and dialup connections for communication with electrocardiographs. Bi-directional communications between the EMS and the HIS (Fig. 1) is supported.

Figure 1. Philips ECG Management System and its interactive components.

3.  Applications

Research. Ownership of ECG databases on a commercial management system has been debated over a few decades. The openness of the industrial standardized XML format, a major breakthrough in the Philips EMS, allows researchers to own the ECG databases. Issues related to proprietary format no longer exist. Researchers in medical and pharmaceutical studies have complete flexibility in storage, searching and accessing patient data, ECG interpretation and the digital ECG waveform in high fidelity.

Education. ECG reading is an art in medical science. ECG variations shown in textbooks are very limited. Without an ECG database management system, training in ECG reading can be very difficult. In the Philips EMS, ECG examples can be searched, retrieved, displayed and printed in seconds. The database can be queried easily using the SQL editor and SQL aware tools.

ECG Management. ECG data aid decision-making. One of the many advantages using an ECG management system is the ability to perform serial ECG comparisons [Ariet, 1991]. The serial comparison algorithm assists cardiologists in ECG diagnosis and evaluating progression or regression of a condition. Furthermore, most hospitals in the United States are experiencing a shortage of experienced ECG professionals. The Philips EMS automates ECG order and reports order fulfillment to offset the resource shortage. The process of downloading ECG orders from the HIS order entry system, entering ECG order fulfillment in the HIS billing system, and reporting confirmed ECG reading to the enterprise clinical information system, can be configured in the workflow layer to maximize process automation.

4.  Conclusions

The system design is focused on the needs of researchers and health care professionals. The system is adaptable and scalable to meet a wide variety of IT and clinical information environments in the United States and worldwide. Although the system is non‑trivial and may be considered a complex clinical information system, it is easily accessible via a standard web-based application, and is flexibly configurable to satisfy the varying requirements of cardiology departments for enhanced patient care.

References

Ariet M, Crevasse L, Caverly S, Leggett S, Greenfield J: Computerized serial comparison of electrocardiograms. J Electrocardiol 23(supp):123-131, 1991.

Gamon R, Cooper A: Applying an ECG database to aid decision-making in the A&E. Accid Emerg Nurs 10(2):62-62, 2002.

Norman JE, Bailey JJ, Berson AS, et al.: NHLBI workshop on the utilization of ECG databases: preservation and use of existing ECG databases and development of future resources. J Electrocardiol, 31(2):83-89, 1998.

Nose Y, Akazawa K, Yokota M, et al.: An electrocardiogram database incorporated into the hospital information system. Med Inform (Lond), 12(1):1-9, 1987.

Wang C, Ohe K, Sakurai T, Nagase T, Kaihara S: An ECG storage and retrieval system embedded in client server HIS utilizing object-oriented DB. J Med Syst 20(10):35-43, 1996.