Assigning Coded Elements

Normally people use identifiers to represent complete long names in order to economise data storage and comprehensibility. The coded element (CE) provides a strict standard that requires the user to fill it in with the identifier, text of the laboratory item, the name of the coding system, alternate identifier, alternate text and name of the alternate coding system (see the AF mark in Figure 7). For example, the data element |530.0^Cough^ICD9| refers to the diagnosis of the common cold with the identifier in Version 9 of the International Coding of Diseases (ICD9) being 530.0. Other commonly used systems in medicine for providing coding include LOINC (Logical Observation Identifiers Names and Codes) and SNOMED (Systematized Nomenclature of Medicine). However, no field of names or coding systems are in common use for LIS. This could poten-tially result in confusion during information exchange. To prevent this, users can assign a constant value for the field and sometimes use the value “Z” to define the locally defined coding system.

Figure 8. Source-and-destination selection wizard, that provides an alternative way of expression for HL7 in Chinese (middle part of the window), and a drag-and-drop way of schema mapping

The bottom part of the window shows the validated mapped fields.

366 The Journal on Information Technology in Healthcare 2003; 1(5): 357–368

Long, Liu & Chang

RESULTS

OSIHA was created for a government sponsored project MIEC (Medical Informa-tion Exchange Center³) and introduced into the clinical laboratory of three hospitals located in Taiwan. OSIHA has been released as open source software and is being adopted by a growing community of users in hospitals on Source Forge (http://sourceforge.net) – the biggest open source community with a General Public License (GPL). The system begins with a source-and-destination selection process (Figure 8). The process can enable users to compare every data element via a drag and drop interface.

After selection of the source and destination process, a data transformative tool-window pops up and enables users to convert their data elements semanti-cally (see Figure 9) with separation (see Figure 10), supplementation, truncation, replacement, concatenation (see Figure 11) and numerical conversion.

Figure 9. The data conversion main window

Figure 10. The separation function, allows the user to separate one field into several elements by setting a start position, an end position and a temporal alias name

The Journal on Information Technology in Healthcare 2003; 1(5): 357–368 367 An Open Source Intelligent HL7 Agent for Integrating Laboratory Information

To generate the required HL7 message at an appropriate time, users must configure the occasion of the triggering event. There are two kinds of trigger event, one is external and one is internal. For example, OSIHA can wait for the external application to send out a query message and then search the database according to the patient identifier, physician identifier or data time range. It then generates the HL7 message in the external triggering status. In the internal trig-gering status, the user configures the scanning time interval to program OSIHA to scan the LIS database at the right time. After the configuration, OSIHA associates the related tables via the referential key and primary key that is designated by the user and generates the HL7 message in a semantically corrected form.

DISCUSSION

Independent projects are best built using freestanding components that leverage the power of freely available open source tools¹¹. An additional advantage of open source tools is that there is less concern about their integrity compared to using a commercial product. Open source tools can be downloaded and inspected by the user and publicly evaluated. In our opinion any national project on exchanging patient information should utilise open source tools.

OSIHA provides a customisable GUI data element transformative procedure that enables users to interface their application with HL7 in a zero-line-program-ming implementation. The method requires the user to have only a very basic knowledge of HL7 control and query and of the schema know-how for their application.

In this study we managed to successfully use OSIHA to convert data from a laboratory database into HL7 messages. Based on the results of our preliminary experience with OSIHA we now plan further work including:

(i) Implementing the methodology into other clinical operating processes such as scheduling, admission, discharge and order entry

Figure 11. The concatenate function

The large arrow designates the new alias name.

368

H ^{T I} J

© The Journal on Information Technology in Healthcare 2003; 1(5): 357–368

Long, Liu & Chang

(ii) Generating more schema mapping, transformative and other program-mable tool sets

(iii) Formalising the methodology and adding more graphical user interfaces to assist with data transformation.

If we are successful, it should be easier and cheaper to introduce HL7 stand-ards into hospitals in Taiwan and consequently aid the process of implementing electronic patient records.

REFERENCES

1 Liu CT, Long AG, Li YC, Tsai KC, Kuo HS. Sharing patient care records over the World Wide Web. International Journal of Medical Informatics 2001; 61: 189–205.

2 Lung AJ, Liu CT. Modeling HL7-Taiwan referral information system. Medical Informatics Symposium, Taipei, 2000, p41. http://mist.med.org.tw.

3 Li YC, Kuo HS, Jian WS et al. Building a generic architecture for medical information exchange among healthcare providers. International Journal of Medical Informatics 2001;

61: 241–46.

4 HL7. http://www.hl7.org.

5 LINKMEDICAL Computing, Inc. http://www.linkmed.com.

6 Symphonia. http://www.symphonia3.com.

7 Microsoft Biztalk. http://www.biztalk.org.

8 NEOTOOL. http://www.neotool.com.

9 HL7 Taiwan. http://www.hl7.org.tw.

10 Ashish N, Knoblock CA. Semi-automatic wrapper generation for internet information sources. Proceedings of Cooperative Information Systems, 1997. http://portal.acm.org/

citation.cfm?id=276330.

11 Jakobovits RM, Rosse C, Brinkley JF. WIRM – an open source toolkit for building bio-medical web applications. J Am Med Inform Assoc 2002; 9: 557–70.

369

H

J

Correspondence and reprint requests: Dr Ping Yu, The Initiative for e-Health, School of Information Technology and Computer Science, The University of Wollongong, Australia. E-mail: ping @uow.edu.au.

© The Journal on Information Technology in Healthcare 2003; 1(5): 369–380

INTRODUCTION

The Electronic Health Record Task Force¹ in Australia defines electronic health