Rob Bart, MD, CMIO, UPMC
In this interview with healthsystemCIO, Dr. Rob Bart, CMIO at UPMC, discusses the challenges and opportunities of moving towards an enterprise imaging environment. UPMC’s strategy involves consolidating various PACS, and then moving those images to a Vendor Neutral Archive (VNA) to ensure they’re readily available for patient care. While the technology exists, the real challenge lies in managing change and gaining buy-in from clinicians. Read on to learn how UPMC is navigating this complex transformation to enhance care delivery and patient experience.
Watch or listen below; scroll down to read
LISTEN HERE USING THE PLAYER BELOW OR SUBSCRIBE THROUGH YOUR FAVORITE PODCASTING SERVICE.
Podcast: Play in new window | Download (Duration: 35:16 — 24.2MB)
Subscribe: Apple Podcasts | Spotify | Android | Pandora | iHeartRadio | Podchaser | Podcast Index | Email | TuneIn | RSS
Anthony: Welcome to healthsystemsCIO’s Interview with Dr. Rob Bart, CMIO at UPMC. I’m Anthony Guerra, Founder and Editor-in-Chief. Dr. Bart, thanks for joining me today.
Dr. Bart: Anthony, thanks for having me and asking me to join you this afternoon.
Anthony: Today, we’re going to talk about Enterprise Imaging. We’re trying to bring some stories to light of people who are doing work in that area, maybe having some successes and working through some challenges. Please tell me a little bit about your organization and your role.
Dr. Bart: I’m the Chief Medical Information Officer here at UPMC. I am responsible for the functional use of any of the clinical systems here at UPMC. UPMC is both a healthcare delivery and a payor organization, about a 50-50 split in revenue. Approximately 28 billion this year, more than 40 hospitals, 700 acute care delivery sites, predominantly in the state of Pennsylvania with small outposts in western New York and western Maryland.
Anthony: Excellent. As I mentioned, what we want to talk about is Enterprise Imaging and push to make images more readily available to clinicians across different specialties and the Enterprise, we know this is an area that can be difficult for patients and clinicians to navigate.
Dr. Bart: Thanks, Anthony. Historically, when healthcare systems talk about Enterprise Imaging, they are predominantly referring to picture archive systems. They’re usually specific for radiology images, occasionally for cardiovascular imaging systems which are frequently centralized. The early definition certainly didn’t include the other specialties that handle diagnostic studies and/or interventional studies that also generate images. Whether its gastroenterology or pulmonary with bronchoscopy or pathology as it moves more and more into the digital world today.
As I think about Enterprise Imaging, it goes back to before I started at UPMC. When I started in the summer of 2017, I really thought about all the images that entail a patient’s imaging jacket, so to speak. If I reflect on the electronic health record world, one of the goals has been to create a single consolidated digital electronic record, that’s text based, of all the information on the care of a patient.
Conceptually, I’ve taken that same thought and applied it to creating a single imaging jacket. A one for one relationship of that patient who is receiving care in the healthcare system and all of the digital images that are related to their care. From a consumer patient perspective, it should be as simple as going to one desk or making one phone call or one patient portal request to obtain a complete record that includes all the text-based documentation as well as all the images. I don’t believe anyone achieves that easily or well.
A lot of people talk about information blocking as it relates to what is traditionally text-based documentation in the EHR. But information blocking requirements also extend to images and making sure that patients and consumers have access to the images of their care. If we’re going to solve that for patients and consumers, we should also be able to solve and improve the workflow for clinicians with the same solution. That’s really the goal here, can we make it easier both for clinicians who need access to this information to effectively and efficiently deliver care, and for patients who need easy access to the information on their care.
Anthony: How do you move forward? It seems like every health system is going to be very different in terms of vendors. You’re going to have different PAC systems, different usage of images in different specialties. How do you start? Is there a general principle that says okay, we want to move towards an Enterprise Imaging paradigm; therefore, here are the 3, 4, 5 things we want to work towards. How do you start?
Dr. Bart: Frst, it’s defining what Enterprise Imaging is and as I said, I think it should be a single container that has all the images that pertain to any patient’s care and in this instance, it will be any patient’s care at UPMC. The other container is a vendor neutral archive which allows us to bring in images from all types of modalities of acquisition regardless of the specialty that uses the device.
I think one of the barriers is the change management on who owns the images. Historically, images have been used for that specific group of clinicians that are involved in the acquisition of them or ordering them. I’m a pediatric intensive clinic by clinical training and I order images and rely on the radiologist for the interpretation, but I also look at them and make sure that interpretation marries with the differential diagnosis I’m working on. That’s a collaborative process. Radiology has always been involved in a very collaborative process.
If you go to the next area where the largest number of images tend to be generated, that’s cardiology. Those images tend to be ordered by the cardiologist who is consulted and then the consumption of those images is either the cardiologist or the cardiac surgeon. That’s a fairly close group of people. But an ICU needs access to those images and the report of those images for the management of a patient in an ICU and so it starts to get larger.
There’s change management in working with a different physician or clinical specialties and they need to get buy in and understand that the images do serve their needs and a broader medical community. That’s why it’s important to move into an Enterprise architected picture management solution like a VNA. You need to get them to understand the why of the approach and what you’re trying to do. I think it also becomes somewhat serving to that group – our gastroenterology group expands a larger geography and so we want to make sure that the images acquired at a single location are available to that group across all the venues of care that we have.
As UPMC and other healthcare systems have grown, access points may not be just the place that they acquire them. Maybe they’re seeing that patient across the state where the patient originally traveled from and they need those images there. We can meet their clinical needs that way but also, it’s really getting them to understand that the broader community of clinicians need to see those images.
I think once you solve that, you move those images into a VNA through the change management approach that I just mentioned. Once you get them into VNA, then serving them up to the broader clinical community as well as being able to serve them up to the patients becomes much more realistic.
Anthony: Obviously clinical care and the needs of clinical care have not changed. As an Enterprise, you want to move into a paradigm clinicians can easily see an image. If we’re going to change the workflow, we need to tell them why.
Dr. Bart: I think how you state that is exactly right. The ultimate goal is to improve the efficiency of clinicians and also the effectiveness of their ability to deliver care. Going back to the cardiology example, early in my career I worked in a pediatric cardiac ICU. I took care of children who needed cardiac surgery. I looked at cardiac imaging like echocardiograms, working very closely with cardiologists. In the early days, I had to go to what was called a heart lab or a heart station or a cardiac lab to actually review the images but could only do it if I was physically side by side with the cardiologist at that time.
Now, I have the ability to view images independently and I can reach out to the cardiologist if I have a question or concern or thought about it. Sometimes, they might not have the full clinical context of the patient and I want to make sure that they’re aware that when this study was done, the patient was on certain medications or was in a certain phase of their illness like in septic shock or something of that nature, and we can have a dialogue. This made me a better clinician, and more valuable to the patient because I can view those images in near real time, in a process that’s relatively convenient for me as opposed to the days when I had to go find the cardiologist and go to the heart station to review images with them.
Anthony: The scenario you just gave, you talked about the old days when you had to sit at the cardiologist’s station, is that the VNA scenario or is that pre-VNA scenario?
Dr. Bart: In today’s world that’s pre-VNA.
Anthony: You’re logging into the PAC system?
Dr. Bart: I’m logging into their cardiovascular imaging system, yes.
Anthony: The VNA is sort of the nirvana state. It’s even better than that, correct?
Dr. Bart: Yeah, it is. It’s even better than that because it’s one-stop shopping for all the images of a given patient. They’re all within that patient’s jacket whether they are obtained by cardiology or gastroenterology or pathology, so the clinician can look at all these images that are important to patient care decisions. I have that now at my hand as well as the context of everything as it relates to the electronic health record.
Having that all together in one plane of glass, so to speak, in front of me, allows me to have a better picture and understanding of a patient and should enable better decision making or more efficient decision making on behalf of the patient. That’s ultimately what we’re trying to do. Simplify the task burden of the clinicians by bringing all the information together and decreasing the task or the cognitive burden so the focus is really on figuring out what is the best approach or the best care for the patient.
Anthony: Would you say that you have your VNA in all the PACS, it comes out of the modality and again, tell me if I’m wrong – the modalities take the images, they go into PAC systems, all different PAC systems, we’re going to get all those images move to the VNA.
Dr. Bart: Correct.
Anthony: Now, you have the VNA for all the images. You got Epic for sort of everything else. Help me understand the integration or if there is any, between the VNA and Epic, from a physician point of view. I understand we want all roads to lead to the EMR. Tell me about the relationship between the VNA and Epic in this case.
Dr. Bart: I think there are a couple different ways you can do that. The simplest way that images tend to be integrated into electronic health records like Epic and others is a context aware launching into that image from the electronic health record. If I want to look at a chest x-ray from today, I might have that report in front of me, I can launch into that. It carries the context of the patient and my log in into either the PACS or the VNA and I can look at that image. But that’s conducive if I need to look at one or two images.
If I want to look at a complex patient with a long history with our healthcare system that has a large imaging jacket with multiple images, we can link through an HL-7 construct called CCL which is again context or a mapping. It allows me to have the electronic health record and a full patient jacket up at the same time. Then, I may want to navigate from the patient jacket and say ‘well, I didn’t know that this pathology result was there. Let me look at that’ or ‘I didn’t know this patient had so many brain MRIs in 2021, that may be pertinent to the care I’m delivering.’
What’s important in either of these cases is making sure that it’s context-aware linking. Let the patient’s context and the authentication of myself happen seamlessly so I don’t need to log in to multiple systems.
Anthony: You used a great expression before, cognitive load. That’s what this is all about. This is all about figuring out the best way to bring information together onto one screen.
What are we planning for? Because if we plan for three giant screens and a clinician is working on a 17-inch laptop, things don’t line up. We’re human beings, we can only take a certain amount of information. You probably have committees figuring out how do we want to see things, how do we want to tweak workflow?
Dr. Bart: Typically, most of our clinical workstations are single monitor workstations with usually a 24-inch-high resolution monitor on them. Many of them are computers on wheels and so there’s a practical size that you can have. But in most of our care delivery nursing units, there is usually at least one location where you have higher quality multiple monitors for reviewing images in that location. Going back to when I started care, I had to go downstairs to radiology because it was film that you were looking at. With Enterprise Imaging, I don’t know if people would view this as three or four generations beyond the old film model.
You need to have a high-quality glass within the framework of where clinicians are delivering care to review certain types of images. Even a high end-laptop monitor might not have the type of contrast you need. A good example is a pneumothorax which is air that is outside the lungs but inside the rib cage. A pneumothorax can be challenging to see depending on the right lighting and the right type of monitor, and we want to make sure that it’s not only efficient and convenient for our clinicians. We want to make sure that we’re providing them with viewing stations to give them that ability.
Anthony: They’re not all the same. If you’re reading an image for a diagnostic purpose and the monitor is not up to the requirements, that’s a problem.
Dr. Bart: You’re absolutely correct, Anthony. I think we need to separate. Diagnostic image reading is really the domain of the specialty that’s acquiring it, radiologists. The highest order of that is in breast imaging. There are strict guidelines as to the type of monitor as well as the amount of ambient light in the room when you’re doing a diagnostic read. You want to make sure that the gastroenterologist, the cardiologist, they have good reading stations that meet their diagnostic requirements.
Most of what I and other clinicians are doing is a referential read. We want to confirm that what is on the monitor and what I read actually lines up. If I don’t quite see it eye to eye, then I need to track down whoever authored the report and review the images on their diagnostic monitors to make sure that I understand what they’re seeing. Make sure I’m interpreting things appropriately to make the best decisions for the patient. But predominantly, what I’m doing is really about referential images reading.
Anthony: You mentioned the VNA as a key strategic decision of moving forward in this image sharing world. Are there other options or is the VNA a no-brainer?
Dr. Bart: When we decided on the VNA there weren’t other options to consider. There were only brands of VNAs to consider. But the idea around the vendor neutral archive is it’s not limited as to the type of image it can import into it for storage. Most people understand in radiology or even in cardiology DICOM as sort of a wrapper that wraps that image.
VNAs can support DICOM images but they can also support non-DICOM images like jpegs, tifs and mpegs. It’s vendor neutral. It’s like a universal archive of imaging. Are there different flavors that will improve over time? I think that’s absolutely true. But I think conceptually, it is probably the broadest and most generic perspective on a system that can be used to collate and acquire the images in healthcare.
Anthony: Where are you in this journey to free these images from the individual PAC systems and migrate to the VNA?
Dr. Bart: We’re extremely early in on it. I think as you’re aware, Anthony, UPMC is partly an organically grown healthcare system but also one that has grown through merger and acquisition. On the radiology side, we’re in the process of implementing an Enterprise PAC system for radiology images. We have more than one PAC system today. There are two things we’re trying to accomplish within radiology, to move everybody to a single PAC system while also moving to a PAC system with a more modern architecture.
We’re also in the process of moving our cardiology imaging repository onto a singular platform. The first step in this is consolidating the acquisition of images within each discipline onto a single platform for acquisition. Then taking the images out of each of those storage modalities and pushing them all into the vendor neutral archive or VNA.
Anthony: That makes a lot of sense. Almost always when an Epic hospital or health system acquires a hospital that is on a different EMR, all that information is going to be migrated to Epic. We see it all the time.
From my understanding, it doesn’t seem to work that way in the PAC space because of the heaviness and the size of the images. But I’ve heard from other health systems, maybe the smaller ones, that’s just not going to work. Sometimes it’s better just to leave them where they are because migration is very expensive, very difficult. Do you have any thoughts around that?
Dr. Bart: It is a challenge. I do agree with the perspective that larger healthcare systems have more resources available to look at the migration of images. We are going through the migration process now. It can be a challenge but at the same time, the goal is to make sure we only have one copy. We don’t want to have the same large image represented in multiple systems. You’re paying the storage cost multiple times and if an image and a patient who had cancer diagnosed a decade ago is still pertinent to what an oncologist is doing for therapies today or if there’s unfortunately a recurrence in the future, and it’s pertinent to have that.
We want to make sure it’s available to our clinicians with ease. Our ultimate goal is to try and migrate everything into single platform but how fast we do that is going to be somewhat limited by both technology and the expense of it. But the goal is, the ideal state is to migrate everything and whether we actually achieve it, there’s some realistic limitations and much of that has to do with finance. The technology to do the migration pretty much exists.
Anthony: I just want to make sure I’m thinking of this right. Step one, a consolidation of PACS to make sure maybe you have one in each specialty as opposed to 3, 4, whatever, and then there will be a second migration to the VNA.
Dr. Bart: Your summary is appropriate. We’re trying to move more and more of what we do clinically when it involves any technology platform into single Enterprise instances whether that’s radiology PAC system, cardiology PAC system, gastroenterology imaging, bronchoscopy, et cetera. Because having a singular platform allows us to also make sure that we’re delivering a single standard of high-quality care across our system, eliminating redundancy and multiple contracts.
On many levels, having a single platform that is the right platform for UPMC really achieves a lot of the quality and operational goals that we have. We meet the immediate imaging needs of our clinicians then the VNA allows broader access for all the clinical community.
Anthony: You alluded to it before. A lot of times the hardest part doesn’t have anything to do with technology, it’s the human beings that we’re asking to do things differently can be challenging. You explained that explaining the why can go a long way to getting buy in.
Any more advice around that that? You’re the CMIO, almost every CIO who is working at an organization of any size will have a CMIO to partner with on this which I’m sure is exceptionally helpful. That’s why the position exists because physicians like to hear from physicians as opposed to techies. But your thoughts there around the change management piece?
Dr. Bart: I spent a lot of time trying to work through the clinicians being impacted, whether they’re physicians, nurses or pharmacists, therapists, trying to understand what’s in it for them as the organization is trying to accomplish some of the operational organizational goals. Understanding what brings of value to them or making sure that the project has inherent value to them and highlighting that, in my experience always eases the burden of the transition.
It still doesn’t mean that there won’t be challenges and you’ll have to work side by side with them through those challenges during the transition, but getting them to understand that upfront frequently results in a conversation, sometimes it’s 6 months, 12 months, two years later where they say, ‘I see why we did this’ or ‘I understand it now.’
The other thing that’s becoming much more apparent in healthcare is we’re trying to meet the needs of two groups of people. We’re trying to meet the needs of those who are delivering care regardless of the degree or initials behind their names, and we’re also trying to meet the needs of the patients who use the healthcare system. Getting people to understand that balance.
Let’s make it easier for the clinicians while also making it easier for the patients and consumers. Getting the clinicians to understand how important it is and how competitive healthcare is today as a service in communities is part of the other piece of the equation.
I look back 20, 25 years ago when we’re building systems, they are frequently to serve the needs of the healthcare systems solely from the clinician perspective and I think today when we are building things and working with different platforms that we’re deploying, we have to look at the lens of the clinician and then also through the lens of the consumer or patient.
Anthony: That’s excellent point, Dr. Bart, and a great point to wrap it on because we’re just about out of time. But I want to thank you so much for your time today.
Dr. Bart: Thank you for having me, Anthony. I enjoyed the discussion.
Share Your Thoughts
You must be logged in to post a comment.