IBM has launched two separate projects with two top health systems. One involves the use of big data to help clinicians provide predictive care at Emory University Hospital. The other is focused on reducing vendor fraud – again by employing big data analytics.

The company announced the details of its work Nov. 4 at IBM's Information On Demand conference.

Emory University Hospital is collaborating with IBM and Excel Medical to employ a novel use of big data to help ICU doctors and nurses provide predictive care for critically ill patients.  

As IBM executives explained the project in a news release, Emory’s "ICU of the future" will identify hidden patterns in data and provide real-time insights so clinicians can spot early warning signs before a critically ill patient turns for the worse.

[See also: IBM targets big data with acquisition, new research.]

Using "streaming analytics" from IBM, clinicians can be alerted and intervene as soon as serious medical complications develop. The sooner doctors and nurses know a patient is at risk, the better their chance for survival.

According to the Society of Critical Care Medicine: Up to 29 percent of the 5 million American adults admitted to ICUs each year will die before leaving the hospital. And sepsis – a leading cause of death in the ICU – carries a 25 to 30 percent mortality rate, but early intervention can boost survival by more than 15 percent and cut hospital stays by five days.

Emory’s collaboration with IBM is the first to apply streaming analytics across the entire ICU system.

Other medical institutions are also using the technology to transform critical care: UCLA physicians are piloting the technology in monitoring traumatic brain injury patients, and Columbia University researchers are exploring IBM streams to detect life-threatening conditions in stroke patients.

[See also: UPMC and IBM keep focus on big data for personalized medicine.]

Financial losses due to healthcare fraud are estimated to range from $70 billion to $236 billion annually, with significant losses tied to vendor billing schemes, according to IBM executives.

Working closely with IBM, Memorial Healthcare System is tackling this problem head on by applying big data analytics to vet its third-party vendor network, IBM officials said in a news release.

Here's how they described the project: Combining a content management system for consolidating accounts payable processes with an intelligent analysis system that checks vendor data against more than 800 internal and external databases, Memorial Healthcare unlocks new insights that manage potential vendor risks, including conflict of interest and even criminal behavior.

In one case, the analytics revealed that three vendors were colluding to price rig a proposal. In addition to rooting out and preventing fraud, the hospital system has also increased vendor invoice processing times leading to vendor discounts of more than $2 million.