When Verdande Technology, a Houston, Texas-based provider of real-time case-based reasoning (CBR) technology, established in the energy sector for oil and gas organizations, connected with The Methodist Hospital System, also in Houston, it didn’t jump out as the healthcare IT marriage made in heaven.

But indeed, two of Houston's major industries—medicine and oil and gas–surprisingly have discovered many similarities in the technologies they use and the challenges they face. This discovery prompted a meeting to examine the parallels and determine if there are crossover technologies that could benefit each.

In 2007, in a session called "Pumps & Pipes 1," a group of petroleum, medical, and imaging experts met at the University of Houston's (UH) Texas Learning and Computation Center to explore their similarities in the hope of sparking solutions to problems inherent to both industries.

While drilling wells might not be the same as pumping hearts, Pumps and Pipes has only grown from that point forward—and now, some of its core principles are being implemented at The Methodist Hospital System, a five-hospital system whose cardiology team is working with Verdande to develop a CBR-driven application to combine historical pre-, peri- and post-operative patient data, bolstered by the cardiovascular team’s knowledge, to guide the design, implementation, and testing of the data mining procedures.

Perioperative blood loss is a well-researched problem in open-heart surgery, but the majority of the existing studies focus on select factors instead of developing a framework combining static, pre-operative information with real-time, dynamic intra and postoperative data, says Alan Lumsden, M.D., medical director of the Methodist DeBakey Heart & Vascular Center in Houston. At this facility, the infrastructure is currently being created, and implementation is coming soon, Lumsden says.  “The idea is to identify problems before problems identify themselves. Then you would say, ‘We have seen this pattern before,’ and that’s where the case-based analysis part comes in. We wanted to apply this in healthcare.”

To this end, Verdande and Methodist’s physicians have been working together to develop a decision-support system to evaluate patient risk and provide real-time, evidence-based decisions for physicians. Case-based reasoning moves beyond predictive analytics, by providing physicians with real-time data, based on past events, to help them make split-second decisions in the operating room (OR).

 Where predictive analytics is based on static data and often comes back with multiple “what if” scenarios that need to be analyzed by physicians, case-based reasoning automatically and continuously compares real-time data with past cases where the data has shown that action should be taken. As a result, the physician is provided with the best course of action.

“We want to get real-time data flowing out of our OR. And then look at real data flowing out of the intensive care unit (ICU). We chose open-heart surgery because a lot of data is collected from those patients. There is a national registry across the U.S. that predicts risk in those patients, and you plug in a whole bunch of variables to find out the mortality risk. So that is done preoperatively, and we have that as a baseline. But the problem is as soon as you start the case, the variables start changing.”

An example of this would be a transfusion reaction, which occurs after a patient receives a transfusion of blood. If something like that were to happen, or say a patient has a heart attack in the recovery room before he/she goes back in for open-heart surgery, that means a massive shift has taken place in the risk, Lumsden explains.

Physicians need data-backed, evidence-based information to leverage their instincts in real time, so they’re not just reacting to a situation, but instead become empowered to make better, faster, real-time decisions. Predictive analytics that are currently available don’t address the real-time aspect of surgery; The Society of Thoracic Surgery (STS) has really developed the predictive algorithm of risk for patients, and while that is as good as we’ve got, it’s by no means perfect, Lumsden says.

“Physicians often solve problems by being reminded of similar situations they experienced in the past. That’s how CBR technology works—it searches for past experiences and integrates that data with the patient’s real-time vital signs to let the physician know that prior events are about to repeat themselves. We have the opportunity to pull the collective wisdom from senior people in our institution, but also beyond the institution. This can be national wisdom.”

Verdande’s platform provides physicians with a visual which quickly communicates the ever-changing level of risk.  “This project is not only a matter of CBR in terms of helping us assimilate that data, but also in presenting that data in a unique format that physicians have not really seen before. It will let us act earlier and more specifically on the changing elements that need our attention. It makes sense to do risk stratification before the case, so it sure as heck makes more sense to re-analyze that and identify real-time problems.”

Lumsden admits he thought the technology part would be a challenge, but that has been overcome. “The other challenge is acceptance. Here is a whole new concept of how we’re going to operate by the bedside, but are we going to believe in it? I think so, because this can completely change how people handle data. I am expecting to improve survival rates.”