Hello and welcome to making virtual teams work. My name is Nadia Khedem. I'm an adult and pediatric neuroradiologist at Emory University and Children's Health Care of Atlanta. I am German, although I couldn't look more different than this beautiful German woman, so here's what I look like. I have a few disclosures, but none of them pertains to what I'm presenting here today. Making virtual teams work before COVID, I would have said, well, only if you must because there are some benefits. You can bring team members together that are dispersed geographically and you don't have to worry about finding a suitable meeting place. You can meet anytime, anywhere. Since we had COVID and we were all forced to engage in virtual teamwork and virtual meetings more frequently, I think many of us have discovered more benefits of virtual teams. Clearly, increased productivity is reported for many I've spoken to probably because of fewer interruptions at home. It saves office space. There's reported in the literature less sick time and absenteeism probably because of better work-life balance. As a result, also less employee turnover. Meetings can be more efficient. Just think of the time you usually spend to get to the meeting place and then walk back to the reading room. You can access talent that is far away. We could have remote radiologists working for our team because they're the smartest and the best, and they wouldn't have to move to our location. There's commuter savings. Just think about what we can do for the environment if we can curb the exhaust from our cars, but also there are cost savings and transportation costs, parking costs, and so forth. There are also big drawbacks though that need to be mentioned, HIPAA and privacy concerns. When I engage in virtual meetings that I have to do from the workstation, I'm typically in the reading room surrounded by other people, and that is not a good environment for privacy or HIPAA concerns. Access to technology can be difficult. Not every home Internet connection performs the same way. There could be differences, there could be changes related to the weather, it may not be reliable, and being dependent on technology becomes a disadvantage. We all lose our physical interactions with each other. Sometimes it's just nice to give a hug or do a high-five when we're in person. It's a little bit more difficult to do this in the virtual world, so trust-building is possible, but it's different and more difficult. The sustained engagement can also be difficult with our trainees. Myself actually, I'm not as engaged in our lecture series participation as I should be because it just seems to be so easy to be absent from a virtual meeting when nobody is looking around in the room and counting faces and noting who is there and who isn't. Here are the four horsemen of procrastination, and I think many people actually suffer from distractions when they work from home, such as napping, snacking, social media, or doing some chores in the house. Not everybody is cut out for this home working environment, but a lot of people are and enjoy it. It's just a matter of introspection and identifying who you are. Certainly, with COVID, it has become very difficult for families to manage childcare and work at the same time when other childcare options were not available. What I want to do is maybe go over the 10 predictors of success for virtual teams and I cannot go over everything in detail, but I want to give some examples about technology and barriers or interesting aspects of it, and talk a little bit about team agreements. Here are the 10 predictors of virtual team success and a lot of them are very straightforward. Having the right technology is very important, and I'll talk about a little bit what we may think is enough technology, but also what we really need to make virtual teams work. We need to hire the right team leaders. You need a lot of confidence to be the leader in a virtual team where you don't get enough feedback maybe as in real life about how you're doing as the leader, and it can be more difficult to reach your members and hold them accountable. Then for the team members, as I've said before, hopefully, they have this introspect where they can tell whether the virtual workspace is suitable to them or whether they're easily distracted, then they should not be a virtual team member. It's important to always establish clear visions and values, especially for virtual teams. Of course, aligning those goals with company goals that applies to any team, having a solid team agreement in place, and we'll talk a little bit about that. Using a communication strategy, there are different communication tools that could be used in different circumstances, and I'll mention that a little bit also later. Agreeing on a process for team workflow, so how do we work together on documents or projects? What is our workflows? In charge of signing roles, all of those is important. Then onboarding strategy is important, and I'll touch on that a little bit on that too, and actively managing executive perceptions. Let's talk a little bit more about the technology. When we may think that we have email to communicate, and now we have video conferencing, and that's it, but there's actually much more to team collaboration than this. You really need collaborative tools, so on a team meeting, if you do video conferencing, is there an option for all the team members on the call to actually work together right on a board that everybody can see, or create a document together on the call? You need project management tools for any team, and including virtual teams. There need to be options for document sharing and creation, so those are usually outside of email, and video conferencing tools. There could be a need for instant messaging, where instead of an email, where it's kind of unclear how fast you have to respond, with instant messaging, it might be more immediate that you get an answer, and that is sometimes necessary. Scheduling and shared calendars become very important, where you just don't run into people to make an appointment with them, or you can just not peek into their office, but you have to do it online, so that becomes very important. And then social networking, building social connections within the group is extremely important as well. So when I was asked to work from home, I had tech savvy, but these were the instructions, and it's a little bit involved here with routing all the cables, and it was a little bit distressing when I had to put this together, but I managed, and I had our IT support on the phone when I ran into issues, so that was very nice. Having IT support for online teams is extremely important. In addition to email, we have a HIPAA-compliant text messaging tool that we can use to communicate our texts. Here I'm communicating in blue with our MRI technologist in red, and so it's about a neck MRI in a child, and we're quickly resolving questions and issues and deciding that we're done here. I couldn't do that by email, and it's better than a phone call because it's faster. I don't have to spend all this time being nice on the phone. I can just say, okay, I'm looking, we're done, and that's enough for this type of communication. We also have a different communication tool where we are a little bit more friendly. We're using that for our virtual huddle every day where we come together at least twice a day in this platform and exchange information of what's going on in the day. We also let each other know if we have to leave and we need coverage for the service for the time that we're gone, so this is a very helpful tool outside of that text messaging tool, and the text messaging goes only to one person at a time. In this application, we can message a whole team at once, so everybody on the team can get this message, and then we have a website, so we have some resources that we share like phone numbers or how to do something, image processing policies and guidelines, something we may have to access throughout the day, and this is on a website for our pediatric neuroradiology team. We put that in Microsoft Teams, so you have to log in to your email account with a password to get to it. On the adult neuroradiology site, you just have a website that's out there on the web. It also has phone numbers and stuff on it, but it's obscure enough that people don't happen to come across it and abuse it, but it's important to have these sharing platforms. How about a team agreement? A team agreement can include all of these items, the team purpose and values, the vision and strategical. Those also are like instant opportunities for team building. It's like a team building exercise to put that together. You can include the budget and resources, and importantly, it should list explicitly what are the measures of success for the team and for individual team members. If needed, you can explicitly say that diversity, equity, and inclusion is valued in the team, and then there can be operating guidelines like how to use the technology. In our case, when do you use the SIPA-compliant text messaging system versus the Teams communication tool? Conflict resolution could be difficult virtually. Some people are probably more likely to avoid confrontation in the virtual world, and so how do you go about that? And then there can be team meeting guidelines and an ethics and code of conduct. I've highlighted the team meeting guidelines because we are participating in Zoom for Any meetings now that I thought it might be worth giving a little bit more detail on that. So meeting guidelines should really include a requirement to turn on the video because that is the only way where you can make sure that folks are on the call and are engaged in the call because you can see them. Unfortunately, a lot of radiology workstations are not currently equipped with video. So if somebody's participating from their workplace, it makes it more difficult. Icebreakers, we use them in in-person meetings, and we may just as well use them in virtual meetings. And here, the point is to break the ice so that people unmute themselves and participate in discussions and contribute, or they use the chat box and write down their ideas. So those are important activities on a meeting call, on a video call. It's good to have the agenda in advance and especially assigned roles in advance. And often on these virtual meetings, it's good to have a dedicated role for someone to monitor the chat box. And then also make sure you develop a culture of taking a break once in a while and asking what's on the chat box to catch up with those people who didn't want to chime in verbally, but they had a great idea in the chat box and they should be considered. And then also consider suitable meeting times. Here in the US, we already have to reconcile East Coast, West Coast, and everything in between. So think about that when you select your meeting times, make it good for everybody. And it could be debated whether professional attire is required. Some people, I guess, have resorted to just wearing something suitable where it's visible and otherwise not. And I don't know, I don't personally have a problem with it, but sometimes it could actually influence your attitude during the meeting. And if you think that's happening for you, then you should probably take the extra step and just dress properly. But if you can behave properly while wearing pajamas, then I think I really don't have a problem with that. Other housekeeping items that are good to just set the rules at the beginning of the meetings and remind folks of turning the camera on and mute themselves when they want to talk and conversely mute themselves and tune out any noise in the background, no alerts from phones and email while they're on the call and remind people that they can use the chat box if that is something that you're monitoring. So the take-home points are that there are many benefits to remote working and virtual teams. It's a great opportunity moving forward. Technology is a big factor and it's worth investing into all the tools that we need to work effectively as virtual teams. So that means not just email and video conferencing, but also other tools for project management, for document sharing, for collaboration, for different types of communications. And many challenges remain, but they can be overcome. There are ways to build trust virtually. It's just maybe takes longer, is a little bit more difficult. If you cannot deal with distractions at home, some people just simply can't deal with it. So then you should try not to be on a virtual team. And then there is the drawback of dependence on technology, which may sometimes not be reliable. So overall, I think a good way to move forward with virtual teams now that we have all experienced it. And I'm very excited about it. Thank you. Hi, my name's Alex Dobin. I'm gonna be talking today about how to make AI part of the team. Unfortunately, I have a little bit of a theoretical task. So I'm gonna be doing things like defining non-interpretative artificial intelligence, describing a framework to consider non-interpretative artificial intelligence, and then provide some examples of how non-interpretative artificial intelligence could improve the radiology team. When I think about artificial intelligence, I tend to think about Star Trek. And with Star Trek, we know that one doctor could take care of the entire ship. So in 2260, Bones was able to use some pretty far out high-tech devices to care for the entire Starship Enterprise. Not much changed. And by 2364, Dr. Crusher still was a single doctor taking care of the entire ship. Her tools changed, her fashion changed a bit, but essentially she was doing the same thing. But fast forward just a few years later, in 2371, a holographic doctor, artificial intelligence, was taking care of the entire ship. There was not an entire human doctor on that ship. So how do we get from 2020, where we're dealing with a relatively yokel type of computer, to 2371, where we have pure artificial intelligence? I would argue that what we're doing now is we're trying to take images of every part of the body and go through and have a computer try to interpret everything. And some of that goal to get to 2371, where we have artificial intelligence, is doing that. It's looking at every part of the body, trying to find every single disease and make a diagnosis. But it's not just happening in radiology. It's happening in opto, GI, ENT, derm, and pathology, as well as every other specialty out there. And it's all happening so that one machine or one intelligence can know everything about the human body. But that doesn't help us for today. In fact, this is pretty close to what we've got today. To get to that point, radiologists or other specialists have to annotate every single image, sort of drawing lots of circles. Then it's writing lots and lots of code. And then it's doing lots and lots of testing. And that's just for one disease or for one imaging study. To get that for every single disease and every single modality, it's lots and lots of circles, lots and lots of coding, lots and lots of testing. And it's doing it over and over and over again. And that gets us to 2371. But how do we get to 2260? Because we're still several hundred years from 2371. But to get to 2260, we still have to deal with the problems of today. So what are our most pressing needs? Well, is our most pressing need something like fracture detection, bone age detection, or brain bleed detection? I would argue the answer is no. I have really good radiologists working in my department in every specialty area that are really good at making all of those diagnoses. So what are our most pressing needs? Well, I would argue our most pressing needs are human interactions. Many of us at the end of the day feel like this or like this. And why do we feel frazzled or angry where we wanna pull out our hair? Because we need help. We need our work to be more simple. And so the goal of everything that we should be trying to do from now on through AI is to simplify. And that really needs to happen everywhere. It can happen in the reading room. And in some ways, our interpretive AI is helping to simplify the reading room. But it should happen in the procedure room. It should happen, we should do things to help simplify the life for ordering providers, for technologists, for patients, and even for the business side of our practices. So what is non-interpretive AI? Well, I define it as an algorithm where the primary output is not a diagnosis. Now, there can be a bit of a slippery slope. And so I'll try to make a little bit of distinction with this. So we think of artificial intelligence on the interpretive side as detecting the subarachnoid hemorrhage. And that subarachnoid hemorrhage, if the output of the algorithm is to say subarachnoid hemorrhage, is interpretive AI. But another purpose to detect the subarachnoid hemorrhage could be to order a work list. In that case, the output is the order of the work list. And so I would argue that, while there is interpretation happening in the background, that this is a non-interpretive function. And I'll give different examples that are not this muddied as far as interpretive versus non-interpretive. So to simplify, I'm gonna start with a few examples. I won't go through each of these areas, but we'll touch on some. And we'll start with the reading room. So this is not my reading room, but my reading room looks very similar to this. And I think one of the first things that you notice about this reading room is all the paper. And what is all this paper for? When I went through, I counted something like 28, 29 different pieces of paper. And I found that there was a total of 28 different pieces of paper hanging on the wall. Well, all of these things are reminders of some sort. They may be a reminder of a phone number to call, a process to follow. There may be diagnostic reminders on here. All this paper basically serves as wallpaper at this point. Don't really notice it. Most times, the thing you need is in front of you, and you don't even know it's there because it's become so much of the backgrounds. So then something like this comes up. And you know that there's something out there, right? There's something, some way to classify it. And if you don't know it, you're not gonna do it. But if you do know you're gonna classify it, you have to search for that wall. You have to look around, try to find that piece of paper. Is it this classification system or this one, this one, this one? You know, some are clearly not applicable. They're not even on the right organ. But all these classification systems are out there. They're all stumbling around in your head if you even know they exist. But what if the computer could come around and say, hey, you need to know about Lyrads. That's the one that's gonna get you. Or actually, no, Lyrads doesn't work here. It's Pretex. This is the one you want. And this is how you do it. Now, radiologists, we're smart enough. We can figure out that Pretex, how to use it. Just teach me that this is what I need, and this is the way to walk through it. So the next step would be to simplify for the business side. This is an example of our schedule in my department. And you could look at the schedule as a human and say, oh my God, that guy Tobin's on ultrasound on Wednesday. The day is gonna be like this. He doesn't answer the phone. So all the phone calls are coming to me. This is gonna be an awful day. On Thursday, October 3rd, I don't see Tobin anywhere on that list. This is gonna be smooth sailing. October 4th, uh-oh, Tobin's on call that time. This is one of those days when we have to scream and yell because everything's gonna be a disaster. And so you could argue that on the business side, a human knows that. Why can artificial intelligence help us to schedule our radiologist resources to use them more effectively to avoid those screaming, poop hitting the fan type days when we have bad people or slow people or people that make the job crazy in one area. If you know that those people make the workspace crazy on those days, either help to schedule around them or schedule people that help to decrease the work effort, make it that smooth sailing to go along with them. Let's look at something from the patient's point of view. Unfortunately, this is many waiting rooms. This is not the waiting room we want. It looks really hectic and crazy. Most people would argue that we want a waiting room like this where there's no one in it. And maybe that we don't even want a waiting room at all. But why do we need a waiting room? Well, we start off nice, right? We have a patient scheduled on a scanner and we have a slot for that scanner on the schedule. And so everything starts off good. The technologist positions the patient. Now, anesthesia is needed. So that slows down our schedule. Anesthesia is doing their job and they're staying on time. But now the technologist has to follow the protocol and look at this protocol. There's like a thousand sequences that they have to run. That's more than they thought they were gonna have to run. That slows us down. They have to call the radiologist for questions. Unfortunately, the radiologist is in the middle of a care conference and is waxing poetic with the care team about what's going on in this chest x-ray. Because he's doing that, he's not able to answer the phone call from the technologist. That adds some wait time. And then they go to give contrast. And unfortunately, there's contrast extravasation. So all of those things lead to delay. And because that's happening on one scanner, this is what's happening in the waiting room. But what if AI could come in and help? And AI knows what's going on or could know what's going on with the schedule. It can predict either that those things are happening or see what's happening, knowing all of the timestamps that are going on in the department. Our official intelligence can say, good morning, we have a backup problem. The waiting room is getting full. We have more patients checking in than checking out and our exams being completed. This is causing backlogs in our schedule. The artificial intelligence could then theoretically send a text message saying, we're running 30 minutes late for your MRR appointment. Would you like to keep your current appointment or arrive later today? I'd like to reschedule and to respond back. No problem, do you mind if we call you now to find a time that would work for you? And that initiates the call. And helps to do good customer service. You could also throw in some traffic. So what if traffic is causing people to be delayed? When accidents been reported, do you want to reschedule because we know traffic's bad? Nope, I'm out and I can take surface roads to get there. Would you like me to send directions? Nope, I'm good. Okay, give us a call if you need help. Again, artificial intelligence can be monitoring all of these things and help our patients have a better experience. And then the final example I'll give is with technologists. So this is something that happens in our department. We have an X-ray of an airway and the technologist gets the X-ray. They go to the radiologist, interrupt what she's doing. The radiologist pulls up the study, says, no, this is not a good study. So the technologist goes back, takes a second picture. She then comes in, interrupts the radiologist again. The radiologist looks at the image and says, nope, this is not a good picture. So the technologist goes back and takes a third picture. She interrupts the radiologist again. The radiologist has to pull up the study and say, oh, we finally got it. You can send the patient. Then the radiologist can go back and finally do her work. That of course leads to anger and frustration on everyone's part. But what if we have the same process? The radiologist is doing her job, the technologist is doing hers. Takes a picture, computer comes in and says, nope, bad picture. Radiotechnologist takes another picture. Computer comes in and says, nope, bad picture. And the final time, the radiotechnologist takes the picture and the computer comes in and says, good job, we're done. The radiologist never looked at those pictures, but the artificial intelligence has said that they're sufficient for interpretation. That leads to everyone being happy with fewer interruptions and fewer time wasted. So those are some theoreticals. Let's get back to the real world. Unfortunately, we have to ask the question, what is real? Is any of this even real? We have problem in that all of these things require a workflow and none of the workflow exists to get the artificial intelligence output to the person that needs it. For example, in the technologist case, we can define and we have defined artificial intelligence to say pass or fail. Unfortunately, the technologist works on the modality, and we need a way to get the output from the machine to the modality in the technologist's space of work. We can come up and create computer pop-ups nearby, but if it's not in the place they're looking, it's useless to them. So all that code that we developed is not going to help unless we can get the workflow to work. So the big why that's out there is why hasn't this happened yet? I will make an argument that we're already going crazy. We're already frazzled. So there is a desire in the department for this stuff. I'll also make an argument that there's a business case for this, a quality case for this, a patient satisfaction case for this, a technologist and radiologist satisfaction case for this, and even an FDA rationale. Interpretive AI needs FDA clearance. Much of this non-interpretive AI does not need FDA clearance. It's not making diagnosis. It's only doing things to make our work easier. So the reason why I think it's not happening is it's not been advocated for. We as radiologists are excited about the computer-making diagnoses, but we're not excited about doing these non-interpretive things. I would argue we need to start advocating for that. So in conclusion, AI can be a productive member of our team. It's really not yet, and it's nowhere near being that member. I believe that we need non-interpretive artificial intelligence solutions now, and I believe the radiology community must advocate for these solutions. I know that's something that I am doing and would love your help in that matter. If you're interested, the American College of Radiology Data Science Institute has a non-interpretive panel, and we'd love your participation. Thank you. Hi. I'm Jennifer Kemp. I'm a prior practice radiologist in Denver with Diversified Radiology, and I'm talking today about stepping out of the reading room. Thanks so much for inviting me to participate in this symposium. So we've been hearing over the years many examples of how and why we should try to step out of the reading room. I first remember this from the ACR Face of Radiology campaign in 2007. Then there was the RCNA Radiology Cares campaign in 2012, and the ACR Imaging 3.0 campaign in 2013. These are all excellent resources to help teach us about reasons why and how we can step out of the reading room. In addition, it seems that nearly every presidential address from both the ACR and the RCNA has discussed specifically how we radiologists can be more patient-centric. But despite this, over the years of discussing this, we haven't made a lot of progress. So we're clearly finding that stepping out of the reading room is easier said than done. So we all know that there are many ways that we radiologists can step out of the reading room and many ways that we can improve the patient-centered experience. The patient experience starts at the time the patient registers and ends at the time they get a bill. And there are many pieces along this road in our radiology world that radiologists can potentially affect. I will be talking specifically about radiology and patient communication and consultation today. So we've already been hearing about teams in radiology. We've heard about highly functioning teams. We've heard about virtual teams, teams in AI, and we will be talking about the business case for teams shortly. And I look forward to that talk as well. I'm going to be taking it much, much smaller. I want to talk about the smallest team we have, and that is a team of two, the patient and radiologist. So the team of two delivering health care at the end of one, the end of one is a phrase coined by a friend of mine, Andrew Dallau, and he further described this as patients want care delivered specific to them, for them, in consultation with them to meet their expectations and on a personal basis by an entity they trust. Patients simply want their care delivered at the end of one. And that's certainly how I want to be treated, and I expect my family members and loved ones to be treated in the health care system. So stepping out of the reading room, why should we do this? It's important for patients. That's really the number one reason. Studies show that there are improved outcomes when patients understand their disease processes, and we are certainly in an excellent position to help explain that to patients when we have their imaging studies to show them. Patients report decreased anxiety after reviewing results with a radiologist, and I'll be discussing that further momentarily. 83% of patients say knowledge of imaging findings helps them take better care of themselves, and 65% of adverse health outcomes are secondary to failed communication. The more we can communicate, the better. The better outcomes will be and the less likelihood of something slipping through the cracks. It's also important for us radiologists. We all know that burnout is very important in the world of medicine these days and important in radiology. The ACR Commission on Human Resources recommends improving burnout by reducing isolation of radiologists, and certainly this can be achieved by stepping out of the reading room to talk to patients, technologists, referring physicians. In addition, 80% of malpractice cases against radiologists are secondary to failed communication. The more we can communicate, the less likely it is that something will slip through the cracks and we might get sued. Radiologists are motivated by autonomy, mastery, and purpose. This comes from one of my favorite books by Daniel Pink called Drive. He didn't study radiologists. He studied highly functioning people and highly functioning teams. But we did study radiologists through a survey done with the RSNA and published in Radiology in 2017 and showed exactly what Daniel Pink's book stated, that radiologists are intrinsically motivated. We asked the question to radiologists, how likely are the following to motivate you to communicate more directly with patients? And the number one reason was personal sense of satisfaction. Over 78% of respondents said that this is the reason they would choose to communicate more directly with patients. It's not because somebody is holding a carrot or stick in front of them or offering five more dollars. It's personal sense of satisfaction. So anytime I talk about radiologists communicating directly with patients, people have concerns about the referring physician. Won't the referring physician be upset? The referring physician might be concerned that they will get increased number of phone calls from their patients. They might get confused patients calling. And is this a real concern? Well studies show that 86% of referring physicians say that patients having access to their results causes either decreased or no change in their workflow. Open notes confirms the same thing. And we know that referring clinicians these days are spending twice as much time doing administrative work as they spend doing direct patient care. So the more that we can help alleviate some of this for our referring physicians, the better off our patients will be. So in my practice, we kind of started this organically. Just like many of you, I would guess my interaction with patients was usually with a fluoroscopic study or maybe an ultrasound, a complicated ultrasound that I need to go and observe or scan myself. I might talk to a patient. And my partners and I found this really rewarding. It didn't happen very often, but when it did, it was really a rewarding part of our day. So we talked about how could we increase these interactions a little bit more. We're in private practice. We can't give every result to a patient. We just have too much of a workload, too many cases to read. So what we decided to do was just with any CT, ultrasound, or nuclear medicine exam, we allowed the technologists to offer patients immediate results when they thought it might be useful. So if they had a patient asking a lot of questions, if they had a patient that seemed quite anxious and also partially dependent on the technologist workload, partly dependent on the technologist personality and comfort level, partly depending on the radiologist workload. So just kind of play by ear, but allowed them to make these calls a few times a day to us. They would say, hey, I've got a patient that I think would be good, who would really like to speak with you. Do you have time? So is this possible in private practice? After talking about our little organic practice of communicating more with patients, talking about it at different ACR and RSNA meetings, talking about it in the ACR Imaging 3.0 case studies, I decided it would be worthwhile to do a research study. To try to prove our findings. And research in private practice is quite lacking, as you might know, yet 80 to 85% of practicing radiologists in the U.S. are in private practice. So it's more relevant and comparable to real radiology. In addition, private practice has 34% higher average RBU production compared to academics. Therefore, a study like this is even more valuable in the private practice setting where seemingly we might be too busy to have these types of communications with patients. So we did a research study that was IRB approved. We had 100 patients. We only offered this to outpatients between the ages of 18 and 89, having a CT, ultrasound, or nuclear medicine study with a history of abdominal pain. Patients were given the option of getting immediate results or traditional results. The patients who chose immediate results took a survey before getting their results to gauge their level of anxiety. The results were then communicated by a radiologist to the patient, either face to face or via phone. And then the patients retook the anxiety survey to see how that communication with the radiologist impacted their level of anxiety. And the radiologist recorded result positivity or negativity. So our results, 78% of patients in our study chose immediate results. I've seen several other studies that ask patients, would you rather have your results communicated to you by your physician that you know or by a radiologist? And time and time again, patients will say, I'd rather talk to my physician. We asked our question in a different way. Basically, do you want immediate or traditional results? And it's not surprising that most patients chose immediate results, given that most patients don't know that we're physicians. Most patients don't know the value that we have to bring. So this allowed us to prove our value. In addition, it's giving care at the end of one, asking each patient, how do you want your results delivered? So looking at anxiety before results, not surprisingly, the majority of patients have some level of anxiety when they come in for an imaging study. And after getting their results from a radiologist, a statistically significant number of patients, 57%, had decreased anxiety, 57% had no change in anxiety, and only 6% had increase in anxiety. So that concern that's frequently brought up that patients may have increased anxiety after talking to a radiologist or getting their radiology results was unfounded by our study. So you might ask, maybe the change in anxiety was dependent on whether the results were positive or negative. The majority of results were negative at 68%, but the decrease in anxiety was similar in both categories. So 38% in the positive results category and 34% in the negative results category had decrease in anxiety. So the decrease in anxiety was irrespective of negative or positive results. Referring clinician response, many people asked me about this as well. Unfortunately, we did not survey our referring physicians, but as an indirect indication, none of the radiologists who participated in this study received any complaints from a referring physician. So the future in my practice, we realize it is feasible to continue to discuss results with a small subset of patients. We will continue to give our technologists the autonomy to decide what patients might benefit from immediate result discussion. And if a patient asks for results, we accommodate this request, giving care at the end of one. So in conclusion, I hope that I have shown that small teams of two can make a big impact at the end of one. Thank you so much. Thank you for joining us today in the QI Symposium, High Functioning Diagnostic Radiology Teams. I'm Ben Wahnke from the University of Rochester, and I will be discussing the business case for teamwork. After hearing this talk, I believe you will be better able to make the case to obtain the funding to implement artificial intelligence or technologies that promote virtual teams into your organization. I will also share my experience in promoting teamwork outside of the reading room. I have three objectives today. First, I will describe the primary elements of a business case, focusing on the concept of return on investment. Next I will share a real world example of business case development and provide some tips on how to make a successful pitch to obtain funding for your QI projects. Finally, I will discuss teamwork and collaboration and share my vision for how these concepts will be important for the future of quality improvement in radiology. I'll start with a familiar sight. Most radiologists work in an environment that looks a bit like this. Large monitors filling a fairly large desk separating you from your colleagues. This is an efficient setup for the task work of reading cases, which makes up most of our day. Communication often leads radiologists to believe that they are working independently. But healthcare is filled with complex tasks. These tasks typically require multiple individuals working together towards the same goal, the definition of teamwork. We work with technologists, sonographers, schedulers, nurses, IT staff, and others, all driven by the common goal of providing consistent high quality care for our patients. But effective teamwork requires communication, and communication in healthcare is often inconsistent. Let's say your team is not performing optimally. You take a moment to reflect, and then it strikes you, an idea that will solve the problem that has been plaguing your team for months. You have to share this great idea with your team immediately. And boom, they are blown away. Let's do that now. So as you sit down to develop the plan for your new project, you realize that you need to convince someone to give you money to make your plan work. That someone is typically in charge of a budget, often an administrator or clinical leader such as your department chair or group president. They are constantly under pressure to rein in costs and improve margins. They also typically have more requests for resource allocation than their budget allows, forcing them to make decisions between competing requests. Their first question when you ask them to fund your new project will most likely be, why should we devote our limited resources to your program? In order for your project to get off the ground, first you need to put yourself in their shoes. You will need to speak the language of health care finance and understand what they value. Financial stakeholders often focus on five elements, efficiency, generating more value with the same resources, cost reduction, reducing expense without losing revenue. Revenue enhancement, increasing revenue or exam volume. While increases in exam volume are associated with increased variable costs, these tend to be lower than fixed costs in radiology, which mean increases in volume often come with a high margin, up to the point at which new equipment is required. Balancing improvements, for example, reductions in liabilities like malpractice claims and intangible improvements such as higher staff engagement that results in lower staff turnover and associated costs of improvement. The most common way to measure the financial impact of a proposed project is to determine the return on investment, or ROI. ROI represents the net project benefits minus the project cost divided by the project cost. The net project benefit is essentially the value added by your efforts. The ROI will result in a ratio that represents how much value you have added per dollar spent on the project. An ROI above one is a positive result, or an ROI above, sorry, an ROI above zero is a positive result, while an ROI above one is optimal, meaning that you have generated more than one dollar of value for every dollar spent. Remember that there are many different stakeholders in healthcare. I have listed a few on the left. These stakeholders often value outcomes that can be more difficult to measure or to provide a specific dollar value to fit into an ROI equation. So how do you avoid having your new project scrapped before it ever gets started? Based on the data, rely on published project results if others have implemented a similar initiative in a similar site. Remember that quality improvement projects are not performed in a laboratory, and local factors can influence to what degree a project will be successful or not. National benchmarks from a registry can also be helpful. If your institution is significantly behind the average institution, you can be sure this problem is solvable. If neither is available, start small with a pilot project. You take a fraction of the resources of rolling out an initiative department-wide, and can be a valuable proof of concept, providing you with the data to make a proper business case in the future. Even a failed pilot project allows you to learn that your idea was not as wonderful as you initially thought, saving time and resources you would have misspent with a broader initial rollout. While making your business case for a quality improvement initiative often relies on data from outside your institution, to be successful you should bring the data home, and focus on the local impact. Start with an audit. You can develop a pre-project baseline while demonstrating room for improvement when comparing to results from other institutions. Closing this gap will be your goal, and provide a measurable estimate of the potential benefits of your project. You should have detailed your plan sufficiently to allow for an estimate of resources needed to operationalize your vision. Be conservative in your estimates. Overestimating the benefits or underestimating the costs will only result in disappointment and ill will towards your project in the future when you do not meet your targets. There is almost always an unanticipated hurdle or two that will reduce the project benefit or increase the cost of your project. While we work in teams every day to provide care for our patients in radiology, departments in most health systems are still predominantly functioning as independent silos. I think the sleepy-looking silo in the shade, second from the left, is probably our department in this picture. But as I said before, complex tasks are best performed in teams. When patients travel from one department to another, optimal care will require effective interactions between members of each department's teams. That is the definition of collaboration, a concept that I view as Teamwork 2.0. I want to share an example of a collaboration QI project to better demonstrate how to build a business case for teamwork. Our safety event reporting system had identified four instances of delayed diagnosis of cancer in a short period of time despite the identification of an incidental finding by the radiologist and the placement of a recommendation for imaging follow-up in the radiology report. A review of the literature revealed that noncompliance with follow-up recommendations is a common problem. We performed a local audit to determine our baseline, 45% compliance. There were no published reports of successful ways to increase this compliance to help us estimate the potential benefits of implementing a program to improve follow-up compliance. So we started with a small pilot project in one affiliated community hospital. Our plan was fairly simple. Flag recommendations for tracking, store them in a database for monitoring, and hire staff to work collaboratively with referrers to improve compliance. We estimated that this would cost $25,000 for a one-year trial. We did not have a sound business plan to bring this to administration, so we considered other stakeholders. Our malpractice insurance provider was looking for ways to reduce risks of expensive malpractice cases, so we submitted a grant proposal. We estimated that we could reduce the number of patients at risk for delayed diagnosis of cancer by at least 50%, cutting the number of severe patient harm events in half. If we could prevent even one malpractice case, it would potentially save millions of dollars for our insurer. This pitch was sufficient to obtain the small amount of funding for our pilot. Our pilot project provided the following results. An increase in follow-up exam completion rate of 20%, a flagging system that identified a recommendation in 1.25% of all reports we dictated, and enough new exam completions to show that the majority of follow-up studies resulted in high-revenue-generating CT exams. We could then apply these rates to the volume of our entire six-hospital health system. This allowed for an estimate of the number of new exam completions we could expect, which when compared to the average revenues for the mix of modalities identified in the pilot project resulted in an estimated $1 million in new revenue. As ROI measures net project benefits, we must first deduct the variable expenses related to generating the additional revenue, which we estimated at $225,000. We had a better sense of the labor and IT costs required for broad implementation based upon our pilot results, which conservatively came in at $200,000. Plugging these estimates into our ROI equation, we get a result of 2.88. That means that for every dollar we spent on this project, we expected to return $2.88 annually. In addition to the positive ROI, our pitch to hospital leadership included estimates of the number of severe patient harm events prevented and the percent of malpractice risk reduced. There would also be additional downstream revenues from biopsies, surgeries, and pathology services, as well as anticipated increase in patient satisfaction. So do you think we were successful in getting the project funded based upon this business plan? My kids think Chuck Norris is the toughest man on the planet. Apparently this is a thing with kids, but even he would approve of this plan. Just as importantly, we were also able to meet or exceed all of our projections in our business plan after implementation. In the last few minutes I have, I want to introduce a relatively new concept to radiology QI. This is what I call Teamwork 3.0, or expanded collaboration between teams of teams. These are called network-based learning health systems, but are often referred to as learning networks or QI collaboratives. These networks are comprised of teams of improvement specialists from many different institutions working together to facilitate more rapid diffusion of knowledge and the development of evidence-based best practices around difficult quality improvement topics. One of the leading examples is the Children's Hospital's Solutions for Patient Safety Network, comprised of 112 sites with the goal of reducing serious harm events by 50% over seven years. This network is designed to focus on prevention, supporting a culture of innovation while sharing data transparently through their network and partnering with other stakeholders. The SPS network was able to reduce catheter-associated UTIs by 59% while reducing variability of care throughout the many participating sites, while the ICN network improved remission rates for inflammatory bowel disease from 58% to 75% in just a few years. The University of Michigan now leads at least 16 different quality improvement collaboratives, but one thing you may notice, not a single collaborative is based in radiology. This is not unique to Michigan. While the quality community in radiology is linked to the learning networking, we can learn from the successes of these collaboratives, some of which have saved tens of millions of dollars in unnecessary health costs through their prevention of harm. In summary, I have defined return on investment, the core building block of most business plans. I have shared my experience in building a business case and setting yourself up to make a project pitch that even Chuck Norris would approve of. And finally, I introduced the concept of learning networks, what I consider Teamwork 3.0. The motto of my institution, the University of Rochester, is Meliora, which means ever better. This is particularly relevant to those of us working in quality improvement who realize that the opportunities for improvement in healthcare are nearly limitless. Please take a few minutes today to think about one way you can improve the way your team cares for its patients in the coming year. I will leave you a quote from Russell Brand. I honestly never thought I would utter those words in a national presentation. The revolution that's required isn't the revolution of radical ideas, but the implementation of ideas we already have. Be willing to study other medical specialties, even other industries. Many of these people have solved problems that we face today. You never know where inspiration will come from. Thanks for your attention and enjoy the rest of the conference.

virtual collaboration

radiology

productivity

AI integration

patient communication

team agreements

technology tools

quality improvement

learning networks