This report originally appeared in the May 2011 issue of DOTmed Business News

By Atul Agarwal

Diagnostic imaging’s move from film to filmless has significantly changed health care. Technology has digitized the processing of images and reports, enabling quicker access to important health care information. Yet the industry continues to struggle with efficiency and cost optimization. Consider today’s challenges in sharing diagnostic images across vendor systems, across multiple locations, among referring clinicians, and among patients themselves.

Today, physicians expect near-immediate access to information about patients, and patients are increasingly asking for ready access to records as well. Lab results and pharmacy information, for example, are available in most electronic health records and personal health records. So why not images? Why do we still see patients carrying image CDs (and sometimes even films) across provider settings? What’s on the radar as we move to newer health information exchange environments?

Most of today’s health care solutions were designed to provide document-centric, physician-to-physician communication. Physicians exchange discharge summary documents, cardiac evidence documents, lab reports, referral notes and other written documents or reports, but not images.

Why? Because image exchange is challenging – images are large; image vendors do not adhere to a single standard format; specialized viewing technology is needed; and many institutions have not added diagnostic images to their primary networks due to the cost and burden of hardware, staff and/or systems. Additionally, challenges arise when considering approaches for aligning a specific patient’s documents (imaging reports) and imaging studies within the same solution.

“Souped up” client-server PACS
Over time, technology has enabled more cost-effective, timely exchange and viewing of diagnostic images. Historically, a radiologist workstation served as the working environment for diagnostic imaging. This is a networked computer with most resources installed locally, rather than distributed over a network. Radiology reading rooms, for example, provided environments for radiologists to work on client-server based PACS solutions. These systems provided “souped-up” capacity to handle massive data volume and advanced image processing.

With the advent of Web technologies, many radiology practices moved to Web-based, “thick client” PACS viewers. This typically describes a scenario in which a Web browser is used to deploy a full application. In this environment, images are not sent to reading workstations ahead of time, thus time and cost savings can be achieved. Because images are retrieved from the PACS archive in real-time; both data duplication and pre-routing complexities are eliminated. For security reasons, special permissions are often required to download a web-based PACS onto a PC. This can present problems, especially in healthcare institutions with strict IT policies to prevent viruses and attacks. Generally, special permissions can be readily managed within the controlled environment of a radiology department. In attempting to get images to “the masses” (i.e., to multiple PCs across disparate organizations), permission management becomes a more complex obstacle.

From a technology standpoint, it’s easy to make something work in a single browser configuration or to install via the Web. But to cover many scenarios (all browser types and versions, a variety of operating systems, various firewall configurations and the like) the capabilities of “Web-based PACS” get hobbled. Work must be done to make the application run in the browser, and this nearly always involves an investment of time and money.

Efficiencies in moving from thick to thin
As the Internet advanced and technology evolved, many radiology practices moved to “thin-client” image distribution. Using a physician portal, a referring doctor simply logs into an application and pulls up an image and report for review. It’s important to note that radiologists and other clinicians who frequently interact with diagnostic images (cardiologists, pathologists, orthopedic surgeons, oncologists, etc.) have different needs from those of some specialist and most primary care physicians. The latter do not need the comprehensive advanced visualization functionality of a PACS. Hence a “thin-client” solution with limited functionality became a very attractive alternative to Web-based thick client PACS for this subset of clinicians.

Thin-client solutions minimize issues related to thick-client solutions but do not eliminate them completely. Thin-client solutions rely on browser-based plug-ins like Flash, Silverlight or applets, to provide viewing capabilities. Challenges can arise, if a Safari browser is used rather than Internet Explorer, or if the user is running an older version of IE. In some instances, applications can be downloaded and readily run on a PC, but not on a Mac.

In many organizations, thin-client technology has helped reduced delays and costs associated with CD distribution. Yet technical hurdles still exist, resulting in couriers or patients themselves serving as the CD transport mechanism.

Zero-client eliminates many challenges
Recently, health care organizations have turned to ”zero-client” solutions. To understand this concept, consider the use of Google Maps. For millions of people, Google offers an easy way to find a location, view surrounding areas, pan to a specific point and zoom to detailed information. With a single click, you can even bring up satellite images – to view a car in someone’s driveway, or catch a bird’s-eye view of a neighbor’s pool.

Now, draw a parallel to a physician who wants to view a CT image -- at a hospital bedside, at his office, at his home or en route to the golf course. Just as the Google Maps application allows people to view and manipulate map images, doctors can use the same type of ”zero-client” technology to view and interact with digital medical images.

Numerous advantages exist with server-side rendering and “zero-install,” standards-based viewing. Since all the processing happens on the server, the client workstation can be a basic machine. Previous bandwidth challenges no longer exist. This approach involves the transfer of only a rendering of a JPEG or PNG view of DICOM data from the server to the client. The amount of data being transferred is significantly reduced, making the solutions suitable for even low-bandwidth networks.

For both thin- and thick-client environments, IT professionals are typically required to perform installation, as well as ongoing maintenance and upgrades. Imagine a hospital with hundreds or even thousands of users, and consider costs for IT staff to routinely visit every workstation. Zero-client technology does away with these labor-intensive installs and upgrades.

In a zero-client imaging model, technology is used to ”serve up” the imaging-specific portion of the medical record to clinicians where and when they need it – without transferring the assets to them. (You could draw a parallel to checking the weather on your local newspaper’s website. The weather data you’re seeing are not housed on the newspaper’s server. Instead, they are served up to the newspaper’s site via the owner of the information, in many instances Weather.com.)

With zero-client environments, privacy concerns are addressed. Since no patient data of any kind reside on a client server and no caching is involved when images are viewed remotely, patient records remain confidential, ensuring compliance with all HIPAA and other industry regulations.

EHRs, HIEs ACOs and the future of image exchange
Early adopters of zero-client viewing are leading the way. Some of today’s EHR and HIE vendors, for example, plug this technology into their solutions, thus eliminating the need for their users to go outside of their working environments to access diagnostic images and also eliminating replication and transfer of huge data sets.

Improvements in health information exchange have the potential to be significant. Today’s organizations adopt technology in response to meaningful use incentives and from the promotion of new models, such as regional health information organizations, accountable care organizations and the like. Savings can be achieved by eliminating CD burning and transportation. The opportunity to improve quality by eliminating delays in image transfer is also significant. The ability of a physician to have timely, hassle-free access to any image, anytime, anywhere not only saves money, but could help save lives as well.

While zero-client solutions are great for access to/distribution of images to a wide variety of stakeholders in the care cycle, there are current limitations in technology (dictation integration, multi-monitor support, etc.) that prevent these solutions from becoming true radiology reading workstations.
Currently, the “thick, thin, zero” client continuum includes tradeoffs of functionality, features, network bandwidth and IT support. With advances in zero-client technology and the advent of new standards like HTML5, we anticipate that increasing capabilities will be available in a zero-client environment. Radiology reading workstations as we know them today will become a thing of the past. They will continue to become thinner and may even become zero-client one day.

Beyond image viewing to true interoperability

Facilitating the exchange of information from ANY content creators to ANY content consumer extends beyond Web-based, zero-client viewing. Accessing an image for display and accessing an image for interoperability have different requirements. Not all modalities and PACS systems create images in a DICOM format the same way.

The newest interoperability solutions neutralize vendor-specific DICOM information so that an image can be exchanged from one PACS to another and then viewed. Additionally, vendor-neutral archiving and storage solutions are advancing to bring together previously siloed data sets.

The newest solutions, such as Merge iConnect, offer an integrated, modular, standards-based approach to image storage, exchange and viewing. With this approach, organizations can readily leverage existing technology investments and build incrementally toward a comprehensive solution that enables any image, anytime, anywhere.

The future of true image interoperability is upon us, where diagnostic imaging data is seamlessly shared -- no matter where the data originated, how it was generated, the processes that invoked it, or the people using it.

Atul Agarwal is the Vice President, R&D for Health IT solutions at Merge Healthcare. Since joining Merge in 2004, Mr. Agarwal has held a variety of roles in operations management, account management, product management and R&D. He has 15 years of experience in delivering IT services and product development and holds a B.E. degree from Indian Institute of Technology and an MBA from Richard Ivey School of Business.