SUNY Upstate: Taking the VNA Route to an Image-enriched EMR
Purchase a vendor-neutral archive (VNA). Design it to store DICOM and non–DICOM images. Integrate it seamlessly with electronic medical record (EMR) system. Watch its users visually enrich their patients’ medical records. That’s the experience that the State University of New York (SUNY) Upstate Medical University is having, but not before doing its due diligence and building a support infrastructure to ensure success, according to Steve Roberts, SUNY’s director of host systems. The VNA not only is providing a centralized repository for patients’ images, but has helped improve the quality of medical records, has eliminated some security risks, and has dramatically slashed data-migration costs for images. SUNY Upstate is central New York’s largest medical center and the only academic medical center in the region. On its main campus in Syracuse, it operates Upstate University Hospital, with more than 400 beds (and with the region’s only level I trauma and burn center); Upstate Golisano Children’s Hospital; and both adult and pediatric cancer centers. In July 2011, it purchased 306-bed Community General Hospital, which has been renamed Upstate University Hospital Community Campus. Together, these facilities perform approximately 350,000 imaging exams each year. With the acquisition, the IT department found itself operating two different PACS and two different software versions of the same RIS. The desire to merge the archives was one factor contributing to a decision to acquire a VNA, but the IT department had investigated this option at least four years earlier and concluded that technology was still emerging and implementation would be premature. “In late 2011, several months after the acquisition of Community Hospital, a consulting company was retained to conduct an enterprise imaging assessment,” Roberts says. “The company recommended that IT operations be centralized and recommended that we take a fresh look at VNA technology.” Another factor that spurred interest in considering a VNA was the implementation of an electronic health record (EHR) system for both inpatient and ambulatory-care use. When IT staff started to discuss this project with various departments, a decision was made to implement this first for ambulatory care.
“With many departments we visited, the reaction was the same: Everybody said great, we can get rid of our charts—but what do we do with our images?”
The EHR software selected did not have an imaging module. The main Upstate University Hospital PACS had been storing all sorts of patient-related images in addition to radiology exams. Its archive served as the repository for all cardiac images; endoscopy images; ultrasound exams performed outside the radiology department (such as those originating in the perinatal center); and images from departments that happened to have a modality, but no storage. Would it have the capacity to store additional DICOM and non–DICOM images from all the other sources at SUNY Upstate? VNA Reconsidered When the IT staff took a fresh look at VNA offerings, it liked what it saw. It began a selection process that started in January 2012. Nine months later, the VNA was operational. The ambulatory-care EHR had been implemented. Images from the Community Campus PACS archives were in the process of being transferred to the main FUJIFILM Synapse PACS archive. A major software upgrade (version 4.2) to the PACS was being planned. The vendors all worked together to develop a transparent EHR–PACS–VNA integration, Roberts says. He was impressed with their performance, and with the deliverable: An enterprise imaging link from the VNA, embedded in the EMR, provides the user with access to diagnostic images. The ability to add and access DICOM images was first made available at SUNY Upstate. The rollout was very basic; people were informed that the link was there. There was practically no training. It really wasn’t necessary, according to Roberts; he says, “The software was integrated so well that when users wanted to see images in a patient’s file, they clicked on the link, and that was that. They didn’t realize that they were opening another application, namely access to images contained in the VNA.” The next step was to add non–DICOM images to the mix. The rollout for this application was slower and more cautious. The software-development team wanted to verify that it worked well and wanted to watch how it was being used (to catch any potential software problems or user-related issues). The departments selected as the first users included the wound clinic and the photography department that supported the operating room. One of the issues to be resolved was the need to create a method by which the medical-records department could include images when it received a request for them. The IT department had to develop a workflow to enable such requests to be processed smoothly, as well as policies for the use of research/teaching images. When the IT department determined that users were uploading images, it needed to develop and implement a campaign to educate users that only images that pertained to a patient’s visit or treatment at SUNY Upstate should be added to patients’ files. The staged rollout allowed documentation protocols and training programs to be developed. The second user phase (consisting of the emergency department) is about to go live. Current users are delighted to be able to add non–DICOM images to the EHR. They feel that they have an official place to store images, and they have stopped using external USB drives, thumb drives, or whatever mechanisms they could find to store images in their departments—a result that improves patients’ privacy and data security. Even if secured and locked, portable image-storage devices have a potential risk of being misplaced or stolen. Creating a Support Infrastructure With so many images expected to be added to the VNA, an image-governance committee was created at the recommendation of the consulting company. The committee is tasked with creating guidelines for adding images and with establishing any necessary protocols; its members include IT directors, the CIO, the CMIO, the chairs of the cardiology and radiology departments, the operating-room nursing administrator, and representatives of the medical-records and legal departments. One protocol in the process of being developed is identification of the images that need to be retained for a specified time. Some radiology exams and other DICOM images need to be retained for 21 years, but what should be done, for example, with a series of dermatology photos showing a rash and its reaction to treatment? Data storage is expensive, and as more people gain the ability to add non–DICOM images, the larger the storage capacity needed will be. The image-governance committee is making the necessary rules. Two positions were created in the IT department after the support structure for the VNA was evaluated. There is now a manager of enterprise imaging and storage management who is responsible for all storage requirements for the institution. Prior to the creation of this position, the management of storage was distributed across several groups. With the creation of the new group, a VNA administrator’s position was also created; this person reports to the manager of enterprise imaging and storage management. “We wanted one team to cover all storage issues and to look formally at trends of data use and growth,” Roberts explains. “This had been done piecemeal, and we decided that if we consolidated individuals into a group team, our collective knowledge would improve our purchasing decisions and strengthen our negotiating power.” The changes have worked out well. A big benefit of having a VNA has been a dramatic reduction in the cost of data migration. One IT staff member is responsible for this, and the process has been ongoing. Roberts says that the cost of data migration (using the VNA) for an existing cardiology system that was being replaced by another vendor’s system was about 25% of the expected amount. In addition to being a very pleasant budgetary surprise, this means that future information-system replacements won’t require data to be migrated again. Once in the VNA, the data will stay there. Migration Strategy The data-migration process, however, is never simple. Regardless of who does the migration, there always will be messy data: mismatched names, duplicate names, and incorrect medical-record numbers. Roberts says that when SUNY Upstate decided to add a VNA, everyone knew that this would be a project with challenges and bumps. The team had to rethink its strategies several time to determine the best path for image flow, as well as what should be linked. Decisions such as whether to link records by accession number or by the study’s unique identifier, for example, needed to be made. One of the lessons learned from the experience, Roberts says, was to plan for this—and to build time into the schedule for dealing with glitches. Doing one’s homework to evaluate the technology and vendor offerings is important. Careful planning and enough staff resources to do the job well (without unnecessary pressure) are also required. Roberts adds that the reasons that this was a successful project were that the hospital administration supported the project, provided the capital-investment funds needed, and the paid for the additional staffing required. “We’ve had a good experience with our VNA, and we are very happy with it,” he concludes. “Everyone is waiting to add and view non–DICOM images. It’s a great feeling to know that all types of clinical images can be added to enrich a patient’s medical record. Everyone appreciates this value.” Cynthia E. Keen is a contributing writer for Radinformatics.com.
—Steve Roberts, SUNY Upstate Medical University, Syracuse, New York