Insurance has always been about quantitative analytics and sifting through oceans of data. What’s different now is the sophistication of tools and the amount of data insurers can analyze. That’s good because healthcare reform is about to steer carriers into uncharted waters.

The terabytes of information held in the data warehouses of health insurers will be crucial resources for navigating the ever-changing healthcare landscape.

But like precious minerals deep in the ground, data must be mined lest its secrets remain as such. With intensifying pressure to contain costs, healthcare insurers such as Hartford, Conn.-based Aetna Inc. are turning to business analytics to better align expenses with their predictive business models.

Business analytics is “taking the past to predict the future,” according to Aetna’s head of medical cost analytics, Marla Pantano. “Any time you get more data, you learn new things, and the more you have, obviously, increases your credibility.”

And Aetna has loads of healthcare data – about 37-38 terabytes containing some 71 billion records (see sidebar: “Aetna’s Vast Data Gold Mine”). With healthcare reform looming and tens of millions of uninsured individuals about to join the rolls of companies such as Aetna, business intelligence and reporting tools will play an even more important role.

For example, Aetna recently found that costs for a certain geography were coming in high for inpatient hospital claims. Instead of reimbursing on a per-diem basis, Aetna discovered it was paying claims based on diagnostic related groups under which hospital stays are predefined.

“We were able to complete audits on these claims and determined that there was an opportunity to take corrective actions through our contracting language as well as utilization and billing review and overpayment recovery,” says Pantano. “We’ve been able to identify several million dollars that we are now recovering, having paid more than we feel was appropriate.”

To perform such audits, Aetna uses a mixed bag of technologies. To generate top level reports and drill into the data, it uses Diver 5.0 from Dimensional Insights, Inc., Burlington, Mass., tools from the SAS Institute, Inc., Cary, N.C., and tools as common as pivot tables in Microsoft Excel.

“[This] allows people to sort things and find out what’s driving the largest variance from our expectations. It allows us to drill down and say ‘oh, okay, it’s not physician costs. It’s hospital costs are materially higher than expected,’” Pantano says.

Aetna also used business analytics to project its costs for the H1N1 flu pandemic in 2009. To forecast costs, Aetna used a variety of data types, such as the flu incidence rate forecast by Centers for Disease Control, the types of drugs used to treat flu symptoms and historical claims data.

“We were able to predict what the expected costs would be and created models based on the immunization rates we predicted would happen. Then we modified that prediction [based on] how many more or less people would get immunized. It was fairly predictable,” says Pantano.

In 2009, the costs turned out to be “pretty darn close” to what the Aetna models predicted, according to Pantano.

The program included educational outreach on where to get immunized and programs to steer patients to their personal care physicians and away from more expensive emergency rooms.

Aetna conducted the same modeling for the flu this year, but the illness was not as widespread as it was in 2009.

That Aetna could accurately map its costs says volumes about how far analytics technology has come – and its value proposition for the future.

“Business analytics today is becoming much more enterprise scale. [A few years ago], I could get away with working with localized data on my desktop, conduct data analysis on a few million folks, summarize the results and put them out there for someone to absorb,” says data mining expert Christopher Scheib, business solutions manager for health and life sciences at SAS.

Now, server-resident data must be analyzed in real time. And information silos must come tumbling down in order to enable insurers to venture deeper into the nature of market drivers. As such, business analytics will be a competitive battleground.

“People do not have the luxury of getting a problem, working on it for six weeks and coming out of the basement with a tablet of answers,” says Scheib, who once analyzed data for a BlueCross BlueShield plan. “The environment is too fast now. Business users have to make changes as information is being generated.

You have to be able to iterate, make changes, get business input and be able to work on three to four things at once.”

“Aetna, for one, has revamped its analytics approach to keep up with the pace of data,” Pantano says.

“We have actuaries (Pantano is one).

We have medical cost analysts who really understand our contracting,” she says.

“We have programmers and people who have very good skills in knowing how to summarize data and create the reporting everyone else uses. We have people who understand clinical coding, types of conditions, drugs and other things that can be grouped together [that determine] how care can be accessed and utilized.”

Liz Boehm, principal analyst at Forrester Research, Inc., Cambridge, Mass., sees many of the same challenges facing healthcare insurers as does Scheib.

In particular, data silos bar them from fully exploiting their data assets.

“They apply analytics in narrow areas where a broader application could really help them change and fix more of the things people find frustrating about their business. A lot [of what they look at] is in silos,” she says. “I see pieces of what they are measuring and what they are not.

Population health management programs often times take a very blunt approach.”

But, can business analytics transform the healthcare insurance business?

“It absolutely can, but it depends on what data they put together,” Boehm says. “They need to put some of that data together even better so they have complete picture and can really start to find the unknown correlations between different healthcare behaviors and what drives costs.”

With healthcare reform, insurers will benefit from as many as 50 million new customers. Accurately pricing programs for today’s uninsured will require heavy use of analytics.

“Healthcare insurers in the past have a lot of data, but have not used it or mined it properly,” says Kunal Pandya, senior analyst on healthcare and insurance at the Aite Group. “There has to be a larger emphasis on how [they] price the variation in their programs if they want to get the 30-50 million uninsured coming on board.”

Pandya credits Aetna, as well as Humana Inc. and Kaiser Permanente with changing that.

SIDEBAR: Aetna’s Vast Data Gold Mine

75 terabytes of storage (half of it occupied) 71 billion records three years of current year-to-date information 4,500 users: (on an average day, about 600 users a day submit about 25,000 queries)

Source: Aetna

SIDEBAR: Aetna’s Healthcare Insurance Rolls

MEMBERSHIP (millions)

  • 18.688 medical members
  • 13.953 dental members
  • 10.410 pharmacy members


  • more than 966,000 health care professionals
  • more than 543,000 primary care doctors and specialists
  • more than 5,200 hospitals

A network of specialist physicians with Aetna’s Aexcel® designation earned through clinical performance and cost efficiency.
Source: Aetna

Register or login for access to this item and much more

All Information 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