"Big Data" is not really solving health problems.

I fear that the public is being oversold on “big data” in healthcare. The fact is that if big data were a drug it would be classified as early stage and highly experimental – still working on saving mice and rats, not transforming human lives in the ways most consultants would like us to believe.

Big data in healthcare is actually three things: Clinical Decision Support (CDS), creating treatment algorithms based on specific patient characteristics to optimize care; Population Health, identifying patterns in how people live, eat, exercise and use of healthcare resources that lead to markedly better or worse overall health; and Disease Insights, finding new ways to identify, predict and track disease that lead to novel treatment and prevention strategies.

The biggest breakthroughs have come in the area of population health, largely through longitudinal studies such as the Framingham Heart Study (which gave us the term “risk factor”), and disease registries, which have helped uncover important insights into the genetics and progression of disease.
The EHR scenario that gets the most attention still has a long way to go because the data is generally organized by treatment and diagnosis codes, so the type of outcomes most people are concerned about must be inferred. We may know that grandma is being treated for Parkinson’s disease and is receiving certain drugs, but in most cases we don’t know if she is spending her day sitting in a chair in the corner or actively engaged with her grandchildren.

This is not to say that big data or the approaches outlined above are bad. We simply need to temper our enthusiasm and focus on what is important. For instance, the biggest obstacle in applying big data to EHR’s may not be so much scientists capable of handling noisy data, as agreement on standardization of EHR data, and central repositories where it is possible to view data across systems to allow for suitable sample sizes. We also need to make better use of the data generated through the hundreds of millions of dollars spent every year on drug trials that companies are not able to share for regulatory reasons.