Q&A: How Ochsner is 'Amazon-izing' itself with big data

By Kate Spies
09:23 AM
Share

Ochsner Health Systems came by its big data organically. That is to say the IT folks did not so much kick-off a big data strategy as begin aggregating internal systems – only to realize “wow, we have a lot of data here.”

Government Health IT Contributor Kate Spies spoke with CIO Chris Belmont (pictured at left), and enterprise information architect Jonathan Stevenson about the providers’ big data plans, how it intends to employ predictive analytics to improve patient care after an Amazon fashion, and the challenges particular to public health.

Q: How has Ochsner been approaching big data thus far?
Belmont:
Ochsner started more or less automating and collecting clinical data with Dr. Lynn Witherspoon - who’s now our Chief Medical Information Officer - who created a database to start storing lab results, and that morphed into an electronic medical system that’s been serving us well probably up until about a year ago. It’s actually still in production today. But we needed to replace it with something more enterprise-based, something that would be a little more agile, a little more flexible and scalable as Ochsner continues to grow. We collected the data, but the issue is getting it back and getting it into the right hands. We also, over the years, have aggregated about 38 different systems that collect clinical data; we have a total of about 225 information systems that we support here at Ochsner. And all of those systems, more or less, keep their data in islands.

The real value is not necessarily reporting out of those individual transactional systems, it’s the ability to aggregate and correlate that data horizontally across those organizations. So for example, if we want to look at labor and how that’s used, it would be nice to say, ‘Okay, we had this many lab tests,’ and then to match that against how many hours worked, which is in another system, and then to say, ‘what does that look like from a productivity perspective?’ So that’s just one example of thousands of opportunities we have out here to look at the data. And again, the big data piece more or less evolved because if you look at the systems that we’re aggregating today in our current systems warehouse, if you slam all that together, plus the number of years that a lot of the system has been running, it just kind of showed up as big data. We didn’t go in with a ‘big data strategy’ – once we drew it altogether, we realized, ‘wow, we have a lot of data here.’ And it’s not just big data in quantity, but it’s big data in complexity as well; the data doesn’t always match when you take it from different systems and put it together. So there are data governance decisions that need to be made around that.

Q: Why is Ochsner working to mobilize big data?
Stevenson:
Healthcare in itself is changing. It’s morphing into a different market essentially, where we’re going to be held responsible not just for the volume of patients that we manage but also for the actual outcomes relative to the volume of patients. So the magic in making that successful for any organization is the data. So why are we undertaking this strategy? It’s because we have to, not because we really have any choice relative to the way the market is moving.

Q: Why is the healthcare industry behind in terms of harnessing big data?
Belmont:
At the end of the day, big data is here to help us take care of our patients better, and also to service our community better. And one of the reasons I think we’re behind has just been this traditional approach of using the data for specific cases, versus using our data for operations, and not more on how do we look to serve our community better, and serve our patients better. Unfortunately, what healthcare does is a lot of retrospective reporting, so we gather up the data, we give it back, and say this is what happened yesterday, or last week, or last year. We’re not doing well at what other industries have done – whether it’s finance or retail, and so on – to do more of the surveillance and predictive stuff.

To read more of our interview continue onto the next page...

When you order a book or an item from Amazon and they say ‘people like you have also ordered these things’ - we’re not doing that progressively in healthcare. I think we need to, more or less, ‘Amazon-ize’ ourselves; we need to get out in front of things and try to understand what’s going on in our organization and stay ahead of the curve and not react to patient events as they occur, but ask, ‘how do we get out in front of them, and prevent them from occurring?’

Q: What have been the challenges you’ve encountered as you work to harness big data? What challenges are specific to public health entities?
Belmont:
That culture of being more retrospective and reactive: as in, ‘give me a report and I’ll act on it;’ that’s so 1990’s, and before. So I guess our culture still is, ‘give me a two-dimensional piece of paper with data on it, and I’ll find out things.’ We have more of a pull-mentality of ‘I’ll go out and get a report, and I’ll pull it when I remember it.’ I think we need to move more toward a push-mentality, so as events occur, we can present it to the right people so they can react to it, whether it’s a physician, or an operator who is managing a piece of the business and so on.

Stevenson (pictured at right): On the data side of things, to us, the technology is not scary. We have all the data, and the technology to be able to pull that data together and smash it, merge it, and make sense of it. To us, that’s the easy part; technology is easy.

We’re a medium to large-sized healthcare organization. Data lives across silos, and every silo has an owner or a hierarchy relative to the ownership of that data. What’s interesting about that is, getting the silos to become a more centralized platform, a more centralized methodology, is kind of easy on the technology side. On the people side - the governing side - is where it’s hard. We’re finding that a lot of our challenges are not necessarily getting the data to make sense, but getting the people to agree that the data makes sense. So going forward, our core focus is really going to be on the people, and not necessarily on the technology.

Q: Looking ahead to the future, what are some things you think all healthcare organizations should be considering right now to harness big data?
Belmont:
I think there are some immediate issues, there are some near-term issues, and I think there’s long-term. So immediately, we’re being faced with things like meaningful-use reporting that’s being put into place. We have to show meaningful use and that’s going to require a lot of data collaborated from a lot of different systems. A lot of that information does not exist in one of the core platforms, so people are going to have to create big data type structures so that they can do this reporting. And remember that meaningful use is not just during the incentive period. In 2014 after the incentives are over and beyond, it will actually be a penalty. So if you’re not a meaningful user after the incentive period, they’ll actually reduce your reimbursement. So I think that’s the immediate thing.

I think, near-term, as healthcare continues to morph, how do we provide data to support the decisions to do the modeling? So if we decide to go into an accountable care organization, how do we have data that supports that, instead of just playing a hunch that says, “we think that’s a good deal and let’s go for it?”

To read the end of our interview please continue onto the next page...

These changes are going to be significant, and not only costly, but changes in the core business operation. So, how do you use data to model those things? Down the road, how do you get predictive – whether it’s genomics, as in, ‘I have a person that’s a member of a family that has a history of diabetes, and how do I start managing that population going forward?’ So, again, I think that’s what we want: Tell me what I can do to make sure I live a long healthy life and have the right outcomes, and how I can head off adverse events using the data that’s out there.

Stevenson: If you look at this from a health IT perspective, and you don’t look at it from a strictly healthcare perspective, what health IT needs to be aware of is a couple of things: you need to be aware of the fact that healthcare is changing and that the model of care is changing. They need to be focused on improving care for the patients, where traditionally health IT has been focused on improving quality for providers. So healthcare is moving toward more of a commodity-based model, kind of like the Wal-Marts of the world or the Amazons of the world. What we can do to differentiate ourselves, not only from a cost perspective, but from a quality-of-life perspective, is going to be very important for our patients. Health IT needs to figure out how we can support the practice to be better providers of care for the community. So an example of that in our instance is: we live in the world of an electronic medical record here. In IT, our core systems are our electronic medical records. How we can improve the experience in the electronic medical record for both our providers and our patients, as well as provide the predictive analytics is going to be core and critical to differentiating ourselves in the future. So I think what health entities need to be focusing on is how to differentiate themselves from other organizations in their area so they can improve the care of their patients and increase their market share.

Belmont: The other thing is we’ve got to stay way out in front of this – right now, we are playing catch-up, by the time we deliver what I’m working on, it’s going to change. So an example of that is, think about the world of ATM’s. So you can go anywhere in the world, and, with your ATM card, you can get cash. But the reality of it is, people are going cashless. So I could get really aggressive and build this really strong ATM network that pushes data out to everybody, and them I’m going to wake up and say, ‘well that was made for yesterday, and now what do I work on?’

So we’ve got to look down the road and get out in front of this, instead of continually playing catch-up. Retail is out in front of us as far as the predictive analytics piece; we’re not doing any of that. By the time we get there, there will be something new. So we’ve got to look at what else is happening in the IT industry and what other industries are doing that can compliment us. Because at the end of the day, we are an information-driven industry; even though you’re taking care of patients and you’re touching humans, it’s all about the information that you’re giving that physician, whether it’s him collecting it at the point of care, or us giving it to him proactively.