Images in the EMR: One Size Does Not Fit All
Image source: Children’s Medical Center of Dallas
As more and more healthcare enterprises deploy and expand EMRs, the issue of clinical image access within the EMR presents itself. For radiologists and referring physicians alike, image access in the EMR can be a game-changer—boosting clinical collaboration, enhancing patient education and care and shaving minutes from patient encounters throughout the day. Although the benefits of image access in the EMR are clear and multifaceted, there is no straightforward formula for a successful integration.

That’s because the roadmap to and model for image access varies among and, oftentimes, within institutions. Some cutting-edge facilities shun customization and employ an EMR-centric model for all image access. Others adopt a more flexible model that allows users to determine their image access method and offer a progression of options to support diverse workflows across the enterprise. This month, Health Imaging & IT visits a variety of enterprises to learn more about their strategies for image access in the EMR and the clinical, workflow and patient care benefits derived.  

HIMSS Stage 7 Sites: A tale of two models

The 52 facilities that have earned Healthcare Information Management Systems and Society (HIMSS) Stage 7 EMR status epitomize paperless healthcare and readily share clinical information via standard electronic transactions with all entities within health information exchange networks. Most also emphasize rich image access mechanisms. However, while one pioneer stresses user-centered flexibility, another focuses on patient-centric standardization across the enterprise.

HIMSS Stage 7 site University of Wisconsin Hospital and Clinics (UW Health) of Madison, is comprised of a 493-bed facility and 85 outpatient clinics. More than 1,200 physicians are affiliated with UW Health and expect image access in the EMR as the health system integrated PACS and the EMR nearly a decade ago.

“We see a wide variety of image access needs across the enterprise,” offers Gary J. Wendt, MD, MBA, vice chair of informatics and enterprise director of medical imaging. The academic medical center strives to support a variety of imaging workflow models among clinical end users. “A lot of image-intense people don’t live in the EMR. A neurosurgeon or orthopedic surgeon may sit in front of a dedicated workstation and work more like a radiologist than a primary care physician,” explains Wendt.  

Children’s Medical Center of Dallas, another Stage 7 institution, a two-hospital system with five specialty centers and 11,000 EMR users, employs a very different tactic. “Customization is a bad word,” offers CMIO Christopher Menzies, MD. That’s because customizations tend to break down with upgrades. Instead, Children’s emphasizes platform-enabled nimbleness and encourages all providers to use the EMR as the single patient chart and access point for all radiology and non-radiology images.

The clinical rationale for maintaining the EMR as the central enterprise hub is simple. “Images don’t make up the patient. They don’t have all of the essential clinical information attached to them. Patient data stored in the EMR, coupled with images, make up the patient,” explains Menzies.

The IT infrastructure for different types of image sets varies slightly; however, Children’s strives for a seamless, standardized viewing experience for the end user. The enterprise established a completely integrated environment; a context-enabled link in the EMR takes users directly to radiology images stored in PACS.  Secure data are delivered through a hosted environment and are available for all users across the enterprise. Other images such as photographs of lesions are stored directly in the EMR, and some are housed in niche image management systems. Although images are stored in multiple systems and users can view images via a variety of systems, Children’s stresses the EMR as the access point.

Menzies recalls an encounter with providers accessing echocardiograms via the cardiology PACS. He asked why they preferred that method over the EMR. When they responded that it was faster to access the studies via the cardiology PACS, Menzies suggested an EMR interface to streamline access via the EMR. The fix worked and edged the medical center closer toward its universal EMR hub model.

The ultimate beneficiaries are patients because the central hub model improves multidisciplinary collaboration among caregivers. When a craniofacial plastic surgeon treats a child with a cleft palate, surgical assessment includes photographs, CT images and hybrid surgical planning images. “These images are helpful to the surgeon, but he [or she] doesn’t operate in a silo,” points out Menzies. Orthodontists, speech pathologists and ENT specialists access the same images via the EMR to develop a comprehensive collaborative clinical plan that includes braces and speech therapy.

UW Health reports similar enhanced clinical collaboration, but uses a multi-pronged image delivery strategy. “As a family practitioner, I’m using images much more [now that they are available in the EMR]. It’s easier to collaborate with radiologists and other specialists, and it can accelerate decision-making,” explains Valerie Gilchrist, MD, chair of family medicine at UW Health. For example, Gilchrist recently connected with a radiologist to interpret a growth plate fracture while the patient remained in her office. She also uses the EMR for patient education, explaining back pain to arthritis patients by showing them the details on their x-rays.

Flexible, open access

UW Health offers a three-tiered model for image access. The basic level consists of a very thin-client web viewer that presents key images first and then allows the provider to drill down into images.  The next level, a 100 percent browser-based middle tier, offers more functionality than an HTML page. A heavier, web-vendible application serves as the third tier. Finally, some users access images via PACS rather than the EMR.

Which model fits where and for whom hinges on both clinical and IT factors. Replicating the PACS model for image-intense ‘ologists is not necessarily the ticket to success. Nor is it wise to assume that all family practice providers can get by with a thin-client web viewer. “It varies more than one would think,” offers Wendt. That’s because some family providers, such as those with a high proportion of arthritis patients, may rely on robust access to images more than others. On the flip side, the occasional neurosurgeon may be satisfied with a thin-client.  

Although provider preference is a consideration in the image access equation at some sites, the end IT platform determines the access model to some degree as an older desktop with 512 Mb of RAM running Windows 98 cannot handle a web-vendible application, says Wendt. However, the underlying infrastructure needs to support customization and flexibility; UW Health utilizes an EMR that allows a multifaceted approach to integration from very thin clients to web-vendible viewing.

Ultimately, UW Health and Children’s Medical Center of Dallas begin and end at the same point—with the patient. UW Health’s flexible, user preference-based model is rooted in patient care, says Wendt. Enabling clinicians to view images according to their preferences encourages them to leverage image data and promotes image-based collaboration, which ultimately benefits patient care. A standard, global model, such as the one employed by Children’s Medical Center of Dallas, ensures all providers have access to the complete patient dataset including history and radiology and non-radiology image sets.

The middle of the road model

Community Hospitals of Indiana in Indianapolis offers complete access to the full image set in the EMR and as an API (Application Programming Interface) to providers across five hospitals, multiple imaging centers and physician groups. The configuration provides access to current exams and relevant priors as well as all priors in a timeline and includes measurement and complete image manipulation tools.

Although the desire to improve patient care drove the PACS-EMR project, physicians across the health system have realized workflow gains because they can look at images quickly and collaborate with other physicians remotely, shares Tammy Leighton, IT manager for medical imaging. Average turn-around time  for exams dropped from 24 hours in the pre-PACS/EMR era to 103 minutes after the project, which accelerates treatment and thus improves patient care, says Leighton.  

The hospital’s model is based on user-driven consistency. “We try to stay with one standard across the enterprise, so we can support users in the most efficient manner,” explains Rick Copple, vice president and CIO. Community Hospitals developed the model in collaboration with its physician groups and radiology providers.

In fact, the PACS deployment played an important role in establishing the infrastructure needed for PACS-EMR integration. Community Hospitals connected its sites by a fiber network when it implemented PACS, ultimately ensuring that the health system has bandwidth sufficient to support image access in the EMR. The health system maintains a private fiber network that currently uses 13 of an available 32 lambdas, between campus nodes. Each lambda supports up to 10 Gigabits.

Although Copple stresses consistency, the IT team recognizes that users’ needs do vary. An open PACS model allows the hospital to add APIs as user needs arise. That openness, coupled with a robust network, means that the system can be customized based on specific user requests.

For example, says Leighton, a future API might support pathology images or critical results notification. With integration between the PACS and EMR, the data would be shared and available in the EMR without a separate log-in to a different system.

Image-enabled efficiency

Deaconess Health System, an Evansville, Ind.-based six-hospital health system serving southern Indiana, southeastern Illinois and western Kentucky, integrated its PACS and EMR in 2009 on the inpatient and ambulatory sides. “The transition from not having anything but PACS and some historical documents to having PACS integrated inside of EMR is a tremendous timesaver,” explains Gregory K. Hindahl, CMIO, family physician and long-time EMR user.

Hindahl estimates that the integration saves a minute or more per each image set viewed over the previous method, which required separate log-ins to two different systems. Ease of use has boosted image review among clinical users, continues Hindahl. “We had images before in PACS, but a lot of physicians weren’t using computers to take care of patients. Plus with the EMR, physicians share images with patients, and it’s easier to communicate with specialists about what we see. It takes a lot of the guesswork out of the radiology part of diagnostic decision-making.”

Like UW Health, Deaconess Health System reports divergent access methods. That is, the PACS workstation serves as the primary image entry point in the ED, ICU and surgical suites, while providers in less image-intense locations access images via the EMR. Hindahl suspects that the PACS-driven image access model may change over time as users become more accustomed to tapping the EMR for image viewing.

In such cases, however, the monitor rather than the platform may be the critical issue as more complex imaging cases are better viewed via a high-resolution monitor typically found on PACS workstations rather than a lower resolution monitor on a non-diagnostic workstation.

How to plan for success

At this stage, there are no hard and fast rules for successful image access in the EMR. As pioneers demonstrate, it’s possible to build success via a number of strategies. A few universal guidelines still apply:

Planning stage: Think large and long-term. Collaborate with and engage all providers who need access to images.
  • Consider the IT infrastructure and support needs associated with the proposed strategy.
  • Plan in five- to 10-year timeframes, particularly with respect to bandwidth.

Decision-making stage: Seamless integration is an elusive and mythical goal (but efficient integration is attainable).
  • A quick way to assess what’s under the integration hood during the site visit is to count log-ins and clicks. Providers popping in and out of different systems is a surefire way to stymie productivity.
  • Take the integrated system for a thorough test drive in the home environment and make sure it functions at clinically reasonable performance level at all relevant tiers.

Implementation process
  • Offer flexible support and provide how-to manuals that address the continuum of image access options.