Analytics, Population Health Deliver a One-Two Punch

Analytics represents knowledge; population health represents action. Too many healthcare organizations make the mistake of simply expecting their health information exchange to provide sufficient analytics to drive a population health initiative, says a senior executive at Optum.

Payers, armed with analytics technology and data, are showing signs of breaking through some of the barriers to true population health progress.

While researching my November cover story on big data, I interviewed A.G. Breitenstein, chief product officer of Optum Analytics and a co-founder of Humedica, an Optum company. Optum is a division of UnitedHealth Group, which is also the parent company for mammoth payer UnitedHealthcare.

Breitenstein says some healthcare institutions are making the mistake of lumping together their analytics and population health initiatives, when instead they should be considered as two separate disciplines.

Analytics represents knowledge, and population health represents action, she says. Too many providers make the mistake of simply expecting their health information exchange to provide sufficient analytics to drive a population health initiative.

"The real problem is with the people who are not in your office and not getting care and [who] really need to be in your office," she says. The most basic pay-for-performance initiative must be able to understand analytically "who is in the population [and] who's attributed to which contract. Each of those contracts is likely to have different quality requirements."

Whatever those requirements, population health technology at its most fundamental identifies the actions associated with getting people at risk in for visits, Breitenstein says.

Providers then need to take "a more advanced proactive management stance relative to say predicting and then preventing hospitalizations and ED visits," she says.

With only so many healthcare dollars to go around, Optum can leverage millions of lives' worth of claims data to rate where a given provider stands in cutting costs through their population health efforts.

Payers, Providers Aligning
"Most of what we do today is pay-for-reporting more than pay-for-performance, so giving people the beginning of a sense of 'where do I stand' relative to other folks who are competing for the same dollars or the same contracts or the same patients" will determine success, she says.

"The world of the payer and the provider and the Berlin Wall separating them, I think are largely gone," she adds.

"And the patient as well, those divergent interests are starting to align, so everything from utilization management and utilization review, clean claims submission, getting the whole process to work better in terms of accelerating time-to-value and time-to-money also comes into this picture, and benchmarks around those as well."

I mused on how CMS earlier this year had warned providers not to get too dependent on pop-up reminders in their electronic health records to do the heavy lifting of population health management.

"It may be my background in public health predisposes me to that anyway, but I absolutely believe that if you are sort of depending upon the physician at the point of care, as an incidental matter, to be managing all of the diversity of issues that come with bearing risk and managing quality and using a pop-up decision support too, you probably miss four fifths of the opportunities to really make a difference in a population," she says.

"And the most important one you're probably missing is the opportunity to more proactively identify and stratify patients based on more predictive analytic capabilities, to design more proactive interventions that can prevent events from occurring sort of before the patient even knows to call you and come in for that visit."

"We are definitely of the Sun Tzu school that [suggests] the battle is usually won or lost before the armies have even taken to the field, and I would say that clinical decision support is the point at which the armies are already on the field."

'Cheaper Than Nurses'
One Optum client, which Breitenstein declined to name, identified several hundred patients out of tens of thousands who had "preventable opportunities" to avert congestive heart failure. "They built a clever little intervention that really focused around health coaches as the primary interventional clinician," Breitenstein says.

"Much cheaper even than nurses, and they were able to reduce the hospitalization, so they indexed hospitalizations [and] took 65% of those out of that pilot population, and they took 30% of all-cause readmissions, 84% of the CHF-specific readmissions out."

"This was really a sea change," she says, "in terms of actively managing the population before they were in the hospital, taking various cost-effective and clinically effective resources, applying it to a small number of patients in a large population and preventing highly preventable and predictable events and significantly taking cost out of the system and keeping patients happier and healthier to boot."

The news may not be good for small practices, however.

"The ability to take risk occurs when a physician is participating in a larger organization that has the shock absorbers to manage risk," Breitenstein says. "That's a very different model from, we take care of the sick people that walk in, and we write a script, and we tell them what they should do and either they do it or they don't do it, and if they show up tomorrow, they end up in the hospital. We just sort of collect more money."

Seeking Clinically Preventable Opportunities
An aggressive organizational change in the mental model of a healthcare organization is required to attack population health opportunities, Breitenstein says.

The focus is on "finding statistically predictable, clinically preventable opportunities, and really teaching and working with organizations to maximize those they attack, the more scale they can add to, if they win in CHF, then they go after diabetes, and then they go after pediatric asthma and then they go after combinations of those different disease states, and if they do that, they can take anywhere from 30% to 40% of those hospitalizations and ED visits out of the system," she says.

There is still the garbage-in, garbage-out problem of structured data analysis missing the clinical narrative, but here too, Optum has a strategy for success.

Rethinking the Medical Model
"I'm one of the cofounders of Humedica, and Humedica's business model was built on addressing the fact that particularly the most valuable data that you find in the EMR is also typically the least structured, the most highly variable, and the most prone to errors," she says.

"Humedica's core value proposition was building a very scaled, but incredibly anal-retentive data cleaning and normalization service, as part of preparing the data for analytics of the type we spoke of."

"Early on, I heard a bunch of people sort of say stuff like, we're just going to take the 1% of high spenders and just assign them a nurse. I think people are starting to realize that that's not actually the answer either. So it's not taking the good old fashioned medical model and applying it to expensive patients."

"It's really rethinking the clinical model in light of a more population-based approach, so I think we're going to see a second phase of what we mean by analytics and population health management which is much more sort of practice and protocol focused, and that I think is going to be really exciting, because then we're going to start to see the fruits of real outcomes and real data to prove analytically what we're doing and how well it works."