GUEST BLOG: Can ‘Big Data’ Save a Life?

Truth, accuracy, precision. Effective analytics have always required that the incoming data be valid and reliable enough to support useful conclusions. That’s nothing new.  What is different now is that with more and more different types of data coming at us faster than ever, big data has created new demands on the methods employed to ensure data veracity, as it specifically relates to the method and its objectives.

In our scenario, I have suggested that linking simple grocery and pharmacy retail data to medical claims will provide new insight into consumption behaviors predictive of health events or outcomes. The first challenge is to match individuals across sources. With consent, it should be possible to effectively match using common identifiers like social security number and address. 

What we need are sources of disparate data that can be searched for this elusive signal hidden within the noise. The data might be grocery store or pharmacy sales records, or text messages, or Google searches, or book sales.  And, we need to earn the confidence of our citizens and our lawmakers and regulators that the security and rigor required to safely wok with this data will be held to the strictest standards. These are not easy hurdles, but we are closer than ever to making the real, lifesaving promise of big data a reality.  

Bob Kelley is senior vice president, analytics, at Truven Health Analytics; bob.kelley@truvenhealth.com