How a Health Insurance Provider Uses Big Data to Predict Patient Needs

    Slide Show

    Questions to Ask Before Implementing a Business Intelligence/Analytics System

    The next time you’re on the phone with your health insurance provider (in the event that you’re one of the lucky ones who actually has health insurance) and you’re struggling to straighten out some billing mix-up or to find out why they refused to pay for some procedure your doctor ordered, take heart. The information your insurance provider gathers during that call might help it to take action to prevent what would have been your next hospitalization.

    It’s all about Big Data and predictive analytics, and if there’s one place predictive analytics vendors see dollar signs, it’s the health care market. Big Data and health care appear to be a match made in heaven for these guys, and it’s not just the health care providers that are being targeted, but health care insurers, as well.

    A case in point is the announcement last week that InsightsOne, a predictive analytics vendor in Santa Clara, is working with Independence Blue Cross (IBC), a health insurance provider in the Philadelphia area, to identify patients in need of care even before those patients recognize the need themselves. I spoke about all this last week with Bob Dutcher, vice president of marketing at InsightsOne, who explained the idea behind the collaboration:

    What we’re doing is getting data from call center notes, such as the procedures that were done with the patient, and we’re also getting information on who’s filing and what the grievances and complaints are being filed about. We then analyze that information,  and we can produce two things. One is a list of patients who are at risk of having an issue, in which case IBC will address those [potential] problems with the individual. Or, what we’re working to provide in the future is to integrate that information into the call center so when the call comes in from a patient, the call can be routed to the appropriate person, who can be aware of the potential issue even if the person doesn’t bring it up on the call, and the call center rep would be able to proactively handle that issue for them.

    If the thought of your insurance provider having all of that medical information about you and doing all sorts of stuff with it creeps you out, you’re probably not alone. Here’s what Dutcher had to say to assuage those fears:

    When we get this patient information, we actually don’t know who the patient is. We have no ability to identify individual patients, so the patient’s identity is secure—that’s completely shielded. Think of it as a large glob of data that we cannot identify any individuals from.

    I wondered where HIPAA compliance fits into all of this, so I asked Dutcher whether this predictive analytics technology ventures into any area that HIPAA might have been unprepared for. He said HIPAA compliance was built into what they’re doing with IBC from the ground up:

    I don’t think it falls into any area that HIPAA would be unprepared for. Any solution that Blue Cross or any of the insurers or providers are going to use have to be HIPAA-compliant. We’ve gone through a HIPAA compliance exercise to make sure that we are compliant. As we’re getting into deployment with Blue Cross, they’re going through an audit of us as well—they would not go into production with us without the HIPAA compliance.

    Dutcher went on to explain that InsightsOne performs the analytical work as a third-party service provider, rather than licensing its technology to IBC for the insurer to use in-house:

    Basically, we get data in, and we send results back. …They’re using the InsightsOne virtual private cloud with the whole HIPAA compliance around it, which brings down the cost of deployment. … IBC has [its own] predictive analytics capability in-house.  But what we were able to do is demonstrate that in these specific cases, we were able to far outperform any capabilities that exist out there that they were familiar with—and they were using some fairly sophisticated stuff—in some cases by 400 percent. So if you’re looking at patients who are at risk of having an issue, being able to more precisely know what the patient [data reflects] so you can better address those issues makes it actionable. Now you can do something about it that has an impact on the patients and on the business.

    Get the Free Newsletter!

    Subscribe to Daily Tech Insider for top news, trends, and analysis.

    Latest Articles