ACOs make progress in using big data to improve care

Accountable care organizations across the country are in sharply different stages of aggregating and using patient data to improve quality of care and reduce costs.

Only a handful of the largest and most sophisticated ACOs have established a “big data” warehouse that will let them pull together information from a variety of sources to help optimize care for individual patients and for their overall enrolled population. But many ACOs are developing the capacity to track patients in real time when they go to the hospital or the emergency department so they can intervene quickly to improve cost and quality outcomes.

One Medicare Pioneer ACO that is well along in using big data analytics is Minneapolis-based Allina Health. Its ACO has about 21,700 attributed Medicare patients and it also operates a commercial ACO. Its data analysts have produced about 60 data dashboards, enabling Allina providers and administrators, with the click of a mouse, to track outcomes against a variety of performance measures and target improvement efforts. For example, Allina's ACOs can track blood glucose, blood pressure and cholesterol levels for patients across the entire Allina system, or by clinic regions, individual clinics and even by individual physicians.

Allina launched its big data effort in 2008, contracting with Salt Lake City-based data storage and analytics systems developer Health Catalyst to build its integrated data warehouse. This multifaceted database gathers information from 42 sources, including clinical data from Allina's electronic health record as well as information about costs, claims and patient demographics.

“With our ability to identify just who these patients are out of the several million who access Allina in a given year, we can define a panel of like patients and match them up with caregivers and care managers who can most impact positively the care they're receiving,” said Ross Gustafson, who until Jan. 1 was vice president of performance resources for Allina. He now has a similar post at Health Catalyst under a new health information technology outsourcing and partnership deal with Allina.

“We're able to get very specific, near real-time information and data on (ACO patients') health and track our performance on key measures that we could be rewarded for,” he said.

But most smaller ACO providers aren't that sophisticated yet. They primarily are using data already pooled in their own and their partner providers' EHR systems, experts say. For many of these ACOs, accessing, sharing and analyzing these smaller pools of data, then figuring out what to do with the results, present their most formidable information challenges.

For example, the Aledade Delaware ACO, which just started this month, offers its physician partners software that lets them search their EHRs, find patients with multiple diagnoses and weigh the severity of their conditions. Physicians then can stratify those patients with the most urgent risks and care needs, said Dr. Horatio Jones, who leads IT efforts at Stoney Batter Family Medicine in Wilmington.

The Aledade ACO has an advantage since Delaware was the first state to develop an operating statewide health information exchange, the Delaware Health Information Network, or DHIN. “Our main objective is we're trying to work on the transition-of-care aspect,” Jones said. Success will depend on timely notifications when patients go the hospital or the emergency department, information that the DHIN exchange will provide, he said.

Most ACOs are still a long way from the holy grail of being able to gather a vast amount of data from wide-ranging sources and use them in real time to improve care. It's envisioned that one day health systems will be able to aggregate and quickly analyze a range of data, including online records of consumer purchases, social media posts, geocoded information showing enrolled members' proximity to parks and other recreational amenities and genomic data.

Gartner, a research firm that watches the technology market, estimates it will take five to 10 more years before big data attains the “plateau of productivity” on the firm's Gartner Hype Cycle and becomes widely adopted in healthcare, said Vice President Robert Booz. The big data phenomenon is now over what his firm calls the “peak of inflated expectations.”

“There is so much left for us to do with regular data, we feel there are years to work there,” said Dan Burton, CEO of Health Catalyst.

Dr. Farzad Mostashari, a former national health IT coordinator, said that for healthcare providers looking to launch an ACO, a good, interoperable EHR system is the place to start. Many of the data-gathering, transmission and analytics tools their ACOs will need are available from a growing number of data analytics software and services providers such as Health Catalyst, as well as from health information exchange vendors and organizations.

Aledade, a Bethesda, Md.-based company that Mostashari co-founded last summer, also supplies interoperability and data analytics services. It partners with its clients—such as the Aledade Delaware ACO—and shares ACO income with them.

ACOs “have to keep getting better,” Mostashari said. Eventually, “unless you can tap into big data, you won't succeed.”

Randy Farmer, chief operating officer for the Delaware exchange, said the DHIN rolled out its Event Notification System in 2013, linking the state's hospitals' admission, discharge and transfer, or ADT, systems to the exchange, and from there, to physician practices that are exchange members. The system promptly notifies ACO doctors if one of their patients has been admitted to, discharged or transferred from a hospital. While the Event Notification System was built in response to the needs of physician practices and insurers that wanted to expedite claims when a patient was released from the hospital, it can fill another important need for groups like the Aledade Delaware ACO.

Mostashari said this type of real-time patient information can help reduce readmissions. “If a person gets discharged from the hospital, the single most powerful predictor of not having a readmission is if they have an appointment scheduled with their primary-care physician within seven days,” he said. Real-time links between the hospital and a patient's primary-care physician enable care coordinators to book those appointments before the patient is discharged. For subscribing provider groups, which pay $4,000 for two years of service, they receive ADT alerts via printer, Web portal inbox or EHR.

But there isn't a large enough ACO market yet to spur health IT developers to create systems that serve ACOs' care-coordination needs, experts say. The upcoming Healthcare Information and Management Systems Society convention has 81 vendors registered as data analytics system developers, but only five tout their ACO capabilities in their HIMSS profiles.

ACO patients remain a “very small” percentage of the healthcare population, said Todd Cozzens, a partner at Sequoia Capital, a venture capital firm. “The real underlying trend that is driving the use of big data is a move toward integrated care,” he said. ACOs are only subsets of the integrated-care movement. “We're in the early stages of (merging) clinical data with claims data,” he added.

Cozzens noted one problem with big data is that “before you can consume it, you've got to organize it.” So when sales representatives from software or services vendors claim their products facilitate big-data analytics, he said ACO leaders need to ask, “Have you served up the data in a way it can be consumed to drive improved care and the cost of care?”

Another problem ACOs face in pulling together patient data from a range of provider organizations is what Mostashari describes as a “tax” from EHR vendors in the form of hefty fees to interface their customers' EHR systems with those from other EHR vendors. “Just the cost of the interfaces is taking up the bulk of the cost of their IT for ACOs,” he said.

Another, more technically complex problem ACOs face is creating what Mostashari calls “a dynamic order set” or a template “that feeds information and pulls information.” Building this order set requires the use of application programming interfaces to unlock data in older EHR systems.

“You need a good EHR that creates a common EHR within the (broader ACO) group,” said Dr. Scott Fowler, CEO of the Kingsport, Tenn.-based Holston Medical Group, which has 110 physicians in 23 locations in Tennessee and Virginia. Most major EHR systems can fill the needs of both a group practice and an ACO, he said. Typically, there are different EHRs from multiple vendors in an ACO, and they “need to be put together to get a common record for the community.”

For example, Qualuable Medical Professionals, a Medicare shared-savings ACO that Holston joined, has roughly 500 providers in 132 practice locations using about 20 different EHR systems. Qualuable built a private health-information exchange to facilitate data sharing, using software and services from dbMotion, a unit of Allscripts Healthcare Solutions.

“You want to be able to bring the data out of that EHR and have it available to that doctor seeing the patient using a different EHR, and you want that to be relatively seamless and bi-directional,” Fowler said. An ACO also needs basic analytics information so providers can “go out into the population and get those falling through the cracks.”

Qualuable uses a data analytics and graphing system from OriginLab Corp. “It's one thing to say the patient has a gap in care,” Fowler said. “We need to push back information into our EHRs and say, 'These are the things that we need to meet our quality marks, and these are the things that are impeding the quality of care.' ”

And ACOs need EHR interoperability with other providers, he said. When Qualuable ACO patients go to the hospital, their primary-care physician is alerted because their EHR is connected to the hospital's computer system for admissions, discharges and transfers. The admission alert pops up on the doctor's EHR and that doctor can collaborate with the ED staff, for example, by sharing a recent imaging test result.

For Qualuable, the results have been outstanding, according to Fowler. “In our first year, we saved (Medicare) almost $14 million and we had high quality scores,” he said. “They gave us $6.7 million back.”

Follow Joseph Conn on Twitter: @MHJConn