In the evolving landscape of healthcare, understanding the needs of populations and deploying effective strategies for better outcomes is of paramount importance. With the ongoing shift towards value-based care and the pressing need to optimize health resources, having the right tools at one's disposal becomes critical. One such tool, the Johns Hopkins ACG® System, is a vanguard in the realm of population health analytics software. Developed over 30 years ago, this system now plays an instrumental role in a myriad of healthcare settings, from commercial health plans and governments to health systems and large employers, both in the US and globally.
The heart of the ACG system's power lies in its versatility and comprehensive approach. It's not simply a data aggregator; it's a keen-eyed scout that combines a multitude of data sources, providing far-reaching insights that surpass mere medical records. By identifying risk and longitudinally tracking patients, it enables healthcare providers and systems to anticipate, plan, and cut costs—key considerations for any risk-bearing entity in healthcare. More than being reactive, the ACG System inspires proactive planning and strategizing, adapting its functions to your population's unique health needs.
But what does this mean in practical terms? How does the ACG System play out in real healthcare settings, and what results does it enable?
In our recent episode of The Health/Tech Edge, a podcast dedicated to the latest trends and innovations transforming healthcare, Sarah Kachur, the Executive Director of Population Analytics Strategy and Solutions at Johns Hopkins, joined Productive Edge’s Raheel Retiwalla, Chief Strategy Officer, and Zoe Jacobs, Sr. UX Content Strategist, to share her expert insight. Listen to the full conversation or keep on reading the summarized transcription of the conversation to gain insight into ACG System's evolution, its application, and how it forms the backbone of successful population health management strategies.
Zoe: Today we're going to talk about population health. But first, we'd like to know more about you and your background in healthcare. Sarah, you're a doctor of pharmacy and you've been working in population health with Johns Hopkins for more than a decade. Currently, you're leading the strategic vision for the Johns Hopkins, ACG system, which is known for being the leading population health analytics software. And prior to that, you were a clinical pharmacist.
What led to your transition from pharmacy to population health analytics?
Sarah: That's a great question Zoe, and thank you for having me today. I started my career as a clinical pharmacist with training in pharmacoeconomics and pharmacoepidemiology. I was spending a lot of time designing and implementing medication adherence programs in implementing medication therapy management programs and measuring those outcomes. I gradually got to spend more and more time with the population health and clinical analytics teams and migrated over to a role leading population health analytics in 2014. It's been an amazing journey and opportunity to learn and use the combination of clinical and data skills to improve populations in Maryland and all over the world.
Raheel: That's great, Sarah. Certainly, Johns Hopkins’ ACG System has made a huge impact in the industry for many years. I think it'd be great for you to explain the ACG system to our audience who may not be familiar with it.
Can you provide an example of how the ACG system is used in a practical healthcare setting, the outcomes it enables, and how your team supports your clients in its use and adoption?
Sarah: The ACG system is a longstanding suite of population health analytics tools. It's developed and maintained by faculty at Johns Hopkins in our Center for Population Health. It has grown consistently over the years to have a very large US and international presence. We've evolved the tool over the years as population health needs have grown. We started as a set of risk adjustment models migrated into predictive analytics, enriched our pharmacy tools as the importance of that data stream has grown, and most recently have layered in tools, in patient segmentation and social determinants of health.
Our customers use the tool to understand their population's needs and very finely programming to improve the health and outcomes of those populations. They can measure variations in the outcomes of those populations, design best practices, and identify high-risk patients for targeted clinical interventions. We’re the backbone of a successful population health improvement strategy, which is becoming more and more important as the healthcare dollar is stretched and we're moving both in the US and internationally towards more value-based and total compensation-based payments to improve outcomes among patients.
Raheel: I can imagine, as you mentioned, how important population health analytics is and how important it will continue to be to drive a better understanding of an organization's ability to successfully deliver on the value-based goals that they have.
How effectively are the organizations able to understand, adapt, and apply predictive analytics and segmentation concepts, particularly to extract value from systems like ACG?
Sarah: Raheel, that's a really excellent question, and I think you and your listeners know better than anyone that data is only data and models are only models until they're translated into action with a certain level of rigor and measurement. Their data points on a spreadsheet are in a data table. Over the past three years, we’ve shifted the approach of our ACG system, customer support, and implementations, so that we're providing not just data tables and predictive model outputs, but we share with our customers best practices for implementing their care management and clinical improvement programs. We aim to scale our knowledge and best practices approach across countries and across our customers as to how we translate these findings into action. It's very common for our customers to want to know the inputs to a population improvement program and we can add more value by providing not only the inputs but how to monitor that program as well as the best process metrics and the outcome metric that the program is enabling.
Raheel: Looking at the big picture, the impact of Johns Hopkins' ACG system isn't just about advanced technology. It's also about aiding process improvement to ensure organizations effectively utilize it. The system's success lies not only in its sophisticated technology but also in the comprehensive guidance and best practices it offers. It's more than just software; it's about handling complexity and fostering maturity.
You mentioned a little bit about segmentation and social determinants of health earlier. I also read about your new version of the ACG system, V13, introducing concepts of patient needs groups. Can you share a little bit more about what they are and how is it changing how predictive analytics works and how you can really target the cohorts to drive the right interventions and care needs?
Sarah: I'll start by defining patient segmentation a little bit. Segmentation is looking at the entire clinical needs of patient groups within the population and grouping patients according to those clinical needs. So we'll contrast that with risk prediction or risk stratification. Many models that you're familiar with, including many of our longstanding ACG models, will categorize a patient's predicted cost or likelihood of readmission as high, medium, or low. That's really excellent when we're thinking about devoting our resources to the highest-risk and highest-need patients. So we'll always have those models out there and they're always a great starting point. We've moved into a segmentation approach because the label of high is not always clinically definitive enough to define what a patient really needs as a holistic person when we think about their healthcare needs.
None of us can be defined by labels of high, medium, or low. So why would we expect our patients to be, it's not a clinical view, right? When we think about segments of a population, which is really now becoming a best practice, particularly for complex populations and those in the Medicare and Medicaid space, we're thinking of patients who have a defined set of needs and how we best meet those needs. Are some patients frail? Do some need designated support for social determinants of health improvements and services? Do others need basic disease management support to manage their diabetes or other chronic conditions? So we've migrated into that approach through the Johns Hopkins patient need groups. It's 11 defined clinical needs groups within the population that allows our clinical partners and those implementing the ACG system to really refine the services and the care pathways that they're providing to certain groups of patients. It enables better outcomes and it enables more transparency when we're asking clinicians to take action.
Raheel: With considerable investment and strategizing going into understanding social health determinants, structuring community programs, and designing targeted interventions, it's challenging to track efficacy and maintain a close feedback loop. The ACG system's support in this is noteworthy. Can you share examples of organizations that have applied these PNGs and seen interesting results?
Sarah: We have some very good partners that have applied the PNGs within their care management strategy and I think the value that they're seeing is that exactly what you described, experiencing challenges in measurement of care management success. Is it a black box? How are we really benefiting patients as we move into the next era of patient outcome support and improvement within a value-based care model? We need to establish those defined pathways, and then we need to measure process points along those defined pathways in order to ultimately achieve our measure of a successful outcome. So what might that look like? Let's say for a frail patient, we wish to ensure that they have all their meds, they have caregiver support, and they have D M E or other in-home services as needed. So we designed that into our clinical pathway for those frail patients. We can measure each of those process steps. We can identify where we better devote our resources, how we mitigate common barriers, and then we can measure the alternate outcomes that we're looking to achieve.
Raheel: That's phenomenal. As organizations adopt and embrace a lot of the best practices that come along from Johns Hopkins, I can't wait to hear the stories coming from that, quantifying the impact. What also excited me, by the way, is a recent blog article that I read on your website that talks about how providers can leverage the ACG system at the point of care. Can you shed some light on how the ACG system enables the application of insights ranging from population, now to an individual ,and how do you imagine that actually happening at the point of care?
Sarah: Absolutely. The migration of ACG and other analytic tools into the front end of the EMR is going to be one of the largest drivers of change that we see over the next few years. We can't expect people to bounce around between systems, and that's especially true of primary care physicians and specialists who are already stretched thin in terms of time and resources.
Many of our Epic-using customers and others who have ACG data available in their EMR see select ACG outputs at the point of care, we don't need to show everyone all the measures. So we've made targeted selections of those that are most relevant to clinicians and other office staff within the outpatient setting. Typically we'll be able to see a measure of likelihood of readmission for those relevant patients as well as surfacing opportunities to improve care that could be addressed in the office setting. That may even be with someone who's registering the patient. You know, have they had a lack of primary care? Do they have an opportunity to improve medication adherence? Do they have known social determinants of health barriers that they've reported to a different provider within the health system? And in some of our more advanced implementations, including our implementation here at Hopkins, some of those flags drive best practice alerts that can trigger certain referrals right in the office setting.
Raheel: That's unbelievable and so great to hear. Out of curiosity, have you worked closely with the EMR organizations to ensure that these insights show up at the right screens in the right places?
Sarah: Yes. We work very closely with a few EMRs. The largest are Epic as well as a nice piece of functionality that's available within eClinicalWorks.
Raheel: It’s that level of collaboration and ensuring that the last mile outcomes are enabled through the type of insights that are being generated that’s so great to see. The ability to influence care through best practices, improved patient engagement, communication, and integration with EMRs is crucial. This also enables smoother processes like prior authorization, further enhancing patient support and care.
I'm curious, what are the unique advantages of the ACG system compared to other population health analytics? There are a lot of emerging models and approaches. I'm curious how you see the market evolve and what you see is a more sustainable approach to drive a lot of the things that we're talking about here?
Sarah: The ACG system is a whole patient clinical approach to modeling. We look not only at dollars and utilization but the overall patient needs frailty, medication needs, and the presence of multiple chronic conditions cannot be understated. It’s really tempting to look at a patient as a diabetic or a patient with hypertension. And we know in clinical practice that that's simply not true, particularly when we think about high-need and high-cost patients, those who have more intense social needs patients, who we tend to see in a Medicare population who might have end-of-life needs, one chronic disease is manageable, two chronic diseases is manageable, but patients get to a certain point and have a certain number of medications mm-hmm. Where their risk just starts to multiply exponentially. So this clinical view is important to understand the holistic needs of the patient and drive action.
There are other models out there that are really excellent financial models. They're good for that purpose but sometimes it's hard for a case manager or a clinician to understand what's driving the model when we're looking solely at dollars. So there's a little bifurcation in the market of the clinically oriented models and the financially oriented models, although there are a lot of great options out there. I think there's a wonderful model and wonderful technology to create very accurate models. I would urge your customers to think about the maintenance load and what's the plan for maintaining and translating that model into action. Many of the larger commercial models are updated quarterly. We have new and changing codes, and practices. If somebody in your organization is building a custom model, but they're not talking about how to maintain it, you have to realize that it has a 12 to 24-month lifespan.
Raheel: Very true. The models have to be upskilled and maintained, as you said. There's no such thing as a model that will just continue to deliver what is expected. You have to understand whether it's shifting its recommendations, whether all the factors that went into it are changing fast enough for it, and whether it needs to be updated and scaled.
Zoe: Could you give us a little insight into the future of the ACG system? Any exciting updates or developments that we can look forward to?
Sarah: Sure. As we continue to evolve the system and have some exciting enhancements planned, we will continue to grow our investment in social determinants of health functionality. We've made some initial movement within our latest release that came out about a year ago that's had really great adoption. It's called ACG GO Health. From that, we've learned a lot from the market and a lot of new customer needs and new customer requests, particularly as we see the direction of CMS and these state Medicaid programs migrating more and more into requiring health plans and health systems to take action on social needs and social determinants. So we'll continue to grow that functionality as well as making it easy for customers to understand the impact of social determinants on their populations.
We’re also making some investments in usability and transparency to better enable clinical teams to take action. Additionally, we continue to see that across the industry. So we'll be making some enhancements in that direction within the next version.
In the continuously evolving landscape of healthcare, our tools must grow and adapt alongside us. As our conversation with Sarah Kachur has highlighted, the Johns Hopkins ACG® System is a clear testament to the power of innovation in meeting our shifting needs.
Its robust, comprehensive approach to population health analytics is not merely about aggregating data, but about enabling healthcare organizations to anticipate, strategize, and adapt to the unique health needs of their populations. Through its integration with EMRs and its focus on holistic, patient-centered modeling, the ACG System helps turn data into actionable insights. It's about much more than identifying risks—it's about proactively managing them and setting a course for better patient outcomes. But while there are a wealth of wonderful tools and models out there, we must be mindful of how we maintain and translate these models into real, sustainable action.
Navigating the complexities of today's healthcare systems requires not just effective strategies and understandings, but also the right tools to execute them. As we move further into an era marked by value-based care and optimization of health resources, the need for powerful, versatile, and adaptive tools like the ACG System will only grow. As we've seen, the right tool can be a game changer—it can transform the way we see, understand, and respond to our population's health needs, driving the future of healthcare into a brighter, healthier tomorrow.
To learn more about how to transfer your clinical data into the cloud and unlock the advanced analytics, artificial intelligence, and machine learning tools you need to innovate, streamline processes, and improve patient and clinician experiences, download the free white paper from our partner, Redox, on Unlocking Hidden Data Roadblocks of Cloud and AI Adoption in Healthcare.