Four Key Technologies for Physician-Led Accountable Care Organizations

The idea of aligning payer, provider, and patient interests around the 3-part aims of better health, better care, and lower costs is at the heart of the accountable care movement. In its most fundamental form, an accountable care organization (ACO) creates financial incentives for providers to try to reduce the total healthcare expenditures for a panel of patients while at the same time improving quality and patient experience metrics. Unlike the unpopular managed care reforms of the 1990s, which focused on achieving cost reductions through limiting patient choice and access to services, the ACO movement seeks to achieve cost savings by providing the most appropriate care for each patient—ultimately resulting in higher-quality care—and in many cases, the patient is not restricted in their choice of providers and services. While these factors introduce greater difficulty in achieving savings compared with managed care, an increasing number of provider organizations are taking on this challenge, at least in part due to the belief that the modern information and data tools that have transformed our economy could be brought to bear in healthcare.

The meaningful use of electronic health records (EHRs) and the structures put into place for patient-centered medical homes form a critical infrastructure for accountable care, but they are not sufficient. In this article, we describe 4 technology-dependent areas of innovation that we believe are key to the success of ACOs. At the Brookings ACO Learning Network, 17 physician- led ACOs are working together in small teams to implement 1 or more of these new competencies, and through an open dialogue, they are sharing their experiences with other workgroup members such that all can benefit from these “experiments” in health delivery. We intend to present the preliminary results of these “Innovation Exchanges” at the National ACO Summit, held June 18-20 in Washington, DC, and hope to publish the key lessons learned from this project in the form of a “road map for physician-led ACOs” later this fall.

1) Risk Stratification: Description > Diagnosis > Prediction > Prescription

The movie recommendations we receive on Netflix, our likelihood of being approved for a loan, our chances of landing a job interview, the advertisements that are targeted to us, how political campaigns speak to us, and in many cases the very screen that loads when we browse an e-commerce website are all customized to us based on a sophisticated analysis of multiple streams of available data, from our historical activities, our social networks, our responses to questions, even the Web browser we use. We are slotted into 1 of a multitude of behavioral segments along multiple dimensions.

By now we are all familiar with the premise of “hot spotting"— the idea that a very small percentage of patients account for the vast majority of dollars spent on healthcare—and that efforts to control costs should focus on managing these patients in a more coordinated fashion. However, it is not enough to focus on patients who are the current high-cost utilizers. Many of these individuals are going through a time-limited crisis which will be resolved without special intervention (for example, a patient with a high-risk pregnancy). For many others, there may be little that can be done to affect the course of illness. What we seek, therefore, is to identify the population of “susceptibles” who are showing a pattern of behavior or characteristics that may indicate impending clinical decompensation and a relative lack of resilience—both physical and psychosocial. Providers who take accountability for the total cost and quality of care will use data and analytics to move beyond the current paradigm of the “risk score” and toward an incremental progression to understand the different pathways by which our patients fail—and by which we fail them—be they end-of-life agonies, social isolation or family dysfunction, substance abuse, or mental health comorbidities. Such information will not be obtained simply by combing through claims data or the EHR, but instead through more creative approaches such as asking physicians and even patients themselves what barriers and challenges are preventing them from achieving their health goals. Through better understanding of our patients, we can target our care management interventions most appropriately, and with the highest chance of success.

While the concept of using predictive analytics to segment patient populations is not new, ACOs provide an opportunity for significant improvement over prior, often health plan—led “disease management” efforts. The ability to supplement utilization/ claims data with clinical impressions and patient interview information collected at the clinic will undoubtedly provide additional predictive power. More critically, the in-person interaction with patients and their caregivers, and the clinical relationship, provide stronger opportunities for effective “prescriptions”.

2) Advanced Network Management

While many efforts at quality improvement and practice transformation focus on improving work flows within the practice, ACOs must also be able to create “flight plans” for their patients that ensure they receive high-value care across the care continuum. This includes crafting a network that provides care at the lowest cost/highest value setting, but also creating (and enforcing) expectations with specialists, ancillary providers, hospitals, and postacute providers. These “compacts” may include mutual commitments to operating according to shared care plans, timely electronic communication, and coordinating more closely with the primary care provider. Being able to craft (and prune) a referral network requires analytics that “tier” existing or potential partners on the basis of cost (eg, facility fees) but also utilization patterns (eg, adhering to “Choosing Wisely” appropriateness guidelines). Enforcing these compacts will require the ability to capture and report on metrics relating to expected communication and coordination behaviors.

3) Event Surveillance

ACOs must understand the actual “flight patterns” of their patients outside of their practice. An essential technology that underpins effective accountable care is the capacity to move from retrospective accounting of “leakage” toward prospective surveillance for high-valence events such as emergency department visits and hospital admissions, discharges, and transfers (ADTs). There are a number of different technical resources for obtaining these notifications, from regional or statewide health information exchanges, to direct interfaces from hospital systems, to emerging commercial providers of this service. Needless to say, the form and manner of the notifications will have to be a fit for how the ACO plans to take action; for example, a discharge alert can enable a primary care practice to ensure that every discharged patient has a telephone follow-up within 48 hours and an office visit within 7 days (which would qualify for a $250 “transition in care” payment from Medicare).

4) Patient Outreach and Engagement

A key difference between accountable care and managed care is that patients are not limited to a given provider, and attribution may even be retrospective, as in the Medicare Shared Savings Program. While patients enjoy the freedom to choose their providers, this limits the ability of the primary care provider to “utilization manage” the patient, and introduces the distinct possibility that the patient may migrate their care elsewhere and no longer be attributed to the ACO for savings calculations. Yet, on the other hand, this will incentivize providers to work hard at engaging their patients and making sure that they reach out to patients to bring them in for primary care visits, including wellness visits, which provide an opportunity for crafting a stronger therapeutic relationship. The techniques used by marketers and electioneers to increase the loyalty of customers and their probability of showing up (at the voting booth) therefore become invaluable to healthcare. This is partly based on the application of the hypotheses of behavioral economics, and partly the product of relentless empirical testing (A/B trials).

Successful accountable care will require many types of innovation, and the competencies highlighted in this article are only a few of the challenges facing organizations that have the courage to move away from fee-for-service health delivery. Only time will tell which solutions in each of these areas will be most effective. At the Brookings ACO Learning Network, we are excited to watch how the experiments that are being implemented as part of our Innovation Exchanges play out. The accountable care movement is still in its early innings, but the new financial incentives that align provider, payer, and patient interests can justify the business and technology innovation needed to transform American healthcare.Author Affiliations: Dr Mostashari is a visiting fellow, Brookings Institution, and former national coordinator for health information technology, US Department of Health and Human Services. Dr Colbert is a hospitalist at Newton-Wellesley Hospital, instructor of medicine at Harvard Medical School, and a consultant for the Brookings Institution ACO Learning Network.

Download PDF: Four Key Technologies for Physician-Led Accountable Care Organizations