TDWI | Training & Research | Business Intelligence, Analytics, Big Data, Data Warehousing

Case Study: Forecasting ROI

Apex tapped a BI infrastructure from SAS to help illuminate its 4 TB healthcare information warehouse.

• By Stephen Swoyer
• September 21, 2005

As a director of econometric risk strategy with Apex Management Group, Dr. Jody Porrazzo knows a thing or two about the insurance business.

And in Porrazzo’s view, companies in every vertical can benefit from the insurance biz’s Golden Rule: namely, that an ounce of prevention is worth a pound of cure. It’s better to identify potential issues before they become full-blown problems, Porrazzo says, just as it almost always makes sense to spend money in the near term to defuse ultimately costly long-term problems.

Of course, actionable insight of this kind presupposes an especial clarity into the murky depths of one’s operational and multi-dimensional data. That’s where business intelligence (BI) comes into the picture, says Porrazzo, whose company taps a mix of data mining, statistical analysis, and BI technologies from SAS Institute Inc. Used intelligently, she argues, BI solutions can help illuminate the murkiest of depths. And in Apex Management Group’s case, SAS’ business intelligence tools have consistently returned both tangible and intangible proceeds. In a health insurance space where an ounce of prevention can mean the identification of an at-risk patient, intangible proceeds amount to perhaps the most rewarding ROI of all.

“Far too many people talk about ROI in terms of dollars and cents, and what I’d like to do is to move away from that,” she comments. “If I catch somebody whose risk factors are at critical mass for a heart attack and we get them into a disease management program and are able to prevent a heart attack, how can you realistically put a price on that? You can’t—although I can try.”

That’s because it’s much cheaper to proactively identify patients who are at risk for heart disease and take whatever steps are necessary to prevent the occurrence of a heart attack, Porrazzo says. “I can give you some benchmarks for cost of treatment and hospitalization and the cost of drugs to treat this, but what about the quality-of-life benefits: how do you quantify those? There are things you cannot quantify, and you have to take those into account, too.”

Apex recently complemented SAS’ vaunted data mining toolset with the Cary, N.C.-based BI giant’s somewhat newer Enterprise BI Server, which it first implemented as part of the SAS8 BI suite. Recently, however, Apex installed the newest version of SAS’ BI suite on top of a new 64-bit Itanium2 server from Hewlett-Packard Co. As it turned out, Apex’ SAS9 implementation was easily a six figure effort, including hardware, but Porrazzo says it’s been worth every penny.

“There’s always the saying that ‘If it ain’t broken, don’t fix it,’ but you have to realize that [the return on investment] will be there; that you’re spending all of this money so you can save even more [money]—and that you’re also going to find ROI in other places, whether it’s improved quality of life for patients or something as drastic as saving someone’s life,” she comments.

It also helped that Porrazzo was able to sell company management on her long-term growth strategy, which she linked to the acquisition of scalable BI software and server hardware. Apex’ data management needs are already prodigious—the company analyzes historical insurance data (consisting of administrative claims and national healthcare costs) to accurately forecast future trends—and they’re expected to grow even more over time.

“When I asked for money for the server to draw up the plans, I told them that the server we’re thinking about is about $95,000, but it’s good, because it’s got 4 TB of [storage] and we’ve got 64 GB of RAM, so we have the room to grow,” she explains. “So I’m also thinking of the future. I’m thinking that in the third or fourth iteration of this analytics server, I’m going to want another 4 TB of [storage], I’m going to want to add more memory, but I showed [management] that the cost [for both] is going to go down [over time].”

Now that Apex is live on SAS9, Porrazzo uses her BI and data mining tools to assess and understand the relationships between the insurance plans, benefit offerings, and bottom-line costs of her company’s clients, which mostly consist of large insurance firms.

“For us, the business issue here is to develop health plans that are cost-wise and cost-effective based upon outcomes. That’s unique—to underwrite a policy based on outcomes,” she explains. “The whole point to me of business intelligence is changing behaviors, modifying behaviors. If you are a physician and you’re seeing that there’s a high trend in prescribing these neuro-psychotic drugs to younger people, the first thing I would do if I were a director [would be] to implement a physician education program so that people understand what they are doing. The minute you’re able to see the future you can change the future. If you’re quick about it, you literally can turn on a dime. That’s where we’re heading.”

In some cases, Apex’ forecasting simply identifies subterranean trends. But this in itself is valuable. For example, says Porrazzo, almost anyone with any knowledge of the healthcare space knows that antidepressants are among the most frequently prescribed of all drugs. But very few healthcare professionals know what class of drugs is prescribed even more frequently than antidepressants. “Drugs to treat rheumatoid arthritis [are] trending out of the world, but we don’t know why yet. We’re just at the point of being able to identify this trend, using our BI [tools],” she comments. “This information that turns up can be made available to a health plan or a healthcare provider in an instant, so once you run the data, it’s up on the Web, there it goes. If you can see what’s trending up, like rheumatoid arthritis, you can develop a program so physicians and patients are aware of what’s going on, of what [doctors] are [prescribing].”

What’s next for Apex and Parrazzo? Text mining, which should deliver value in the form of both text analysis capabilities and base-level ETL functionality, courtesy of SAS’ Enterprise Miner and Text Miner solutions.