Will Big Data Fade into Transparency in 2016? We aren't convinced.

IDC Health Insights recently published its U.S. Life Science Top 10 Market Trends for 2016 paper.

In this document, IDC focuses on “upcoming trends that will impact the overall life science ecosystem, including developments surrounding both business and technology in 2016.”

IDC’s Top 10 trends covered topics from risk-based monitoring to international reference pricing to big data. Big Data was given the #5 slot, and was titled “Big Data Will Fade into Transparency in 2016.”

I must admit – that as the CMO of a company specializing in big data health analytics – that one caught my eye.

Here is the bulk of the text that accompanied the top 10 assertion:

“The growth of new data sources (e.g. genomics, EMR, imaging data, and tracking of serialized products) [has] finally raised data needs to the level of “big data” in the life sciences. Technology advances, along with externalization of most IT functions to external partners, has enabled leading life science companies to keep up with ongoing data growth and focus most of their efforts on extracting value from data available to them. Big data management and analyses are currently being effectively managed in the background with the help of agility conferred by the cloud. As a result, IDC expects that only IT management will need to focus on big data issues over the next few years, leaving researchers and management to focus on science and business.”

Hmm… keep up with ongoing data growth and focus most of their efforts on extracting value from data available to them…

I am not so sure that is what we are seeing at Vencore. We routinely find that pharma and biotech – both large and small – are struggling with extracting meaningful value from the data they aggregate and purchase. They are used to working with one data vendor, who charges too much for their entire database, so they buy data “cuts” – one ICD 9/10 code at a time. This prevents them from ever having a truly comprehensive view of the surrounding landscape. They aren’t working with big data – they are working with pieces of big data.

We believe that in order to use big data, you have to be able to handle big data – terabytes and petabytes at a time. You need to be able to run analyses on 150 million patients – on all of their ICD 9 codes, CPT codes, NDC codes etc. If you decide ahead of time which codes are relevant – and then buy your data accordingly -- you have biased your results from the outset. Those who work in big data must always select a data set to answer a question – we cannot formulate a question based on the data available.

Tara Grabowsky, MD
Chief Medical Officer, Vencore