Many hospitals and health delivery systems are rapidly moving toward fully electronic record keeping, the goal being to put patient-centered informationradiology reports and images, laboratory results, pharmacy information, and any other relevant dataat the fingertips of authorized clinical users across the enterprise. Doing so requires not only that the disparate systems containing this information work together harmoniously, but that they be able to interoperate fluidly, so that users can access the various data (including images) on a single display monitor. This is known as desktop integration.

In settings where desktop integration has been achieved, users save many minutes throughout the work day. They do not have to sign in to one system to obtain the electronic patient record, execute a patient search, and drill down to the encounter level or deeper, only to find it necessary to sign in on a second or even third unrelated system operated by the different departments within the enterprise system to obtain other data such as radiology reports (each of which will entail executing the patient search all over again, drilling down as before). Instead, users sign in just once; the disparate systems then extract contextual information on the patient and display all that has been requested through one interface.

Perhaps the best way to provide users with this kind of access to information is through a web-portal application. Because of its design characteristics, a web portal can serve as a mechanism for bringing disparate systems together easily into a single, patient-centered view (with a single sign-in process and continuous provision of the correct contexts). Web-portal applications conventionally come with an entire toolbox of methods for extracting information from other systems, from SOAP/XML or SQL queries (direct database access) right down to screen scraping (extracting information from the displayed screen and using it to emulate interaction with another system).

A web portal can accomplish two things: it can drive a path to the information that the user is trying to obtain and it can manage that process in the quickest way possible. With Agfa's IMPAX® product, for example, this transpires by means of a an image-viewing module embedded within the electronic patient record. This module, called a Java" bean, employs the Java programming language; like an applet, it contains its own set of functions and display, but only some of the applet's functions are present in the bean. With the Java bean embedded, the user is able to retrieve the electronic patient record, search for the patient, and find the desired data. Without ever having to go to another monitor or sign in separately, the user can navigate to a page containing a list of available radiology reports for the particular patient. The user can open the needed report and read it.

If, and only if, accompanying images are available, a button icon will appear. Clicking that button will cause the images to be retrieved and displayed. The Web 1000 product is structured so that it can notify the electronic patient record any time that radiology images specific to that patient have been completed at the departmental level and are available for viewing by the user. It is this notification that enables the button icon for image retrieval to appear. Configuring Web 1000 in this manner avoids the problem of users expectantly clicking the icon, only to discover that no images are available. Because the icon only launches when there are actually images available, users have confidence that clicking the icon will be productive, not a time-wasting dead end. Meanwhile, all of the operations of the various other systems responsible for providing the requested reports and images remain in context to the patient.

Given that desktop integration is relatively new, no studies have yet been performed to quantify the cost savings represented by a successful desktop-integration effort. Those who currently work in environments where this has been achieved would be likely to agree, however, that integration at least opens the door to improvements in quality of care. With information more readily and rapidly accessible, physicians and support staff can devote more time to caring for patients. Moreover, they can make faster and better decisions, leading to earlier intervention and, ultimately, improved outcomes. This carries the added and obvious benefit of leaving patients better satisfied by their experience with the enterprise.

DEPARTMENT-LEVEL INTEGRATION

It is vital to note, too, that desktop integration can be undertaken at a departmental level, as well as enterprise wide. The benefits derived are the same in both environments; however, the primary reason for integrating the desktop in this narrower setting is not so much to provide access to patient-centered information as to streamline work flow. A department is a prime candidate for desktop integration if multiple informatics systems are in use and there is the desire to create more efficient work flow across those multiple systems.

Consider, as an example, the effect of integration on a hypothetical radiologist sitting at a picture archiving and communications system (PACS) workstation. As he or she works, the radiologist will, of course, want to have information from the radiology information system (RIS) immediately available; this information will include the patient's history and previous reports. With desktop integration, a context would be developed so that the information on the RIS could be synchronized with the PACS workstation. The radiologist could then flip back and forth, effortlessly, between an image window and a patient-history window. He or she does not have to sign in on the separate RIS and search from the top for the corresponding patient information to support the images on the PACS monitor. Similarly, when the radiologist is ready to dictate, he or she need only click an on-screen button on the PACS workstation in order to launch the dictation system. The dictation screen comes up, already fully in context to the right patient and the right study. Almost immediately, then, the radiologist can proceed with dictation. Later, when the dictation is ready to be saved, the system stores it while remaining fully in context with that particular patient and study. In short, thanks to desktop integration, all three separate systems (PACS, RIS, and dictation) act and operate as a single application from just one workstation. A similar work flow could also be driven from the RIS.

VENDOR COOPERATION

Whether one contemplates a desktop-integration effort at the departmental level or the enterprise level, cooperation among the companies that supply the various systems that are to be made interoperative is absolutely essential. Without vendor cooperation, desktop integration simply will not happen. Whenever two or more systems from different vendors attempt to communicate with one another, there are going to be idiosyncrasies that only the respective manufacturers, working in concert, will be able to surmount.

Vendors have been adhering to communications standards (such as the Health Level 7 and Digital Imaging and Communications in Medicine standards) to facilitate integration of their systems through the enabling of data exchange. Now, with the need to integrate by displaying data from disparate systems on a single monitor, the same vendors are adopting the fairly new communication standard of the Clinical Context Working Group (CCOW) (see story, page 5).

Those messages are dispatched to a context manager, the function of which is to notify the other registered applications instantly of the changes. Once so notified, those registered applications react either by reestablishing their context to reflect what is on the current application or by ignoring the changes, depending on what the message is and whether there is a prior agreement to react to that message in the first place. CCOW allows a fair amount of flexibility in defining context so as to help ensure the relevance of the context to any given pairing (or grouping) of applications.

Agfa's presence in the desktop-integration movement long predates the advent of CCOW. For this reason, it was necessary to create a proprietary communications methodology when the company began blazing a trail toward desktop integration. The Agfa solution embraced several strategies closely resembling those currently found in the CCOW standard; that fact afforded Agfa a great advantage as it moved toward the adoption of CCOW for all of its future offerings.

Beyond vendor cooperation, little else is required to achieve desktop integration. Strengthening the fiber-optic backbone is not necessary; neither is the addition of more routers and hubs. In fact, departmental-level customers may actually be able to eliminate pieces of equipment in the course of establishing integration at the desktop. Since desktop integration entails using a single display to gain access to multiple applications, one no longer needs separate terminals for RIS and dictation-system use. The PACS monitor alone will be sufficient. This is especially welcome news for radiology departments in which space is at a premium; the more pieces of hardware that these departments can remove, the better.

It should be clear that there are numerous benefits to be gained through desktop integration, whether it is implemented at the departmental level or at the enterprise level. Radiology departments should certainly think twice about putting off a desktop-integration effort.

James Herrewynen is market segment manager of enterprise connectivity, Agfa Inc, Waterloo, Ontario, Canada.