Montefiore Health System is serious about keeping its patient population healthy-so serious that it's working with a dozen New York City public schools in its service area to get kids active and eating their vegetables to reduce their chances of showing up later on its diabetes registry.

Mony Wechsler, senior director of application strategy at the Bronx-based health system, is developing plans to engage students. He's shopping for the right gadget to help them track their activity and calories. He's looking for something that's not too expensive, able to run for the whole school year on a regular watch battery, and ready to download the kids' data into Montefiore's database when they come to school.

Wechsler hopes fitness trackers will harness children's natural competitiveness to get them moving and developing good eating habits, which then can carry over into adulthood, while also giving Montefiore the data it needs to flag potential health problems early. "This is not just a stunt," he says. "We want to make sure we can get data we can make sense of."

Hard work ahead

Montefiore is years ahead of most providers in making sense of population health data and acting as an accountable care organization. It takes care of about 30,000 Medicare patients as part of the Center for Medicare and Medicaid Services' Pioneer ACO Model program, plus another 350,000 patients covered by its own health plan.

Because Montefiore has been taking on financial risk for a significant percentage of its patients for years, it has been in the vanguard of using data to manage care. Its homegrown data warehouse draws from its electronic health record, ancillary systems and other sources. The organization has developed and patented a software tool called Clinical Looking Glass to analyze the data, and those analyses have anchored dozens of scholarly publications by Montefiore researchers, while helping its clinicians target their preventive efforts to the patients who need them most.

Wechsler says his colleagues in health IT should prepare for years of hard work to get their data-crunching capabilities up to par for accountable care, and not simply assume they can just go out and buy what they need.

"Lots of vendors claim they have an ACO solution, but it's not true, and most of them don't understand what an ACO is," he says.

Population health and accountable care require a fundamentally different approach from traditional get-the-bills-out health IT. In fact, if payment reform continues on its current trajectory-away from fee-for-service and toward quality-based risk-sharing contracts-the day will come when getting bills out is irrelevant.

Instead, providers will need to track costs to the penny, and minimize use of the most costly services and care settings by keeping patients healthier. To get a full picture of a patient's health risks, they'll need data not found in their EHR or a claims database. For example, they'll need to know whether patients have someone in their house or neighborhood who looks out for them, whether they can afford their medication co-pays, whether they have an air conditioner or a car or snow removal service. Providers' IT departments also will have to handle streams of data from wearable health trackers like the ones Wechsler hopes to give to New York City students.

Where IT needs to start

Many organizations begin by asking their IT departments to develop population health analytics capabilities, without first determining exactly what they need to analyze, says Bob Schwyn, a principal consultant with Aspen Advisors who works with providers on both strategic planning and developing IT for population health.

"Executive teams think IT is fairly easy to develop, so they give a directive to go build the capabilities they need to run an ACO," he says. "They aren't clear about what specific capabilities they need, and which existing investments they may be able to leverage."

Moreover, Schwyn says, most organizations give short shrift to data governance: making sure data is collected in a consistent and standardized way that makes it easy to feed into the analytics and care management tools that accountable care requires. Without clean, reliable data, no tool will produce usable knowledge.

"If you know that in the next year you'll be focusing on asthma and diabetes, you need to start to hone your efforts around data governance" for those conditions, he says.

The market for population health software and services is already crowded and becoming more so. It includes add-ons from most large EHR vendors, customer relationship software adapted from other industries by companies like SAP and IBM, and new entrants with purpose-built product lines. Some leading providers have leveraged their advanced population health experience into start-up companies that they either own outright or have invested in, including Explorys (the Cleveland Clinic), Health Catalyst (Intermountain Healthcare), xG Health Solutions (Geisinger Health System), and Evolent Health (University of Pittsburgh Medical Center).

Joe Van De Graaff, research director at KLAS Enterprises, predicts that most providers will start with their EHR vendors, which have at least theoretically already done the heavy lifting of integrating their systems with their population health tools. If those tools prove inadequate, providers may bring in outside vendors for analytics or care management capabilities, creating a complicated environment and still not producing all the necessary knowledge.

"The whole financial side of population health management is very little understood," Van De Graaff says. "A lot of the vendors can bring data together and segment populations [to identify particular patient groups that will benefit most from care management], but we have not heard that any vendor is really leading on understanding reimbursement. Those products may not exist yet."

Using what you have

Some experienced providers advise starting quick and dirty. Whether they're participating in federal accountable care programs or taking on risk-based contracts with private insurers, ACOs typically have access to at least two sets of data: the clinical and billing information they generate internally, and claims data from the payer on the patients covered by the contract.

That's enough to make some significant progress on understanding your patient population and identifying care gaps, says Larry Garber, MD, medical director for informatics at Reliant Medical Group, Worcester, Mass. The multi-specialty physician group cares for about 250,000 patients under risk-sharing contracts with half a dozen payers, including CMS through the Medicare Advantage program.

"Claims data is the poor man's health information exchange," Garber says.

Reliant has incorporated its payers' claims data directly into its Epic EHR, which Garber says was relatively easy to do. The claims data captures care delivered by other providers, so that the EHR's decision support rules can take those services into account. If a patient has already received a flu shot at a pharmacy, the EHR doesn't pester Reliant's care team to make sure the patient gets a flu shot. Medication claims are loaded every day, and other types of claims are loaded weekly.

Reliant has used the data to built some sophisticated patient management tools. For example, the EHR automatically checks claims data three days after a hospital discharge, notes which medications the patient has picked up, and notifies physicians of any problems, such as possible interactions with one of the patient's other medications, or the need for a blood test to check whether medication dosages are correct.

Rodger Prong, executive director of Physician Direct ACO, Sylvan Lake, Mich., initially struggled with the claims data he receives from CMS for the 8,900 patients for whom the ACO is caring under the Medicare Shared Savings Program. The file containing claims data is complex and a challenge to use, he says. "Everyone who gets it has to either develop some sophisticated system to deal with it or pass it on to a third party." Physician Direct contracts with Lightbeam Health Solutions, Irving, Texas, to turn the CMS data into actionable information for the ACO's 450 physicians, most of whom work in practices with fewer than 40 physicians.

Physician Direct also has a homegrown data warehouse that incorporates data from labs, hospitals, and the EHRs and registries of its physicians, although Prong is considering whether to enlist Lightbeam's aid for that data as well.

Costly interfaces have been the bane of Prong's existence, as the warehouse grapples with too many different EHRs. "We started out saying we'd support any EHR, but now we have a favored group we're more willing to work with," Prong says. To his relief, many of the organization's physicians are shopping for new EHRs, and actually asking him which one to pick. He directs them toward NextGen Healthcare, athenahealth, Practice Fusion or Greenway Health, all of which can feed data smoothly into the warehouse.

Connectivity is essential

Integrated ACO, a physician-led organization in Austin, Texas, used $1.5 million in advance funding from CMS to create connectivity among 50 small primary care practices spread across a rural area between Austin and San Antonio. It was one of only a handful of Medicare Shared Savings ACOs to get money in advance, and CEO Eric Weaver says that without that jump start, the organization could never have afforded to connect 10 disparate EHRs and create its data warehouse.

"We used a lot of it for HL-7 connectivity," he says. "The physicians didn't want to pay their vendors $10,000 each [for interfaces]." If a given EHR didn't support HL-7 standards, the organization did screen captures and mapped the data elements back into the database.

Some vital information was still on paper, particularly from long-term care facilities that didn't receive federal incentive payments to adopt EHRs, and had to be captured via optical character recognition.

Currently, the ACO cares for about 13,000 Medicare beneficiaries, and has developed analytics capabilities in cooperation with Loop360 Solutions, a software company created to enable Integrated's activities and then sell the resulting products and services to other ACOs.

Managing future events

Vipul Mankad, MD, one of the founders of both Integrated and Loop360, is particularly proud of the predictive analytics they have developed. "It's easy to find out the patients who generated the costs last year, but that won't tell you anything about who will generate the costs next year," he says. "If a patient has a myocardial infarction and a bypass, then you've spent that $100,000 and probably won't have to spend it again." It's more valuable for the system to predict who's going to have the next heart attack, and Mankad says computers can pick up on subtleties that even experienced physicians may miss.

"Clinicians make predictions every day," Mankad says, "They see that a person is 85 years old and short of breath, and they know he's going to go to the hospital soon [for congestive heart failure]. Humans can handle three to five factors at once, but a computer can handle 500." He estimates that the organization's analytics can identify 900 out of a thousand patients who are likely to be hospitalized for CHF, using that information to prevent admissions with timely care management.

Heritage California ACO, another CMS Pioneer ACO, also has learned to cope with disparate EHRs, says Mark Wagar, president of Heritage Medical Systems, the physician group that provides care under the ACO contract. Some 3,000 employed physicians and 35,000 independent ones care for more than a million patients, and the ACO has refrained from limiting which EHRs the independent physicians use. You can't make progress on value-based care if you're going to require that every provider must use the same system, Wagar says.

Heritage's data warehouse includes claims, EHR data, labs and miscellaneous information provided by patients and their families. The goal is to create as clear a picture as possible of the patient's circumstances, not just his or her health status. "We want to know who's your trusted family member, or a neighborhood person that you look to for help, and we will ask that person to call us if they're worried about you," Wagar says. "Oftentimes, people don't want to go to the doctor and don't want to burden the family or others with their issues."

Wagar says that complete information can help an ACO figure out when it makes sense to make interventions that would never happen under a fee-for-service model. For example, the ACO might come out ahead by paying the entire cost of certain patients' medications, to increase the odds that they'll stay on the regimen that's keeping them out of the hospital. If icy sidewalks are keeping a patient from his daily walk, a care manager might help him find a place to exercise indoors.

"Sometimes the most important pieces of information have nothing to do with injections or surgery, but with someone's life circumstances," Wagar says. "When you're paid entirely on value, you will do whatever it takes."

Register or login for access to this item and much more

All Health Data Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access