The Transformational Effects of Informatics on the Practice of Radiology: A Roundtable Discussion
On October 6, 2009, four physicians gathered in Stamford, Connecticut, to participate in a discussion moderated by Cheryl Proval, Radinformatics.com editorial director.
Robert Lipman, MD,is a radiologist at Straub Clinic and Hospital, Honolulu, Hawaii.
Terence Matalon, MD, FACR, FSIR,is chair, department of radiology, Albert Einstein Medical Center, Philadelphia, Pennsylvania.
Greg Mogel, MD,is medical director, imaging, at USC University Hospital, Los Angeles, California.
Andrew Wells, MD, is a radiologist at Margaret Pardee Memorial Hospital, Hendersonville, North Carolina.
Robert Lipman, MD,is a radiologist at Straub Clinic and Hospital, Honolulu, Hawaii.
Terence Matalon, MD, FACR, FSIR,is chair, department of radiology, Albert Einstein Medical Center, Philadelphia, Pennsylvania.
Greg Mogel, MD,is medical director, imaging, at USC University Hospital, Los Angeles, California.
Andrew Wells, MD, is a radiologist at Margaret Pardee Memorial Hospital, Hendersonville, North Carolina.
Radinformatics: PACS has turned out to be quite evolutionary in its effect on the way in which radiology is delivered to the marketplace. What have been the greatest positive effects on your radiology service of this ability to move images? What are the limiting factors of the software and technology? Wells: The greatest positive impact is that the historical priors are available. You have to remember that in the past, the image location was dictated by the physical location of the film, jacket, and report. The cycle time, from image acquisition to reporting with comparison using a result from earlier today or from yesterday, has dramatically improved. That’s a basic, fundamental issue, but it’s very important as a clear advance. The limiting factor is that there’s no perfect delivery platform to get you the clinical information to make that interpretation, but it’s better. Lipman: PACS multiplies the number of copies of the image that can be seen at any time, which is a very important feature. The radiologist can be interpreting it and the surgeon can be operating on the patient while looking at that same image. To show how important that is, in my system, we have satellite clinics; half of them have CR that gets sent to the radiologists immediately and half of them have film, and the ones that don’t have CR are clamoring for it. They’re asking for stat readings on their films, which is kind of ridiculous because it took a day and a half to get the film to us, but they’re demanding that as soon as it rolls in the door, we read it immediately. What they really want is CR. Matalon: PACS has had a tremendous impact. Obviously, there’s been a marked ability to improve productivity for the radiologists, and it’s also driven major changes in the subspecialties and how we deliver imaging interpretations. The ability to read remotely has had a big impact on staffing, and it’s also driven what we’re currently seeing, which is subspecialization of image interpretation and the ability to send the image and allow it to be interpreted by the proper person, regardless of where he or she may physically be. The other major change has happened because the images are now digital: The ability to manipulate images is something we didn’t have with film. Mogel: It’s hard to mention the limiting factors because I find it hard to envision a future of radiology that doesn’t have PACS. To me, PACS and radiology are inherently tied together. There is no separating them anymore. While there are fewer and fewer practices that don’t use PACS, there are almost no new ways of imaging the human body that don’t require you to have digital presentation and storage. It is impossible to imagine imaging in its present form, and certainly in any of its future forms, without PACS. Its limitations are those inherent in medical imaging: the fact that we’re developing imaging studies that are larger and larger, and more and more complex. Those are the promises and the limitations. The first huge breakthrough that changed everything was the fact that you had the stored image and the clinical information all in one piece. Before PACS, the limitation was that if you lost it, it was gone; only one person could see it at a time, and it was not at the point of care. Everyone had to trudge down to the radiology reading room. To most younger physicians today, this is almost unimaginable because of PACS. The limitations are the same as for any IT: the absence of standards and the tug of war that exists between customization and standardization. Now that we can have images almost anywhere we want at any time, those rare circumstances where we can’t become intolerable. Radinformatics: Aside from PACS, three other information systems are fundamental to the interpretation of radiology exams: RIS, advanced visualization, and reporting. What is the potential impact of these four interfaced systems on radiologists’ productivity? What has been the experience (both positive and negative) of your facility? Matalon: The potential is to use information across systems, and information initially delivered on the RIS side will drive exam protocols, should drive reading protocols, and should drive the advanced visualization presentation. All of those things are happening; they’re just happening very slowly. At our institution, we’ve seen much more use of advanced visualization over the past few years. It’s uncommon to read a CT exam without using some form of 2D or 3D toolset to enhance the interpretation process. It’s become a basic part of how we interpret almost every exam. Wells: Reporting is the opportunity where changes have not created more value, but have disrupted the process. Dictating for a transcriptionist is a process that has a cycle time and has limitations and errors, but it also provides a certain ability to concentrate on the image, whereas voice recognition or structured reporting—where I am, almost at the same moment, analyzing the image and analyzing the report generation—has been quite disruptive. Part of that is technology, but part of it is having more than one focus. Reporting systems are an opportunity to improve not just the process of generating the report, but encoding value in the report. How significant or new is the finding? What is the urgency of follow-up? Reporting systems are a great opportunity to do that, and they have not, to this point, done it. Lipman: RIS and PACS developed at a similar time, and I remember being told not to get a PACS without a RIS, but I know someone who’s had a PACS for a decade without a RIS and has derived value from it for all those years. Mogel: A RIS is an important backroom thing that should be as invisible to the radiologist as possible. It’s the engine of the car, not the leather seats or the chassis. It had better work, but it’s not sufficient to make a radiologist happy. Advanced visualization, for some of it, isn’t even advanced any more—it’s absolutely core. There are imaging technologies that exist today that are not perceptible by human beings without advanced visualization. Thousands of images pour off these devices to study one small part of one patient’s body. Just the compilation and rendering of those images is impossible without advanced visualization. Reporting, though, is the most exciting and compelling thing to work on, because in the end, what do we do? We tell other people what we know, and they make decisions based on that. The way we transmit that information to them has changed far more slowly than anything else, in part because displaying an image is technologically simpler than voice recognition. The report is only now changing, but in the end, that’s our product. The images really are for us, and the advanced visualization is a tool to help us, but the information we provide is what changes clinical decisions. Radinformatics: Do you foresee having to invest in complementary imaging systems in order to meet the necessary standards to qualify for funds under the American Recovery and Reinvestment Act (ARRA)? Do you hope to use ARRA funds toward advancing imaging informatics? Wells: In western North Carolina, there is a consortium of hospitals that agreed to share certain data within a secure environment, and this allows patients to be seen at any of the 16 hospitals. I can securely log on and identify laboratory reports, imaging reports, history/physical reports, and other information with an encrypted, HIPAA-compliant mechanism. The process of delivering smarter, nonredundant health care through funding these kinds of efforts will be an opportunity to cut down on waste, and perhaps to approach fraud and abuse. Empowering that would be a great way to improve health care delivery not by changing the delivery units, but by supporting their independence. Whether or not our group is successful in getting ARRA funds for this kind of thing is uncertain at this time, but the group is aware of the possibility. Mogel: Our specialty is in much better shape than most, but it would be good if there were some organizing body that would help us recognize how we could effectively apply for those funds. Most specialties want to get stimulus funds, but they are hamstrung by the meaningful-use concept. The federal government said that applicants for these funds had to demonstrate meaningful use of the technology, but it was not ready to define meaningful use when potential applicants asked. Everything that’s coming out about what those definitions will be, though, is already what we do: privacy protection, portability, changing behavior, outcomes, and data collection. We’re in a good position to do this, but it’s an unformed picture. Matalon: A national effort to support regional health information organizations (RHIOs) would certainly be beneficial. It would have a positive social and economic impact. While we aren’t doing that independently at our institution, we certainly applied for stimulus funds from the broader perspective of a hospital information system (HIS), but not from a radiology-specific perspective. We have done some preliminary work with RHIOs, but each institution generally fails to see the value to itself of that activity. That’s why national support is a positive idea. Radinformatics: Stories abound about health information systems that have not delivered on their promises. What do you see as the single most significant barrier to the successful deployment of health IT? Matalon: The major barrier is cost. What it costs to buy a reasonably integrated HIS is a staggering amount of money. At my institution, we’re looking at more than $50 million over just a couple of years to implement a HIS that will not even answer all of our specific needs, but will get us to a point where we feel we need to be. You basically are struggling between the needs of customization and standardization. Wells: I absolutely agree that cost is a profound factor. I have the experience of an institution that changed its HIS/RIS recently; PACS was already deployed, so that brought all sorts of adventures in nomenclature, medical-record numbers, patient-identification numbers, and the tremendous challenge of changing vendors. Health IT is constrained by vendors who perceive their value in a silo. They deliver a product, and you’re supposed to use all of their applications within that. If you have a PACS, laboratory information system, or legacy ECG monitoring system that doesn’t plug in and play, then you have additional frustration, cost, and perhaps limitations on how to extract value from the application and interact with that new foundation product. I see a great opportunity for companies that develop tools to link across systems that don’t really want to talk to one another. Some vendors don’t see the value: they sell functions in a silo, and they say, “Do it our way or no way,” and that’s just not valuable. Throwing out a multimillion-dollar choice to get something more adaptable is very expensive, but I think intermediary players are going to become more valuable. Lipman: The money is a limiting factor, but imagination/creativity is, too. Mogel: I had the experience of working with an organization not known for being constrained by resource limitations or by region: the US Department of Defense. When you look at federal systems, what they have done with health IT is better, but it’s not that much better. They haven’t overcome all the hurdles, even though they haven’t been nearly as cost sensitive as the private sector, so there’s something other than price that constrains health IT. There’s something else—something cultural and deep. A lot of the IT that went into health care for the first decade or two was focused on administrative (not clinical) tasks. Physicians saw IT as a threat to their autonomy, because it was used that way, and now physicians and patients have ended up with the least modern information systems. They are still somewhere in the 1980s or 1990s because they were reluctant to adopt IT when it was growing. Radinformatics: How will the wider adoption of IT throughout the health care system affect the future practice of radiology? What impact might this wider adoption of IT have on patient care and on cost? Mogel: No technology was ever widely adopted by human beings because it saved money. We adopt technologies because they do something else that satisfies us. Wells: It’s important to realize, though, that you will change your behavior and use a process if you get more value. Health IT must create greater value for the stakeholders throughout the enterprise. In radiology, that means pulling clinical laboratory reports, pathology reports, clinical notes, and radiology reports together and delivering them to me, wherever I am sitting. Matalon: There is going to be a wider adoption of IT; I hope that will be done in concert with improved standardization of how information can be moved across platforms, particularly from both inpatient and outpatient perspectives. Most people think about a HIS and forget that a lot of the care is delivered in an outpatient system that may not be part of the HIS. The need to combine those will deliver improved historical and current patient information, and that probably will give both better outcomes and lower cost. The increasing use of intelligent physician order entry systems that recognize historical use of imaging and prompt the ordering physician about current clinical need based on signs and symptoms, as well as historical use of similar services, is going to have a positive impact as well. Lipman: Medicine can learn from other industries. If the hospital industry was run the way that the electronics industry is run, we would have been replaced by something else by now. We’re very inefficient. We do unnecessary things, and we do them in wasteful ways. Mogel: Technology amplifies social structures, and if medicine is illogical in how it is delivered because it developed over thousands of years, technology won’t eliminate those illogical aspects. It may shine a light on them and let us eliminate them, though. Technology can’t make us practice medicine better unless we decide to do that first, and then use the technology that way. Faster CT scanners haven’t made the CT scans that are ordered more appropriate. Wells: Health IT should empower business analytics. If you know the resources involved in doing something, you can manage it. In health care, the filters of resources used, hospitalization time, infection rate, readmission rate, major morbidity, and death should form a funnel of quality that can be brought to bear. Business analytics should help drive higher-quality, lower-cost care. Lipman: A classic illustration of this effect, in radiology, is that if you do more abdominal CT exams for people in the emergency department to rule out appendicitis, you’ll do fewer appendectomies for people who don’t really have appendicitis. Mogel: In the same way, presurgical planning based on imaging leads to shorter stays, better outcomes, and higher patient satisfaction. Radinformatics: What do you see as radiology’s biggest challenge over the next three to five years? Lipman: It will be to limit the amount of imaging done. We’re going to get better, faster, and maybe lower in radiation dose, but we can’t keep expanding what we do. We have to do it when it’s worthwhile. Matalon: Controlling imaging is very complex. We can have an impact on the imaging we do, but imaging done in offices or entrepreneurial setting where there are economic incentives to do imaging is something we have no ability to affect or control. Our biggest challenge is to continue to show value, and we’re going to have to become much more efficient and to employ more business tools to reduce waste and reduce our own inefficiency. Mogel: Our challenges are in fighting the urge toward commoditization of imaging and in making radiology a high-value service. In too many places, radiologists are judged based on how many studies they can get through, rather than on whether those studies are important, and to whom. Wells: Radiology’s challenge is to advance how we deliver value and where we can do it, as well as serving the patient, the enterprise, and the referring physician. In a PACS environment, you don’t own the image anymore. You owned the image when you had a sheet of film in your hand. Now that you don’t own it, you have to prove that you deliver value through that image and fight to keep it.Edited by Kris Kyes, technical editor of Radinformatics.com.