About three years ago, the physician leadership of Cornerstone Health Care, a large physician group with more than 360 doctors based in High Point, N.C., decided that the future of patient health was outcome-based. That meant the solution to keeping people healthier was not to do more testing, but rather to better analyze patient data to identify patterns of illness and suitable remedies.
This was a somewhat radical concept at the time.
"The science of prediction modeling has been around for a while in the payer world [of insurance companies], but it is relatively new to providers," said Dr. Dale Eric Green, an M.D. and Cornerstone's chief medical information officer. "The big issue is data overload."
The reasons for using such modeling are compelling. New technologies have made it more feasible to gather and analyze large sets of data from various sources. In the past decade or so, health-care providers and insurers have converted to digital patient records, while the federal government has made publicly available troves of data from clinical trials and patients. Drugmakers also have aggregated electronic data culled from their research work.
(Read more: When the Boss Is Big Data)
Together, these vast data sets can point to unexpected patterns and help increase effectiveness and reduce costs. The health-care industry's implementation of big data has trailed other sectors, such as financial services and retail, but demand is growing. Using advanced analytics could reduce health-care spending by $300 billion to $450 billion a year, according to a report from consulting firm McKinsey & Co.
But some health-care technology professionals are skeptical about the industry's ability to fully realize that promise.
Steve Huffman, vice president and chief information officer at Memorial Health System of South Bend, in Indiana, said providers already use data to help improve care for high-risk patients, for example, in obstetrics. Organizations must identify their needs before increasing budgets for so-called big data solutions, he said.
"We need to start with a business problem, not a buzzword," he added.
(Read more: Retail Goes Shopping Through Big Data)
At this early stage, it's more important for organizations to ensure any system they implement is flexible enough to incorporate big data strategies, he said. "In health care, we fall into this best-of-breed mentality," Huffman said. "Meanwhile, we're trying to unwind ourselves from all our data."
On the other hand, some major insurers appear to be luxuriating in the torrents of fresh data coming their way.
Michael Palmer, head of innovation at Aetna, said combining newer sources of information, such as lab results, and prescription and clinical data, with existing patient records provides his company with opportunities to answer questions they hadn't thought to ask.
"We're determining what needles to look for in the haystack," Palmer said.
For example, Aetna has been able to identify patients at risk for metabolic syndrome, based on a number of common risk factors. According to Palmer, the goal is to provide as much information as possible to clinicians and physicians, who can then recommend behavior modifications to patients backed up with particular risk information.
"We're trying to understand whether giving patients much more specific data about their individual factors will have an impact," Palmer said. "If we can say, 'In the next year, you are 87 percent likely to go out of range for your glucose level,' we believe it can have an impact."
(Read more: Big Data Becomes a Market Darling)
Palmer said that Aetna is exploring pharmacogenomics, where big data can help connect the right drugs and the right genomic markers, allowing physicians "to manage the best medication for a specific patient."
And while patient privacy is always a concern, Palmer said that because of the Affordable Care Act—which as of Jan. 1, 2014, will prohibit insurers from discriminating based on preexisting conditions—insurers soon will be able to "take some handcuffs off some of the data" that can lead to innovations and insights.
Stephen Blackwelder, Cornerstone's chief data scientist, says his group's big data strategy relies on "pairing with insurance companies and using their data." But certain data sets remain that the company cannot incorporate into its analysis.
"We can't see what is outside of our network," Blackwelder said. If a patient goes to the emergency room between doctor visits or is under psychiatric care, that information will not be factored into the overall analysis of his or her condition.
Nevertheless, as providers struggle to cast a wider net to capture more patient data, they are continually surprised by the insights made available by big data analyses.
"The insight I did not quite expect was the impact of pharmaceuticals [on patient health]," Dr. Green said. For example, he added, "the drugs used to treat Alzheimer's are pretty dangerous medications."
Read More: Big Banks Cashing in on Your Feelings