Plans must translate big data into actionable intelligence

July 1, 2012

IT analytics help plans identify patient populations and anticipate member behavior.

Key Points

IT analytics are beginning to take hold in healthcare, and payers must consider how they will approach the trend. They must structure administrative workflow and technology systems around big data, and then use the results to create value. The key concept is to leverage historic member information on a forward-looking, predictive basis.

"Big data comes from the notion that we generate a large quantity of information about ourselves," says Joseph C. Kvedar, MD, director of the Center for Connected Health, Partners HealthCare. "We leave transactional and informational 'bread crumbs' around every day."

"The technology part is not futuristic," says Kvedar. "The application in healthcare will take some time, but it will get there. Most health plan executives have an analytic mind and an analytic strategy."

Being able to scale big data is key. Some health plans are large enough to draw data from a significant population, while the smaller plan might need to pull in data from other sources to create a complete picture. With sophisticated analytics, a variety of source data can be extrapolated into actionable business intelligence.

The promise of big data hinges on being able to find patterns that predict expected outcomes and the influence of member decision making.

"Good decision making is a huge part of healthcare costs and outcomes," says Ahmed Albaiti, CEO of IT consulting firm Medullan. "So that's the big deal for big data: being able to create that decision-making opportunity before actions are taken-before specific tests are ordered, a diagnosis is made or treatments are prescribed."

The opportunity to influence comes at the decision point, which means the timing of the outreach is important. Although review of claims data, for example, provides member-decision information, the knowledge comes after the fact. Big data will facilitate faster responses on the part of health plans that want to prevent downstream costs.

Albaiti says plans have access to relevant claims information at virtually any moment. However, the time lag in the analysis doesn't provide an opportunity to change the dynamic quickly. Specifically, systems must move from today's less frequent batch processing to ongoing real-time processing.

"We're at a point with technology where big data can be instantly discovered and navigated without having to move it into a database," says Shawn Dolley, vice president of big data, healthcare and life sciences at IBM. "This means healthcare organizations can now look at any and all data, and decide what to analyze to improve research and patient care, and better manage operational costs. Big data changes how a health plan can do business."

The difference, compared with even four years ago, is that such ambitious analytics can now be performed cost effectively. And the technology can scale to a population.

"Big data is aggregate, comparative-effectiveness, outcomes analytics," says Dolley.

He says improved-outcomes measurement and cost components should be part of a plan's analytics system.