So, this is the research abstract section, Pioneering Pathways to Improve Quality. We're going to hear about some fantastic work being presented by some of our fellows and junior faculty here. If you're in the wrong section, you can sneak out. No judgment here. It's okay. There are a number of these sessions, and it's easy to sneak into the wrong room. I'm Amy Morris. I am the chair of the Education Committee for two more months, and it is my pleasure to introduce these speakers today. So we are going to get started with Pooja Shekhar presenting her project, yes, and I'm just going to turn it straight over to you. We're going to run this fairly tight, so the presentations are going to be eight minutes. There will be two minutes for Q&A, so speakers, I will give you a one-minute warning and I'll stop you if you're running over. I want to make sure that everybody has plenty of time. If we're missing a speaker, we may be missing a speaker, so we'll go to the next person, but we're going to keep that tight timeline just in case they're running from something and they get here late, okay? All right. Any questions, concerns? Good haikus before we get started? All right. I'm going to turn it over to Dr. Shekhar. All right. Thank you. Good morning, everyone. My name is Pooja Shekhar. I'm a third-year internal medicine resident from UConn. So today we're going to be talking about a quality improvement project that was performed last year, increasing PrEP-naught 20 vaccination rate in an inner-city community-based primary clinic, and I have nothing to disclose. So the main objectives for today would be to talk about quality improvement, a model for improvement, the purpose, aims, metrics, analysis, results, and also talk about the limitations and the conclusions of my study. So we know that pneumococcal disease caused by streptococcal pneumonia can range from your sinus infections to including CNS infections, and we estimate 150,000 hospitalizations every year in the United States. It does have a large cost burden. $35,000 have been spent every year for acute hospitalization for severe pneumonia, highest being between the ages 50 to 64. We also know that there are two types of pneumococcal vaccine. One is the polysaccharide vaccine and the conjugate vaccine. This is the slide that's showing the data nationally and also in Connecticut. A lot of people, they do not complete their pneumococcal vaccination the entire series, and it's mostly seen in people of lower educational status, lower income, and also certain ethnic groups. Similarly, in Connecticut, we found these health disparities, and we also identified that women had higher vaccination rates in comparison to the other groups. So we decided to take it as a project to increase the vaccination rates in our clinic, and the site description of the clinic involved predominantly people of multiple ethnicities and also with socioeconomically underserved areas. Less than 10% had commercial insurance. So our aim of the study was basically to increase 20% vaccination rates with PREV-NA20 in a duration of six weeks in our inner city resident-run primary clinic. So we went about the QI metrics. We mostly focused on the outcomes to determine how many percentage of people are vaccinated at the end of our study. So when we looked at our problem analysis, we had multiple barriers. One is the resident barriers. They were not aware of the latest updates of vaccination. The patient barriers, they were just tired of getting multiple vaccines, and this was when we had the COVID boosters, the flu vaccine, multiple vaccines, and they just didn't want to take any more. And we also had some clinic barriers with respect to staffing and also scheduling their visits and not enough time for the residents to actually discuss about pneumococcal vaccine. The first thing that we did was basically initiated a ticket in our EMR to basically update the healthcare modifiers to update whatever the guidelines were with respect to pneumococcal vaccines. The second thing that we did was a pre- and a post-test. So we did the pre-test, we educated our residents, and we did a post-test to see how they learned about the pneumococcal vaccines. And the third thing that we did was this PneumoRx Vax Advisor app. It is a CDC-guided app that is an interactive pathway, and when we click the appropriate fields, it basically gives us the recommendations based on the latest updates by the CDC. And after the pre- and the post-test, we realized that a lot of residents did not know much about the pneumococcal vaccine, and after our educations were at least slightly updated, and we can see this in the response, even though that a number of residents who did take up this test after the education was slightly lower. And finally, I think one of the best methods that basically helped us achieve our goal was these reminders that we sent to the resident BCPs that their patients qualified for the pneumococcal vaccine, and to consider talking about this to the patients when they came in. And we finally collected the data, and we studied around 428 people who were scheduled to come to clinic, and 197 patients did not qualify for the pneumococcal vaccine, and quite a lot of patients did qualify, around 185, and 46 of them did get the vaccine as a result of our quality improvement project. And when we looked at the pneumococcal vaccination rates, we identified at least around 63.3% of the population who did not have a single dose of the pneumococcal vaccine, and as a result of our study, we identified at least around 53.3% of the population that did qualify, and after talking to them, 4.7% of the people had at least like one dose of the previous vaccine, 4.7% of the patients were vaccinated with both PCV13 and the PPSV23, and this was the result. And we identified a significant increase in the percentage of pneumococcal vaccine from the baseline and also with respect to the PREVNA 20. So going into the outcomes, we did achieve 19.9% in a span of six weeks, and we did see a significant increase in female population who did get the vaccine, and the numbers are in the slide. I'm just going to go a little fast. Sorry about that. But looking at the limitations of our study, one of the main things was the no-show rate, which happens to a lot of us in clinic, and a lot of them declined the vaccines, mainly because they were just tired of taking like multiple vaccines, and the other thing was we couldn't do an ethnic group-based study, mainly because in the epic modifiers, we didn't have a lot of races listed, and it was either one or the other. So coming to the conclusion of the study, 19.9% increase in the vaccination rate from baseline, and we did have major screening among our populations, and the no-show rate was indeed like higher in females in comparison to the males, one of the secondary outcomes. Thank you. Thank you for that. That sounds like a lot of work, a big project. I'm going to pause and ask for questions from the audience. We do have two minutes after each. And I'm going to pause and ask for questions from the audience. abstract, what questions do folks have for Dr. Shachar? And if folks are still thinking, I actually do have one. So I was curious about the reasons why the screen patients did not get vaccinated. And so you mentioned, was that anecdotal that they were just too many vaccines in one spot? What could you add to that just a little bit? So I think we're all interested in, we have a lot of patients who would screen positive for somebody who should be eligible for this. And so what are the reasons that they don't, even once we identify that they could be, it's not that we're overlooking it, right? So what's the next step in breaking down those barriers? So I think one of the main things that we faced with our patient population was their mistrust in vaccination because of multiple vaccines that had just released, the COVID pandemic, the COVID vaccines. There were a lot of people that didn't believe in vaccines. That was one thing. And another quite frequent thing that we were facing was because this was performed in October and November of last year, it was also the flu season. And that was also the time when the COVID booster was out. So explaining to the patients that they are supposed to be taking three different vaccines in one visit was a little too much. And they felt that they didn't wanna subject themselves to multiple vaccines. And having them come back for like a nurse visit was another thing that we did try, but a lot of them did not show up for the follow-up visits just to get like a vaccine. So some of those factors potentially being modifiable at a different time, a different season even. Thank you very much. Let's move on to our next presenter so we can make sure everybody gets time. Yeah, we'll go ahead and skip it. Thank you. Let's do another round of applause. Everybody gets two rounds of applause. All right. So second speaker on our list is already walking to the opponent. Dr. Chan is gonna talk to us about her project on IVC filter retrieval. Good morning, everyone. My name is Kathy Chan. Thanks for coming to my presentation this morning. I will be talking about a project I did regarding the impact of an EMR-integrated automated IVC filter tracking system on retrieval rates across a large U.S. healthcare region. I did this project when I was in residency at Kaiser Permanente Santa Clara. I'm currently a second-year Palm Crit fellow at UT Health in Houston, and I have no financial disclosures. As a brief background, between 1979 and 2012, IVC filter use dramatically increased in the United States. In 2010, the FDA issued an advisory recommending removal of IVC filter once protection was no longer indicated. This was ideally between one and three months. And this came as a result of several studies that highlighted the complication of IVC filters. And although filter placements have declined, retrieval rates remain suboptimal. And rates can vary depending on a number of factors. Usually, literature ranges anywhere between 2.6 to 66%. However, retrieval rates that are in the higher range tend to be in single, large academic institutions with dedicated IVC filter retrieval clinics. So for the purpose of this study, we utilized the 2011 American College of Radiology and Society of Interventional Radiology and the 2012 American College of Chest Physician Guidelines. Although these guidelines are slightly outdated, in order to compare to our previous study, these were the guidelines we had to use. Because that study was in 2013-2014. And as we can see, indications can vary and cause a lot of confusion. And really, the only consensus at that time for guidelines for filter placement was proximal DVT or PE with a contraindication to anticoagulation. There are a number of filter complications, including filter migration, filter fracture, filter embolization, vena cava thrombosis or stenosis, and filter perforation. So in 2016, our group published this study regarding filter retrieval rates and utilization. In our previous study, we showed that with physician education and implementation of a novel web-based tracking system, we were able to statistically significantly decrease the utilization of filters from about 13.3 to 8.7 per 100K members. And statistically significantly increase retrieval rates from 38.9% to 54% post-intervention. So the aim of our study was to evaluate the impact of an automated EMR integrated filter tracking system on filter use and retrieval rates. Our primary endpoint was filter retrieval rates, and we had a number of interest points, including number of filters deployed, filters placed for primary prophylactic indications, and filtered dwell time. So in November of 2018, we implemented this EMR integrated system across 22 different northern Kaiser facilities, and we collected patient data who had filters placed between 2019 and 2020. We did exclude patients who died within 12 weeks of filter placement or patients who had permanent type filters placed. So this is what it would look like in the EMR at time of filter placement. The nurse would scan in the barcode for the filter, and a lot of this documentation at the top would automatically insert itself in the EMR. And then in the bottom portion, the physician could fill out any documentation they needed, including the type of filter, the plan for the filter, and the follow-up facility. And this is what our automated IVC filter tracking workbench looked like. It included basically all the information the physician would need to know regarding the filter in their patient, such as the filter-specific data, when the filter was placed, how long the filter had been in there, what the plan for the filter was, and if the patient was on anticoagulation or not. Additionally, in this section, there was automatic one-click patient-physician communication ability. In table one, we can see that in 2019, in the bottom row, despite increasing member enrollment across the facilities, we continue to see a decreased trend in filter deployment. And in table two, we can see that in 2019, there was an increased trend towards IVCs placed that were for ACCP and SRR compliance to about 90%. And then concordantly, we saw a decrease in filters that were placed for only SRR compliance or neither ACCP nor SRR compliance. In table three, we can see that in 2013, first attempt retrieval rates were about 38.9%, increased in 2014 with education and a web-based tracking system to 54%. This further increased to 62.1% in 2019 with the automated EMR integrated tracking system. And although this increase between 2014 and 2019 was not statistically significant, our P-value was 0.06 on chi-squared analysis. This was a trend towards improvement. And we did see about a similar rate of first attempt success at retrieval. And we saw a slightly longer dwell time, about 10.2, 10.8 to 14.9 weeks in 2019. And so we see that with an automated EMR integrated tracking system, we increased filter retrieval rates to 62.1%. However, on further sub-analysis, we did see that of filters placed that were eligible for filter retrieval, our true first attempt retrieval rate was 80.1%. So for example, we had patients who had filters placed during this time period who moved out of state and out of the area who were not eligible for filter retrieval at that time. Additionally, we had patients who had documented decision-making between them and a physician who decided to leave their retrievable filters permanent for a number of reasons, such as they moved towards hospice and they didn't want any further aggressive interventions, such as filter retrieval. And we think this increase in filter retrieval rates was due to a number of reasons, including increased compliance with documentation, reduction in filter documentation errors due to an automated barcoding system, and improved physician guideline compliance with filter retrieval time due to the comprehensive reporting physician workbench and a simplified physician-patient contact interface within an existing physician workflow. Additionally, we saw a decrease in filter utilization, increase in guideline compliance, and a dwell time of 14.9 weeks on average. So in conclusion, we see that implementation of an EMR-integrated filter tracking system was independently associated with increased filter retrieval rates and attempted retrieval rates. We suspect that with further education regarding updated filter guidelines and education regarding utilization of this EMR system, we will be able to further reduce filter utilization and further improve IVC filter retrieval rates. And lastly, this EMR-integrated tracking system was implemented across 22 different facilities, not a single one facility. And it was implemented in a widely used EMR system. So this infers that there is a potential to replicate this tracking system and improve filter retrievals across other U.S. healthcare systems. Thank you, I will take any questions. Thank you. to actually follow up like the scheduling for the removal? Was there a coordinator, a nurse, a physician who looked through the database to get that scheduled and taken care of? Yeah. So in our, in the Kaiser facilities, the interventional radiology department puts all the filters and are in charge with taking out all of those filters. They did have a coordinator to help set up the scheduling. Awesome. Thank you. Other questions? Yeah. So that answer is actually one of my big questions too, was who else was involved in the process? And then the other question I had sort of related to that was it seemed like with the data input initially with the form, there were, there was data that the provider who was inserting the filter had to insert. And then that triggered the follow-up. Am I right about that? Sort of intent to retrieve and things like that. Was there anybody evaluating that or overlooking that part of the initial process, the initiation of entering them into, or was it, is that a chance for somebody to slip through the cracks if the physician didn't put that in right in the first place? Does that question make sense? It does make sense. There were, when we looked through the data, some people who didn't have actually a plan in place, but they still with the automated barcoding system would populate into the tracking workbench. And then each facility had its own interventional radiologist who was essentially the lead at that facility who would be in charge of looking at that. So layers on top, I think that's really the point of both of these questions, not just the program itself, but it takes people and layers to make that work. Excellent. Thank you. Thank you. All right. Our next speaker is Benjamin Parker, who's going to be talking to us about O2 range project. I'll give you a warning, it's seven minutes. Thank you. Okay. We're doing a great job staying on time, everybody. Keep those questions coming. Well, good morning, everybody. I'm excited to have the opportunity to share two pieces of our work this morning. I'm going to first be talking about the development of our Oxygen in Range, or ORANGE guidelines, which was an effort to improve evidence-based inpatient oxygen administration at our center. My name is Ben Parker. I'm a third-year pulmonary and critical care fellow at the University of Pennsylvania, and I have nothing to disclose. So over the next few minutes, I'm going to briefly talk about the risks of hyperoxemia and inpatient oxygen management. I'm going to talk about the barriers that we uncovered at our center to achieving evidence-based inpatient oxygen management, and then I'm going to talk about our Oxygen in Range guidelines, which developed a novel pathway for evidence-based oxygen titration and weaning. So as a word of background, inpatient oxygen management has traditionally focused on avoidance of hypoxemia, and as a result of this, liberal administration of supplemental oxygen to inpatients has become the norm at many centers. Well, in addition to not being necessarily the best use of resources, a number of studies have also shown mechanistic and clinical risks associated with liberal oxygen administration and hyperoxemia. Of course, many of us are familiar with the risk of inducing hypercapnia in patients with high-risk conditions such as COPD, other obstructive lung disease, chronic hypercarbic respiratory failure, OHS, and others. And then perhaps most provocatively, a study in the Lancet in 2018 showed an association between mortality and liberal oxygen administration. So I've excerpted their figure with their mortality outcomes here below. They were comparing conservative versus liberal supplemental oxygen strategies, and you can see that each of the diamonds lead toward favoring less oxygen with relative risks ranging from 1.1 to 1.21 for mortality in the liberal oxygen administration group. To better understand this problem at our center, we performed a retrospective analysis of electronic health record data to better understand oxygen practices. This is data from inpatients in receiving supplemental oxygen in one inpatient medicine unit for one month. What we found, and here we define hypercapnic risk as presence of one of those high-risk conditions. What we found is that compared to incidence of hypoxemic events, we had a much higher incidence of hyperoxemic events ranging from 27% of shifts for the hyperoxemic event to up to near 87% in patients at high risk for hypercapnia. And then additionally, a good number of patients who were receiving supplemental oxygen didn't even have an order for it or based on their hypercapnic risk had an inappropriate oximetry goal selected. So based on this background and this data, we aim to develop a set of evidence-driven oxygen in range guidelines to improve our evidence-based oxygen administration and to empower bedside nurses and respiratory therapists to really titrate both up and down oxygen. To do this, we, with the support of executive leadership at our health system, we built a large interdisciplinary team and launched a multi-year effort to explore this. We started by better understanding our baseline state and using quality improvement tools. We also did a structured literature review to better understand best practices for oxygen initiation, titration, weaning, and oximetry goals. We then developed our orange intervention, which I'll talk about in a moment, and then ultimately we piloted this in two units in the 2019-2020 range and then went live across five hospitals in our large academic health system in 2021, and I'm just going to highlight here, we use a number of QI methods throughout this, and I'm just going to highlight here a couple that we'll talk about in the next couple of slides. So to better understand pain points in this process, we performed an analysis of root causes, which included a survey of various groups of clinicians across our health system, and while we identified a number of pain points, we converged toward three sort of primary targets for intervention here. So the first was we identified significant variation in practice. This included variation in understanding about the evidence of oxygen administration, including variation in assessment for need for supplemental oxygen, hypercapnic risk, oximetry goals, and even ordering practices around supplemental oxygen. This is demonstrated with this figure here on the right. This is one of the questions we asked in our clinician survey where we provided three clinical scenarios and asked, when is it appropriate to administer supplemental oxygen? The different blue bars represent different provider groups, nursing, ordering providers, respiratory, and what I want to focus on is not the right or wrong of each of those scenarios, but just the significant variation at baseline between and within groups about the appropriateness of oxygen use. We also found that interdisciplinary communication between clinicians was limited and varied in amount, venue, and whether it happened at all. And then finally, our institution had no guidelines that focused on the weeding or titration, especially the titration down of supplemental oxygen, which led to sort of a diffusion of responsibility and risk factors. A diffusion of responsibility and non-uniform occurrence of this. So to address this, we developed a set of oxygen and range guidelines. This was a set of formal guidelines that discussed assessment for supplemental oxygen, initiation of supplemental oxygen, determination of oximetry goal, and assessment of hypercapnic risk, and then guidelines guiding oxygen titration, both up and down. And we incorporated interventions that targeted those three pain points. So in order to standardize clinical practice, we developed a set of set range-based goals, 94 to 98%, or 80 to 92% for hypercapnic risk. And note these both have lower bounds and upper bounds. We developed clinical decision support in our EHR order for selecting these ranges. We redesigned our electronic health record oxygen order to be a central source of truth for communicating target range, hypercapnic risk, HOMO2. And finally, we designed these guidelines to empower our bedside nurses and respiratory therapists to titrate with support from these guidelines within the ranges provided by ordering providers. In order to implement this, we recruited a group of clinical champions from across our health system, and we developed a complex dissemination strategy. I've included some examples of what we used here. And just to highlight them, to educate broadly about these new guidelines and practices, we designed role-based training modules that were based on provider roles in our health system and rolled those out. We also made visual representations of these guidelines as a pathway that went into our Penn Pathway central repository for these, and they're all flowchart and easy to follow through. I mentioned the clinical decision support in our new electronic health record order, and you can see a little bit of that there. And then finally, we also developed some eye-catching visual reminders, including these flow meter hang tags to provide just-in-time reminders of what someone's oximetry goal should be while at bedside. In our six months after initiation, this was used pretty extensively with 170,000 uses across our health system. Primarily, that normal goal range was selected 90% of the time, with the remaining 10% being the lower goal range for hypercapnic risk. Qualitatively, we've received very positive feedback from our clinical champions and clinicians, and we're in the process of analyzing quantitative outcomes and efficacy here. We've had some delays due to data availability in our health system. One minute, Ben, you're doing great. And so with that, I just want to acknowledge that there was quite a large team that was part of this, including many people not pictured here, many clinical champions, health system leadership. And I want to just thank all of them for their hard work in making this a reality. And with that, I want to thank you for your attention and welcome any questions, comments, thoughts. I know you all have questions. This is a problem we all deal with, right? So I want to, my first question actually is, how many alveoli did you have to draw explaining hypoxic vasoconstriction? I'm just guessing millions of times throughout the, right? I'm sure. But, so what struck me actually in that, the pretest data, that survey was that 70, only 70% of physicians identified that none of those three situations would be ideal. So plans for a post-test assessment and thoughts about was there discussion about why that is? Because the goal obviously was to focus on RN and RT education, right? But clearly there's some MD education that needs to happen too. Can you talk about that a little bit? No, I think you highlight a great point that I think supplemental oxygen is not as widely understood as we wish it would. There's a lot of variation, I'm sure, even between specialties and providers and education around that. And so I think you're totally right that doing a post-test would be really useful after we've implemented this new order with this, you know, really well-described pathway. So this education, see if that improves and maybe that's an opportunity for us to better understand where we can better educate providers about appropriate oxygen administration. And do you have data on what type of MDs were responding to that question? I don't think we do, but we might. I'll look into that. That's a great question. Cool. All right. Thank you. All right. So we're moving, well done. Yeah, perfect. You got us there. So we're, oh wait, nope, that's you again. Wait, is that you? Yep. Oh, it's you again. Yes. Yes, it is. All right. Whew, okay. Ben doesn't get to sit down. Excellent, all right, well, I don't need to introduce Dr. Parker again, so I will just let him move on. Ben's been busy with QA projects here. So, excellent, thanks for coming twice. Well, thank you for having me. I'm excited to be able to share this work too. So, I'll be describing a slightly different project now. This is, I'm describing our Penn Rescue Project. This was a rapid access pathway that we designed to ensure access for inter-hospital transfer patients in need of specialty care at our center amid capacity strain. I'm again, Ben Parker, a third year fellow at University of Pennsylvania with nothing to disclose. During this talk, I'm gonna be talking about the importance of rapid access to specialty care for patients with critical illness and severe end organ injury, a population that overlaps many of who we take care of as pulmonary and critical care physicians. I'll also be describing our novel rescue program that improved transfer access times for hepatology patients amid persistent hospital capacity strain. And finally, I'll end by describing some of the methods and measures that we think are critical to generalizing this to other disease states and centers. So, as a brief word of background, we first developed this focusing on patients with acute and chronic liver failure. Acute and chronic liver failure is a decompensated state that occurs in patients with pre-existing liver disease or cirrhosis with a superimposed acute insult. This leads to multi or a single organ failure as well as systemic inflammation. And we know from multiple studies that compared to patients with chronic disease alone, those that have an acute decompensation have significantly increased mortality. We know that this acute decompensated state rapidly evolves over the matter of days. And we know that early transplantation for those with moderate to severe disease saves lives. And therefore, we believe that prompt access to specialty care and transplant evaluation for these patients early in their clinical course is paramount. To understand hepatology access at our center, we performed retrospective analysis of transplant center data for inter-hospital transfers as well as bed request data for ED admissions to our hepatology service. This is all focusing on patients transferred to hepatology. And we noticed a number of statistically significant trends. So first, we noticed a significant rise in demand for hepatology. This was manifested both as transfer requests for inter-hospital transfers to hepatology as well as ED admissions to hepatology. This was superimposed on a constrained supply of hepatology beds during this time. So, perhaps unsurprisingly, we also saw a marked rise in waiting time for beds, both for our inter-hospital transfer patients and our ED admissions to hepatology during this time. And we saw a marked decline in transfer completion rates for hepatology immediately preceding the implementation of our pathway. For those hepatology inter-hospital transfers we deemed most urgent, they had a median transfer waiting time of just under 38 hours. And under 51% of these are being completed within 24 hours. So with this in mind, we aim to develop a rapid access pathway for this vulnerable cohort using standard QI tools as well as standardized measures of efficacy and safety that could be generalized to other disease states. To do this, we employed a number of methods to better understand the process for inter-hospital transfer access to hepatology. We employed standard QI and process improvement tools. And from these, we identified three key targets for intervention. And those interventions became the backbone of our Penn Rescue Pathway. Our Penn Liver Rescue Pathway went live in July of 2020. And to assess its efficacy and safety, we performed a before and after study focusing on outcomes of number and percent completed within 24 hours of hepatology transfers as well as transfer time for inter-hospital hepatology transfer patients. And then we looked at waiting time for ED admissions to hepatology as a balancing metric. We analyzed this using interrupted time series analysis with a level and slope change model that I'll talk about in just a slide or two here. So as mentioned, we performed root cause analysis and identified several key targets for intervention. Those three specific targets that we identified were one, variable use of clinical criteria for identifying which patients were most in need of rapid access. Two, delays related to insurance and transplant benefit review. And three, delays related to lack of team or bed capacity. And so I'm illustrating here one of the analyses we did. This is analysis of our inter-hospital transfer data looking at patients where transfer was not completed and looking at the reason for that that was captured. And as you can see in this Pareto chart, several of those bubble to the top where you see insurance delays and bed delays that number two and number three reasons accounting for 42% of lost transfers. And then in our analyses including voice of the clinician and our process mapping, we identified those same pain points that I mentioned. I don't expect you to see all the details of this process map here, but I just wanna highlight that we laid our three key countermeasures on top of those three pain points in understanding the inter-hospital transfer process. And so the three interventions that we incorporated into Penn Liver Rescue were one, standardized clinical criteria for identifying patients who are most in need of rapid access. We did this with the help of our hepatology transplant experts and we focused on data that would be available to us at the time of inter-hospital transfer request. So for this group, that was MELD sodium score or the presence of acute liver injury. And patients then who met those criteria were labeled with a standard nomenclature, Penn Liver Rescue patients. We then worked with our case management and transplant center teams, our transfer center teams to identify ways that we could expedite insurance review for these patients, particularly on evenings and weekends. And then finally, we developed a set of protocolized solutions for situations in which hepatology beds or hepatology team were full in order to ensure that those were not delays, those didn't impose delays for patients coming into our center. In its first 36 weeks after Go Live, Penn Liver Rescue was activated for 107 patients, yielding 85 transfers to our center. And here are those outcome measures and interrupted time series analyses that I mentioned. So just to briefly orient you, these graphs on the left show the pre-intervention period and then the shaded period on the right shows the post-intervention period. The solid line is linear regression pre and post of the outcome measures. And then the dotted line shows the counterfactual or the prediction forward of the pre-intervention trend to show that difference pre-post-intervention. And what we found is that we found statistically significant increases in the percent and number of patients transferred immediately in the percent of patients transferred and then reaching statistical significance by 36 weeks for the number transferred within 24 hours in this vulnerable group. And we also saw a marked decline in transfer waiting time, an immediate decline of 30 hours that was sustained over the study period. Looking at our ED wait time, which we said was a balancing metric, we saw an immediate rise of four hours that then reached statistical non-significance by the end of the study period. And we plan to continue to monitor that as a balancing metric. Clinically, we are in the process of evaluating outcomes. However, anecdotally, the liver transplant team has had its highest transplant volume in history by a large amount in the first year of this pathway being present. So in sum, with this rapid access pathway, we were able to identify and provide access for this vulnerable cohort of patients in need of specialty care. And we're excited to think about how we can adapt this to other disease states that need rapid access to specialty care. We're starting to do so, including pulmonary disease. And we've boiled this down to what we believe are four key components to a successful rescue program. The first is interdisciplinary engagement with stakeholders, including executive leadership, transfer and capacity management leadership, clinical experts who can provide context, and then experts in QI methodology and others that are other necessary stakeholders as well. The second is precise clinical criteria to be able to identify this vulnerable group with some large amount of sensitivity and specificity. Aberration either way risks either under-identifying these patients who need access, or in setting capacity strain, risking access to other vulnerable patient groups. And so I think that's really important, thinking about what data's available at the time of transfer request. Application of QI methods allows understanding of bottlenecks. And here, we've had to innovate some unique capacity solutions, including unique inter-service partnerships and thinking about different places that we can appropriately care for these patients in our hospital in order to ensure team and bed capacity. And then finally, standardized metrics allow for evaluation of efficacy and safety and detection of untoward effects and refinement of your pathways as necessary. And again, there's been a huge team that has worked on this, and I'm honored to have the privilege of presenting on their behalf. And so I just wanna thank the many people on this slide and many others that are not listed. And again, thank you for your attention and open to questions, feedback, and thoughts. Thank you. So this is a really common, so I'm at University of Washington in Seattle and we are at a liver transplant center as well. We have a slightly different set of challenges because our referral center is quite far. So our decision making is a little bit different. The question that I had was about the, I was, the 43% of transfers that were essentially the outside hospital canceled, right? Curious whether there was insight into why they canceled. So was the delay part of that problem? Meaning that there's a, you know, iteration there in that? Right, we hypothesized that probably so. It's difficult to tell because sadly that's like all the data that was captured was, we may be able to go back and do some manual review of those records to see if there's more captured in the unstructured comments, which is something that could be really useful. But yes, we hypothesized that there's probably, that delay is probably a good piece of that bar. And then here, just to note, we filtered out cancellations that were canceled for medically appropriate reasons like this person doesn't need a transfer or this person clinically improved. So this is just patients who met criteria for transfer that didn't end up being transferred. Oh, I see. So that does not include patients who decompensated and died and no longer needed. Yeah, okay, all right. That makes sense. And so, and just for folks who might be interested in this kind of a process, so do you have a, you have a centralized transfer center, like an RN who takes the calls and then feeds it out to the MDs? So there's some layers behind this that were pre-existing. Okay. And then we have, oh, it is time. I had another question, but dang it, it's 9-10. Oh, so many good questions for you guys. You guys are all doing such great work, such cool projects. Thank you so much, I appreciate it. Oh, thank you. One more round of applause for Dr. Barger. All right. And then next coming to the podium, Rhea Rubin, yes? Okay. Who's going to be talking to us about COPD ambulatory practices in primary care? And I'll give you a one-minute warning at seven minutes. I'll just kind of lean over like that so I don't distract you too much. Sounds good. So I'm going to be talking about revitalizing COPD ambulatory practice, which we have affectionately named ReCap in our ambulatory setting. My name is Rhea Rubin. I'm a brand-new assistant professor at the University of Cincinnati, where I just completed my fellowship training and this was really my fellowship baby. But more important than me are really all these people who are the residents who kind of make up our clinic, as well as our nursing staff and MAs. I have no disclosures. So I'm going to talk about the role of improvement science and redesign and how we can improve ambulatory COPD care, talk about the difference between process and outcome measures, and also show how small tests of change can lead to significant improvements over time. So why did we care? Why did we choose this project? We all know COPD is bad. It's the fourth leading cause of death with over 150,000 deaths per year attributed to COPD. It costs the U.S. a lot of money, $15.5 billion. Seventy percent of those are actually from hospitalizations defined as severe exacerbations. And we know that every time somebody has a severe exacerbation, they have a permanent loss of lung function. They have an increased mortality within five years. So we know that these are really poor outcomes. So wouldn't it be great if we could stop people from ever coming into the hospital with kind of these severe exacerbations? So what do we know about ambulatory COPD care? We know that every year we come out with these gold guidelines, which are the best practices within COPD care that have both pharmacologic and non-pharmacologic practices. And we know that when we adhere to them, we actually decrease healthcare costs. We decrease exacerbations. Pulmonary rehab, which is one of our non-pharmacologic treatments, actually helps with symptoms. But we also know we're really bad at adhering to these practices. In a study that was done in 2021, they actually found that in those labeled as gold A and B, only 30 percent really were adherent. And in C and D, only about 50 percent were adherent. So our theory was that if we can kind of look at the gaps within our system, we can stop people from falling through the cracks and kind of embed it within. So every single patient that has COPD ends up getting these best practices within COPD care. So a little bit about our context. So we are a downtown Cincinnati clinic. We chose resident clinic because it's predominantly an urban population with the underserved and also the most kind of inexperienced providers. So it's really a big kind of portion that we could kind of help out. They have 4,141 patients within their clinic. 847 of those carry a diagnosis of COPD. And our clinic is made up of one NP who stays forever. There are 33 different residents. And our system is a little bit different. And every November, our entire residency class changes over. And they're brand new pulmonary care providers. And they stay a PCP for an entire year. There are five medical assistants which have also had a bunch of turnover like every other institution within the past year or two years. And there are 14 different faculty members who staff this half day every, they alternate half days. So our global aim was to maximize outcomes in this ambulatory COPD setting. And if we could prove that we could do it in our resident clinic, then we could prove to the system that we can really do it anywhere. So how did we start in doing this? So first we started off by creating a team. So we had our NP. We had three resident physicians. We had an MA. We had our chief resident. We had the clinic director. We had a clinical informatics fellow. And then we had a quality improvement expert. So, and then we looked at the data. So we looked and tried to find where exactly our opportunities were. So we did a literature review looking at the gold guidelines, but also the data that supported the gold guidelines. We talked to patients and families and asked what they thought that we could do better. And we talked to providers to ask them, you know, where are your gaps? Are there gaps in knowledge? Why are we not kind of doing these best practices in COPD care? And then finally we looked at our system to see what we could really redesign within our microsystem structure so that we could build on these best practices. We then did a random sampling of 30 patients to figure out if this was really a problem for us. Spoiler alert, we found that it was. So we found that in those 30 patients, only 53.3 were actually adherent with inhalers based off the gold guidelines. 3.3% of providers actually talked about inhaler technique. No one discussed pulmonary rehab. And only 18.1 actually talked about smoking cessation. We then did a Pareto chart in where we talked to both providers and family members and patients to figure out why were we, what were our biggest barriers and where would we really focus on. And we found that lack of education as well as also kind of this incorrect diagnosis of COPD which kind of fell into this education bucket were one of our biggest failures. We then did a failure mode effects analysis chart which I don't, you don't need to read, but essentially went through every single process or every single step within the process and came up with failures for each of those steps as well as potential interventions to figure out where we wanted to focus. And then we kind of decided with all of these things what we wanted to actually follow in the short term so that we would know that we would have bigger kind of improvements in the out term. So we focused on appropriate inhalers based off the gold guidelines, pulmonary rehab, as well as smoking cessation. And if we were able to adhere to these things within the short term which were our process measures, then our thought was that we would be able to have these outcome measures which we thought would be increased pulmonary rehab participation as well as increased smoking cessation which everybody knows improves kind of outcomes within this population. Our SMART aim was to increase on a weekly basis from those three metrics and the number of opportunities from 23.2 which was our baseline to 90% which was based off the level of reliability. We came up with key drivers in terms of where we should focus all of these aspects and what are the factors that we thought would help us implement change. And then we used the model for improvement to really help us kind of do this whole process and kind of work in small acts of change. So this was, this is a P chart that kind of shows our story, but we did a baseline that showed that we had 23.2% adherent for the first 30 kind of days or weeks I guess. Our first PDSA was a SMART phrase that enhanced education as well as kind of directed the resident of what to do and kind of all these practices. And then on the very end of this is the long block where there's a whole new PCP class that starts on November 1st. So an entire new class comes in who are brand new primary care providers. Our second one was an educational blitz where we had a half day that just focused on COPD care, why do we care, what are we doing, to really focus on providers and MAs as well as our nursing staff. And then our third was a patient sheet that was actually MA driven. And so during the time that the MA was started when they first got into the room, but then it turned into when they were staffing, would provide a patient with a sheet of paper that they would then be able to fill out regarding their symptoms, regarding what inhalers they were on, and also had an information sheet regarding pulmonary rehab. And then lastly, we started a PDSA that provided weekly feedback to each of the providers so that they were then able to see how they were doing with each of these metrics. And through this entire thing, we were able to go from a baseline of 23.2 to increase our average to about 68%. For outcome measures, we found that those who actually received the intervention were more likely to quit smoking, which was statistically significant. And we also found that those who actually received the intervention were more likely to participate in pulmonary rehab. So our overall conclusions were that we were able to have a multidisciplinary approach. We can redesign our system so that every single patient is able to kind of get appropriate inhaler use, pulmonary rehab participation, and also stop smoking. And that we can embed these practices within our system so that it doesn't become the responsibility of the individual, but rather within the system itself. And we still need to do a lot of work to actually sustain this and make it to our goal based on their level of reliability to 90%. And then I have an appendix of like a thousand things, so I will hold off. Thank you. Please. So I'm a lung transplant pulmonologist, so I come at things from an odd perspective. It's not odd. Transplant's not odd. That's not what I've been told. Yeah. Oh, oh, all right, okay. So is there a place within, this is, I mean, this is spectacular and a ton of work that you did in order to accomplish this, is there a place within this education forum to add in things like referral to advanced methods of treating COPD, lung transplantation, bronchoscopic lung volume reduction within this for the patients who potentially meet criteria? Yeah. So we're actually working on a smart set that actually pops up when the CMS diagnosis of COPD comes up and it has all the inhalers and it has the referral to pulmonary rehab. We do, we're not a lung transplant center, but we do have lung volume reduction and that would be one of the referrals as well as to pulmonary to try and figure that out. But our goal was really just to focus on these measures and then, but that's part of our smart set that's actually being deployed. Awesome. Because this is a primary care clinic. This is a primary care clinic. And so our thought was that if we, and if we can do it in this clinic, we can really do it in any clinic within our system. Excellent. Excellent. I'm hearing so cool back there. Yeah, please. We got time for a quick one. I can project, but that was a really impressive. A lot of, so we learned a lot of them didn't ask about it at all and so a lot of it was actually the sheet what also had smoking, are you still smoking, are you interested in quitting, and just as like a trigger to the provider to say, oh this is something that is like being brought up to the forefront. So that was one of the things and then we just continued to go in every, like I think having the data there that they could see, we did not do anything else other than those two things that really brought that intervention. But we found that basically no provider was really thinking of smoking cessation. Sometimes the simplest interventions can have really big change or lead to really big changes. Actually, and so my question was, it is time, I was just going to say, so food for thought, I noticed that one of the biggest changes was after you did the educational blitz, which that takes effort, right, and time, and so it's always a question of how sustainable are these efforts, and so do you have plans to study this effect over time? Yeah, so we did a pre and a post test, but we wanted to do the post test further out so that there wasn't some sort of bias in there as well, but our plan is to kind of do it quarterly and have a quarterly kind of blitz, and then we also have, we have a weird education structure in general within our residency program, but we have a half day that's just dedicated to COPD for all the interns as well, before they even get there. So there's some ongoing pressure on the education piece, great, okay, well thank you, I want to make sure our last speaker does not get shortchanged here, so let's do a round of applause and bring up our last speaker here. Oh yeah, so if you escape, here, just escape. Should take you back to the list. Got it. So, Hollyann Louie is going to be talking to us, bringing it home as the last speaker for the session. Thanks, everybody, for sticking around. This means a lot. Okay. Dr. Louie, I'll give you a little warning at seven minutes, okay? Okay. All right. Hi. Good morning, everyone. Thank you for being here and listening to my presentation entitled Improving Spirometry Rates in Patients with a COPD Diagnosis. My name is Hollyann Louie. I'm a first-year pulmonary critical care fellow at UCLA. I have nothing to disclose. Jumping right in, COPD is often misdiagnosed or undiagnosed, and spirometry is required for the diagnosis of COPD, but worldwide, it remains underutilized. Increased spirometry utilization can improve diagnostic accuracy and facilitate better disease management of COPD, and this was the foundation upon which my QI initiative was born. I started off by doing a current state analysis, and using Tableau, we identified 10,901 patients who had COPD listed on their discharge diagnosis or their problem list and who had interfaced with our healthcare system in the last three years. Sixty percent of these patients had no documented spirometry or PFDs in the previous five years. After controlling for patients who were still living at the time of analysis, we were left with 56 percent, or 5,528 patients. My goal was to increase spirometry rates in our patients with a COPD diagnosis by 5 percent in one year, which in retrospect was a very enthusiastic goal for a pilot study. So jumping right in, this was my root cause analysis, and obviously there are a variety of different factors that can contribute to low rates of spirometry, but I wanted to hone in on the following four factors that I really devised targeted interventions for, and these included the fact that there were limited appointments for PFDs, inadequate quality spirometry being performed, that there was no standardized process in our medical system for identifying and referring these patients for spirometry, and the fact that many patients were being diagnosed with COPD based off of vague respiratory symptoms alone and response to empiric treatments without ever being referred for spirometry. These are my tested solutions. So starting first with limited spirometry appointments and poor quality spirometry, we created a dedicated screening spirometry clinic at one of our pulmonary locations. It was open for nine weeks on just Wednesday afternoons. Next to address the lack of a standardized screening process by which these patients could be identified and referred to spirometry, we sorted through, you know, the literature and looked for a variety of different surveys and screening tools and ultimately adopted the already existing COPD diagnostic questionnaire, the CDQ, that we translated into an online survey for patients to complete via the patient portal. And I'll talk more about this in a couple slides. Lastly, to address the fact that we didn't have a standardized process for monitoring our COPD patients within our healthcare system, we are working with our pulmonary quality improvement team to develop a COPD dashboard so that we can better track our patients and focus in on their care gaps. Once we identified our target population, we reached out to these patients by telephone to schedule appointments for screening spirometry at our clinic. Once patients arrived to our clinic, they underwent spirometry using an EZ1 spirometer with a trained respiratory therapist. We used prebronchodilator measurements just to facilitate ease of the screening test. Patients also completed the CDQ at the time of their visit. So this slide demonstrates the COPD diagnostic questionnaire or the CDQ. It is a screening tool that's used to select patients at high risk for COPD and identify those that would be appropriate for referral to spirometry. It breaks patients down into three groups, high risk, low risk, and intermediate risk. The high risk patients should all definitely be referred to spirometry. Low risk patients do not need to be referred. And the patients that fall in between may or may not be referred to spirometry just depending on the availability of this limited resource. It is externally validated in several international studies, including from Australia, Belgium, and Japan. In the nine weeks that our clinic was open, we reached out to 119 patients, 21 patients scheduled spirometry appointments, and eight patients successfully completed spirometry. So we improved our rates of spirometry from 44.3% to 44.4% in that short time. Five patients did demonstrate an obstructive ventilatory pattern on their spirometry consistent with COPD, but the remaining patients did not. And all of these patients had a positive CDQ score indicating that they were appropriate for referral to spirometry. This slide focuses on some barriers to spirometry, both logistically and patient identified. So I already discussed how we reached out to 119 patients, but few patients scheduled spirometry and even fewer patients completed spirometry. Some of the large overarching themes that they identified were that some patients, actually quite a few patients were no longer seeking care within our healthcare system, and many had completed spirometry out of network hospitals. Many other patients were preoccupied with other active medical issues, and several were even currently hospitalized. And transportation issues arose time and time again for our patients as well. And in terms of logistic barriers, since this was a small pilot study, it was really just three individuals reaching out to these patients to ask them to come into our clinic. It was myself, my mentor, and our dedicated respiratory therapist. Moreover, we didn't have self-scheduling set up at this time, so there wasn't a good process by which patients could reach out to us to reschedule these appointments if they needed to make a change. And I think this led to a lot of no-shows. And lastly, because this was a pilot study, our clinic was only open for a short amount of time in a very dedicated time frame, just Wednesday afternoons, and a lot of patients just weren't able to accommodate this limited availability. In conclusion, COPD should be considered in patients with dyspnea, chronic cough, and history of exposure to risk factors. Spirometry is essential to make the diagnosis of COPD, but it remains underutilized worldwide. The purpose of this quality improvement initiative was to see if we could improve our spirometry rates with a dedicated screening clinic, but also a screening tool, the CDQ. And the CDQ did confirm that all of the patients that performed spirometry at our clinic would have flagged as high risk for COPD and were appropriately referred for spirometry. Of the patients who completed spirometry in our study, 62% did demonstrate an obstructive pattern on spirometry consistent with COPD, but the remaining patients did not, which just goes to highlight the importance of spirometry in making this diagnosis. And our study does confirm that more work is needed to accurately diagnose these patients with respiratory symptoms, but substantial barriers do persist. Moving forward, so I have some exciting news. We recently had a meeting with our project team, and we're actually going to roll out self-scheduling bulk orders on November 8th. So I'm very, very excited for this, and we're going to start reaching out to our target population first in small groups of 50 to 100 patients, but hopefully thereafter we'll be able to increase those increments based off of patients' interest in completing spirometry and also our clinic's ability to keep up with the demand. Thereafter, once we're able to reach our target population, we hope to reach out to our larger UCLA population as a whole and deploy the CDQ via the electronic health record patient portal. Patients will be able to take the survey if they screen positive, then we'll invite them to schedule screening spirometry appointments at a variety of different PFT locations. These are my references. Thank you very much for your time and attention.

spirometry rates

COPD diagnosis

underutilized worldwide

current state analysis

pulmonary function tests

quality spirometry

standardized process

misdiagnosis of COPD

interventions