ACO = HIE + Analytics: Enabling Population Health Management

Editor's Notes

• #2: I want to thank each of you for attending today. Accountable Care Organizations and the changes required to implement them is a top of mind issue in healthcare. My goal today is to examine the idea that Accountable Care Organizations require a combination of a Health Information Exchange and Analytics to be successful. One of the challenges of the first generation HIEs was a viable financial model due to the limited capabilities offered by those designs. The inclusion of advanced analytics to manage population health, develop risk models and examine clinical outcomes based on their cost will drive a stronger value proposition and enable accountable care.
• #5: Enterprise PortalsBusiness IntelligenceInteractive Design
• #7: Our agenda today is to quickly define what an accountable care organization is, review it’s key components, and define what will make an ACO a success. In addition, we’ll examine the use cases for an ACO, especially the HIE portion – today and in the future. Next, we’ll look at the challenges to adoption and, most importantly, we’ll jump into the architecture of combining the HIE with Analytics to power the successful ACO.
• #8: Let’s start with a definition. The challenge is that often the focus of healthcare organizations is bullet #2 – collaboration and coordination of patient care and they are struggling with the payment and quality responsibility part. The shift from traditional fee for service model to the ACO has many physicians worrying about a return to capitation versus risk and cost sharing. The key idea is managing the health of a fixed population of people, generally on a regional basis. In order to manage population health, lots of data analysis will be required – not the traditional retrospective reporting but a predictive, project the trend approach.
• #9: Notice that when we look at this list of key components of an ACO taken from CMS literature that Healthcare IT is NOT specifically called out but is implied in nearly every item listed. My favorite is the next to the last bullet – Care Coordination ability and mechanisms to support – clearly that is a Health Information Exchange. The last item is interesting as well – payer partnerships. This component could be challenging due to the lack of data exchange between payers and providers in the past other than billing and claims. It is implied in this component that the provider community will be able to produce quality, efficiency and patient satisfaction analyses to make the ACO work – thus HC analytics.
• #10: The reality is that each one of these organizations will need data from outside their organization and the ability to analyze that information. No one will be able to operate independently any more – it is not adequate in an accountable care environment. Today, most ACO discussions are discussing the issue of collaboration internal to the organization and not focused on the external collaboration environment. The debate is in full swing – should the ACO be hospital centric, anchored by a large multi-specialty group practice, led by physicians, independent or, my favorite, virtual. The point is this: there is a place for each type of model, but the key will be the ability to gather, analyze and act on evidence-based clinical information. Many of the current examples are leaders in Healthcare IT and Geisinger Health is well known for their efforts in developing an integrated clinical data repository with a strong data model. Each ACO model has their own idea of how to achieve care coordination and cost effectiveness. Finally, each of these models will do their part to reduce chaos and fragmentation in the current healthcare delivery system.
• #12: Let’s not forget the patients!
• #13: According to Donald Berwick, John Whittington, and Tom Nolan of the Institute for Healthcare Improvement the strategy for improving the U.S. health care system requires simultaneous pursuit of three aims: improving the experience of care, improving the health of populations, and reducing per capita costs of health care. Preconditions for this include the enrollment of an identified population, a commitment to universality for its members, and the existence of an organization (an "integrator, or HIE") that accepts responsibility for all three aims for that population. The integrator’s role includes at least five components: partnership with individuals and families, redesign of primary care, population health management, financial management, and macro system integration. The integrator required is a “Pay for Population Health Performance System,” which goes beyond medical care to include financial incentives for the equally essential non-medical care determinants of population health. The challenges are even greater than Berwick and colleagues realize for the Triple Aim within medical care; they include agreement on population health measures, financial incentives, ways to avoid unintended consequences, coordination across sectors, and resistance to the reallocation of resources – all of which must have set goals, tracking of actual performance and actionable information for continuous improvement. Setting metrics and driving to the metrics will be the determinant of ACO success – the enabling technology will be using business intelligence tools.
• #14: OK, so if we agree that a Health Information Exchange is a key component of an ACO, then what is required? The use case diagram outlines our initial understanding of expectations around an HIE as primarily a clearinghouse of information moving from one organization to another. Note in our earlier view of HIEs, the exchange of population health information was viewed as a Public Health Authority role. While an HIE must meet these basic use cases, there is broader emerging use cases to support the ACO. One of the key trends in HIEs is the shift in focus from state-level efforts to build HIEs to private HIEs built by large IDNs or collaborations of a series of regional IDNs. These regional alliances for HIEs have a common interest – managing their population under the Accountable Care Organization models. Let’s shift from this original view to the current view.
• #15: The current view of the Health Information Exchange moves from data interoperability to broadening the data collection about regional populations to support the type of analysis needed to coordinate care and manage costs. Notice the focus on drawing in external databases as well as the clinical information in the previous slide. The major shift is the collection of data about variables impacting on healthcare costs: environmental factors, social/economic conditions, shifts in lifestyle and behavior, health assessments and biometrics. First, note that data will be oriented toward quantifying the variables impacting an individual’s health to provide predictive assessments to avoid costs. In addition, the amount of data collected and analyzed will be exponentially greater – far below the level of HL7 transactions today. Think of capturing your exercise data and it’s impact on blood sugar, your weight and even mood – like Nike + today and then harnessing this remote monitoring to avoid expensive trips to the emergency room or potential hospitalizations. The challenge will be business intelligence systems that can handle “BIG” data, analyze it and recommend a course of action.
• #16: The Emerging ACO view is that HC analytics will reside at the HIE level, thus ACO = HIE + Analytics. The key is building analytics on the population being managing by the ACO versus today’s individual silos of care settings. The key in this diagram is the focus on risk and predictive modeling versus the traditional retrospective reporting in healthcare. The shift in focus to preventive medicine and holistic view of the person under population health management versus the patient in the past.
• #19: Discussion of Oracle’s “integrated stack” to address these architectural layers.
• #23: Population Health Management in an ACO is a platform that aggregates health data across the continuum of care from disparate systems, creating a longitudinal patient record with decision support, quality measurement, and analytics for population management. Upon this platform, programs are built for specific purpose of managing chronic conditions and high priority population health management objectives. The programs feed knowledge and decision support into applications to manage workflow of existing clinical roles and new roles such as health coach, care coordinator and VP of Population Health.
• #26: Note that key idea that basic HIE data can be combined with demographic or environmental data to create public health views of the managed population and then use the BPM/SOA capabilities to “take action.”