Welcome and thank you for allowing us to participate in the ACC Quality Summit. Good day. My name is Kenneth Schaefer. I'm a pediatric cardiologist in Austin, Texas, and I work with Pediatrics Cardiology as part of their quality collaborative. I also sit on the ACPC QNET steering committee and was asked today to present some of the information on QNET and how we have leveraged technology to improve quality improvement. Developing, implementing, and reporting quality measures for ACPC QNET has been an iterative process for many, if not all, of our partner institutions. Matt Park and I have had a fair amount of personal experience with leveraging electronic health records to improve compliance with quality metrics, both through better documentation and smart reminders. Technology has also had the promise of decreasing administrative burden for reporting on these quality measures. Today I will discuss how QNET participating sites have attempted using technology and how implementing these changes have been a QI process in and of itself. The benefit these institutions are seeing in metric compliance and the ability to report on increasing numbers of metrics appears to justify any growth experience. With limitations of today's format, however, I will be presenting on behalf of Matt as well as myself. I also want to extend the thanks to our QNET partner sites who have provided their experiences, which I will describe later. Today we will begin once again with reviewing the Adult Congenital and Pediatric Cardiology Section Quality Network Program, QNET. I will discuss a little bit about its history and development. I would like to review our experiences from a recent learning session and then discuss additional experiences with EHR modifications, both across the program and within our individual institutions. We will finally wrap things up by discussing how these experiences will allow us to build a better mousetrap. Our goals are simple. We hope you will understand our previously documented best practices for submitting data to QNET, but also appreciate how technology advances can improve quality as well as quality metric submissions. I figured we would start with a bit of a QNET refresher. QNET developed out of the Quality Improvement Work Group Activities for ACPC. The ACPC leadership aimed to develop a series of quality improvement metrics that would be specialty-specific. We aimed for items that reflected congenital issues and were age-specific. We appreciated that ambulatory cardiologists are a large part of our provider population and that IMPACT, SDS, and other registries and QI initiatives did not assess or reflect their work. But most importantly, we felt that if we present and validate metrics, we would not be told by third parties what quality is. The experts set the expectations. So ACPC started with a quality work group to develop quality metrics. These precursors of QNET metrics were developed under rigid evaluation tools, implemented to vet the full set of metrics that would become QNET. Our early team was appointed in 2015, and since 2016, many of the institutions contributing today began to submit data on a self-selected subset of our initial 24 metrics. Growth from the early pilot sites was aggressive over the first couple years, and this program entered the NCDR portfolio a little bit more than two years ago. At our peak before COVID, that is the year that we entered NCDR, we hit 53 participating sites across the United States. There was some attrition during the pandemic, but programs have begun to expand once again, including international participants. As stated previously, the metrics were developed based on clinical importance to potentially any person who cares for patients with congenital heart disease. Our target was for metrics important to large programs, small programs, academic, and private ambulatory only practices. Early targets included common indications for cardiologists evaluations like chest pain, diseases that had very specific guidelines published like Kawasaki disease, or even defects that were commonly managed by practices across the scope of congenital care. Data collection was meant to be simple, numerators and denominators. A site could either submit a random sample or data from the total population as a denominator, and with this straightforward formula, the number of patient touch points grew rapidly. Again, the COVID pandemic had altered the trajectory of the curve, but early 2022 data appears to be on track to approach or surpass the 100,000 touch point mark. And with growth of the program, we have developed a mechanism to identify high performers. The participation awards were first awarded in March of 2021 and identified institutions that were able to submit an increasing number of individual metric data. Since QNET inception, however, we have leveraged high-performing institutions to share best practices with sites that are looking forward to improvement. The QNET learning sessions are great examples of how we have worked within our program to improve results. A theme of ACPC, as well as QNET, has always been to share our successes across participating sites. This seamlessly share, shamelessly steal approach has a format of creating learning sessions where we can discuss and implement experiences from all of our high performing practices. At our most recent learning session, we focused on metric 20 and heard from leaders of several participating sites as listed above. Multiple QNET partners presented their experiences that included manipulation of EHR and technology to optimize metric compliance for data collecting or reporting. All participants presented best practices that have led to improvement in quality metric compliance. Many institutions use the above key driver diagram to develop interventions. At the learning session, each described its own set of enhancements and how they affected the process for metric 20. For today's presentation, I will briefly describe and review some of their experiences. Dr. Susan Salib, the moderator of this session, described how Boston has experienced early improvement in metric compliance and a steady increase over the series of quarters since they began in 2016, with their most recent data harvested in 2021. There were many digital as well as non-digital interventions. We have a couple key items to share. With early chart reviews, it was identified that much of the data for the 22Q11 deletion truly did exist, but identifying its presence and location remained difficult in spite of a robust EHR system. Early EHR modifications included simple improvements that re-identified data points to enhance extraction. Check boxes could then be used for data search to better identify numerators. These enhancements, in turn, allowed improved clinical communication by reporting data and capturing the geneticists' narrative. Simple check boxes could then be added to standardized patient touch points during the evaluation and management process in order to increase awareness and compliance with the goals of the metric, genetic testing in Tetralogy of Fallot patients. Dr. Eason in South Florida had similar experience in his practice, which was also identified as a high performer for this metric. In their case, they had early success in comparison to the QNET average. While experiences with the EPIC EHR system have been troublesome for some practices because of the limitations in its ability to modify the product, his team was able to build smart lists embedded in their core templates, which automated genetic testing as an option as the documentation developed. Dr. Reballedo in Memphis described his ability to demonstrate early success with enhancements that focused on standardized documentation. Manipulations within the Cerner EHR product led to development of a nursing ad hoc form to collect data and or notify the clinicians of the need for genetic testing. This provided clinical documentation as well as discrete fields for data extraction. They did reach limitations, however, on the ability to modify standard templates within the Cerner system, and although the target was a discrete field for genetic testing, the only potential solution was a free text box and education for providers to document in a substantial or similar way each time. And within the system, the genetic results would appear with other important medical history. Dr. Sabati's experience in Phoenix demonstrated additional issues with EHR modification. Sometimes the most simple asks for our IT support team can become quite difficult. In this case, they requested a straightforward pop-up to identify patients with Tetralogy of Fallot. Finding the correct people to implement this change was very limiting to his team, and having IT support for the process was a limitation in and of itself. As we examined EHR modifications more broadly across the QNET community, certain experiences define themselves as more common between different institutions. EPIC, for example, has an extensive market preference, market presence, but it's, but the inability to modify it significantly made it difficult to manipulate. Best practice alerts were one of the few items successful at Cincinnati Children's, for example. Cerner at Children's National has been limited to modifications that included dot phrases with the goal eventually of developing pop-up reminders. A couple institutions, however, have been successful at modifying ecosystems, and in fact our institutions have personal success in partnering with the VitaStar reading system to automate the ability to identify QI candidates, as well as to collect data and complete assessment providing output for the QNET program. So what we have seen is that our many participating institutions have many brands of electronic health records, and collaboration is needed at the general design level, since integration into various EHR platforms will have different approaches. Our superior and active participants seem to be the ones that are most successful at leveraging these tools more extensively. To explore a bit how EHR enhancements truly can help, we have taken advantage of participating practices that have common EHRs. Pediatrics is a group of practices that share economy of scale related to their EHR. All practices participate in development and enhancement of a single EHR platform using the NextGen product. With this, I will focus on some personal experience within our institution and our sister institutions in pediatrics. Pediatrics cardiology is a series of multiple independent practices that are integrated with a national group in order to leverage economy of scale. One of these opportunities that presented early was a system-wide standardized EHR that could be enhanced as needed for congenital heart care specialty. NextGen EHR was selected because of our ability to have internal programmers and IT specialists who could implement enhancement requests and modify the product specifically for our practices. With respect to quality improvement programs, we use the model of Plan-Do-Study-Act to implement enhancements that improve data collection, data management, and reporting, or all three. The Pediatrics Cardiology Quality Collaborative, formerly known as MedNax, includes 17 cardiology practices and over 130 clinicians. This year there are nine sites enrolled in QNET, although 17 practices are part of our QI program. In 2016, during the early years of QNET, our collaborative met and ranked all available metrics for EHR development. We used a grid that included the ranking of clinical interest, any potential impact of the data once it was available, previous performance in the metric, predicted interest of other agencies like payers, and the potential for data extraction ease and management. Over the first six quarters of QNET data submission, our process defined itself as successful by increasing the average number of metrics submitted per practice from our affiliated programs and surpassing the average number of metrics submitted from other programs in QNET. Our EHR modifications leveraged some key opportunities. We looked for discrete data fields that could be entered and measured. Many of the congenital heart disease diagnoses were variable and difficult to set as denominators. We leveraged both smart reminders and hard stops to improve data input. Data reports were to be agile and available to practice leaders as well as quality improvement team members. And finally, for data fields that were too variable, we instituted denominator reports that would allow us to identify patients and generate random chart reviews for data submission. We did experience many other issues, however. Our early successes existed because of a hybrid approach of manual data collection using denominator reports and some automated collection related to the EHR enhancements we implemented. Through the years, though, we've experienced uncontrolled factors that affected our ability to measure and report data as well as our ability to enhance the EHR. About a year and a half ago, the entire EHR was rebuilt and all resources for enhancements related to quality improvement were sacrificed for this major project. Recently, however, there has been redeployment of these resources and with that we have developed an elegant dashboard for reporting as well as enhancements for recording data itself. Since we were previously discussing Quality Metric 20, I will share with you our personal results. Our initial experience with this metric related to problems with documentation and it did not necessarily translate into the clinical presence or absence of 22Q11 testing in our patients. Our documentation issues placed our practice at a very low performer range in comparison to the Q-net average, a approximately 30% compliance. A series of EHR enhancements allowed us to approach average, but as you see in 2020 there was an extreme variance. This was the time that the EHR was reconstructed, factoring all data collection improvements that had been made over the previous years. Fortunately, previous experience with enhancements allowed us to build rapid recovery. As we emerged from both the pandemic and our EHR rebuild, we can identify certain steps that were important in maintaining success within our program. First, we have weekly, sometimes bi-weekly, meetings that allow us to reassess ongoing enhancements. With that, our enhancements have to be relatively agile to make these meetings important. Our IT partners have had to learn details of the Q-net metric specifications and begin to speak our language, and we have learned how to make requests speaking their language. This has provided us with some out-of-the-box solutions. For example, recently we were able to leverage natural language processing in free text fields to search for the word fever, febrile, and afebrile in Kawasaki patients during follow-up visits. One new, very important item has been the development of dashboards that provide results at the discretion of the provider and provide immediate feedback if needed. This has rapidly decreased our PDSA cycle lengths to improve our enhancement times. With this format, it may be somewhat difficult for me to walk you through our dashboard. The dashboard is on the right side of the screen and can be accessible during any patient encounter. We currently have 10 metrics that have been developed for EHR data collection and reporting within the environment. For any given patient, compliance with the metric during the visit can be visualized so that potential EHR discrete fields can be modified for compliance, or as we look at the metric over time, we can change background algorithms so that the clinician modifications don't have to become too cumbersome. We also have the grid at the bottom that reports the previous quarter available numerator and denominator, allowing us to speed up our submission process at the end of the quarter. So with these examples, I believe we have demonstrated a couple key items. In working with EHR enhancements, they are most successful when you understand how your providers are documenting in the EHR and you attempt to make EHR data collection and reporting reflect their natural processes. The work with our IT partners for EHR enhancement will frequently require engaging these experts by speaking their language and making them stakeholders in the process such that out-of-the-box ideas can be shared openly. Next, as usual, the most elegant solutions are generally the most simple ones. And finally, the process needs to be a stepwise, iterative process that leads to small changes with each enhancement. QNET has had a fair amount of success with our learning session model and will continue to use high-performing sites to share their pathway to success. This will be even more important as EHR modifications can be extended across different EHR platforms. Previously, we have attempted a workgroup model with the idea of sharing modifications between different groups. Unfortunately, this is very difficult with the variability in EHR products. Potentially, an enhancement repository can be developed where an enhancement in one EHR product could be used to demonstrate possibilities in another EHR product. This, of course, would require IT development involvement. One final potential would be some type of certificate of EHR products as a QNET or NCDR preferred product based on their ability to be modified or willingness to be modified. Hopefully, this will encourage you to work on building your own better mousetrap. As we work within the quality improvement processes for ambulatory as well as inpatient cardiology practices, leveraging the EHR and figuring out how to leverage the EHR will become important for data collection as well as data reporting. Thank you very much.

EHR

quality improvement

compliance

documentation

collaboration

technology