Dr. Siegel is Chief of Imaging, VA Maryland Healthcare System, and Professor and Vice Chairman, Department of Diagnostic Radiology, University of Maryland. Baltimore, MD; and Lead Imaging Informatics, National Cancer Institute, Bethesda, MD.
When attempting to understand the value of information technology in everyday practice, the question to ask is “What does every radiologist need to know about imaging IT and imaging informatics?” It is much more than just whether or not your PC is working or whether your picture archiving and communications system (PACS) is working.
In this supplement to Applied Radiology, we will discuss the critical role of imaging informatics as a way to improve productivity. We are going to address how you can streamline your practice, in an era of increasing study volume and complexity. We are going to focus on ways to get real-time feedback on radiology department performance, which is going to be critically important in a pay-for-performance environment. We are also going to talk about reinventing the radiology reading room, and how that’s related to imaging informatics. We are working in an era where radiologists are tasked with reading large data sets and so we will discuss the benefits of volumetric imaging in thin-client, or server-site rendering, models. We are also going to discuss how to re-invent the radiology report, and the increasing role of structured reporting, particularly in how we communicate and track potentially important and urgent findings. As an information technology or imaging informatics challenge, it is perhaps the greatest challenge we face.
The future of radiology
With regard to preparing for radiology in 2015, there’s an impending and probably already-arrived productivity crisis. At the University of Maryland, we have had multidetector computed tomography (CT) since September 2001 and we have seen a precipitous rise in the number of CT studies per thousand patient visits in our emergency department since then. We have also seen a sharp rise in the utilization of CT studies for outpatients, as well. So if we average about 800 images per study. And if a typical radiologist reads about 40 studies per day, times 3 window/ level settings, times 3 planes, with a current and a previous study, we total about .58 million images. You would have to look at 6.6 images per second, for 24 hours, without eating or sleeping in order to be able to keep up with that number. So we need to think about using information technology to reinvent the way that we review these individual slices.
Before multidetector CT was introduced, we thought that we would be able to keep up with the volume in CT. And as we have gone to higher study volume with more images per study, I think we are starting to feel like Lucy and Ethel, from the classic I Love Lucy candy factory episode, as the assembly line started moving faster. Many of us feel as though we are fighting a losing battle.
So how do we build a better candy factory? Radiology departments are not traditionally run like an efficient assembly line. There’s often limited, if any, workflow analysis, and there are limited quantitative tools for analyzing our workflow. IT has a role in helping us to analyze and improve workflow. By example, we did a workflow study in 1989 and found that it took 59 steps to complete an inpatient chest radiograph. From start to finish, the procedure involved multiple different personnel. How are we able to use IT to reinvent our workflow and reduce the huge number of steps down to a manageable few? No single tool is the answer. PACS must not be an island, but needs to be able to communicate with other hospital information systems. At the University of Maryland, we enabled communication between our various information systems, and in doing so, we automated the process and were able to significantly decrease those 59 steps down to 8 steps.
Our efficiencies are not solely attributable to transitioning to filmless imaging, they were realized because we have integrated our various information systems and enabled them to intercommunicate. The next big challenge for informatics and information technology is going to come from pay-for-performance.Government and reimbursement agencies are increasingly asking healthcare providers to assess the quality of the services that they provide. We have only just seen the start of these initiatives. We are going to increasingly be asked to perform benchmarking and to clarify the level of quality and the level of performance that our imaging departments provide.
At the University of Maryland, we have created a cornucopia of different dashboard tools to help us analyze quality, productivity and performance. On a real-time basis, our dashboards allow us to monitor image orders and patient arrival; image acquisition; interpretation parameters, including the number of studies that are undictated at any particular time; and, we even have measures of CT scanner availability, andwe can track hand hygiene in the department. We are able to look at performance on a monthly basis, a daily basis, or even on a minute-to-minute basis. We are also using IT for real-time assessment and feedback on image quality. We can track such measures as repeat and discard rates for general radiography. We also use our tools for education, by generating technologist report cards that provide feedback for parameters such as positioning, data quality, radiation, use of markers, and several other criteria.
This supplement features the work of William W. Boonn, MD, Chief, 3D and Advanced Imaging at the Hospital of the University of Pennsylvania (Philadelphia, PA). Dr. Boonn will address how we can use IT to cope with complex studies. There is a key paradigm shift that comes with the way we view images from thin-section CT. It is similar to the paradigm shift we saw when first going from film to digital, where we changed the way we acquired data, we changed the way we stored data, and we changed the way we displayed that data.
For example, even though we scan patients in the axial plane, as we acquire CT, we now have the capability to use isotropic imaging and thus go from acquiring slices to acquiring volumes. This gives us the ability to look at images in any plane and in any perspective. Rather than being passive consumers or recipients of image slices, we now have the capability of navigating through more complex data sets. It doesn’t matter how many images we acquire, because that number determines the level of in-formation, and complexity of the information, in a volumetric data set. We have effectively disentangled the process of acquisition from the process of image review. We can think of a CT study as an interactive movie, where the radiologist has complete control over factors like slice thickness, plane of view, lighting, etc.
The other change we are seeing is that “advanced visualization” has matured. The term used to apply to reconstructions in the coronal or sagittal planes, or being able to do maximum-intensity projection for pulmonary nodules. These have become routine aspects of the image interpretation process. We are now taking the tools that were once thought to be associated with an advanced visualization, and putting them into daily, routine practice.
With this has come another paradigm shift in terms of having the capability of remotely interacting with these large data sets. Instead of waiting for large data sets to be transferred to our PCs we can now work with only the visual subset of information that we are interested in at the time.
Benefits of CAD
Another major impact of IT on how we practice comes from computer-aided detection or diagnosis (CAD). Besides relatively established applications in mammography, chest X-ray and CT colonography, CAD is working its way into musculoskeletal imaging, and additional organ-specific uses. In the future, I believe we will be interacting with CAD in a way that’s vastly different from what we do today. Rather than CAD serving as a second reader it will help us render a primary interpretation.
It is important for radiologists to understand both the strengths and the weakness of CAD and other automated decision support tools in order for us to be able to use them as effectively and as intelligently as possible.
No discussion of the value of IT as it pertains to everyday radiology workflow would be complete without discussing its benefits on radiology reporting. In this supplement, we address some tools associated with structured reporting and with speech recognition. First, Daniel L. Rubin, MD, MS, Assistant Professor of Radiology and of Medicine, Department of Radiology, Stanford University, (Stanford, CA) will discuss controlled terminology for reporting and new emerging standards.
Second, Rasu B. Shrestha, MD, MBA, Medical Director for Digital Imaging Informatics and Chief, Division of Radiology Informatics at the University of Pittsburgh Medical Center (Pittsburgh, PA) will discuss how to optimize a voice recognition system from a workflow and usability perspective.
And so, there’s the capability of being able to look at an image, generate ideas about that image, and then be able to put those ideas down, in a structured way, and report that information in many different ways.
In the future, we will use IT to create personalized interpretation dashboards that allow radiologists to be able to create their own specific workspaces with all the information and tools they need at the time of image interpretation.
Communicating radiology reports
One of the most important challenges of all is communication. The radiology report has not changed significantly since the late 1800s.With the introduction of IT and imaging informatics, we are beginning to make significant strides in the way we report. Development of speech recognition systems is moving forward rapidly. And we are also making some major strides, for the first time, in structured reporting. BIRADS has been one example of a successful implementation of structured reporting which has had major clinical benefits. The RSNA is currently involved in a large-scale effort to create a number of structured templates outside of mammography, in many different areas, including oncology. As many as 60 templates could be available by the end of 2009.
In my opinion, structured reporting is going to pave the way towards a whole new way that we think of medicine. It will enable us to provide our image interpretation and cross-correlate it with genomic and proteomic information. So I think that probably the biggest challenge in diagnostic imaging is closing the communication loop. It is not just about dictating and signing off on our reports, but making sure they are received and understood. It would also be helpful to track whether or not our recommendations are being acted upon. We do not really do a good job of that, today. And I think IT, in the next few years, will allow us to help close that communication loop.
With regard to education and training, I think that we are going to use IT to reinvent the way that we train and educate not only our residents, but ourselves. Maintenance of certification should be something that we do as part and parcel of our workday. By that I mean that we should get credit for all the times that we look up information, because these are times when we are learning in the course of a normal work day.
As it pertains to our residents, we are changing the nature of our residency programs with IT. These changes allow us to make the education process more active. We can use social networking tools like blogs, Wikis, and document tagging to enhance our educational material. Our residents are using laptops, and we are creating an environment where they can be much more interactive, and where they can learn in a much more dynamic way.
Reading room design
Another significant challenge in informatics has been the human-machine interface—specifically as it relates to the design of a radiology reading room. This supplement includes a piece from Khan M. Siddiqui, MD, Principle Program Manager at Microsoft (Redmond, WA).The issue is significant as radiologists, technologists and administrators commonly experience eye strain, back pain, and carpal and cubital tunnel syndromes. Still, radiologist reading room re-design has received surprisingly little attention. Dr. Siddiqui will look at optimizing our use of lighting and acoustics, and he will address the impact of those modifications on a radiologist’s performance.
Image and data sharing
Finally, we can use IT to change the way we transfer data amongst (or within) facilities. The problem has gotten so bad that our colleagues in the American Medical Association recently passed a resolution that detailed and decried the problems associated with the use of CT or magnetic resonance imaging (MRI), complaining that it is harder to look at their patients’ images in a digital environment than it was with film.
We are going to see IT solutions that allow us to electronically, and in a much more seamless way, be able to exchange information. And, in many cases, we’re going to change from the model of information being siloed in specific hospital environments to having information associated with the patients. And we are going to see more of the patient-centered record.
I don’t think we give ourselves credit for the many things that we, as radiologists, do well when it comes to leveraging IT to its fullest capabilities. For example, we are ahead of the curve when it comes to archiving images. We take responsibility for storage and backup of those images. Ask a radiologist, why she keeps images archived as long as she does, and she will respond that it is for medical-legal reasons. Then ask a gastrointestinal endoscopist (and I have asked this question) why he does not archive his images, even though the studies could be archived as easily as we archive ours, and he will respond that it is for medical-legal reasons.
Radiologists have effectively taken full responsibility for their images. And that means that we are also responsible for sharing those images and making them accessible outside of the radiology department. It also means that we have taken responsibility for: integrating images into the electronic medical record; optimization of image displays; security and access; education of our colleagues; comparison with previous studies; review of studies for appropriateness; presentations at conferences; and timely reporting of studies to other physicians.
All of those responsibilities rely heavily on IT, and those responsibilities are what radiology provides as value-added services in the healthcare chain. I am very optimistic about the future of diagnostic radiology as a strong and enduring specialty: as long as radiologists continue to provide added value above and beyond our diagnostic interpretations, and as long as we continue to provide a leadership role when it comes to our expertise in image acquisition, management analysis, and computer systems implementation. I think radiologists are far ahead of the IT curve compared with other specialties and I think it is critical that our specialty maintain that lead.