As any accountable care organization (ACO) can attest, there were a million things on their to-do list in the early days of formation: establishing governance and administrative structures, creating a clinically integrated network, developing a data-sharing strategy and much more. This article acknowledges those many challenges, but will focus on a different piece of the puzzle: using data analytics to help ACOs make sound decisions that affect their clinical, financial and performance management operations.
Can we make better decisions through analytics? The proposition sounds straightforward, but for most providers, this is yet one more element of a seismic shift in how they have been doing business for most of their careers. No longer will they be asked to fill beds, do more procedures and keep the MRI machine humming. In this new world, care will be pushed out into the community. Patients will monitor their health in the comfort of their own homes. Most importantly, disease will be managed, and even prevented, through proactive outreach. Granted, we’re a long way from that world now, but there is a vision emerging.
Physicians are now being asked to look beyond the 20 patients reading People magazine in their waiting room and begin worrying about the thousands who rarely make it into the office – the emergency department (ED) frequent fliers, the chronically ill but noncompliant, even the relatively well who could benefit from a better diet and more exercise. Borrowing a page from economics, I see this part of the ACO’s business as the demand side: understanding where the demand for healthcare is likely to come from across a population and what interventions may reduce unnecessary demand.
Meanwhile, ACOs will need to ensure that their networks are populated with the most clinically effective and efficient providers in order to meet quality standards and survive economically in an environment in which better – not more – care is rewarded. In this supply/demand framing, the critical tasks of creating and managing a high-performing network represent the supply side. To succeed, ACOs must closely monitor and understand both the supply and demand sides of their business. This is where comprehensive data analytics, powered by predictive modeling and risk adjustment technologies, comes into the picture.
To be accountable for an entire population, providers must understand their patients in ways that they’ve never had to before. By using claims data to feed population health analytics and predictive modeling tools, ACOs can identify and stratify health risk, in turn forecasting, among other things, which patients will likely be high-cost, top utilizers of medical resources in the near term. Although claims data are often dismissed as being too dated to be analytically useful in some situations, it provides the most comprehensive record of the interactions that a patient has had across the healthcare system, rather than in a given setting where EMR data may be timely and rich, but misses the experience of the patient outside of that particular setting. Claims data have the additional advantage of relative standardization and longitudinal history.
While a critical tool in the analytic process, predictive modeling alone cannot provide the ACO with the holistic view of patients needed to deliver proactive, effective care. To truly understand patient needs – both individually and collectively – predictive risk scores must be augmented with clinical data, diagnoses, complications, gaps in care and so forth to illuminate actionable opportunities and filter high-priority populations. That’s the 360-degree view of patients that causes the physician, for example, to rethink prescribing a beta blocker to a 60-year-old diabetic male with heart disease and high predicted risk once he sees that the patient is also asthmatic. (Certain beta blockers can exacerbate airway constriction, so the provider would choose a cardioselective option that targets only beta receptors in the cardiovascular system.)
A key element of demand management involves segmenting the population into those patients who can be helped in the near, medium and long terms. It’s the first group, those who should be targeted in the near term, who are caught up in what some have dubbed “the vortex of consumption” or “the diagnostic-therapeutic cascade.” Examples of these patients are the diabetics who have been using the ED as a primary care office or the elderly with nonspecific complaints who pinball from specialist to specialist without getting better. These are the folks who reveal where the system is breaking down and who require immediate, targeted interventions. Reducing ED visits, hospitalizations and excessive use of specialists are the primary goals with this group; much of this can be achieved through improved care coordination powered by 360-degree analytics.
On the other end of the spectrum are the majority of most ACO patients – individuals for whom the healthcare system is more or less working, but for whom a few broad-based interventions will yield not only better health, but also the potential for significant savings down the line. For this stratum, the focus is on closing care gaps and changing behavior (e.g., ensuring that diabetics are getting regular HbA1c tests; motivating people to exercise, eat better and stop smoking).
Art complements the science
Acquiring that 360-degree profile of patients will take an ACO only so far. This is where the art comes in, in the form of so-called physician extenders. As risk-sharing payment models continue to penetrate the market, ACOs will need these clinical/analytical hybrids, adept at comprehending what the data reveals about individual patients, cohorts and populations and then determining the most effective interventions. Actions might include reaching out to the diabetic with elevated glucose levels to schedule an appointment or running a weight-loss group for patients who are morbidly obese. Because these types of activities have traditionally been in the jurisdiction of insurers, I anticipate many ACOs will seek out employees with these skill sets from the payer world.
In addition to having a thorough understanding of their patients, the most effective ACOs will also have a firm grasp on the supply side of the equation: their network providers. Analytic tools are also vital for measuring and monitoring provider efficiency, answering the fundamental question: How are providers performing once you take away the “my patients are sicker” argument?
Risk adjustment levels the playing field by explaining how much of cost and utilization is likely driven by illness burden and how much by physicians’ practice patterns. Time is not wasted chasing down supposed outliers when there are legitimate reasons for the variation. In the early days of ACO formation, risk adjustment can enable planners to make better decisions about which provider groups to invite into the risk-bearing network and how to compensate them; later, it can be a critical tool for objectively evaluating ongoing performance for reasons of quality assurance and compensation, as well as pinpointing both best practices and those that need a closer look (e.g., high observed to expected ratios of imaging studies or ambulatory care-sensitive admissions).
Risk adjustment should also play a role when it comes to compensation for primary care physicians (PCPs). In the ACO model, PCPs are required to manage health and wellness and improve care coordination to a degree that they haven’t in the past, not merely treat disease – for individuals and entire patient panels. And providers have long lamented that this improved care coordination effort is not reimbursed in the current system.
Providing PCPs with base payments that are risk adjusted to account for their panel’s disease burden will help support activities such as outreach to patients who are overdue for preventative care, coordination with specialists and, in some cases, even the hiring of nutritionists or wellness coaches. Without such base payments, PCPs will have a tough time meeting performance and quality metrics.
To succeed in a value-based system, ACOs must have both a close-up and a 10,000-foot view of their patients and their providers. I believe that we’re just at the tip of the iceberg in terms of using analytics to help us better understand risk, and thus engage and manage individuals and populations more successfully. If all of the hype related to the health information exchange, as well as new sources of patient data from innovations like remote monitors and mobile applications, is true, we can expect these analytical tools to become even more powerful in the future.