Fed Ghali: Thank you so much for having me. Yeah, it's a real privilege to speak with you and to share some of the work that we've been working on. So really, really grateful to be here. Thanks. I'm going to share my screen and tell you a little bit about the development of this Electronic Health Record-Based Care Pathway that we've implemented at our institution, and just share some of our experience using it. So the first thing I want to do is I want to put an acknowledgement slide right up front. These are some of the folks who really were the brains behind this project. Pat Kenney is just a fantastic surgeon. And like I say, a lot of the brains behind this project. And Shayan is a medical student who's applying for residency now, just a really impressive young man. And Soum is our former resident who's now an SUO fellow at Hopkins who really drove this project forward. So just want to acknowledge them right off the bat.
So the background will be no surprise to your audience, but non-muscle-invasive bladder cancer is common. We see somewhere around 84,000 new diagnoses of bladder cancer annually in America, and about three quarters of those are non-muscle-invasive bladder cancer. So we're all seeing this in our practice and seeing them in our patients. And the patients, the providers who see non-muscle-invasive bladder cancer very widely in terms of their experience with non-muscle-invasive disease and urologic cancers in general. And so there's quite a wide variation in terms of familiarity with the guidelines and just how often this diagnosis is seen by the person taking care of patients. On top of that, we are increasingly becoming more and more nuanced in our classifications. The guidelines are becoming more complex and they're updated frequently. Even the AUA guidelines, even if the whole guideline's not being updated, we definitely are having rapidly evolving landscapes with new medications. And also fluctuations in old medications like BCG shortages, which of course Dr. Chang knows quite well and has been instrumental in helping us manage. So these are some real challenges in the day-to-day care of these patients.
And so the need we were trying to address here is a scalable point-of-care decision support aid. And hopefully this would be embedded with the tools that we use in clinic regularly, in this case, the electronic health record. So this is our EHR-Integrated Pathway. Like I say, it's integrated because it's built into Epic. And so this is not a separate app or a separate website that one has to go out and seek. And we basically attempted to focus on non-muscle-invasive bladder cancer. So evaluation and staging, low-risk/intermediate-risk, high-risk and very high-risk, and then persistent or recurrent disease. And so this is sort of how we broke it down. And already those of us who take care of these patients regularly can appreciate, there's a lot of information just in these four categories and it's quite complex. So this is an interactive decision-tree model. And so I'll show you some screenshots here in a second, but basically a provider is able to pull up the pathway and then click through certain... We can choose where their patient is. And it will sort of take them down a decision tree about what maybe next best steps are. And this brings up automated documentation into the notes. It brings up bundled order sets, things like staging imaging for high-risk and very high-risk patients. It'll bring up treatment pathways like intravesical therapy options that providers can click through, and education materials that pull into after-visit summaries.
It should be noted that this did not come just from us, the providers or the physicians, I should say. There were 16 stakeholders, which included urologic oncologists, residents, pathologists, APRNs, and patients. And the group was led by an APRN and one of our urologic oncologists. So that's sort of the idea behind it. And I just took a screen grab here. And what you can see, I don't know if you can see my cursor very well, but right under our practice advisories, you could see launch NMIBC pathways that auto-populates for patients who come in with a bladder cancer diagnosis with an NMIBC diagnosis. And when you click that, it sort of launches a page within the chart, and this is just a screen grab of that. And across the top here, you can see that you can choose evaluation and staging or low-risk/intermediate-risk, high-risk, persistent, and recurrent. And then this is just one part that sort of goes over basic staging. And the provider is able to scan through this document, click on yes, no. It'll sort of take you down a decision tree. And some of the nodes, like I say, sort of include order sets. Others at some point, some say refer to a urologic oncologist. For some of these patients, they're really best served by seeing somebody who spends a lot of time thinking and taking care of bladder cancer patients specifically. And so that's an endpoint for some of these nodes as well. This was some of the initial data that we looked at. So 412 pathway uses over this period, October 2021 to May 2024.
This was over 250 unique patients. And then here's just basic demographics about the patients representing the NMIBC population that we expect. And we saw this pathway used in multiple settings for both in the office, procedure suites in the hospital. And as you can see here, this was rolled out not just within our academic hospital, but throughout the community centers that use our shared health system as well. And so you could see that approximately a fourth of the time it was used in the community, and then three quarters of the time it was used here. Almost 70 providers utilized it. And you could see here physicians, nurses, APRNs, pharmacists all accessed this pathway. So it was used across a broad diverse set of provider experiences and expertise. So as far as what we learned from this, and I'm hoping we'll get to dive into this a little bit more as well, but this shows the feasibility of embedding a guideline-directed care pathway into the electronic health record. As those of us who've tried to do things like this that are embedded in Epic will know that this is not an easy lift. And so I really applaud the efforts of the team that put this together. This was adopted by almost 70 providers, diverse backgrounds.
It does allow for rapid dissemination of updated guidelines, rapid dissemination of updates for new agents like we're seeing in the BCG-unresponsive space, and now even in the low-risk and intermediate-risk space for low-grade disease. And as we see updates that happen frequently, like BCG levels change even multiple times in a year, we're able to quickly roll out changes in how we're using BCG. For example, are we using diluted BCG? Are we doing maintenance? These types of things that come up frequently for BCG shortage times. But, and I think this is multiple things to think about here. Overall, we kind of recognized that this is lower uptake than we'd hoped. This is over a three-year period and we had 400 uses, whereas the denominator was much higher than that. There's probably less of a third of the uses. Now, this is not a pathway that we expect to use every encounter with every non-muscle-invasive bladder cancer patient, but it is worth noting that this is a little bit lower uptake than we'd hoped. And the other thing, of course, is that ultimately we'd want to see if the use of this actually was changing outcomes. And this is for me the most important and most difficult question to answer is, does the use of this actually improved outcomes in non-muscle-invasive bladder cancer patients? And I will highlight that although the pathway is embedded and it can be used seamlessly, if somebody really doesn't want to use it, it's fairly easy to avoid, especially if there's maybe some fatigue from multiple similar decision tree options that pop up. Or maybe some alerts that come up that people just sort of ignore so frequently, they've now just have...
Their eyes glaze over by them. And so these are some real challenges with building, incorporating these pathways into Epic, because if they're optional, which I think for numerous reasons they should be, then they're easy to sort of not use and avoid. And so these are all things we're continuing to work on. And with that, I will stop and we can talk through some of the details and so on.
Sam Chang: Fed, great work. I think that's a great start. There are so many challenges with integrating anything into Epic, just as you say. And I think when we initially rolled out Epic, or it has rolled out or people use it, there's Epic overload. And at some point people start turning off this alert, or they just get used to seeing that pop-up and just X'ing that pop-up out because there are so many pop-ups available. And I think you're right to say that, okay, there's multiple facets and multiple levers that we can pull, or hopefully we can push to hopefully encourage use and trying to determine what works and what doesn't. And I think you made some really, really important points. I think just in some general comments, and I know the time to roll out, I love the idea of getting those stakeholders from the get-go, representatives from each of those groups. I think that's fantastic. I think the idea of trying to integrate this into the normal workflow. That's going to be important. And I think right now in that initial stage, that's absolutely critical.
I still think that there probably will be a majority of individuals who do not use it. And I think that's where I struggle a little bit. Is that we still have to reach that individual. We still have to reach that provider. And so I think, and I know you talk about this in the discussion, like having educational sessions, and the question is the electronic health education versus the in-person education versus the marketing campaigns, email blasts. All those different steps, clearly all of them will be helpful and probably help remind people that this is there. I love the idea of someone presenting in a multidisciplinary setting of, here are three interesting cases. Let me show you what the pathway will do. Here, click this, click that, click that. Just as you say, rather than just didactic of saying, this is available, use it. But actually application into some real-world cases. Now, I know there are going to be some challenges. And I think the most obvious challenge is, I think most people, you included, you like to think you understand the guidelines and you feel you're applying the correct guidelines, but then you still have to now go to the pathway because you want to... I think this is an initial step because just as you say, there is Epic overload, Epic fatigue, Epic, I don't want to use the word avoidance because it's a necessity.
But there are definitely times where you want to do the minimal possible to close a chart, get out of that chart. So I think all those are concerns that we have, but at the same time, we are so dependent upon Epic. Tell me, as you see this pathway, as you see that clicking of the button, I think your most recent patient was quite telling. You had seven or eight different pathways to be clicked upon. Can you see in terms of improving quality that you start having a metric of, okay, every patient with a diagnosis of bladder cancer, just number one, determining if that pathway is clicked? Because if you set that up as, look, we need to hit 80% of bladder cancer patients. If not, you're going to get dinged. Tell me how you think we should best... Right now, I know you sent out email blasts and different things of, okay, this is what we have available. Tell me how you think you all are going to next try to increase its usage? And then we'll talk about guidelines, but what about just using that pathway more? What are the strategies you guys are thinking about? Because obviously, look, there's no right or wrong answer, and it's something that we all struggle with when it comes to guidelines. We have these guidelines, nobody uses them. We have these new guidelines, nobody knows about them, et cetera. So tell me what you guys are thinking next.
Fed Ghali: Yeah, like you say, it's a real challenge. I think one thing to say is that I'm not interested in use of the pathway for the sake of use of the pathway. I'm interested in improving ultimately the care that a patient receives on the other side of it. And so we've been thinking about how do you target the alerts and target the pathways when there's high risk or you can detect that there's a deviation from what you'd expect for a patient? In other words, if a provider is comfortable and has already incorporated all of these guidelines into their practice, for me, I'm not interested in just having them click the button so that we get another click. But instead, we're rapidly seeing some of these machine learning and artificial intelligence corporations into Epic and some of our charting and so on. And one can very easily imagine, well, one can easily imagine a world where some of these guidelines are detected by Epic. Or detected by saying, "Hey, it's been over a year since this very high-risk patient has had staging imaging."
Just a little nudge. It's just little alerts and maybe point them back to the pathway that way. The other thing is I think that it's key that the pathway pops up when patients are being seen for their bladder cancer and doesn't pop up when patients are being seen for other things. It's probably not very helpful for their rheumatologist or their pulmonologist to see an alert for their bladder cancer pathway. Whereas if they are seeing a urologist... And so there's a little bit of, again, we're sort of expecting a lot from the electronic health record to set up these alerts. But these are ways we can leverage technology to make this the most high yield, I think, without just having every possible pathway populate on the side. Because I'm as guilty as anybody of having my eyes glaze over as soon as I see a full page of alerts. So these are some of the things we're thinking about. And I just want to highlight just one last thing, which is I kind of mentioned it at the end, but at the end of the day, I am interested in the question of is use of this actually improving guideline-based care for patients? And you know as well as anybody, that guideline-based care is very difficult to detect and determine. And I don't want to be prescriptive to physicians who maybe have good reason to deviate. Guidelines are just that.
Sam Chang: Are guidelines. Exactly. Yeah.
Fed Ghali: And so these things make simple narratives about this more complex. But we've been thinking hard about the right markers to use for that to detect guideline-based care and really significant deviations. And so that's the next era of this, is trying to incorporate some of these maybe more advanced technologies in EHR, as well as picking up some of these endpoint quality metrics.
Sam Chang: Yeah, I think your point regarding ultimately the outcomes and improved care, obviously that is the outcome that we all want. That's the end goal. The issue is all these steps along the way, how do we measure that we've actually gotten to that point? Your point about not just counting clicks, I think is a really important one. What I don't know, what I don't struggle is, we've got this varying level of education, usage, familiarity with the guidelines. Where just the initial evaluation may be important for the majority of individuals, not so much for someone like you. On the flip side, just clicking, you're right, exactly, it doesn't equal quality. It's just another click.
Fed Ghali: That's right.
Sam Chang: So I personally struggle just as you say of, do you understand what the guidelines are? Do you actually then think about that for that particular patient and then what do you incorporate in the care? And it's so multifactorial. And I also worry about overtreatment just as much as undertreatment. Of, oh, you've not clicked, you've actually over-treating patients that shouldn't get this. Or utilize BCG when we're in a critical short supply. So it's a lot to consider. But as you look at different ways to measure, what is your next step now? I mean, because ultimately, obviously improvement in care and outcomes, and we hope actually that the guidelines would actually improve that. And we have an idea that the risk stratification is effective, and we have an idea that some of the outcomes that we've talked about are hopefully helpful for patients. But tell me what you guys are going to do next in terms of the EHR.
Fed Ghali: Well, one part of it is we've been working to just remind folks within our department and have opportunities to remind folks that this is available as a resource. So it's clear that not as many people know about it as we would've hoped. So one is just simple, let people know about it. The other thing is we've been working with our data analysis team that's sort of a central hospital data analysis team to see to basically pull routine, attempts at routine metrics. So common ones are time from diagnosis of hematuria to urology referral. Another one we've been very interested in is in imaging. I think that we probably under-utilize cross-sectional imaging in high-risk NMIBC patients. And again, I don't need to tell you, but these are high-risk patients and this is something that is an important part of really taking care of them. And so getting access to information about how frequently scans are done, each one of these is a whole day's meeting with Epic to try to figure out. But these are the types of things we're trying to work on now.
Sam Chang: Yeah. I think hugely important because that ability for us to recognize and then to document and then to see ultimately what ends up happening, I think is really, really important. And I will piggyback on your statement regarding the lack of use of imaging for our higher-risk patients. I think decades ago, I think historically it was really, really something that was underutilized and in fact not utilized. And so understanding that risk and that importance of evaluation is just another example of how hopefully we can, with some of the work that you and your colleagues are doing, will help us better identify these are areas of need and what we need to do next. So, Dr. Ghali, thank you again so much. And as always, we look forward to seeing you again, and I'm sure you'll have many, many more contributions that we'll be talking about.
Fed Ghali: Thanks so much. It's such a treat to chat with you, so I really appreciate it.