And I think that my task today is both to really talk about what constitutes value, what constitutes value in healthcare, and how do we leverage that for equity for the various patient populations that we serve? And I like to think about value first and just in terms of our own experiences as consumers and what products do you find most valuable? And one I like to highlight from the pandemic was a bit of a surprise to me and that's Mario Kart from Nintendo. Now, I have no stock in Nintendo, but, and I really wasn't a gamer growing up, maybe the wrong generation, but with kids living overseas during the pandemic, we began really getting together on Sunday afternoons to play Mario Kart over the internet. And I've been impressed by this. This has been around for many decades, lots and lots of copies sold, adapted over time to appeal to all age groups, such as myself. And when you think about the value equation for Mario Kart, value meaning quality and experience relative to costs, I was impressed, amazing graphics, clever race courses. And in fact, you can make a little icon of yourself as my daughter did of me apparently. Hopefully that doesn't look too much like me, but suffice it to say that's a unique feature in terms of quality. But experience is where I think some of these tools differentiate, the ability to accommodate for limited skills, such as myself, ability to play with friends and family online from around the world. These are really value added experiences that I think those kinds of things equate to really delivering equitable care as we'll come to in a few minutes. And it's all about cost. And in this case, the cost of the device and the game didn't seem prohibitive relative to the benefits that it delivers. And if we translate those principles to healthcare, I'm actually very humbled and privileged to be on the panel today with Dr. Pam Johnson, who's published much more extensively in this space than I have. And this paper she wrote with Yoshimi Anzai from 2021, I think was a call to action for us to really move from being stewards of the imaging resources that we are responsible for to really true leaders in which we are members of care coordination teams to really be true participants at the center of medical care delivery. And that's what really, I think at the end of the day is what delivering value-based care is all about. And if we wanna think further about what exactly is value-based radiology, I often like to think that sometimes it's easier to think in the converse and think about, well, let's, we know we can talk about promoting high value imaging in a minute, but let's just talk about eliminating low value imaging as a starting point, because perhaps that's the lowest hanging fruit for us. And I like to think in terms of the principles espoused by Oakes and Radomsky and JAMA from a couple of years ago, in which they focused on the fact that low value care, whether it be imaging or otherwise, really is care that in which the harms or costs really outweigh the benefits. And so often it's based on local lorries or practices that have been introduced into local medical cultures, if you will. And the challenge is really how to guide de-adoption efforts, how to get a group of practitioners to stop doing low value things. And so often it's based on motivations that come from relationships with patients who may be demanding low value things because of the individual perception of what constitutes good medical care. It's really these cognitive biases and attitudes that have to be overcome. And I know, for example, when I have low back pain, I may know all the guidance in the world, but if it's my back that's hurting, I really want that MRI to show me what's going on, even though it may be of low value. And so tools that we have available to us to help guide de-adoption really occur both before and after imaging. And so before imaging, it's really about guiding practitioners to the most appropriate imaging exam for their patients. Hopefully, as in my example, not recommending a lumbar MRI for every twinge of back pain, for example. And there are many tools available out there, but one that I think is perhaps best known is based on the ACR appropriateness criteria, ACR Select, which has been spread across the globe to, at least in Europe, to ESRI Guide. And when combined with guidance from other organizations, such as the National Comprehensive Cancer Network, American College of Cardiology, and Society of Nuclear Medicine and Molecular Imaging, really fairly comprehensive guidance can be developed for high-cost imaging, over 12,000 rules, for example, that can help guide practitioners to the appropriate exam. And I won't go into further detail about this today, other than to say that if we use tools such as this, we can often cut the incidence of low-value imaging, so-called red zone imaging, if you will, in half. And various experiences from around the world, both, for example, in Barcelona, as well as at University of Virginia, out-of-the-box use of tools like this can result in a red rate, if you will, or operating in an inappropriate zone of about 5%. At MGH, having used tools like this for many years, we've been able to see a red zone ordering rate down in the 1% to 2% range, and clearly a real benefit to guiding de-adoption of low-value imaging. After imaging, we can also use tools and decision-support tools to help guide us to the appropriate recommendations that we make consequent to the findings that we detect. And I often think about experiences I had in clinical practice, in which, for example, a liver surgeon might come into our reading room and announce, or kind of complain, why is it that you all are recommending something different for that hyper-enhancing one-centimeter liver lesion? One person wants the biopsy, one person wants the follow-up, one person wants an MRI, and so forth. And really honing in on what are the most appropriate recommendations can sometimes be challenging for a group of radiologists. And here, I'm very fond of tools that Dr. Terry Alcazar has developed and popularized, in which guidance is provided to us at the point of care where we're reporting cases. And here's an example of where, if we might say there's an intrapartic lesion, natural language processing recognizes that we have guidance available for a liver lesion on CT. And one can call up a guidance window to help select the features of that lesion to standardize not just the language for reporting it, but most importantly, the recommendation we make consequent to that finding. And I think by standardizing and making that best practice guidance available to us, we can ensure that we're eliminating the low-value recommendations we might make in my example, again, of maybe recommending a lumbar spine MRI for inconsequential symptomatology. And of course, those findings and recommendations can be imported directly into the radiology report. The use of such a tool has been shown to really promote best practice guidance in our reporting. And this paper from Michael Liu from some time ago showed that when we used point of care decision support for reporting the guidelines, for example, for what to recommend for incidental lung nodules on abdominal CTs, increased from about 40 to 50% concordance with best practice guidance to over 95% concordance, highlighting the value of making that guidance available to us as we report our cases. Let's shift now to not just eliminating low-value imaging, but how do we really maximize high-value imaging? And let's return to that value equation again. And this time, think about it in terms of quality, the quality of the reporting that we do in terms of accuracy of our diagnoses or the precision of our measurements, combined with the experience that we deliver, being timely service, convenient service, particularly for socioeconomically disadvantaged patients, for example, integrated care to make sure that patients don't fall through care coordination cracks, if you will. And all doing that at an affordable cost really maximizes the value of care we deliver. And it's really in the experience and cost side that we can really ensure that we're serving the needs for health equity best by making sure that we're availing our high quality, if you will, to everyone who is in need of it at a cost that they can afford. Let's start on the quality side of the equation. And there's a whole host of things I could talk about and what we can do to improve accuracy and precision and so forth, but I'll just focus very briefly on what automation or semi-automation as perhaps provided through use of AI tools might hold for us in that regard. And my first list is quite a laundry list. And I thank Jeff Rubin for in a dialogue he and I were having in my email a month or two ago about what AI might mean. And I thought his email was so good, I put it in a slide, but basically there's a whole host of ranges, areas in which we can leverage automation or semi-automation to help us improve the accuracy and precision of our reporting. And even the disposition, leveraging the medical record to compute diagnostic probabilities for various treatment decisions based on our imaging results, they really terrifically add value in the disposition consequent to not just the findings, but also the recommendations we make. But perhaps focusing a bit first on just a bit more pedestrian opportunities with regard to just how we report things, we know that there's a heck of a lot more information in the images that we review than we often capture in our reporting. Simply, but sometimes from pure exhaustion of the detail level. I remember reading out with one of my teachers decades ago who after we got to two or three findings, would just say, that's enough. Meaning there's, you know, there's like as if there was an upper limit to what we could physically report. And I think that with semi-automation, perhaps from AI, we can take sort of traditional text-based reporting, which is awfully detailed. And if we leverage AI to help us just hone in on whether there's moderate or severe stenosis in the right frame and the central canal or the left frame and in a lumbar spine MRI, and compute that for our review and approval, encoded not only in very detailed texts, but even in a more tabular fashion so that the referring physician can see very clearly where is the problem in this patient's lumbar spine and even overlaying a set of images that correspond to each of these findings can let the referring physician quickly correlate our reporting to images that may correspond to them. And then even further overlaying arrows that might indicate disease progression in certain areas. And so I think, again, we can do a host of things to improve quality. And the presumption of course, is that as we do that, that'll be uniformly applied to all of our populations that we serve, whether they be from socioeconomically disadvantaged areas to the more affluent societies that we serve. I will say though, that our ability to really serve and deliver in the health equity promise is really on the experience side of the equation. And I'll highlight sort of four areas, if you will, in which I think that we can really work hard to make sure that the services we provide, the expertise we provide, the quality we deliver benefits all elements of our society. And we'll talk about patient engagement, virtual care, ambulatory access, and care coordination. And with regard to patient engagement, let's start first with just reducing missed care opportunities and improving compliance with best practice screening recommendations. And here I'm gonna feature a lot of the work from our moderator, Dr. Efren Flores, who's in our MGH department and has done just a terrific job of really drilling down things that can be done to help with, make sure that we're delivering care equitably. One of Efren's earliest projects was to really just look at missed care opportunities, what were formerly somewhat pejoratively called no-shows, we've renamed as missed imaging care opportunities, because most of the time these are not, these are often due to no fault of the patient. Various factors often lead to their inability to make their imaging appointments, whether it be based on transportation or other exigencies. And Efren looked at unemployment data and household income across the Boston metropolitan region, and then developed various prediction tools that would help predict which patients might miss appointments. And ultimately developed an intervention for those patients in the form of a transportation assistance program, which we've introduced that helps patients get to their imaging appointments if they have otherwise no means to get there. And this has made a big difference for serving equitably the needs of our underserved populations, and big credit to Efren for both the background work to determine who might benefit from this program and the intervention to help benefit those patients. Another important project that highlight the need to really focus on these communities is in lung cancer screening. And I'm proud to show over the past several years how our lung cancer screening volume has really, had really increased over time. And it was thanks to many initiatives, whether it be a marketing and practice support outreach. But some of the more equitable initiatives I thought were important, I'll highlight here. And again, I believe Efren had a lot to do with these. First was just producing infographics that patients could easily understand with the eligibility criteria, both in English, but also in other languages. And I still am always proud to see these in our waiting rooms as I did just very recently. And not just to make sure that patients understand regardless of their native language, but also to provide access for patients who might otherwise not have it. And so I'm very proud of our same day lung cancer screening program, which is targeted, available to the Boston Health Care for the Homeless Program at Mass General, as well as one of the underserved community health centers that we practice in. And patients in these settings who may look at our brochure, see that they qualify, can simply declare that they would like to participate in lung cancer screening and we will accommodate them on the same day. Very helpful for vulnerable patients. Let's turn now to virtual care. And here, this is about really enabling virtual consults and e-consults for patients who may be in their doctor's office, for example. And rather than necessarily having to defer the patients to a later visit or another time to really review in detail the findings on their imaging studies, we've enabled physicians and their patients to review studies with radiologists virtually and in real time. And a paper by Dania Day that reviewed our progress with this showed, I think, important statistics which show, of course, the popularity of the program, but perhaps the most important is that the number of patients who would try to read the report themselves really plummeted to zero after being provided with this opportunity. The point being that from a technical equity or tech equity, patients may not always have access to the tools that would allow them to access their radiology reports through patient gateways and so forth. And at least providing this kind of a tool can enable patients and their physicians to review imaging studies together and in real time without necessarily having to defer to technologies that may not be available to the patient subsequent to their visit. Ambulatory access is another important principle that I think really helps to address health equity. And this can be either through increasing throughput and decreasing visits to the ER, for example, for imaging, or simply just decreasing visits to the ER, improving, decreasing length of stay in the ER to enable more patients who need to get to the ER, whether it be imaging or for other purposes, to at least have the capacity there to do so. And I want to highlight a rather technically complex project, but one that we've been so impressed with that was led by our medical analytics group, Oleg Pianik and Steven Gietron, who did a very elaborate AI-based, operations AI project to really look at length of stay in our emergency room. They did a lot of data collection by actually walking through the ED, collecting a lot of data, doing a lot of data mining from our data warehouses and discussions with ED personnel to develop a variety of machine learning models that really determine length of stay in the emergency room. And they had a variety of models that were created, no different than predicting hurricanes in the ocean, the European model, the American model, and so forth. And here, I'm not showing you what each of these are, but suffice it to say that they combined four different AI models into a single predictor of length of stay in a very simple to understand a graphic such that in this variable, which is not shown on this graph, but whatever variable this is, the highlight being that between low to high values, it has about a 72-minute impact on length of stay over the range of patients that are served. And so, for example, if you look at arrival time of day, the AI model shows that the time of day from midnight to 11 p.m. has an impact of about 42 minutes, the AI model being much easier to understand than the data snowstorm and the raw data space and the correlation coefficient not really showing that impact at all. So how AI can really help us operationally to understand a length of stay. And Oleg and his team did this for a whole host of different factors that might drive length of stay in an emergency room, only a few of which have to do with imaging, numbers of pending CT and MR exams, for example. And by combining the number of patients that get those studies or are affected by that variable times the impact that I showed, one can come up with an impact score to really help direct us to what are the most important variables that we might tackle to try and improve length of stay in the ER to increase access for all patients, including those vulnerable patients that need to be seen in the emergency room, either for imaging or for other purposes. And next I'll talk briefly about one last initiative I'm extremely pleased and proud of, which is our Care Coordination Initiative, which is really integrating with the clinical care team to ensure seamless end-to-end imaging care. And this is a program developed at the Brigham and Women's Hospital, Nina Kapoor, Sonali Desai, and mentored by Ramin Khorasani, developed the Addressing Radiologist Recommendation Collaboratively, a program which ensures that clinically necessary diagnostic follow-up recommendations are executed, modified, or dismissed by the ordering provider. This is a phenomenal program that I give a huge credit for because it really plugs the holes or seals the cracks through which patients sometimes fall, even when we make good diagnoses and make good recommendations, that the care, that handoff doesn't occur and the patients don't get the follow-up imaging or procedure that's necessary. This does require some dedicated unique eye tools and collaboration among multiple care teams. Again, the goal is timely performance of the necessary follow-up that we recommend through collaborative care plans. The program creates a collaborative care plan between the radiologist and the ordering provider, but it really leverages three teams to ensure that it works. The first is a radiology care coordination team, which are really about schedulers in our department that assist with ordering and scheduling of follow-up recommendations. If an ordering provider concurs, then we do the scheduling automatically for those recommendations. There's a safety net team for patients that fall outside of our network or in whom their doctor is unknown, and there's an operations team that tracks all the data and makes sure all the loops are closed. The tool looks something like this, where radiologists can bring up the tool, indicate that a follow-up is necessary, indicate the exam that's going to be followed, the timeframe and other commentary as to why. That message is passed to the ordering provider both by email and in their Epic in-basket, and they are asked to either agree with the recommendations, modify them, or transfer the care to another provider. If they agree, then basically we will automatically schedule those patients for the follow-up exam. It's a terrific way to make sure that patients that might otherwise be lost to follow-up from vulnerable communities are not, to really help seal those cracks through which they might fall. Last, I want to speak briefly about cost, because cost also is a very important feature. It's the third variable in that value equation, and I'll say in general that quality and experience are more universal than cost, particularly as we start looking beyond our country. When I talk about this in other countries, of course, this becomes extremely challenging in that cost is really, like politics, very much local. I'm very fond of this paper from Yoshimi Anzai and her team, Dr. Peckman and colleagues, that showed that just by simply shifting a care from a hospital-based practice to a community-based practice can really dramatically lower the cost, as in the case for these spine interventional procedures. It's really an important principle that we've been espousing at Mass General Brigham. This paper from the Harvard Business Review, I think, highlights the fact that when you start functioning like a health system, as opposed to a collection of individual assets, you can really ensure that patients and their families have seamless and consistent experience across the system with the intent to affirmatively control the delivered cost of care through a variety of tools, including some that I've already spoken about. At the end of the day, this is about really still ensuring that we provide local services, but widening our expectations to be more universal, such that we give a uniform quality of care with consistent patient experiences, but do it at a lower cost based on site of service so that we can better support our vulnerable communities. And in our case, this required a lot of work that was led, in our part, by Paripindar Pandey, who's now a chair at Ohio State, but she, as our associate chair the past few years, did a terrific job of helping us move in this direction by first defining some specialized practice models on the quality side to enable enterprise-wide scheduling for maximizing patient experience. She did this by forming a host of task forces across all subspecialties in radiology to define what exams really required certain subspecialty expertise and the qualifications for all practitioners across our enterprise, including some of our very peripheral community hospitals to be delivering the same level of some specialized care regardless of location. Some of these hospitals are up to two hours away or even out on the islands of the Atlantic Ocean in the form of Nantucket and Martha's Vineyard. And once she did this and we adopted a uniform approach to quality, then we were able to turn on scheduling such that patients could go anywhere in our system and get the same level of care and quality. And of course, the benefit being that the strategy is that if we move secondary care to community hospitals and ambulatory care settings, we can affirmatively lower that cost of care based on the site of service. And so effectively, we've looked at the value equation for many respects, quality, of course, being all about maximizing the accuracy of our diagnoses and the precision of our measurements, but really the ability to leverage value for all communities we serve, including those that are more disadvantaged, comes into play mostly from the experience side, making sure that we're delivering timely and convenient services, access of our care, integrating our care with those of their care teams, and doing it at an affordable cost. And if we get all this right, then hopefully our services are really truly priceless. Thank you very much for your attention. Thank you so much, Dr. Brink, for this excellent discussion on what constitutes value in healthcare, especially the Mario Kart analogy, also a huge hit in the Spoluto household, Spoluto-Brink-Mario Kart. Next, we'll hear from Dr. Pamela Johnson. Dr. Johnson is the Vice Chair of Quality and Safety for Johns Hopkins Radiology and the Vice President of Care Transformation for Johns Hopkins Health System. In this role, she directs initiatives to improve the efficacy, efficiency, consistency, and affordability of care. Dr. Johnson will discuss how technology can improve and advance patient care and equity while keeping costs affordable for patients and healthcare systems. Welcome, Dr. Johnson. Sorry, thank you. I had to unmute. Thank you so much for inviting me. Thank you for the opportunity to present here today and to follow Dr. Brink, who has inspired me since the beginning of my career in radiology through his research and innovation, and for the opportunity to speak to all of you today about patient-centered, high-value care. My disclosure relates to potential licensure of intellectual property. The objectives of my talk are to broaden the value lens from a single medical decision-making about appropriateness of a test or treatment to thinking in terms of horizontal outcomes across the population and longitudinal outcomes, the downstream resource utilization and downstream patient experience and outcomes that we play a role in in radiology, to discuss some of the opportunities to improve equitable delivery of high-value care, which begins by collaborating with our colleagues in other specialties, and then to describe how structured report impressions and IT platforms can assist us by improving the reliability of the systems that we have in place for early diagnosis and risk stratification of patients. So I became involved in this because of the understanding that, excuse me one second, that healthcare debt has broken our promise to do no harm. It's the leading cause of bankruptcy, home foreclosure in the United States. This affects both insured and uninsured patients, but uninsured patients to an even greater extent can wipe out their personal savings, and it's driving care avoidance where patients are not attending to screening or recommended treatments and procedures because of their concerns about the cost. It's important to understand what are the leading causes of contributors to patients' debt in addressing how we improve the value, and these are emergency department visits, hospital admissions, dental care, and then imaging tests like MRI. So this is one of the reasons why I became involved in it, knowing that every unnecessary imaging test is potentially creating a financial burden for a patient, and we have a very patient-centered definition of value, which is improving clinical effectiveness and efficiency, patient and provider experience, and patient outcomes, and the numerator has expanded over the years as we've been building on the portfolio of work that we're doing, but the denominator has not changed, and that is reducing the personal and financial costs for patients because it's not just the cost of these tests and the financial burden, it's the potential loss of days worked, the effect on a family, the economic burden on the entire family that can result from low-value care. So interesting paper published in JAMA detailed the waste and expenditure in the United States that may exceed $900 billion a year and the different elements that contribute to this three of which are really physician accountability. So it's failure of care delivery, failure of care coordination, and low-value care. And in radiology in particular, even though we are not the primary care providers, we play an important role in care delivery and avoiding the failure of care delivery for evidence-based cancer and cardiovascular disease, screening and surveillance, and in stewardship to reduce unnecessary imaging and all of the potential downstream resource utilization that follows the imaging care that we deliver. So I'd like to discuss two pillars of value in population health improvement that we can address in radiology where we have the tools in place already, and a lot of this pertains to performance improvement in terms of our interpretation, reporting, and building high-reliability systems to ensure follow-up like Dr. Brink described that was done by Ramin Khorasani, really remarkable work. And so I'm going to focus primarily on cardiovascular disease and cancer, but there are other opportunities as well in terms of medication overuse where we play a role that we may not recognize, as well as the overuse of consults and procedures where we have an opportunity to improve the longitudinal value for patients. So beginning with coronary artery disease, coronary CTA is one of the exams that has recently gone from only the European best practice guidelines to now in the U.S. best practice guidelines because it's been shown in many robust trials to improve longitudinal outcomes both in stable chest pain and in acute chest pain. So in the stable chest pain patients, there is evidence of five-year reductions in mortality, reductions in major cardiac events, higher quality of life, all at a lower cost because coronary artery disease is diagnosed earlier and medical management is initiated before patients develop obstructive disease. In the emergency department, it shortens the ED length of stay. Coronary CTA, it's really the primary driver of the chest pain guidelines that should be in all ERs at this point because we are able to help patients safely be discharged at Hopkins. We've reduced one-day admissions for patients with chest pain, but there's also evidence showing that downstream there's more appropriate PCI and reduction in major cardiac events. So a very valuable test in our armamentarium. More payers are starting to reimburse for this. We were even able to convince our own health plan who was not reimbursing that this is a really critical test for patients, much better than treadmill testing, and our current guidelines recommend that we have a chest pain guideline in every ambulatory and ED setting, and that we're using the high-sensitivity troponins in there as well, so these may need some modification at this point. But it's just an important tool. And if we focus on this, if we improve our ability to diagnose coronary artery disease early and begin medical management, we can reduce the incidence of events. We can reduce coronary catheterization, PCI, and reduce mortality. And so this work, though, requires partnership with payers, which is another important element to this, and we've been doing that with our own payers and with some of the larger health plans so that we can understand what happens to the patients if they're outside of our own health system. So I'd like to show a few publications, evidence from a few publications about coronary artery calcification and the importance of diagnosing, of identifying coronary calcification, both through the scoring studies that we do, but also as an incidental finding. So just one of many studies showing that coronary calcification from a gated chest CT improves outcomes in terms of decreasing lipids and by driving increased statin use. And so this is a well-established, but not every patient is sent for a coronary calcium score. They may not have access to robust screening services. And so we have an opportunity on every non-gated chest CT that we do to make a diagnosis of coronary calcification. I tell this to the residents all the time, but they still, we don't have uniform practice where we put it in the impression of a report. So here's some evidence that supports doing that. Beginning with this study that showed that they looked at a large number of non-gated chest CTs retrospectively, looked at the baseline risk factors and the outcomes, and they found that the presence of coronary artery calcification on a chest CT for another reason was associated with a rate of myocardial infarction and rebascularization. But surprisingly, 63% of the patients had not been diagnosed with coronary artery disease. So this is our opportunity to just, I know we see it so frequently and it's ubiquitous, but these patients may not have been diagnosed. They may not be on aspirin. They may not be on a statin. There's evidence that incidental coronary artery disease increases the rate of major adverse cardiac events or it's associated with it. In this retrospective review, it was shown that patients with coronary calcification had a higher incidence of MACE, and many of these occurred within 18 months. So again, our opportunity to improve the health of all patients that come through for imaging, both for myocardial and other cardiovascular outcomes. Identifying coronary artery calcification incidentally increases the administration of medications, which we know improve outcomes. So what we're doing in our own practice is we have this in the lung cancer screening template that we have a structured coronary calcification grading system that can be later mined to evaluate outcomes. We are going to advance this beyond just the lung cancer screening to all of our chest CTs so that we can improve the diagnosis for these patients. Another opportunity is for abdominal aortic aneurysm surveillance. This was great work that was done by radiology partners, a research institute, where they improved the recommendation for abdominal aortic aneurysm surveillance based on the Society of Vascular Surgery guidelines by putting these structured reports in their impression. And they did some mathematical modeling and predicted that if all of the patients that they recommended follow-up on were engaged in the follow-up, that they would potentially avert 250 ruptures and the associated poor outcomes related to this. So here's an opportunity where we in radiology can really serve a large number of patients. The lack of surveillance is the leading contributor to aneurysm rupture in these patients. As Dr. Brink pointed out, there are gaps in delivery of the downstream imaging that we need for patients, and so we built high reliability systems. The system that they have is great because it goes beyond Epic. In our own health system, we have a large radiology practice with hospitals and imaging centers, and we also have a very large community practice. So what we did was to build a system in Epic where our surveillance recommendations are coded by Epic into follow-up recommendations and then filtered to a report for each of the different practices that we serve. And so all the imaging findings, all the incidental findings then go to a report that can be managed by the office administrator, who can see what the follow-up exam is and what the due date is. And the patient does not fall off of the list until the examination is done. We've rolled this out for lung nodules and abdominal aortic aneurysm surveillance and thyroid nodules, and next will be lung cancer screening. I think the advantage of this is that the decision-making is made by the physician, and it can be made at a time that they have the time to focus on, you know, do I want to take care of this incidental finding? But there's a backup where the administrator has a list of all the patients in that office with all the incidental findings that they can filter by date and look at at any time. In the process of building this, we learned that there are actually challenges for providers when they're trying to order diagnostic mammograms from screening mammograms, that it's just not as easy as you would think. And if the patient has ordered through MyChart, which they do, then we need an order for the diagnostic study. So facilitating the ability to make that flow helps ensure that these patients get the imaging that they need, and that the barriers that sometimes the electronic medical record creates don't contribute to failure of care delivery. I just want to just ensure that this is not blocking the screen. Okay, so for early cancer diagnosis, we want to think about our evidence-based screening algorithms and then recognize that we have opportunities for opportunistic screening. We do 80 million abdominal CT scans a year, and we can make the diagnosis of transitional cell cancer, pancreatic cancer, renal cell, gallbladder, colon cancer before the diagnoses are…the patients are symptomatic. And so, each imaging test that we do in this setting is an opportunity for opportunistic screening. The Lung RADS is a very important program, but the real-world evidence shows that the follow-up adherence rate is much lower than it was in the trials. It's only about 50 to 65 percent, but the use of a reminder has been shown to be a helpful tool by reminding physicians that the patients have to come back, because after the first lung cancer screening, if they have a low score, a 1 or a 2, the likelihood that they're going to come back is very much lower than if they have a higher score. So, that is an area where we need to improve our ability. We can use IT tools, but that's not the only solution. We have to have better community engagement, and we have a tobacco cessation clinic where the patients can come in. They can get the lung cancer screening that day. They may be…you know, they can talk to the physician to reduce the anxiety. And so, we need to have a number of different resources in place. An area where we're driving over diagnosis of cancer, thyroid nodules, important to have incorporated the ACR and ATA recommendations for follow-up when they're seen on CT and MRI, and also the TIRADS or other scoring systems. And when we did that at Hopkins, when the adherence rate for radiologists on CT and MRI hit 90 percent, we were able to reduce unnecessary thyroid ultrasound and biopsy. And then, the residents did a project where they mapped out all of the incidental nodules that went to surgery, and we were able to demonstrate that there was a reduction in the number…the percentage of benign thyroidectomies for benign incidental nodules, which is one of the concerns about us with these nodules, because sometimes the biopsies are indeterminate and the patient goes to surgery unnecessarily. Other work that's being done by Hannah Zafar, as one example, encoding the likelihood of a cancer is going to be very helpful in reducing over-imaging to prove benignity of different lesions, and something that we can use to build data from. What other areas are we contributing to overuse that we may not be realizing? One is antibiotic overuse, and this is a paper that was published at Hopkins that showed inpatients who received antibiotics, 20 percent of them had an adverse event, and 97 percent of these were significant. So, it wasn't just a rash or hives. Some of them had to be readmitted after discharge. Some of them had to be…had a prolonged hospitalization, additional testing. And so, when they looked back even further, they discovered that almost 20 percent of the antibiotics that were administered were not clinically indicated, and those were also driving adverse reactions. So, looking at the scenarios where this happens, now you may start to see where we're playing a role. So, asymptomatic bacteria and non-infectious lower respiratory tract conditions. When one of my…when I was the program director, one of the residents first year said, do you know that every time you say perinephric stranding, we assume it's pyelonephritis and give antibiotics? And I think that this is something that there's a big divide between understanding…between the clinician understanding what we mean and how we report, and gaining insight into how our reports drive care is a really important area where we need to be doing some more research. So, we did our own study with that perinephric stranding. It turned out that it contributed to overdiagnosis of urinary tract infection, of the clinical urinary tract infection, and overuse of antibiotics. We have other work at Hopkins that's shown that when we have ambiguous reports on chest CTs and x-rays, they…that we drive…that we contribute to overadministration of antibiotics. And this is not just adalectasis versus infiltrate. This is also…if we say things like…it's more the descriptive findings, like bronchiolitis that could be infectious or inflammatory. When we put those types of terms in our reports, the receiving physicians often assume that that is a bacterial infection that requires antibiotics. So, another opportunity to improve the quality of care we're delivering is to improve our reporting accuracy and help our clinicians understand what we mean. This is another clinical area where this has become apparent is in the MSK service line, so back pain, shoulder pain, hip pain, knee pain. We're doing a lot of work with our orthopedic surgeons, beginning with appropriate selection for imaging, and then…but there are other areas where we…as radiologists, and that's on the quality of the imaging and helping the ordering physician understand the result. And by doing that, we can draw…we can reduce overuse of opioids, consults, surgeries, and improve the longitudinal functional outcomes. So, appropriate selection, number one, and that's…there's evidence that early spine MRI, not only does it drive up costs, it actually contributes to an increase in back surgery and opioid use. More reasons why we need to be stewards of imaging and reduce low-value imaging, higher costs, and higher final pain score. So, these are areas where, you know, we don't really have that data for us, so the research that's coming out is really helpful to guide us in practice. And so, one area that there is more data evolving is the role…the importance of MR quality in terms of protocols, scanners, interpretations, diagnostic confidence, and report wording, as well as helping the ordering physicians understand. So, one study where a patient was sent for 10 MRIs within a short period of time, and then they were reviewed by experts. They were able to show that there was a very high error rate, both false negatives and false positives, and poor agreement, so opportunities for improvement. A study in the UK showed that the general practitioner receiving the report, 70% of the time, does not understand the findings and over…overperceived the significance, so…and sent patients for consults with specialists that they really didn't need because these are common findings related to age. So, there's work being done to add phrases that help…that help our providers understand that these findings are common for age, and some studies being done looking at the downstream care, one study showed a small decrease in opioid prescriptions. These are ways that we can contribute to a reduction in overuse of procedures, consults, and…and opioid use. Another study showed that better reporting resulted in lower referrals, lower repeat imaging, with no effect on narcotics. So, early work, more work to do, but great opportunities. So, how can we do this? What can we do going forward to investigate and to reduce overuse of the resources downstream from the care that we deliver? Well, we need to investigate which terminologies that we're using that are driving overuse for patients and…and contributing to low-value care, and what can we do to mitigate that? What type of structured impressions can we use? And the use of…of impressions and terminology and reports where we can code this is going to help us to build the evidence that…as we start to measure longitudinally as to what is effective and what is not effective. Again, partnering with your referring physicians on this work is critical, as well as the…the payers who have longitudinal outcomes and also have an interest in delivering better care at a lower cost. So, in concluding, high-value care really hinges on accurate diagnosis. It…imaging stewardship from appropriate ordering, but also from the…from the recommendations that we make, and stewardship for resource use longitudinally after we've served a role in the patient's care. We have a lot of opportunity to serve patients who may not have access to screening services with…by helping diagnose cardiovascular disease and cancer earlier in our imaging, and especially in CT imaging, and building out the high-reliability resources so that the patients don't slip through the cracks are…are also very important. So, with that, I will thank you for your time and happy to take any questions.

healthcare value

radiology equity

high-value imaging

value-based care

AI in healthcare

patient engagement

cost management

health equity