Follow the Data: How Anyone Can Ride the Big Data Healthcare Wave
In the age of big data, information often grows exponentially. I am simplifying here for brevity, but – It starts with a data void. Next someone discovers how to collect that data available. In the next step, the quality and amount of information starts changing. Then, suddenly, there is a tsunami of data, but early on, much of that data is noise. The next few steps, however are crucial – the aggregation, the organizational and analytical innovation steps. Many people will come up with ways to aggregate and organize the information – essentially liberating the data. Once this is achieved the information now becomes actionable and/or put in a format that allows for different forms of analysis. Once actionable and/or analyzable – businesses can be created and/or destroyed.
That’s where the sweet spot, and where the money is.
Look at the evolution of location based services. During the “void” there were no good means to even determine the distance between two points on the globe. Distance was a crude method of measurement and did not allow the consideration of additional information, including directions and travel time. Global positioning (courtesy of the US Government) came along, but at first created much more data than a consumer could use, it also didn’t include the data that were most relevant to them. Then companies such as NavTech and TeleAtlas created digital map sources. Once we established that the next embodiment was the car navigation system. Now, you could find the shortest travel route or choose to stick to back roads if you wanted. All the beginnings of what would eventually permeate every part of modern life (more than you can possibly imagine). Once Google and others got a hold of the data in its now aggregated and organized form they then synthesized these reams of data and devised ways to display it in a consumer friendly way. Further innovation has made GPS a part of daily living, as people use it to avoid traffic snarls, find shops, get deals on products and even connect with friends on a night out. Whether you know it or not – location based services ($40B by 2019) are involved in every part of your life to make it “better” and answer those questions relevant to you and your coordinates. And frankly, some of the best and most useful applications are yet to come.
For businesses the most important period of time for decision-making is that moment I’ll call the “data quake” – when there is much more noise than signal. Sophisticated analytical types, such as “The Signal and the Noise” author Nate Silver; create complex models that help them make predictions during these times. But you really don’t need to have those skills to prosper from a data quake. You do need to be aware of how the data is changing, being organized and what its implications are. Because the use of the data is likely to change dramatically once people have figured out how to aggregate and parse it usefully. The question you must always ask – how will the change in the data affect my business and/or the business model?
The Noise however, can have a paralyzing effect on people. This is also known as “the paradox of choice.” There’s the classic example at 401ks. Employees given a large variety of investment options are more likely to choose less complex funds or not to invest at all. Another classic study found that consumers were ten times more likely to purchase jam if they were offered just six varieties rather than 24. Making decisions is stressful and fatiguing, people are hungry for information that is useful, not a ton of data that they need to sort through.
Another powerful example of this in health care is companion diagnostics. These are tests that help physicians determine when to use a particular drug (i.e. The test is a companion to a drug). The FDA has allowed over a hundred such tests to be mentioned in drug labels and more are on the way. How many physicians are likely to remember all of these tests, let alone the diagnostic information from over a hundred different sources? Companion diagnostics are still in the “noise” phase and it will be the companies who make it simple and convenient for doctors to select among tests who will capitalize most on this emerging field. The key phrase here is Clinical Decision Support (CDSS).
It’s also critical to understand that different fields move at very different rates. Because there are so many regulations in healthcare, particularly around safety and patient privacy, innovation is much more challenging. Circumstances can also speed or slow data evolution. Accelerants include first the digitization of data. Then how it is quantified. And finally comes the leap – the analysis. But the breakthrough step is – what if I combine one data set with another. That is where the magic happens. As an example, think Yelp (Mapping, Restaurants, Ratings, etc.). The key question in healthcare – Is the outcome relevant and therefore worthy of interest? Does it improve patient outcomes? And does it help in being cost effective.
Evolution is slowed when there is a lot of regulation, there is high risk associated with providing incorrect information, or the data is disjointed – does it have to be translated or otherwise heavily manipulated to be useful?
One type of data can also undergo multiple rounds of evolution. Molecular Diagnostics, for example, began with Gregor Mendel’s crucial discovery of laws of heredity in peas. Watson and Crick made another huge leap forward by determining that DNA was the fundamental blueprint of life. But while we knew what DNA was and much about how it worked, we were still confronting 3 billion little bits of information in the entire human genome. The Human Genome Project created an explosion of DNA and software programs like BLAST eventually made it possible for us to begin parsing that data. Finally, the discoveries about particular genes started making their way into the clinic.
Myriad, for example, sells a test that can tell women with a family history of breast cancer whether they have inherited a predisposition for the gene. If they find they do have the gene (BRCA), they can either undergo more screenings or take steps, such as prophylactic surgery, to try and avoid the disease.
But now, researchers are starting to realize that that while the BRCA gene is important, it is by no means the only one that influences someone’s cancer risk. In fact, we are undergoing an explosion in the number of known cancer genes. But many of these diagnostics aren’t making their way into doctors’ offices. Instead a new type of company, Foundation Medicine, for example, has emerged to aggregate the data about all these tests and help physicians more easily decide which one their patient needs.
There is so much more to write, but I’ll put a stake in the ground here – So what is the take away from all this? Every professional should be constantly analyzing the data in their field. We should be asking:
What data is relevant to my work now? Which is the most relevant data? What could it be tomorrow?
How have the data in my field evolved and/or is it being digitized? At what point of maturity is my field?
How can I add value to that data? Or how can I use it to increase the value of the services I provide?
How feasible is that? And what is the estimated return on investment for the effort required?
You can either diligently track the big data explosion or be the dinosaur that may be destroyed in a data quake. It’s really just beginning, but moving fast. Once you digitize anything – the change is exponential.
More to come on, how data are changing the practice of medicine…. Stay tuned.