C. Martin Harris, M.D., chief information officer of the Cleveland Clinic in Ohio, has this piece of advice for provider organizations that are making the shift from paper to electronic medical records (EMRs): when moving data from paper to an online model, make sure that you include all information—basic clinical data and imaging—in that shift. Imaging needs to be part of the strategy in a way that integrates the two types of information together, he says.
About three years ago, Cleveland Clinic embarked on a project to integrate its enterprise imaging data into its electronic medical record, tightly tying its imaging data to textual clinical data. Harris shared details of the project in July as part of a presentation on enterprise imaging, hosted by the Health Information Management Systems Society (HIMSS) and sponsored by Agfa Healthcare. Since implementing its integrated imaging approach, Cleveland Clinic has started a program, called MyPractice Imaging Solutions, to advise other provider organizations on imaging strategies.
Harris spoke of a paradigm in the traditional practice of medicine as practiced in this nation: make observations, run tests, analyze and diagnose, come up with a treatment plan and follow the patient over time. “When we think about caring for the patient, that is the process that is driving us,” he said. Historically, the information was captured on a paper chart and was separate from the image data, he said.
Now that paradigm is shifting, as basic data moves to the EMR at the same time there has been a migration toward imaging. Those parallel trends go to Harris’s point that provider organizations moving to electronic data need to also decide what their imaging strategy is going to be. “The management and delivery of great care to patients means managing all of the information—not just basic data, but imaging data as well,” he said.
The Case for Integration
Harris observed that images today are often captured distinct from one another, and in unique encounters that are related to a specialty. “If you can name a medical specialty over the next five to 10 years, I will guarantee you that they will have an imaging solution that is specific to their specialty,” he said. Those distinct images are linked by capturing some sense of an encounter before being moved to the EMR.
This can be done in either an interfaced way or an integrated way, he said. In an interfaced EMR, imaging modalities often require dedicated systems, displaying images apart from, and not linked to, EMR encounters. An integrated EMR displays images through a single viewer and incorporates images into the patient’s EMR encounter.
Source: Cleveland Clinic
Source: Cleveland Clinic
In Harris’ view, true integration that ties images with clinical data is more effective in delivering good care to patients. Integration allows various types of clinical information to appear seamlessly in an EMR, so the physician can better direct all of his or her attention on the patient and the information at hand, he said. “In my institution, the physician doesn’t have any idea of what the name of the laboratory information system is; and the reason is, he doesn’t need to,” Harris says. “All he really needs to know is the result, and whether or not it is in the normal range, and if it’s valid.”
As imaging moves forward, “We have to make image management as effective and as easy and as integrated for the patient as we are for basic data like laboratory information,” he said. Seeing a chest x-ray and having the image taking over the computer isn’t helpful to the physician; what is helpful is the ability to view the imaging data in the context of the patient data, he explained. “The EMR should effectively manage all of the data relating to the patient, maintaining the context, delivering the basic data and the image data, so I can keep the focus on caring for that particular patient,” he said. Ideally, without leaving the window for the patient, the physician has done acute care and screen work, all of which is driven by basic data and imaging, which appears in a single patient encounter.
Harris said that an integrated approach means getting rid of the distinction between text and images, and moving toward an integrated model. “The way to get there is to think about imaging as part of the EMR strategy from the very beginning, so you are constantly building a comprehensive medical record system that will allow your clinicians to deliver good care.”
Imaging Strategy and the Enterprise
How to make this work? At Cleveland Clinic, imaging grew up in radiology, a large department that had a picture archiving and communication system (PACS) early on, which was matched by a practice model. “The PACS system was designed fundamentally for the radiologist, making diagnostic interpretations,” Harris said. The same thing was true of cardiology. In each example, the specialty was thought of separately, so neither had a shared plan going forward. ”They were doing what they were used to, which was, ‘I’m a sub-specialty, and my goal is to optimize my component of the practice,’” Harris said.
Originally, Cleveland Clinic used an interfaced model, or a siloed approach, in building out its systems. For radiology, it built out workflow ability, storage and retrieval, with a viewer linked to radiology. The pop-up viewer, which was accessible from the EMR (which at Cleveland Clinic is supplied by Epic, Verona, Wis.), took the clinician outside the EMR as it presented radiology images of the patient.
Those radiology images were presented separately from cardiology images, which used the same siloed approach, although with a slightly different viewer. The same was true of other clinical specialties as well. In women’s health, fetal ultrasounds had the same model; that is, it was possible to view all of the ultrasound history or all of the radiology history, but not the ultrasound history and radiology history at the same time, Harris said.
Alluding to his prediction that every practice will have imaging, Harris suggested that an iPhone camera will give physicians the ability to take pictures, and they will want to get those images into the patient record and will expect other clinicians to act on them. “We had to figure out a way to get this all to work,” he said, “and for us, that meant driving an integration model, not a interfacing model, going forward.” The goal would be a single place where doctors could go to see all of the information—basic data and imaging data, and to create content—in a way that would allow them to manage all of the information in the application.
The Road to Integration at Cleveland Clinic
Harris described the transition process at Cleveland Clinic. Before embarking on that project, the head of the imaging department was primarily concerned about image management for radiology, but not imaging management for the enterprise, Harris said. Once the project was set in motion, the radiology team took on the task of setting up an enterprise image knowledge set and library, enterprise viewing capability, and enterprise storage capability. That meant a basic integrated image management approach, or a library of all images that is tied to the other clinical data that is related to that image. After building that layer, it distributed responsibility for it in the organization.
Today at the Cleveland Clinic, representatives from the imaging, cardiology and women’s health departments sit on a council where they are all have equal status, Harris said. They work together, functioning as departmental systems with responsibility for the same things, including the acquisition of the technology they need, and the acquisition and interpretation of the images. Each member is also responsible for managing the workflow inside their department. “They know the clinical pace; they know the turnaround times; they know what storage they need to have locally in order to meet the service level for the patients when they are servicing them directly,” Harris said.
The information technology department has responsibility for how the images are distributed, indexed and coordinated. “Information technology owns that layer in the enterprise,” Harris said. “I sit on the council with the other leaders and we think through how those images will be acquired so we can maintain the basic integrated platform going forward.”
The upshot is that Cleveland Clinic has made a transition to a smart library depository that gets presented to the end users through a common platform. In this model, the physician goes to a single place to access the text information within the EMR, and can click to find all of the images by department in the same context. “This is the fundamental distinction between the interfaced model and the integrated model, and it is absolutely consistent with all of the other information inside the EMR,” Harris said.
Harris also noted images are also easier to manage in the integrated model. In the past, the interfaced model required a team of people managing the interfaces for each department. “I had a whole team of people who knew when each one of those departmental systems was going to go down, when it was going to come up, and when it was going to be upgraded,” Harris said. “I tested and retested each one of those interface is to keep them working.”
In the new integrated model, a master system does the indexing and creates the “smarts,” he said. Using the integrated viewer and storage mechanism, the service levels are managed from a single place out to the end-user device and have a single interface going back and forth from that image distribution system to the end-user system. “It has brought us a practice model that is consistent with everything else we do, and it has also brought technical simplification going forward,” Harris said.
Harris predicted that in the next five to 10 years, imaging information will likely dominate medical practice information going forward. Today images give information about the structure of an organ, he said; in the future, images will give information about the structure as well as the function of those organs. “It is going to deliver a picture, but it is also going to deliver digital data that tells you how well it’s functioning,” he said. “Having all of those things separate from all of the other engines of the EMR is a strategic mistake when you think about efficient and effective decision making by clinicians who are going to be asked to make decisions faster and expected to make them more accurately going forward.”
The Transition is a Continuous Journey
Harris described two challenges to making the transition. One was conceptualization of the model with regard to the distinction between acquisition, interpretation and distribution of the images. The other major challenge was the human relationships piece. “The relationship is the trust that’s necessary to implement an integrated model, because some people are picking up responsibility and others are giving up certain responsibilities they had before,” Harris said. With the integrated model, in which the enterprise is responsible for the distribution of the images, group understanding and shared goals among the stakeholders are crucial, he said.
Those stakeholders are the physician leadership and the larger imaging departments. At Cleveland Clinic, all of the larger departments have made the transition, but Harris also observed that the universe is a lot bigger than originally thought—meaning that there are some sub-sections of departments creating images that are used almost solely by the people in those department sections. Currently, Cleveland Clinic is focusing on images in the operating room.
The last big challenge is going to be the iPhone and similar imaging devices, as well as the hand-held ultrasound machine, Harris said. “They are going to grow in numbers,” he said.
In fact, Harris regards the transition to integrated imaging as a journey, in which there will always be something new. Even images that are acquired and integrated into the EMR will change over time, he said. “It’s not just that it’s a DICOM image, but these images will more and more contain information about function, so you are going to have to change the way the integration happens.” For example, the image is not just telling the physician that a liver is normal size, but that it is normal size yet is functioning below capacity. “That’s important to know,” he said.