Brad Lawson can’t do his job without data. As director of managed care finance with Atrius Health, he’s in charge of estimating expenses for a network of six medical groups serving nearly 1 million patients in eastern and central Massachusetts. Decisions about investments, such as whether to build a new MRI suite, depend not only on the medical benefits, but also knowing what they’ll cost.
Only now, executives, managers and clinicians at the non-profit health care provider want answers faster—about what they’re spending, and about patient outcomes and clinical workflows that drive performance against national health care benchmarks. Federal and state agencies want data, too. “We’re being bombarded constantly with new requests,” says Lawson. And the questions are getting more complex, requiring analysts to dig deeper into a growing volume of data. Atrius anticipates it will have 6 terabytes of data—double what it has now—within 12 months.
Like many organizations, Atrius needed a better way to deliver data and analysis to business decision makers. The drive is certainly strong in the health care industry, and in Massachusetts which in 2006 state enacted health insurance that became the model for the national Affordable Care Act. Both the state and national laws put a priority on data-driven medicine. When it comes to using data to manage patient care, Atrius is already considered an innovator. Last December, the Center for Medicare and Medicaid Services designated the medical group a Pioneer Accountable Care Organization—one of 32 health care organizations nationwide selected to demonstrate ways to reduce health care costs while improving health outcomes.
Atrius consolidated its fragmented, technology-focused analytics function and put a doctor in charge. The result: a more collaborative team whose work, including the investments it makes in new tools, is more closely aligned with business needs and strategy.
“We consider ourselves to be a professional knowledge worker organization that likes to do data-driven adaptive learning,” says Joe Kimura, who leads the Atrius analytics team as medical director for analytics and reporting systems. “There’s a lot of capability out there that we could use. When the practice is ready, we’re set up to give it to them the day before and not the day after.”
Here’s what Atrius is doing to bring deep data analysis closer to decision makers:
1. Put business leaders in charge
Kimura trained to be a physician, but after a fellowship in health services research, he shifted to clinical operations. For two years, he led a department at Atrius responsible for measuring clinical quality. He became a top consumer of corporate data, working closely with analysts to study workflows and develop quality metrics. (For example, Atrius tracks whether its doctors prescribe antibiotics correctly when patients have bronchitis and how often they monitor patients who take medication for chronic illnesses, such as high blood pressure.) “After I complained for years, it was time to put up or shut up,” Kimura says about being tapped to lead the new team. He reports to Atrius’ chief medical officer.
Two and a half years ago, before Atrius created the group Kimura now leads, its core analytics team was attached to the IT department—part of the group that ran the company’s data warehouse. Kimura says the analysts were comfortable working with business users and in some cases had experience on “the medical side.” But when it came to decisions about investing in new tools, he says, the group tended to choose what power users wanted. Adapting those tools for other segments of users, such as executives, managers and clinicians often wasn’t appropriate.
“Each segment has different needs around granularity, timing, breadth and analytical power,” says Kimura. Tools built for a department director were “too busy, too complicated, too powerful” for frontline clinicians, who would reach the wrong conclusions when they clicked buttons without understanding what they were getting.
Putting medical leaders in charge of analytics and reporting “shifted our technology stack and focus and skill sets to cater to those segments,” says Kimura. He maintains a “smaller, but significant” budget for R&D, so that when users need, say, natural language processing tools or the ability to mine unstructured data, his team will be ready to meet the demand.
2. Bring analysts—and analysis—together.
Before the revamp, each medical practice and key business function had its own specialized analysts, and they still do. But now all 38 analysts meet monthly with Kimura’s team of 11 to share how they use data as well as learn new technologies. Kimura is responsible for professional development and training (a pending decision: how best to build expertise in usability, so they can design reports that are easier to navigate) and for data stewardship.
Understanding, for example, how changing clinical practices are reflected in data for surgical admissions helps Lawson and his team make better predictions about future spending. Kimura’s team and the IT data warehouse group work together to standardize data across departments and to create data marts that incorporate corporate standards for measurement and analysis. Having common data and the right tools makes it easier to collaborate. “It’s much more coordinated now,” Lawson says. In the past, “there was never really one person in charge to make decisions about how data flows and is structured.”
When a business leader wants a new set of reports, Kimura looks for ways he can build on work that’s already been done. “We’ll talk it through and have a lot of back and forth with business leaders,” he says. “It’s guided listening, because we understand all the reports and analytics that the other departments are doing.” A report produced for the cardiology department might be able to be repurposed, with different data, for radiology.
3. Tie investment decisions to business strategy
A data analytics steering committee, whose members include Kimura, several C-level executives and other key business leaders, meets monthly to assess opportunities for investing in new analytics capabilities. Whether an investment helps to advance business strategy guides their decisions.
For example, last year, Atrius decided to collaborate with software vendor Verisk to adapt its Sightlines Medical Intelligence platform for mining its claims data. Although there are tools to support “major elements in our business,” Kimura says, “we still do not have the tools to explore every nook and cranny of claims data.” Users can log into SMI to study claims in a specific area to prioritize issues they want to address. If they want to make operational improvements, Kimura says, they would need to ask his team for more detailed data from Atrius’ electronic medical records system. But SMI helps reduce the number of data requests.
In the past, Kimura believes, such projects probably didn’t make the cut very often. “We never have enough resources. When new opportunities like that came forth, until we understood where we wanted to go with the business, it was difficult to make an educated guess about whether doing that kind of collaboration made sense.”
Last year, the analytics group developed a report incorporating operational information such as whether hospitalized patients were enrolled in disease management programs. Hospital admissions are a big driver of health care costs, so Atrius’ medical teams want to make improvements that will keep patients healthy at home. The data is also important to tracking Atrius’ performance as a Pioneer ACO.
The report was deployed in January. “Much of this work is aligned with our core operating principles,” says Kimura. “In the past we would have tackled reports when asked by the business to produce those reports. Today we anticipate where the business is trying to go.”
To Take Analytics to New Heights, Don’t Just Change the Org Chart
Most organizations see the value of creating, as Atrius Health did, a central group to manage analytics. Twenty-seven percent of respondents to a recent Gartner survey said they have established what Gartner calls a “Business Intelligence Competency Center,” that brings together business, analytic and IT skills. Another 35 percent are planning one.
It’s one thing, however, to unite analysts, establish data standards, create common analytics processes and develop an enterprise technology blueprint. To get the benefits, business users have to think about data and analytics from an enterprise perspective, which can be difficult.
“One of the challenges, even when business people are involved, is that they’re thinking from a very insular perspective—from their own jobs and not across lines of business,” says Douglas Laney, vice president of research for business analytics and information management with Gartner. At many companies, “the thing that usually lags is the culture, becoming a fact-based culture,” in which leaders use data to innovate, improve processes and step up performance. Many companies, he says, still use data only to learn what happened in the past, rather than to prescribe future actions.
Elana Varon is contributing editor with Data Informed. Tell her your stories about leading and managing data driven organizations at email@example.com. Follow her on Twitter @elanavaron.