Portability of patient care among health care providers and facilities has long been a hallmark and strength of the US health care system. As a result, clinical information has been scattered across entire regions challenging providers’ abilities to deliver the highest quality care. At long last, the solution to this problem is within grasp: health information exchange (HIE) on a regional (and eventually national) basis that securely links a patient’s clinical data from multiple medical facilities, even competing ones.

Nowhere else in medicine are such exchanges more practical and realizable today than for diagnostic imaging where a 20-year head start on acquiring and storing clinical data in digital and interoperable standard form (ie, Digital Imaging and Communications in Medicine) means that health information from PACS and RIS can feasibly be exchanged today. In order for the meaningful exchange of laboratory results, provider notes, waveform data (eg, ECGs), and other essential pieces of the medical record to occur, widespread adoption of EMRs will have to take place. To date, EMR adoption has been slow and uneven. Moreover, even when EMRs are present in the hospitals and practices of a given region, interoperability between vendors’ products remains absent.

Likewise, nowhere else in medicine is the impact of health information exchange as likely to be so great as with diagnostic imaging. Recent data suggest that, in up to 20% of imaging cases, relevant prior images that are desired by radiologists for comparison with current images are in fact beyond the borders of their own institution.1 At a minimum, imaging data exchange would increase the frequency of side-by-side comparisons, a practice that is the standard of care2 and well known to dramatically improve diagnostic accuracy.3,4

Equally important is the ability of HIE to enhance the collaborative care that is an essential component of the interaction between referring physicians and radiologists. Imaging-focused exchanges permit virtual consults and case conferences that span organizational boundaries. Exchanges support the online distribution of results back to referring physicians and can provide a consolidated view across multiple performing institutions. They also enable efficient cross-coverage of multiple facilities by radiologists, supplanting the need for two, three, or more PACS workstations or web clients, each typically dedicated to a single institution’s system and vendor.


Elliot D. Menschik, MD, PhD

These ideas are not just a theoretical fantasy. They have been implemented and put into practice in Philadelphia, the fifth largest health care market in the United States and one of the most fiercely competitive. The Philadelphia Health Information Exchange (PHIE) is a growing network of unaffiliated imaging providers that serves as a case study funded by the National Institutes of Health. The exchange links multiple, multi-vendor PACS and RIS into a distributed, secure, web-accessible network of diagnostic images and reports. The PHIE currently spans 8 years of patient history and includes imaging results from both inpatient and outpatient settings. Current participants include the University of Pennsylvania Health System, Thomas Jefferson University Hospitals, Children’s Hospital of Philadelphia (CHOP), and Albert Einstein Healthcare Network.

The exchange was designed to accommodate the most demanding of health care environments: multiple competing facilities, multiple vendors, both legacy and new PACS/RIS, 24x7x365 uptime, and volumes of imaging as large as 400,000 examinations per year for some participating sites. The exchange was designed to scale to an unlimited number of participants and to accommodate later expansion into other forms of clinical data. Future development will also allow diagnostic images and results to be integrated into existing clinical applications (eg, EMRs, portals) to more readily fit within the practice workflow. Most important, the PHIE was built to be easily replicable in other regions, based on a minimal set of components that are now commercially available.


The PHIE currently serves as both a clinical and research platform. On the clinical side, the exchange has enabled providers to rapidly tap into expertise not available in-house, particularly in emergent situations. For example, when a motor vehicle accident brought a pediatric head trauma case to the emergency department at Pennsylvania Hospital (a community hospital that is part of the University of Pennsylvania Health System), a neurosurgeon at unaffiliated Children’s Hospital was emergently consulted by phone. While he did not have an account on the PACS at Pennsylvania Hospital, using the exchange’s web portal from his office across the city, he was able to remotely view diagnostic quality images acquired at Pennsylvania Hospital and suggest the appropriate course of action within minutes. Prior to the existence of the exchange, such an interaction would have been impossible with potentially lethal consequences for the patient.

In other less dramatic but more frequent cases, the exchange enables timely and efficient collaboration on patients shared across institutional borders. For example, patients of CHOP’s fetal therapy practice frequently undergo MR imaging in CHOP’s facilities but undergo ultrasonography at the Hospital of the University of Pennsylvania. Prior to the establishment of the exchange, films would have to be hand-carried between the separate institutions to enable the necessary radiologists and surgeons to collaborate and compare findings, resulting in delays. Today, the exchange’s secure web portal provides on-demand access to diagnostic quality images and reports that can span PACS, RIS, institutional borders, and even the absence of a unique patient numbering system among the network’s participants.

The network has also brought administrative efficiencies to the participants in the form of a single portal for on-call coverage of multiple facilities with differing PACS products. It has also functioned as a backup to the PACS allowing diagnostic reading to proceed during scheduled and unscheduled PACS downtime at participating sites.

On the research side, the network is enabling investigators to ask and answer a wide range of questions at the forefront of health services and medical informatics research. For example, researchers spanning competing facilities are collaborating to quantify the extent to which relevant priors are available across a region, the extent to which duplicative imaging unnecessarily exposes patients to risk while adding to spiraling health care costs, and the extent to which imaging-specific HIE can improve patient outcomes across different clinical conditions and specialties.


In designing and implementing the Philadelphia exchange, the key to success has been following a distributed (aka federated or peer-to-peer) architecture. Its alternative, the central archive model, has largely been credited with the failure of the community health information networks (CHINs), precursors to today’s health information exchanges.5 In short, not only are providers reluctant to relinquish their core business asset to a common archive outside their control, but patients themselves appear reluctant if not unwilling to trust any one entity to collect all their medical records.

In contrast, the distributed architecture for imaging exchange keeps the patient data at each imaging provider site, behind its firewall, and still under the site’s control. At the same time, the approach enables each imaging site to field electronic requests for its patients’ data that originate from outsiders via the Internet, provides the means for such outsiders to rapidly meet the site’s disclosure procedures, and enables the site to securely transmit the digital images and/or reports directly to the requestor’s web browser or clinical application.

A common question is how can imaging exchanges even be feasible in the post-HIPAA age? The reality is that imaging providers already share films and CDs with outsiders—even competitors—on a daily basis. All of these organizations have established policies for disclosing protected health information (PHI) that meet if not exceed Health Insurance Portability and Accountability Act and other regulatory or business requirements. The specifics of these policies vary from site to site, but share a common inefficiency rooted in their paper-based implementation and execution.

A distributed model for imaging exchange can take these existing policies and procedures, and replicate them digitally, thereby enforcing each site’s own standard operating procedures at its border, maintaining the status quo as far as legal counsel is concerned, and eliminating what would otherwise be a substantial barrier to participation in a regional exchange. More generally, by design the distributed approach to health information exchange addresses the concerns of even the most conservative potential contributors to the network by maintaining their individuality, independence, and “ownership” of the clinical data.

It is also worth noting that the distributed model effectively accommodates sites that use or are contemplating use of an off-site archive (such as an ASP storage provider). Such storage providers can simply serve as additional nodes on the network.

In summary, the distributed model represents the first realizable approach for building health information exchanges in the real world that satisfies the political, social, and legal prerequisites for permitting on-demand online access to results by others caring for a given patient. As it happens, the distributed model is also technically optimal in that it minimizes storage requirements and bandwidth consumption across a given region while still providing the requisite scalability, availability, and reliability.


Reaping the full potential of an HIE ultimately requires integration of the regional infrastructure into the existing clinical applications and workflow of practicing health care providers.

Today, users of the PHIE access images and results through a globally accessible web portal application. In many practices, this represents another application (with another user name and password to remember) that a provider and/or their staff need to log into to perform their work.

A far more powerful and efficient approach would be to integrate the information exchange functions into the relevant PACS, RIS, order entry system, or EMR. In the case of radiologists, such integration would permit efficient comparisons to priors using their favored diagnostic workstation and application, as well as additional features such as prefetching of priors, which could be extended from local searches to cross-enterprise ones in anticipation of scheduled examinations for a particular patient.

Achieving this level of integration would require either numerous custom software development engagements with each vendor of each clinical application in the region, or the ability to hook into a set of open standards already built into these applications. Clearly, the latter option is the preferable one.

Fortunately, the work to develop and promulgate such interoperability standards has already begun under the stewardship of the Integrating the Healthcare Environment (IHE) initiative. IHE is a multi-year, multi-stakeholder interoperability initiative sponsored by the Radiological Society of North America (RSNA), the Healthcare Information and Management Systems Society (HIMSS), and the American College of Cardiology (ACC). It is intended to enable plug-and-play interoperability between clinical information systems from different vendors. Embraced by most of the major health care IT vendors and increasingly finding its way into their products, IHE harnesses existing health care and IT standards (eg, DICOM, HL7, HTTP, SAML) into integration profiles, technical recipes for how such standards must be implemented by vendors to ensure interoperability for specific tasks and problems faced daily by health care providers.

Each year, competing vendors collaborate to develop new integration profiles, implement those new profiles in upcoming versions of their products, and convene in annual connectathons to extensively test these profiles and prove such interoperability.

For those interested in unlocking the full potential of cross-enterprise imaging exchange, this year’s North American IHE connectathon (held on January 15-20, 2006) is of particular interest as leading health care IT vendors will demonstrate for the first time their products’ compliance with the new cross-enterprise document sharing for imaging (XDS-I) integration profile.

In essence, XDS-I provides a set of components that can be combined to create a regional imaging exchange on the basis of open standards. Not only does XDS-I support a distributed architecture for imaging exchange, but any PACS that complies as an XDS-I “consumer” would, with minimal configuration, be able to hook into such a regional infrastructure to natively share and receive images and reports from other participating sites in the region.

XDS-I is anticipated to add further momentum to the construction of imaging exchanges on both regional and national bases. At the regional level, the Philadelphia Health Information Exchange plans to be the first real-world implementation of XDS-I. At the national level, XDS-I figures prominently in the diagnostic imaging component of Canada’s nationwide network, which is in the early stages of construction.

In the meantime, for those interested in learning more, additional information can be found at www.ihe.net. Furthermore, hands-on demonstrations of IHE functionality in clinical applications such as PACS and RIS will be featured at both the annual HIMSS and RSNA meetings.


Cross-enterprise imaging exchange represents the next step in the evolution of digital imaging and, more generally, patient care. For the innovators and early adopters who would like to see its benefits in their community as soon as possible, there are at least two key actions that can be taken today.

First, work to start an exchange in your region. Find like-minded champions at neighboring institutions and practices. Ask your current IT vendors about what they are doing to help you build, operate, and support such a network in your region today. Likewise, approach those vendors who are dedicated entirely to HIE construction and operation who have production-ready products and services that can be leveraged to build an exchange today. (See Web version of story for a list of those vendors.) Develop your own or engage potential HIE vendors to provide a business model, customized to your community, for how to launch and sustainably finance such a network. If a regional health information organization (RHIO) exists for your community, engage the RHIO to make imaging exchange a top priority, using the image exchange business model to support your case and secure financial commitments. If a RHIO does not yet exist (as was the case for the PHIE), leverage your preferred HIE vendor to act as the convener of sites and as the logistical organizer of the exchange.

Second, for those already with PACS and/or RIS, investigate when you can expect support in your systems for IHE’s XDS-I consumer “actor.” For those looking to purchase a new PACS or RIS, require XDS-I compliance from prospective vendors to ensure that your investment will have the ability to make the most of an exchange when one is built in your community.

Imaging is in a unique position in the clinical spectrum by virtue of its decades-old investments in digitization and standardization and the expected clinical and financial impact of imaging exchange. Imaging providers can and should take advantage of these characteristics today to establish regional exchanges that extend the abilities of their current or anticipated PACS/RIS investments, maximizing clinical benefits for their patients and financial gains for their practices.


  1. Lakhani P, Menschik ED, Goldszal AF, Murray JP, Weiner MG, Langlotz CP. Development and validation of queries using Structured Query Language (SQL) to determine the utilization of comparison imaging in radiology reports stored on PACS. J Digit Imaging. September 2, 2005. Epub ahead of print.
  2. American College of Radiology. ACR standard for communication: diagnostic radiology. In: Standards. Reston, Va: American College of Radiology; 1995.
  3. White K, Berbaum K, Smith WL. The role of previous radiographs and reports in the interpretation of current radiographs. Invest Radiol. 1994;29:263-5.
  4. Aideyan UO, Berbaum K, Smith WL. Influence of prior radiologic information on the interpretation of radiographic examinations. Acad Radiol. 1995;2:205-8.
  5. Lorenzi N. Strategies for creating successful local health information infrastructure initiatives. In: Vanderbilt University Technical Report. Nashville, Tenn: Vanderbilt University; 2003. Available at: aspe.hhs.gov/sp/nhii/LHII-Lorenzi-12.16.03.pdf.

Elliot D. Menschik, MD, PhD, is the founder and chief executive officer of Hx Technologies (www.hxti.com), a health care services and IT company that builds and operates health information exchanges.