Thank you very much. I'm thrilled to start the session with the topic I'm very passionate about. I chose this framework, which is very simple, the four C's of patient-centered care, and they are culture, care, communication, and collaboration. And this presentation is by no means comprehensive in what we can do around radiology reporting to contribute to patient-centered care, but I do want to leave you at least with a few thoughts you can ponder when you go home. So let's start with the culture. We're currently still, at least in the U.S., practicing in an environment where paternalism informs the way that we interact with patients. And so what that means is that the mindset in general is that we tell patients what to do, and we expect patients to just take our recommendations and follow them. And you may wonder how that applies to radiology, because certainly that is a concept or behavior that prevails between clinical providers and patients, but I do believe this also has permeated our radiology practice. So for example, you cannot see your report before your doctor has. In the U.S., we have decided at some point that we're withholding reports from patient's view until the provider had sufficient time to review them first and decide what actions they want to take. And that may have been appropriate at the time when this was implemented, but nowadays there are more and more patients who want to see their reports right away, and we have to think about whether we have the right to withhold these reports from them. Talk to your doctor to explain the report. So we're writing reports that are highly technical and full of medical jargon, and we have no interest whatsoever to make these reports accessible to patients. What we believe is if they do want to know more about the report, they can talk to their clinicians. And here's something else I observed, and this comes from conversation with my colleagues. So it's not worth mentioning XYZ in the report. So I am, by training, a neuroradiologist, and we do a lot of brain imaging. We see these nonspecific gliotic foci in the brain, and there's kind of two schools of neuroradiologists. One of them always diligently report these findings, and other people say it's not worth mentioning. There's a broad differential for what causes this clinically irrelevant. But I do have a story of a friend who was a smoker and had a brain MRI for headaches, and they found these lesions, and she asked me, could this somehow be related to the smoking? And I said, sure, maybe. And as a result, she quit smoking. So there was actually an effect on an action taken by the patient on a finding that some of my colleagues deemed not worth reporting. So if we think about the system, if we're not supposed to practice paternalism, what are we supposed to do? And certainly, sessions like this and the title, it's patient-centered care, patient-centered reporting, that's a step towards the correct goal. But what we really should be working towards is viewing patients as a partner in care, where we support the patient along the way of their experience with the healthcare system and provide them with the tools and information they need to make their healthcare decisions. And this may seem like a joke. In this cartoon, the patient says, I already diagnosed myself on the internet. I'm only here for a second opinion. It seems funny. But when you do talk to clinicians and patients alike, you will find that this is very true today, and it's an increasing behavior we see in practice. Moving on to care, there's certainly a lot of overlap with care and communication, because the way we care for patients is through our report communication, but I have kind of divided these topics up with some different thoughts in these areas. So, for example, if we think of our care as being timely, of course, we care about turnaround times, and raise your hand if that is a performance metric at your institution. Okay. If you're from a foreign country, raise your hand if that is a metric at your institution. Okay. So, it's outside the U.S. that's also practiced. So, that's good that we care about that, but there's recently some wave, and this is a report from Belgium, and I've heard this from my U.S. colleagues as well, is when this thought about clinician turnaround time. Why should I be reading this study on Saturday when the clinician doesn't get to it until Monday, right? We are all stressed out. The volumes are increasing. Why should we do this? And it breaks my heart a little bit because if I think of the future where patients have access to the reports right away, while we're thinking we can wait for the clinician to look at it, they are sitting at home being very anxious about what the results might be, and the sooner we can make these results available to them, the better it would be. Also, we shouldn't underestimate the potential of smartphone applications. We now have health products on smartphones that link directly to the portal. The law says that these can also connect directly to the medical record system. That's currently not done, but if it were, then there wouldn't be any embargo times or times between signing the report and making it available to patients. So, we have to think about this coming down the pipe in the future. How about report accuracy? So, we're still struggling with cognitive error, and in the meantime, we still have errors of grammar and spelling in the report. And even though there are tools out there, dictionaries that find some misspelling, but there's often words that are spelled correctly, but they have the wrong meaning, and we do not have software solutions to really help us with that. And this is a report from 2019. You can see the percentage of these errors is still up to 36 percent, and that's unacceptable, but it does take a lot of time to proofread a report and detect these errors manually. So, we do need more clinical and technical support to find solutions for accuracy. Another important component of or feature of our reports is clarity. And there's this great book by Curtis Langlod, it's a few years old, but still very, very pertinent, the Radiology Report, and it contains a wealth of thoughts and information that inform how we should write the ideal report, and a lot of ideas pertaining to clarity. And I can tell you how much time I spent trying to teach our residents to not say, no acute intracranial abnormality in a head CT, and instead say, normal head CT if it's truly normal. And it seems to be very difficult to get to say that, but when you are patient and you're reading, you have a normal head CT that sends a very different message from no acute intracranial abnormality. And also, if you go down those lists to what's the bottom, there's this issue of uncertainty versus vagueness. And what that means is in uncertainty we give a differential, there's a finding, I don't know what it is, it could be A, B, or C. But vagueness would be, there's a finding, I don't know what it is, it could be A through Z, basically everything in the differential. So we should try to avoid that. But uncertainty is something we have to deal with because there is uncertainty, a lot of uncertainty in medicine. And so how should we go about communicating uncertainty? And there's another great quote in this book by Curtis Langlod, and he said, report uncertainty with consistent terms. So how can we do this? Here's a great example from my current colleague, Brent Weinbeck, and former colleague, Mike Hoke. So they looked at the reporting for brain tumor imaging, follow-up imaging, where there's a lot of uncertainty, whether changes we're seeing are related to therapy, or whether the worst thing that we're seeing is a true progression versus pseudo-progression. So how can we communicate these uncertainties in a standardized fashion? So they came up with this scoring system that you see in the table. I don't know if you can read it, but it's a score from one to four. Each score is defined, and when you look at it, you can see that even though you may not be sure if it's true improvement or medication effect, the management for the patient doesn't really change, versus down here, if it's really worsened, and you think it could be due to some progression, and there might be a change in management. So it helps us focus our attention on those differentials that matter. And also, this table is included in all of these reports, so both the referring physicians and the patients can read what it means when we put a certain score in the report, and it has been very, very well perceived by all parties involved. Moving on to communication. So there is a health literacy model, and it involves the understanding and use of medical information. It involves first having access to this information, then comprehending what the language says, understanding the meaning of this language, and then taking this information and making informed decisions about healthcare. So coming back to this issue of the embargo times where we release reports to patients after a certain time frame, I found when I queried informally several institutions in the U.S. that there are no two institutions that have the same practice. And maybe the answer is that there is no best practice, because the best practice would be to simply release these reports right away to patients. And what we forget is that the patient can decide whether they want to look at this report immediately or not. So if they know that they're highly anxious and they can choose not to go into the portal and look up the report, or they can talk to their clinician and say, I'm not going to look up the report. I expect you to call me with the finding. But that should be the patient's decision. When we think about comprehension, we have a great model of lay letters. And Dr. Zafar, after my talk, will talk a little bit more about creating lay reports and radiology. And these are driven by the law. And there is now also a new law since 2018 in Pennsylvania mandating that patients need to be informed when there are significant findings that require follow-up within three months. So we really have to think about what we can do in terms of providing lay language reports to our patients. And that means at the written, at the fourth to sixth grade level. And when this was done here for recall letters and mammography, the rate of understanding rose from 50% to 95%, which is really significant. And Dr. Zafar will talk a little bit more about some of the techniques we have, like hyperlinks and reports, or the idea of inserting images into the report, which almost 90% of referrers would like to see that we do that. Now, if you think about really translating medical terminology into something more understandable for patients, it could mean that the patient goes, ah, now I know what's going on. But isn't there still this question, huh, but this illness, is this acute or chronic? Does it need treatment? Is it going to kill me or not? So that comes to the next step in the comprehension or in the understanding of the health literacy model, which is the appraisal of what you do with the information. We designed this module that you can add to the report, where if there's a normal report or an inconsequential incidental finding, this language can be added. And what we found is that patients would, without this language, call their providers about results in 23%. And when they can read this message, that goes down to 9%, so actually decreases the call burden to providers, and also patients are less anxious, and almost 90% want to see something like this in the report. And then, of course, there's embedded clinics, where we have a model at Emory in the ENT section, where we actually offer consultations between radiologists and patients. And then finally, how is this applied in healthcare? So in the old paternalistic model, the information would be taken and the patient would be told what to do, and now it's an information exchange where the patient can chime in with their own values and preferences and make decisions about which one of the recommendations we make they actually want to follow up on. And this is something I wanted to show you. This is a patient decision aid. This is for the use of statins to prevent heart attack. But this is a great model that can be adopted into imaging. You can see here on the left-hand side. You can pick what risk profile you have. You can see what the statins can do for you in terms of preventing heart attack. Then you're all in favor of taking them. You can see about the side effects and then see if you still want to take them. And then you make a choice here as a patient if you want to take the pill, don't take the pill, or have more questions answered. So this has been very effective, and you can maybe think about how this could be used in imaging when patients have to make imaging decisions. In terms of collaboration, I'll make this very brief. I think that was mentioned this morning over and over that we need to work with patient and family advisors. And one question I often get is, how do you find these people? And so the best way to find volunteers that serve as advisors to you is to look into your complaint system or your own email and contact those people who care enough to send you a message that something went wrong. And they sent that message to you because they believe that you didn't want that to go wrong and that you want to fix it. And those are the prime targets. You should thank them for their feedback and give them this opportunity to serve as a patient and family advisor. We have 140 at Emory that we have recruited in this fashion. And then, of course, we need to work with our peers to get away from the paternalistic attitude. And if there's any way to anything significant to take away from this presentation, it is to have this new perception that we are serving as partners to our patients. Thank you. I want to thank Drs. Brooke and Dr. Seward for the invitation to speak today. I think most of us in the room recognize that the traditional radiology report is truly written for a physician, for our referring colleagues. That's the language that we use, the format, et cetera. But increasingly, patients are gaining direct access to reports through portals and other mechanisms. I know that Dr. McBee is going to speak about this a little bit later, but the proportion of patients who access portals doubled between 2013 and 2015. It's now at almost 90% of patients across the country have access to reports, which means they are seeing the reports without their physician there for the first time. And this is a key point that I just want to dwell on for a moment. In radiology, we spend a lot of time talking about how do we get patients access to reports. And that is important, but physically having access to a report does not mean that a patient understands the report. Access and comprehension are two different things. And that's really what I want to talk a little bit more about today is comprehension of the report. Last year, Steve Eberhardt and Marta Heilbrunn printed a monograph about the value of the radiology report in Radiographics. And their article was structured around value to the referring physician. I love this model, and I think it can be very easily adapted to talk about how do we explore improving the value of the radiology report for patients. We're going to focus a little bit on the numerator, starting off with clarity and conciseness of the report. There was a study that was done several years ago now where patients at a large cancer center were shown multiple phrases commonly used in radiology reports. The same phrases were shown to a group of radiologists. And both groups were asked, which phrase most strongly conveys to you that you have malignancy or metastatic disease? For the radiologists, the answer was diagnostic for. That is the strongest language we can use to convey that we think this patient has something troublesome. For patients, it was the word probably. Very interesting. So just pointing out the disconnect in common phrases that we use, even when we think we're being clear, we may not, in fact, be communicating accurately to our patients our level of concern. Not surprisingly, this is a different study, but when you speak to patients directly and you ask them what is the most frequently cited problem in the radiology report, it is unclear language. By far and away, this is the most frequently cited problem. Some of that can be attributable to reading grade level. So I think many in the room are familiar with the fact that the average U.S. reading grade level is eighth grade. That's typically the statistic that we're cited. Both the NIH as well as the American Medical Association advocate that any written patient material should, in fact, be printed at a sixth grade level. This is a really neat study that just came out this year that looked at about 110, I believe, MRI lumbar spine reports, ran them through about five different readability indices. And you can see the results here. They all came out greater than a 12th grade level. Again, just reinforcing this point that our reports are written in language and in ways that is not easily comprehensible to the average patient. So there's two roads ahead of us, right, about what can we do if we want to try to improve patient comprehension in addition to getting patients increased access to report. The first option would be let's make the existing radiology report language easier to understand for patients. And this is clearly the most digestible and the most approachable way for us in the room to embrace this issue because the reality is we are all dealing with high volumes of radiology reports, competing pressures on our times. So what can we do to address this? Brevity is equal to clarity. Again, going back to that same study where they asked patients to read radiology reports and grade them on their comprehension. So being succinct is important. And the best way to do this, I think, is through structured reporting. Here we have the same radiology report shown to you using prose on the left and as a structured report on the right. Your eye is immediately drawn to how much clearer and concise and succinct the report on the right is. Your eye is also drawn very quickly to where the abnormality is located. It is in the kidney. It just jumps out at you because it's the largest block of text in the sea of normal, so to speak. And the impression is a little bit more decisive as well. I know that Dr. Cottom touched on this in her lecture as well, but I would just encourage I, in the interest of full disclosure, I trained with Kurt Langlotz as well, who is a very strong advocate for the word normal. Normal is not a dirty word, and it doesn't mean that you didn't care, and it doesn't mean that you didn't look at the structure. It just means that you're saying it's normal, which, in general, is very easy for most people to understand. I would also add that whether you spend five sentences talking about all of the negatives in the kidney or you put the word normal next to the kidney, it does not protect you in a court of law if you didn't see the abnormality or the mass or the finding in the kidney. Either way, you are still liable. Most patients want to understand the radiology report language more clearly. They want a dictionary, if you will. At our institution, Drs. Kahn, Drs. Cook, and several other team members have worked on a system called PORTER. This is a particular paper that they published where they had about 100 patients who had knee MRIs, and they invited them to go online and to use this interactive report where they could hover over terms in order to get definitions of what common radiology phrases mean. They were also able to get definitions of anatomic structures, and it even brought up Wikipedia pages where they could see pictures, illustrations of where these structures were located. Does anybody in the room want to guess what the most commonly viewed definition or word was within the radiology report? I'm sorry? I thought I heard something. It was normal. So again, people interacted a lot with these different terms, and I say this a little bit tongue-in-cheek because I just pointed out the benefits of the word normal. I don't know. We never queried these patients as to why they wanted the definition of normal. Maybe they wanted to know that our definition of normal was the same as their definition of normal. I'm not entirely sure, but I'm just putting it out there. So it turns out that patients, as we know, people, are heterogeneous. So you can query, there are some papers that have shown that patients want access to the full detailed radiology report. There's also literature that shows that patients want to have a lay summary of the report in addition to the detailed report. They want the benefit of all of the color, but then they want the succinct statement that really conveys to them what we're trying to say in language that they use. There's been a real proliferation of standardized assessment coding systems. I think many of us at many of our institutions are using one, if not more, of these systems that I've put out here. And Dr. Cottom spoke a little bit about her InfoRADS project, and I think one of the benefits or one of the easy ways for us to try to improve comprehension is to say, if you are using one of these existing coding systems, it's very easy to provide a translation or to dovetail that into coming up with patient-tailored language, which ideally would be developed at your institution with the engagement of your patient population to use words that make sense to patients, right, so that they know when they should be worried and when they shouldn't be worried. I put up this slide because last year I was at the ACR quality and safety meeting, and I met the CMO of the Jamaica Queens Medical Center. And this is just a map of the area of Queens, the catchment area that her hospitals serve. Does anybody in the room want to guess how many languages are spoken in the catchment area of this two-hospital health system? All right, let's do this in a more democratic way. How many people want to say it was less than 50 languages? No? Okay, I've got a couple. Fifty to 100. All right. A hundred to 150. Yeah, it was the last group. It was 135 languages spoken in her hospital system. That's overwhelming when you think about trying to communicate results to patients in language that they understand. That's also a lot of AT&T operator conversations. But one of the benefits of using these kinds of standardized summaries is that you can translate those into multiple languages that are used by your patient population. It is much easier to do that than it is to translate the entire radiology report. So let's talk about our second approach. What if we were going to try to think more creatively outside the box, and we have more time and more resources? So I'm going to acknowledge this is a little bit of a thought exercise, but people are starting to explore in this area. So let's talk about some alternative methods of report content. There's a strong body of literature that says that patients prefer to have key images within reports. There's a stronger preference when the results are abnormal than normal, but still, that's what patients want. So why is that? Well, when I was preparing for this lecture, I started to go on the internet and think a little bit about how we process words and information. So it turns out that the part of the brain where we read words and recognize them is essentially the same part of the brain which, in non-readers, simply recognizes pictures. And in the left hemisphere, that part of the brain gets co-opted in order to start recognizing words. So this is a big, fancy way of saying, when we look at words on a page and we don't think about it this way, we see pictures. We don't actually see words. We see an image that comes up. This really resonated with me because I happen to have a six-year-old, and I remember the entire projection or trajectory of how she learned to read. And when she first learned to read, I didn't open up Dostoevsky, right? We had these board books with pictures. She wanted to look at pictures. She didn't want to look at words. Slowly, you start to recognize the words that are associated with the pictures, and that's how our brains start to work. And so we've forgotten about this because it's so long ago, but that's actually how we learn to read. Just as a separate note, this was my favorite book when she was a baby, and I highly recommend it to anyone who has babies in the room. So when you think about it that way, it makes sense to start thinking about how can we remake the radiology report so that it's more visually appealing, right? Dr. Krishna Raj down at UVA has done some really great work in this space. And once again, if you look at the report on the left and the report on the right, they both contain the same information, but it's much easier to digest. I would even say this is easier for me as a radiologist, and probably for a referring physician as well, to understand the key components of what you're trying to communicate when you use these images. So when we read, we recognize words as pictures, and what's interesting is we also hear them spoken out loud. So there's a separate study that was done out of a group of researchers in Italy, and they found that when patients... This was done on patients who were undergoing brain surgery who were awake. So this is how they were able to do this experiment. It turns out that when we read words silently in our minds... Just right now, you're reading these words on a screen. You actually hear them spoken out loud in your brain. So Broca's area, which recognizes speech, responds to silent reading the same way that auditory neurons respond to when you are hearing me, for example, speaking to you. So we're actually engaging multiple senses, even though we are just looking at words on a page. So we're debunking that myth, excuse me. You don't just learn as a visual learner, as an auditory learner, or as a whatever learner. You engage multiple senses when you are reading. This is a paper that just came out from a group of researchers in Brazil who created these brief audio-visual reports. So they were two minutes each, a bunch of MSK studies that came out of the ER. What they did is they showed key sequences from, for example, an eMRI, and then the radiologist would record their voice running through the pertinent negatives and positives of that report. And you can see that dramatic improvement in the level of comprehension for a referring provider. And I would advocate probably something that could be extrapolated to patients as well. This is a different study that came out of Duke with Dr. Taylor as the senior author, where they showed patients three different methods of, or three different layouts of a BI-RADS zero radiology report, and they used Amazon Turk, which is a great online mechanism to engage multiple people who may not be patients directly at your institution. So we've got the traditional radiology report here. We have the MQSA letter in the center. And then they created this interactive report that actually allowed the patient to click on different parts of the image, to choose your own mystery, ask different questions that they may have had, et cetera. What they found is that there was much higher comprehension of the same content that was being communicated in these reports if you used either the patient letter or the interactive web report. This next part I love. Patients who use the web-based interactive report actually thought that the radiologist was more empathetic and more competent than patients who just read the traditional radiology report. And that's a little bit funny. Nobody laughed here. But to me it was a little bit funny, but then it made sense, because when you understand what somebody is trying to say to you, yes, you think they are competent. You have trust in them. And that's really, I think, what this study was showing. Dr. Cottom touched on this briefly, but I live in Pennsylvania, and so we are dealing with the realities of Act 112 in a very real way. So direct notification of patients, outpatients, excuse me, just to be clear about that, who have any finding that requires follow-up within three months. This is the language in the legislature that they recommend that is communicated to patients. That is a hunk of text. It is barely comprehensible to me, quite frankly, let alone the average patient. And yet that's what they're recommending in the legislature. Tomorrow afternoon one of our residents, Dr. Middle, will be presenting. We tried to play around a little bit with this, touching on some of the themes in this lecture. What are the differences between giving patients the same information in an infographic versus a letter, using 12th grade versus 6th grade information. The results are surprising, but I'll allow him to have his big reveal tomorrow afternoon. The last thing I want to touch on just very quickly with my remaining time is the usefulness of the radiology report. Ultimately, when a patient opens up a report, yes, they want to have physical access to it. They want to understand, comprehend the report. Really on a fundamental level, they want to digest, but what does this mean for me? You're saying I have an abnormality in my body, but what does that mean for me with my comorbidities, with my patient history, with my profile, my unique profile? And I think one of the benefits of going back again to those standardized coding systems is that it allows us to start to create this kind of tailored information for patients. We have a standardized system that we use at the University of Pennsylvania. We did a study where we looked at all of these patients who had indeterminate liver lesions, correlated them with ICD-10 codes and demographic data in order to try to come up with these answers for our patients. So it's very different to say, here's your radiology report. You have an indeterminate lesion. And to say, you have an indeterminate lesion in your liver, which 2% of all patients who receive ultrasound images at our institution have. It turns out that 7% of these are likely to be cancer. But if you are less than 61 years of age and you don't have cirrhosis or a known malignancy, this lesion is overwhelmingly likely to be benign. And that's really, I think, the kind of information patients want. I know many of us are familiar with ratings, getting rated, et cetera. I envision a day in the future where patients might actually be able to rate their radiology report and give us some direct feedback. And I know that, again, Dr. Krishna Raj's group is starting to play with this a little bit down at UVA. So in closing, I just want to say, I think there's a real shift in the patient perspective of a radiology report. It's less about just giving physical access to patients, but how do we help them engage in shared medical decision making after comprehending the information that we put in the report and understanding what it means to patients like them? Thank you very much for your time and attention. Before I get started, I just wanted to put a thank you out here to these people. A significant portion of this talk is going to focus on a project that I was involved with while I was at Cincinnati Children's, and I couldn't have done it without these people. So I can't talk about patients communicating with radiologists directly without mentioning patient-centered care. This is a whole symposium on patient-centered care. And this has gone over several times, so I'll keep this brief. But basically, the patient is put in the center of the health care delivery model. And if I ever give this talk again, now I know that there's a better one that has the provider partnering with the patient. So make a little mental note there and redo that next time. But patient-centered care has become a fixture in radiology and several efforts, including ACR and RCNA, to facilitate that. And if any of you were fortunate enough to attend the opening session on Sunday for RCNA, they made a really great case for increasing patient contact by radiologists. So since this is a session on quality, I would be remiss if I didn't mention quality. The Institute of Medicine defines six domains of health care quality, and of course, one of those is patient-centered care. And within patient-centered care, the Institute for Patient and Family-Centered Care has four core concepts. And really, all four of these concepts can be applied to radiologists communicating directly with patients. One method of practicing patient-centered care in radiology is increasing our patient engagement via our patient portals. So patient portals have been around for a long time, since the 1990s, but they really didn't see widespread adoption until around 2006 because of meaningful use incentives. And radiologists are inherently digital physicians since the advent of PACS. Most of us sit on computers for most of the day. But despite that, most health care systems do not have a system in place that allows patients to digitally communicate with radiologists. Despite that, over 90% of U.S. hospitals offer the ability to view medical records online, and I'm sure this is just going up and up every year. This is from American Hospital Survey Association in 2016-2017. And 68% of hospitals allow patients to send messages to their health care providers. And the number of hospitals that enabled patients to use patient portals actually grew sevenfold between 2013 and 2015. And in analysis at the University of Michigan, which is a very large academic health center, they looked at over 50,000 patient-initiated messages in their patient portal system, and they found that about 3% of their patient-initiated messages were related to radiology studies. So this was a study done in 2014, so this could have potentially increased since then. So about 3%. And here are the types of questions that they received. And I'll blow up the most common ones here. So about a third, the most common question that was received was, what are my imaging results? And this kind of touches on the embargo period and everything that we talked about earlier. The next most common question with 12% was, what is the next best step for some finding in the report? And then the third most common was, what does this finding mean in my report? And while I personally think this is a great thing, we're enabling patients to have greater say in their care. We do need to be careful with the increased utilization of patient portals because of potential biases. So the uptake rate or adoption rate of patient portals among patients is around 52%. So that means about 48% of our patients aren't utilizing the patient portals. And things like age, race, sex, and income all have an influence on the adoption of patient portals. We have lots of room for improvement there. And actually, I think at the end of the session, there'll be a talk about improving care for vulnerable patient populations. So defining the problem, if patient portals are ubiquitous and patients are using them, and they're using them to message their health care providers, we as radiologists are health care providers, then why are they not using them to message radiologists? The answer is pretty simple, because they can't. We in radiology have done a very good job over the years of kind of separating ourselves from the bedside. And it's not really one thing that we did that kind of led us to this point, but it's just the entire system has just kind of been evolving along, and this is the situation we're now in. So there's been lots of publications on radiology patient portals, and multiple different articles have addressed the value proposition of patient portals in one way or another. We heard about translating into lay language and making reports more understandable. So those are some different solutions to try to increase patient engagement and make things more understandable for patients. Several articles even mention the possibility of patients being able to directly communicate with radiologists, mostly through providing contact information within the report for patients to either call or email the radiologist. But none actually address the patients being able to directly communicate with radiologists in the portal. So that was our solution, was to develop a system that allowed patients to directly send messages to radiologists through the patient portal. And like I said, this took place when I was at Cincinnati Children's Hospital, which is a large academic children's hospital. We already had an integrated patient portal system, and patients were already able to view radiology reports, but we have a 48-hour embargo period for CT, MRI, and PET-CT studies, and other studies were released immediately. So what we did was came up with a working group of several people involved, including radiology, informatics, and the EMR patient portal team members. And we came up with specifications and kind of developed the system. So what we came up with was a list of seven final requirements that we wanted to have in our system. So we wanted it to be available for both completed and upcoming studies. We wanted to require minimal data entry by patients because we figure if patients have to spend a lot of time manually inputting data, then you'd probably see less uptake. And we wanted the ability to initiate questions directly from the report. Part of the development was we wanted questions to go to a pool of radiologists and not specifically the radiologist who rendered the report. Different places may want to do that differently, but in our minds, we figured a lot of the questions would be able to be answered by any radiologist. And we didn't want to have to deal with certain radiologists being out of town and having to check their inbox when they're away and things like that. We wanted the message and the response to be available within the EMR for other providers to be able to view it. We didn't want to be practicing this in a vacuum. And we wanted the ability to document phone calls. And then finally, we wanted to be able to forward the questions to other providers outside of radiology in case it's something outside of our area of expertise. And so in the end, in our system, we were able to satisfy five of these seven requirements. We're not able to have the patients initiate the question directly from the report, but it does happen from the patient portal where they view the reports. And there is a little bit of manual data entry that patients have to undergo. So this is a screenshot of what the portal looks like. So this is the test results section where they view everything from lab results to radiology. Radiology reports are kind of lumped in with lab results, be that as it may. And then down in the right-hand corner here you can see this, ask the radiologist a question. If you have a question regarding a recent or future radiology study you may send a question to the radiologist. And when they click on that link, they're taken to this form. And this form, they have to enter some things. I said it required a little bit of manual data entry. They have to type in the name of the radiology study. They have to say whether it was completed or in the future, the date, how they'd like to be contacted, whether that's by phone or through the patient portal, phone number if that's their preferred method, and then their question. So once they click the send button, the message then goes to the inbox of all the radiologists in the pool. And it's not related to who read the study like I mentioned earlier. But individual radiologists are able to claim a message so that multiple people aren't kind of working on it and duplicating work. So then once it gets into the inbox, this is what it looks like for the radiologist. So there is a list of messages up here. And then once a message is selected, all the questions shows up in the box below. So I'll blow that up. And this one, you can see the name of the study, it's a spine x-ray. It was performed in the past on this date. They'd like to be contacted through the patient portal. And their question was, I can see from the result that there are no fractures caused by my child's osteopenia. This is a children's hospital. However it seems that her scoliosis has worsened. Do you recommend any follow-up tests? And so once the radiologist reads the message, they can either click the reply button and just basically shoot off a message to the patient. It will go to their patient portal. They'll get an email about it and they can log in and view it. The other option is to document a phone call which was one of our requirements. So once that is done by the radiologist, there's an encounter created of a radiology patient email type and everything is automatically placed into a note. So here's an example of the question from the patient and then here's the response from the radiologist, kind of in a text message back and forth kind of system. So once we implement it we then tracked a couple of metrics including the number of questions, the type of questions, the radiologist responsible and the response time. We went live with the system on October 3rd and actually got our first question the very next day. It was very exciting. And we had several promotions trying to increase awareness of the system. We placed an announcement on the landing portal, on the landing page of the patient portal. We had a post on our departmental blog. We had some social media campaigns and just a general awareness campaign in the reading room. So what we found was that over the period of time from October 2017 to July 2018 when we submitted this for publication, we had received a total of 88 messages in a large radiology department. This is a very small number relative to the number of patients that we've seen. And almost at 47% were results when they will be available. The next most common question was a question about or clarification of imaging findings which is really what we were hoping for. So that was about a quarter of the questions. And then 15% were requests to view images. So this is a chart with the number of messages that we received over time. The red shows the cumulative total number and then the blue bar charts show the number of messages received per day. So these right here were our little promotions which as you can see didn't seem to really have any effect but it kind of stayed relatively steady over time. So at the time of publication we had 33 different radiologists respond to messages. So there was good uptake throughout the department. We didn't really have any incentives. We didn't force people to participate in the system and a lot of radiologists anecdotally said that they appreciated the ability to be able to communicate with patients and got a lot of job satisfaction from that. Our median turnaround time was 5.1 hours and as I said earlier the number of questions that we received is very small relative to the number of studies performed. So it was not very burdensome. It was a small number of questions and we didn't really have very many inappropriate questions. They were all questions that radiologists were easily able to answer. But obviously there's lots of room for more outreach and marketing to patients and their families. So I'm just going to show a couple example questions that we received, some of the more interesting ones. My pediatrician was unsure what was meant by likely represents muscular insertion sites. Could you please elaborate? The pediatrician did say it was nothing to necessarily worry about and that we should watch it for now. So I thought this one was interesting because this is a great opportunity for a radiologist to add value here because the patient family member had already spoken to the pediatrician and the pediatrician wasn't really sure what this meant. So a really good opportunity for the radiologist to come in and say, well, you know, actually this is what it means. Another example, family member questioning, what does it mean ASL images demonstrate asymmetry and perfusion within the basal ganglia asymmetrically increased on the right in relation to the left? Is this normal? So this goes back to the talk earlier, using very heavy medical jargon, patients don't understand this. And then finally, I would like to view the ultrasound images from my son's hip from yesterday's exam. And actually, since this system was implemented, the portal system was image enabled so that patients and their family members are able to view the images. So wrapping it up, patients are reading our radiology reports and they expect timely results and that institution needs better education regarding the result embargo period. And I guess we can question whether or not that's actually necessary. Approximately a quarter of the questions were for clarification of findings and this is what we really wanted. And patients are interested in seeing their images. So in conclusion, EMR-based patient portals can be constructed to enable patient communication with radiologists and are a great way to increase patient engagement. And importantly, patients use it and seem to appreciate it and they ask appropriate questions. Thank you. All right. So today I'll be talking about closing the loop on follow-up recommendations. Well, some articles discuss radiology recommendations include all types of recommendations, including clinical follow-ups such as procedures like colonoscopy. For this talk, I will be focusing on recommendations for additional imaging. I will call these RAIs. RAIs are created to monitor potentially malignant findings or to ensure stability or resolution of potentially serious diseases such as aneurysm. RAIs are common, seen in about 11 to 21 percent of radiology reports and are used more frequently now than 20 years ago. This is probably in part due to the proliferation of guidance documents and maybe in part due to higher resolution imaging. Unfortunately, compliance with radiology recommendations remains relatively low, around 50 percent in most institutions. This poor compliance places patients at risk for delayed diagnosis, which raises the question, what value do our recommendations have when they are not followed? The term closing the loop has historically referred to closed-loop communication. This requires the sender of a communication to obtain confirmation of message receipt and demonstrate comprehension from the receiver. The sender then documents that this has occurred. This remains the standard of care for non-routine communication such as abnormal radiology test results that warrant a recommendation for additional imaging. Closed-loop communication can be improved systematically, as demonstrated by the RADCAT3 system at Lifespan Health. They use manual radiologist trigger of a macro in their dictation client when an RAI is dictated. This macro pushes the RAI to the communication desk where assistants perform and document the closed-loop communication. Over the course of two years, adoption of this fairly basic workflow was well-received by radiologists, improving quality and efficiency. Documentation of closed-loop communications increased from 250 events per year to over 9,000 events per year, and they were able to successfully document closed-loop communication for 99.7% of abnormal test results. While RADCAT3 is effective at improving communication at transitions of care, it does not address many of the other causes associated with inconsistent follow-up of RAIs, such as inadequate test result management systems in outpatient offices, failure to notify patients of test results, and diffusion of responsibility when a patient is cared for by multiple healthcare providers. Regardless of the cause, the impact of a delayed diagnosis can be significant for a patient. In the setting of a failed lung nodule follow-up, the most common form of delayed diagnosis related to RAI noncompliance, this often means the difference between treating a stage one lung cancer and treating a stage four metastatic malignancy. Patients diagnosed with early stage lung cancer have a six times or greater chance of survival at both 5 and 10 years. Health systems and providers also face significant legal ramifications of delayed diagnosis events. With all the potential causes of inconsistent follow-up, is it even possible to eliminate these events? Well, the answer is yes, which is why I'm here today. But doing so requires moving beyond the current episodic care model of most radiology practices to that of a more coordinated care model at the health system level. This requires changing your mindset and thinking like a high reliability organization. Yet not all recommendations are created equal, and not all recommendations can be tracked. At the University of Rochester, we track what I call the two A's, recommendations that are both actionable and absolute. These recommendations have defined endpoints, which allow staff without medical training or a computer system to monitor for recommendation compliance. The University of Washington used natural language processing to review 4,800 recommendations for lung, thyroid, and adrenal nodules. Nearly 80 percent of these recommendations did not specify a due date, and about half of their recommendations did not specify an imaging modality. Please note the SAM designation at the top right of this slide. At the University of Rochester, we performed a similar NLP-based assessment of over 24,000 recommendations. This showed that only 18 percent of our RAIs were both actionable and absolute. Even though we can only track a subset of the total recommendations being made, we have found that these high quality recommendations are more commonly related to evidence or consensus-based guidelines and are more commonly present with our highest risk incidental findings. Now that we have defined recommendations that can be tracked, let's discuss how to develop a tracking system. You will need three main components. First, a mechanism to identify RAIs and to enter them into a database, then a database to store them with time-based alerts that prompt review and potential action, and one or more interventions that attempt to improve recommendation compliance. Reports can be reviewed manually, typically by a radiologist by flagging a report at the time of dictation, potentially using a macro, or through NLP. Structured elements then need to be pulled out of reports with actionable recommendations. And structured data entered into a database. The database holds these recommendations for a period of time, then generates an alert when the recommendation is ready for review. Alerts trigger electronic or manual review of the patient record. If compliance is identified, it is documented. If compliance is not identified, an intervention is attempted, which hopefully leads to compliance. As you can imagine, reality is more complicated than my clipart design. The design and implementation of medical informatics tools to accomplish recommendation tracking efficiently and consistently is still a heavy lift at the scale of large health systems. Implementation can be performed more easily at smaller sites, which can tolerate a less efficient, more manual process. That being said, there are solutions being demonstrated in the exhibit hall that can help fast-track the implementation of a tracking system. And once systems are in place for about six to 12 months, a steady state is typically reached, which dramatically reduces the need for project oversight and development. At the University of Rochester, our manual entry mechanism requires five clicks to identify the structured elements from an RAI. Simpler report flagging systems are also available in most voice recognition products. We also utilize a natural language processing analytics tool to identify recommendations that were not manually entered. Recommendation software systems have begun utilizing natural language understanding to identify recommendations in real time, to assist in improving the quality of recommendations being made, and to prompt entry into a tracking database. Our database pulls in the name of the primary care provider, as well as contact information of the patient, primary care provider, and ordering provider to facilitate interventions. It also has multiple tiers of alerts and documentation, and uses the structured data to automatically generate patient and recommendation-specific standard communication documents to both the patient and the healthcare providers. Electronic medical record systems are developing recommendation tracking databases as well, although they are currently less sophisticated and less customizable than dedicated commercial products. Our intervention process is a multi-stage cascade beginning when a recommendation is one month overdue. It consists of a series of notifications and phone calls in an attempt to build multiple layers of defense against human and system error. Our staff send letters, faxes, emails, and make phone calls. Other sites are creating best practice advisory alerts within EMRs, some of which help facilitate recommended exam ordering. We are currently rolling out EMR messaging to ordering providers and to patients about their abnormal test results and recommendations. So what do the results of a tracking system look like? In our six hospital health care system with about 800,000 non-mammographic exams annually, we identify and track about 7,000 exams per year. Compliance with our RAIs was found in just over half of these recommendations, and we intervened upon the other half. A one-in-six recommendations track led to a new exam completion, about 1,200 exams, the vast majority representing an advanced imaging modality such as CT or MRI. We also identify and document appropriate non-imaging care for one-in-five patients. This may be a biopsy or excision or risk-benefit discussion in an elderly patient with limited life expectancy. The number of patients at risk for delayed diagnosis has been reduced by 80%. We are regularly finding cases where outcomes are directly impacted by tracking system intervention, such as this 80-year-old female who came to our emergency department that was shortness of breath. Our recommendation for chest CT based upon a suspected lung nodule was not performed. After two interventions, a CT was obtained identifying a speculated nodule. Subsequent PET confirmed a treatable stage one lung cancer. Allowing patients to continue falling through the cracks of our healthcare system with resultant preventable harm will not be tolerated by the public. A state legislator in Pennsylvania personally knew of two or three individuals who suffered from delayed diagnosis of malignancy, which led to the enactment of Act 112, which it seems like everyone has talked about today. This requires radiology departments to notify patients within 20 days of any abnormal radiology findings which require follow-up within three months. While I believe this law is poorly written to prevent additional occurrence of delayed diagnosis in the future, for instance, it would not have prevented the delayed workup in the lung nodule I just showed you because it was found on an x-ray in the emergency department. But it could be the first of many similar laws in other states if we do not address this problem internally. We are currently studying the impact of early direct patient notification of abnormal test results supported by an ACR innovation grant. Another definition of closed loop is found in engineering. This is basically any process that utilizes feedback to modify its process over time. A common version of a closed loop system is an air conditioning system where a thermostat measures room temperature and compares this result to an input at desired temperature to regulate an HVAC unit. The current standard of closed loop communication of RAIs is an open loop process in the engineering sense. Without measurement of an outcome and without feedback to the radiologist or communication desk regarding the impact of the communication. In the setting of a recommendation tracking system, RAIs generated by the radiologist are monitored by the system navigator who performs one or more interventions aimed at changing an outcome. An outcome of new exam completion rate can be used to optimize the tracking interventions being performed. An outcome of the rate of early malignancy diagnosed versus the false positive rate of lesion workup could help modify the use of recommendations at the radiologist level similar to the current practice in mammography. Or be used to in aggregate to impact modifications to consensus guidance documents. Finally, a list of seven ways recommendation tracking systems can be used to add or measure value to your health system. You can measure the impact of RAIs on patient outcomes and study the cost benefit ratio of the RAIs you make. You can measure the effectiveness of tracking interventions and associated revenue generated to determine the project ROI to support your business case for resource allocation. In my experience, the ROI on tracking systems fall somewhere between two and four X depending upon the technology used in the size of your organization. You can provide feedback on RAI use rates down to the radiologist level and compare to other radiologists reading similar types of cases. You can measure the quality of recommendations or at least their consistency with current guidance documents. Measuring radiology performance is proven elusive. Metrics around recommendations could be integrated into your OPPE process and can be used to reduce the high variability in their use rates currently observed between radiologists. Poor quality recommendations or excessive recommendation use can result in unnecessary imaging, increasing patient anxiety, and wasting healthcare resources. Shining a light on our recommendation practices is the first step towards improving our performance and demonstrating our value. Eliminating preventable delay in diagnosis provides our patients with their best chance for a cure. The technology exists to make this a reality today. I encourage you to embrace it. Thank you for your attention. For the last few minutes of the session this morning, I'm going to be shifting gears slightly and talking about how we can improve the care that we provide to vulnerable patients. We'll start by defining some important concepts. So vulnerable patients are defined as those who are at greatest risk for poor health status or healthcare access, and that may be due to physical, psychological, or social characteristics. These folks may face barriers to care, which could include difficulties with communication, understanding how to navigate the system, feeling welcomed, respected, or safe, and financial constraints as well. A few more important concepts here. Patient-centeredness describes the belief that each patient must be understood as a unique human being. And cultural responsivity emphasizes the importance of providing care that meets the personal perceptions, needs, and expectations of a diverse patient population. The goal here by applying patient-centeredness and cultural responsivity to the care that we provide is to shift the dial from illness-oriented medicine or paternalistic medicine of times past towards patient-oriented medicine, and that's really the goal here. Traditionally, many people have thought that the concept of diversity encompasses things like gender, ethnicity, and race. But in fact, as I think is creatively depicted in this graph here, this figure here, diversity encompasses a much more broad range of concepts that define who a patient is, much more than what we may think we see on the surface. And I think it's really important to understand this range of diversity when we're interacting with our patients and our colleagues. As we consider the spectrum of diversity as it applies to radiology, it's sort of helpful to think about three major groups. One would be physical abilities contributing to diversity, which may include patients who have impaired hearing or vision, impaired mobility, those with acute or chronic medical illnesses, infants, children, elderly patients. There's also psychological and cognitive status contributing to diversity, which could relate to acute or chronic mental illness or folks suffering with alcohol or substance abuse issues. And then, of course, there's social, cultural, and ethnic diversity, which includes LGBTQ folks, women, homeless patients, refugees, et cetera. Now, it's convenient to have three different groups, but I think it's very important to remember that patients often may span more than one group, but the same overarching principles are important to apply to all the care that we provide to all patients. Go through some more definitions here. I think as we consider how we communicate with and about our patients, it's important to be familiar with these concepts. And I think generalizations and stereotypes are frequently confused because they both involve applying broad statements to groups of people. Stereotypes are defined as a widely held but fixed and oversimplified image or idea of a particular type of person or thing. Now, I'm from New England, and an example of a stereotype that you might hear about people in New England is that everyone in New England is an obnoxious, rabid Patriots fan, okay? As we can see from this silly example, a stereotype is limited. It's fixed. It's judgmental. It's an endpoint, right? The conversation doesn't really go on from there. A generalization, on the other hand, is defined as a general statement or concept obtained by inference from specific cases. So another silly example would be that New Englanders tend to be enthusiastic and loyal sports fans. We can see here from this example that a generalization is more of a descriptive and flexible concept. It's based on data or patterns of behavior. It's a place to begin a conversation, understanding that more information is going to be necessary to determine if that generalization applies to the patient that you're speaking with. Now, bias is something that we talk about a lot in medicine. And bias is defined as prejudice in favor of or against one thing, person, or group compared with another, usually in a way considered to be unfair. Implicit bias has to do with our attitudes towards people or stereotypes without conscious knowledge. And we frequently talk about the impact of implicit bias in medical decision making. So implicit bias is subconscious. It's often unintentional. An example of an implicit bias is that you may consciously believe that men and women are equally good at science. But you may subconsciously associate more strongly being a doctor with being a man, okay? And that would be considered an implicit male doctor bias. Implicit biases and stereotypes are important to be aware of because these can interfere with the healthcare provider's attitude and approach towards a patient and can hinder the effectiveness of care. All right? Implicit biases have been shown to impact communication, treatment, and pain management. And I think it's really important to understand and identify our own implicit biases that we all have to help limit our treatment bias in medical decision making. And to that end, I would encourage you all to check out this website, Project Implicit. You can go online and do several different modules where you can better understand your own implicit biases to become more informed. Okay, and as we kind of round out our discussion about how to speak with and about patients, I'd like to encourage everybody to try using something called People First Language. So this is a method of communication that emphasizes a person rather than a disability or limitation. It acknowledges that the disability or limitation is present and is more of a respectful and objective method of communication. For example, if we were to say that we had a person named Jane, we would describe her as a person with a disability rather than saying that Jane is disabled or Jane is handicapped. Okay, this is a more respectful and objective method of communication. As we consider the cross-cultural clinical skills that are important for all of us to practice, I'd like to remind us to try to be informed and compassionate, to always be respectful, to try to be aware of your own implicit biases, to try to avoid stereotypes, and to work to validate differing values, cultures, and beliefs. Okay, so now we're going to go through a series of common experiences in radiology which pertain to patient-centered care. And as we're going through these scenarios, I'd like you all to reflect on your own experiences. Is there anything that you do that's similar? Is there anything that you do that's different? Is any of this information something that you could use to help improve the care that you provide? Okay, we'll start by talking about patients who have impaired mobility. As we consider the imaging environment in radiology, there's a pretty significant physical accessibility of many of our imaging suites. I'm a breast imager, and I often think about the stereotactic biopsy table, where patients have to sort of go on their hands and knees and wiggle around in order to get positioned in the right place. It's a pretty challenging environment. The fluoro table is an obvious example as well, which is a hard table. It's at a fixed height with often no handholds or pads. Fortunately, there are several easy and expensive human factors considerations that we can employ in the imaging environment to help increase accessibility for our patients who have impaired mobility. And that may be as simple as applying fixed handholds to the side of the table to grab onto, or fixed pads that aren't going to slip and move around when a patient is being positioned. Also, always providing steps with railings to patients so they can feel secure and safe as they're climbing up onto these devices. Patients with limited English proficiency, that's also an interesting topic, I think. Estimates based on data from the Census Bureau suggest that over 61 million Americans speak a language other than English at home. And approximately 25 million people are classified as having limited English proficiency by the U.S. Department of Health and Human Services. And so in certain areas, this represents a very large portion of the patient population, as we discussed earlier. In an ideal world, we would be able to provide language-concordant encounters where the language that the provider speaks is concordant with the preferred language that the patient speaks. And that encounter can be performed throughout the entire time in that language. Data suggests that this results in improved patient satisfaction and clinical outcomes. Unfortunately, that's not always a possibility. And in those instances, we're often faced with using an interpreter to try to communicate with our patients. As we're using an interpreter to communicate with our patients, I think it's really important to try to take the extra time that's necessary to listen, to empathize with the patient, and to work to establish rapport. When we communicate with patients through interpreters, patients often report that they feel less comfortable and less respected. And that can be very dangerous if we're considering consenting a patient for a biopsy or something like that. As we consider how we use interpreters, I think it's really important to follow some of these simple rules. I'd encourage you to try to always avoid using a minor to interpret unless it's a life-threatening medical emergency. Also, try to avoid using an adult family member or friend to interpret, again, unless this is a life-threatening emergency or it's at the patient's specific request. If possible, always try to avoid using a bilingual or multilingual staff to interpret unless they are officially qualified, because it's really important to remember that being an interpreter is more than just being fluent in a language. There's a large amount of training and extra credentialing that goes in to being an interpreter. When you're using an interpreter, it's important to always ensure appropriate positioning of the patient to promote communication between the healthcare provider, the interpreter, and the patient. This is particularly important when we're communicating with American Sign Language. And remember, please, if the patient declines an interpreter, you are always allowed to request the presence and assistance of an interpreter to communicate with your patients. For the remainder of my talk, I'd like to talk about healthcare barriers for LGBTQ patients. Members of the LGBTQ population face persistent challenges when it comes to healthcare, and that can include discrimination and lack of access. This is a graph here that shows answers to questions regarding fears and concerns about accessing healthcare, and the respondents were members of the LGBT community and those living with HIV. Highlighted here in red are some of the common answers about what stems from the fears and concerns about accessing healthcare, and that includes that medical personnel may be treating me differently, that there are not enough healthcare professionals who are adequately trained, and that there's a community fear or dislike. And I think that these answers really highlight how much we need to work to improve the care that we provide LGBTQ patients. So what can we do? I think it's important to try to start by connecting with LGBTQ patients through patient satisfaction surveys or community outreach projects and partnering with community health centers, anything we can do to try to create a welcome, safe space in which to receive care. Consider what your imaging environment looks like, right? What does your waiting room look like? Always very easily display your nondiscrimination policy or posters of acceptance. It's easy to provide relevant educational materials and literature in the waiting rooms. And always try to use inclusive language and correct terminology. Along those lines, we'll spend some time going through terminology surrounding transgender patients. So transgender is a term that's used to describe people whose gender identity is different from the sex they were assigned at birth. An estimate suggests that approximately 0.5% of Americans are transgender, which equates to about 1.5 million transgender adults in the United States. Although this estimate is likely underrepresenting the actual number because of underreporting and gender fluidity. A transgender woman is a person who lives as a woman today but was thought to be male at birth. And a transgender man lives as a man today but was thought to be female at birth. And I want to point out that here I'm using the nouns man and woman and the adjectives male and female to refer to gender identity, okay? It's important to keep in mind that some people do not identify as either a woman or a man or identify as a combination of the two. Gender identity is defined as internal knowledge of your own gender, whereas gender expression is the external presentation of your gender. And it's really important to remember that making assumptions about a patient based on their outward appearance can be harmful. So when we're caring for transgender patients, it's important to inquire and use preferred name and gender pronouns. Ask the patients what their chosen name is, what their chosen pronouns are, what their current gender identity is. Follow the patient's lead. How are they self-identifying? It's always important to avoid gender-specific terms such as miss, sir, ma'am. And if you make a mistake, just apologize, move on, and make sure you don't make that mistake again. When transgender patients come to the imaging environment, it's important to focus on the reason for their visit rather than the fact that the patient is transgender. Please always only respectfully inquire about sexual organs if it's relevant to the exam, such as a pregnancy status if the patient's coming in for a CT. And always be aware of your implicit biases and try to avoid judgment. The American Journal of Rankinology published a series of screening recommendations for transgender women and transgender men, which I'd like to include here briefly. And in conclusion, cultural responsivity and patient-centeredness are essential concepts in the delivery of effective patient-oriented medicine. And I'd like to just remind you to reflect on your own experiences caring for patients. Thank you.

patient-centered care

radiology

four C's

paternalistic approach

patient partnership

timely care

smartphone applications

report clarity

natural language processing

implicit biases

patient portals