Creating the Virtual Laboratory

INTRODUCTION

In the increasingly busy modern working environment, the clinician requires simple and rapid access to laboratory results. Most modern, cost-conscious laboratories no longer attempt to provide every diagnostic test “in house” and much work is referred out to specialist reference laboratories. Point of care instruments are now providing a significant and growing proportion of laboratory data. Integrating results from these various sources and making them readily available to the clinician is an important challenge.

The number of computers in the world continues to grow at exponential rates, currently fuelled by the Internet. (Moore determined that processor power doubles every 18 months and this has been true for over 20 years). The Internet has facilitated connectivity between computers and this can be harnessed in the creation of the Virtual Laboratory. Indeed, utilising this all-conquering technology makes good sense and can take advantage of user familiarity with the ubiquitous browser. In addition, utilising the hyperlink capabilities of this communications medium can be advantageous.

For the purposes of this discussion, the virtual laboratory is defined as a single source of information for the clinician integrating results from a variety of instruments, sites and laboratories (Figure 1).

Provision of these data in a timely manner provides benefits in efficiency and reduced training. These goals are mandatory to meet clinical consumer expectations.

IMPLEMENTING THE VIRTUAL LABORATORY

What information should we be attempting to provide and how shall we present it?

Result Presentation

The first and most important task is to present the results of laboratory investigations. An example of such a system, in use at the Waitemata Hospital in New Zealand, can be seen in Figure 2 (1). This system was developed by the Galen Software Group and is installed in this 450 bed hospital served by an external pathology provider. Data is transferred by a 2 megabit dedicated line between the laboratory and hospital. A pentium PII 350 computer acts as the web server running SQL server.

Figure 2.

Cumulative biochemistry report displayed in browser. Note the icons to link to graphic display showing change of analyte over time.

Presentation of data is facilitated in the graphics-rich environment of the browser so that time series, eg. glucose tolerance test results, can be graphed. A great benefit of this web-based approach is the ability to present data from a combination of legacy databases. The browser can provide a front end to several existing databases that have been optimised over years of use. This can be an extremely cost effective way of modernising the interface to the laboratory.

Laboratory Handbooks

Internet/intranet based handbooks offer several advantages. The usual range of information may be provided, including reference ranges, specimen collection details and telephone numbers for contacting advisors. Use can be made of hyperlinks to order-entry screens if electronic order entry has been implemented. The laboratory gains from the utilisation of a single up-to-date source, which is readily updated and maintained. A measure of document control is also provided, since with hard copy handbooks one is never sure whether users are using the latest version or some older hoarded copy. An example of our on-line handbook at Royal Melbourne Hospital, which is available on the hospital intranet, is shown in Figure 3 This is provided using Microsoft Active Server Pages and searches a live Microsoft Access database using Open Data Base Connectivity.

Figure 3.

Laboratory handbook showing search results of tests matching search string “ac^*”.

Other opportunities for assisting users are in the provision of graphical on-line tutorials on the use of point of care instruments, eg. ward glucose meters (Figure 4). The asynchronous nature of this communication means that this information is available at any time of the day or night. The ready incorporation of images allows unskilled users to step through procedures on unfamiliar instruments.

Clinical Interpretation

The possibility of providing interpretive information as an adjunct to the data presents interesting challenges. The opportunity is provided by the hyperlinking capabilities of the browser, so that it is possible to offer click-through explanations of abnormal results. Consideration needs to be given to the sophistication of the target audience, as different information is required by consultants compared to that by junior clinical staff or general practitioners. We have made several pilot attempts at this (Figure 5) and realised that essentially several textbooks need to be produced, one for each target group. Thus, the task is not trivial but remains a worthwhile goal.

Figure 5.

Interpretative page for urea:creatinine ratio showing hyperlinks and graph.

MARKUP

The transmission of laboratory results between computer systems requires standard ways to encode the data including method, reference range and units as well as the result itself. Naturally, it is preferable to utilise existing systems such as HL7 messaging or LOINC codes (Logical Observation Identifiers, Names and Codes). These provide a common language between databases. In addition to this coding system, a markup language is required for browser display of the results.

The standard markup language in use on the world wide web is HyperText Markup Language (HTML). This is a formatting language, which uses tags to instruct the browser how to display text. Whilst this is adequate for presenting general text documents, it creates weaknesses when used for database work. HTML browsers are non-validating so that the parsing engine simply presents the data and there is no syntax check.

In contrast, the eXtensible Markup Language (XML) uses a structured approach based on a Document Type Definition (DTD) that defines the elements in the documents (2). Thus, it focuses on the data itself and allows data in a database to be represented. As a consequence, machine as well as human reading and data processing are facilitated. See Figure 6 for a comparison of data formatted as HTML compared to XML.

Figure 6.

An example of HTML formatting. Tags indicate size of text and paragraph breaks. XML example includes element tags, showing representation of data structure as defined in DTD.

Recently, the combination of HL7 with XML has been demonstrated for the transmission of clinical data (3). This combination of messaging and structured markup offers a very powerful toolset which we can expect to see much used for future applications.

PALMTOPS

Computer devices are no longer confined to large deskbound systems. Palmtop computers or Personal Digital Assistants (PDAs) now incorporate powerful computer chips. The possibilities of using these devices to provide laboratory results at the bedside have been intriguing a number of clinical biochemists (4). Tom Hartley at the Royal Hobart Hospital has been using the Palm Pilot as an order entry device and we at Royal Melbourne Hospital have used one to download results from our LIS (Figure 7). Our approach takes advantage of the ability of these devices to display web pages (HTML) and uses a free utility program (Avant Go) to transfer the data to the palmtop.

Figure 7.

Palmtop computer displaying laboratory results downloaded using HTML.

The usual way for these devices to update their data is via a hot sync cradle. This allows synchronisation with the latest information and in some cases offers recharging at the same time. User profiles can be maintained by the devices, enabling an individual clinician's subset of patients to be stored, rather than requiring the whole database to be downloaded. These devices are capable of continuous updating using infra red or wireless networking, which offers benefits compared to static synchronisation.

CONCLUSIONS

The Communications Revolution offers many possibilities for enhancing the performance of pathology laboratories. The Internet is an important means of communicating and provides standard tools that we can utilise. There are great possibilities in using XML output from point of care and other devices that will improve interfacing these devices. The Virtual Laboratory is a concept that represents a stepping stone on the route to the Electronic Patient Record. It will greatly assist clinicians and it is important that laboratory workers begin implementing these developments now.

ACKNOWLEDGMENTS

Tim Coles, Todd Somervell and Sujiva Ratnaike provided much invaluable discussion of these concepts.