I saw a Quote recently. Big Data is like teenage sex. Everyone talks about it, but nobody's done it.
I personally like to think of Big Data as a re-branding of things we already did, such as epedimiologic studies, registries ext. but it aspires to be much more with an integrative component.
Alot of Big Data is centered on understanding Patient data to include predictors of disease Progression, benefits of Interventions, safety risks, cost benefits etc. The Big Data Comes in with respect to the various sources of data, desire to pool and integrate the data, and spit out Analysis. So for example, there is talk in Europe about linking electronic medical record data bases, or connect Patient registries, or pooling individual Patient clincal Trial data. Alot of this is trying to get the question on value for medicine and in some cases leverage for policy setting or coverage decisions etc. (is the Intervention worth it, to whom, and Overall benefit and risk).
The Problem i see which is something that would Need to be addresses is pooling of data and how to do that there are data-sets that are of Quality and thus useful. Other barriers include Patient privacy or data protection, disharmonized data collection and adjudication/coding, etc. etc. etc. The question is what's big? What do you do with it. Who's done it right? I'm not sure anyone.
If you take the US VA Hospital System, that's pretty big data set, but to what extent ist valuealbe is a function of the question you're trying to ask. But a great data set nontheless yielding alot of nice epi studies. There are Claims data from insurance companies, so they're getting on the bandwagon and working wiht pharma companies too and so and so on.
Personally, I see Big Data either in the realm of Industry, Gov collaborations, and a bit of academic talk, so then i think this beomces specific to the companies with deep pockets - usually playing in the high value chronic disease, i.e. Oncology, Inflammatory Disease (i.e. RA, IBD) and to a lesser extent cardiovascular/metabolic diseases. If you're in a smaller Company - bah - you do what you Need to do commercialize a product, leave the Big data stuff to the academics and big spenders.
Now if you want to get a pulse of how Big Data is being used effectively and I mean really big data, look to the agricultural industry - John Deere is a good start, see what you find there. They got so much data - that can tell you where on your 1000 square mile farm in Russia you may have dry soil and they can direct your tractor and a move Combine to meet up with that tractor to manage that automatically (all without Drivers). All in real time, and they can tell you when a part of the tractor is about to malfunction in real time and do a Software fix to manage until a automatically deployed technician Shows up on the Scene, to meet that tractor automatically in route and have that technician avoid an isolated 100 meter square foot hail storm and a 3 foot pothole by 5 inches and tell you if your pigs Need feeding all in real time...now that's big data working.