Will big data leave game-changing insights in its wake, or will it overwhelm an already overburdened healthcare system?
Big data promises to bring never-before-seen insights that can change the course of history in medicine. At the same time, it brings security risks and inefficiencies that can crumble entire institutions and derail healthcare’s forward momentum.
Effective navigation of the big data sea change depends on our priorities. Will we focus on collecting the right data? Will we make the investment into the tools needed to refine it? How can we make big data work for us? Our expert panel weighs in:
A Healthcare Analytics News® Peer Exchange®
Kevin R. Campbell, M.D.: I want to bring us back to talking about big data. We talked about controversies in big data. We talked about who owns the data. We talked about how we can share it. How might big data harm patients? Geeta, what do you think?
Geeta Nayyar, M.D., MBA: Sure. So first of all, it’s not a silver bullet. I think we’ve put a lot on big data, this idea that it’s all out there, but it’s not. Secondly, the accuracy of the data. We mentioned the data silos, right? So how accurate is the data? All of us as doctors using EHRs [electronic health records], they’re full of junk.
Kevin R. Campbell, M.D.: Garbage in equals garbage out. We’ve said that many times.
Geeta Nayyar, M.D., MBA: Correct. And fundamentally this is big, sensitive data, right? So, someone’s HIV status, mental health issues — if there is a security breach, we’ve seen it so many times in the industry, there are really serious ramifications for that individual — that person, that data has been exposed. So, I think that, again, organizations that are dealing with this, which is everyone in the industry today — good governance, good protection, cyber security protection — is very, important. They have to be the first steps in training, right? A lot of the clinical stuff — you walk into any number of doctor’s offices, there’s a user name and login for your PAC [picture archiving and communication] system, your EHR, your portal. And where’s the user name and password? On the sticky note right on the computer.
John Nosta, B.A.: I agree and disagree, as usual. You know I think that, yes, that’s a concern. My HIV status, my LDL [low-density lipoproteins]— that personal information that I don’t want shared. But there’s another domain that people consider just as personal, and that’s their financial data, how much money I make, what I have in the bank. And that network is seen as an extraordinarily secure network these days.
So, again, I see a duality. On one hand we say, “Oh, it’s very fragile, it’s data, we have to be careful.” Yet FinTech has done an extraordinarily good job at securing financial information. Do I worry about my 401(k) data being released? I know that it’s more in complex data set, but I think that there’s an interesting duality there that financial data is largely considered to be quite secure and well managed. Why can’t we do that with medicine?
Kevin R. Campbell, M.D.: I agree. I don’t think we’re there with healthcare. You think about how many hospitals have had their EMR [electronic medical record] highjacked with malware and they’ve had to pay a ransom to get it back. This happened at Blue Cross Blue Shield; their database got hacked. What if you’re a patient and your employer, or potential employer, finds out that you have these six chronic diseases and they decide they don’t want to hire you because you’re going to be out all the time?
David E. Albert, M.D.: That’s where it’s going to be expensive.
Kevin R. Campbell, M.D.: Or be expensive to the group cost.
John Nosta, B.A.: What about if you’ve been bankrupt? If you declared bankruptcy three times? What about if you’re a felon?
Kevin R. Campbell, M.D.: I’m not arguing that point.
John Nosta, B.A.: If I have a criminal record, if I have driving record, if I have bankruptcy issues, what would an employer do there? That may be as sensitive if not more.
Kevin R. Campbell, M.D.: But I don’t think the security in healthcare is near what it is in financial institutions.
Geeta Nayyar, M.D., MBA: There are lessons learned. There are lessons we can learn from the financial industry.
Kevin R. Campbell, M.D.: I think that the dataset — you know how much money I have in the bank is X, but we’re looking at a complex clinical history over time, there’s a lot of stuff that’s much more sensitive. So, I think it is because it’s a more difficult problem. But again, I think we should act like Picasso and steal.
Kevin R. Campbell, M.D.: All right. So, this has been a great discussion. What I want to do now is go around the table and get final thoughts; just a quick thought on what makes big data the most important thing, or the least important thing on our list for innovation in healthcare today. Let me start with John and we’ll just go down the table.
John Nosta, B.A.: Sure. I like Volvo as a car because their marketing is built around one word, and that’s “safety.” I believe that the future will be built, both societally and in medicine, around data — one word. But I think data will give us unique insights into clinical data points connectivity, between which we never saw a relationship. I think it’s going to be a game changer.
Kevin R. Campbell, M.D.: Geeta?
Geeta Nayyar, M.D., MBA: I agree with everything John said. I always say big data is an opportunity. It’s not a silver bullet, it’s an opportunity for us to get more accurate information on our patients, deliver better healthcare — and even better, more accurate, precise medicine. Predictive analytics is a huge opportunity for us to actually predict when someone is about to get sepsis; someone who’s about to have a fall. So, there’s immense opportunity, but I don’t think it’s a silver bullet, and it has to be safe and secure, period, end of story, or the story doesn’t begin.
Kevin R. Campbell, M.D.: Love it. Dr. Albert?
David E. Albert, M.D.: Today AI [artificial intelligence], our next topic, is at the peak of the Gartner Hype Cycle. For AI, this really translates in terms of the great strides we’ve seen in voice recognition, image recognition, and into what are calling deep neural networks. In order to train high functioning deep neural networks, you need lots and lots of data. Big data is an absolute requirement and it must be clean, accurate and well labeled. In my mind and in my domain, and in the domains of radiology, cardiology, neurology, oncology, big data will give us the ability to train neural networks, which will give us the ability to develop predictive analytics. And in medicine, prediction means prevention.
Kevin R. Campbell, M.D.: I couldn’t agree more. For me, the big data is all a big dichotomy, because we’re using big data — big population datasets — to create personalized treatment and personalized care. So, on the one side we have the big, big, big datasets, and then we have the precise antibiotic that you need based on your DNA.
Get the best executive insights directly to your inbox. Sign up for our daily newsletter.