OR WAIT null SECS
In part 2 of this week's C-Suite Q&A, we continue our conversation with David Delaney, MD.
In part 2 of this week's C-Suite Q&A, we continue our conversation with David Delaney, MD, the Chief Medical Officer at SAP Health. The conversation covered the imperative of value-based care efforts, the technology behind SAP's analytics platforms, and the relationship the enterprise giant has with the exploding field of startups in health tech. Part 1 can be read here.
We were recently at the Philadelphia HIT Summit and it seemed like a theme that people were coming on stage and literally saying “MACRA sucks.” You’re trying to provide analytics to ease this move to the value-based models. What’s the demand like, and can you describe the mood of the providers who are now coming to you in this position?
There’s a healthy degree of skepticism in coming along with just another analytics solution and promising all sorts of impact. There’s been a lot of selling in the market place, things have been overpromised and underdelivered. The challenge the organizations face is that many are doing some good work in analytics, and I travel around and meet with them. The challenge is that it simply doesn’t scale, they’re coming at this in a way that’s very labor intensive. You look at how a data scientist or an analyst spends their time, about 80% of it is spent on data prep work, essentially nudging data together in Excel, cleansing it, normalizing it, and moving it into the new data set. Only 10% or 20% is actually spent on creating value or insight. That becomes a choke point for organizations: they can do very good work, but it’s very people-intensive, and the good work that they’re doing simply can’t scale with the number of bodies they have.
Data scientists are expensive, and they’re also scarce, as are good analysts. Even if they have the budget to put 10 more of them in there, they couldn’t even hire them. It’s a resource constraint thing, because it’s very inefficient. Our approach using native memory computing is really that we don’t say you need to pull out all the systems you’ve already invested in. You can leverage and extend your current data assets, and you can use this essentially to augment and really grow the ability of your existing data scientists and analysts to be productive. You can dramatically accelerate their throughput and you can begin to invert the equation, so that a far higher percentage of time is spent on data science and a far lesser percent on the whole data engineering piece. So we dramatically simplify the process of pulling data sets together, running analysis on them, and creating insight.
We do this not only for the data scientists and the analysts but we have some graphical use interface tools that allow domain experts to begin to do self-service. One of the challenges is that organizations have this tremendous backlog of requests by their various end users, whether they be clinicians or administrators, and it really impacts the data scientists’ time left to do data science. By allowing self-service for these end users so they can do their own ad hoc queries and reporting themselves, and do some data discovery, it really helps free up time for data scientists and analysts to begin to look deeper into the data.
That’s really our value proposition, the ability to do data analytics at-scale across an organization, largely using the analysts you already have in place.
Going further on these ad hoc end user queries, can you be specific about what types we’re talking about? The sorts of questions a doctor or administrator might be punching in?
A good example might be a patient who comes in with an unusual type of cancer and wants to know what other patients have done in that given institution, and provider can sit down using a graphical tool and immediately hone down to a group of patients similar to the one at hand and see what therapies have been used and what outcomes have happened. Maybe a clinician notices that a person who has received a new med is having worsening renal function a few weeks out, and they think that it seems similar to another patient they saw a few weeks before. Is that kind of a happenstance observation, or is there something more to it? The ability to go and use this graphical tool and drill down to answer that question for the patient population.
It’s really the ability to have a question and to drill down. People are using it in terms of supply chain, too, looking at utilization of supplies and find places where there is waste and improve the efficiency of surgical procedures, for instance. We have some good data from the folks at Mercy Hospital in St. Louis, they were a Davies Award winner at HIMSS this past year, and they’re using the tool to optimize their OR supply chain. They’re focused a lot on the knee replacement area and doing cost takeout there, and they were able to do some really interesting identification of it. They were able to switch to our platform and have all 43 of their hospitals’ data in for all of the ORs at all times and do the analysis.
Here’s a question in a different vein. Tech, and health tech, is sort of a startup party, a lot of new, small, innovative companies coming into the fold. SAP is obviously not new or small, it has a reputation. What’s it like to be kind of a big fish, or the old guard, or whatever you want to call it, in a field that’s increasingly fresh-faced?
Well the HANA database, the in-memory platform this is built on, was actually very, very disruptive to the marketplace. We launched this probably about 6 years ago now, and at the time we were just a bit player in the database market, by virtue of our Sybase acquisition. There were some interesting comments in the press from some of our rivals saying that ‘SAP should stick to its accounting’ and that we had no business in memory databases. From 6 years ago, and pretty rapidly through the next several years, all the major competitors were responding with in-memory components for their own databases.
The technology itself is very disruptive, in terms of approach. You look at how people classically look at data in disk-based systems, the relational systems were largely written and first deployed in the early ‘80’s. The challenge they face in pulling data from a spinning media drive into main memory, or even with newer versions using solid state drives, it’s orders of magnitude slower than the ability to pull data from main memory, particularly in this era of multiprocessors and multi-cores. Using that technology to run data in a massively multi-parallel fashion into every single core of a blade and be able to scale that out has actually been very disruptive to the industry.
We are a disruptor to the major incumbents, in terms of scale. The technology approach is very different and it really powers the ability to do a lot of what we’ve been talking about. In terms of innovation within the startup ecosystem, we welcome it. There was some news recently that we partnered with Startup Health, the accelerator program, and we’re in the process of forming collaborations with some of the earlier startups using our technology. Organizations that have no preconceptions, that are just pursuing time to value with customers, they create a vibrant ecosystem which can create great innovation. We are looking to partner with that.
This is a massive, massive opportunity when you look at the value unlock that can be achieved by actually running organizations more efficiently through their data, it’s just huge. This is not a sector that’s going to have one player, even just two players: there’s a lot of opportunities for a lot of companies. We very much welcome the vibrancy that comes from these small disruptive players.
Closing question: what do you personally get out of all of this?
I love technology. I did my clinical training and practiced for 14 years as a critical care physician, but I also did a medical informatics fellowship and coded for quite a few years, and recently restarted as a hobbyist because I find it’s great to just stay deep in the technology. I love the evolution of it, just the incredible capabilities we now see in hardware and software, that part I love.
The other piece of it is that I’m incredibly fortunate to spend a lot of my time out working with really amazing customers, partners, and some people that we talk to who even might never become customers, but who are doing some really incredible things. I get to talk to really smart people throughout the week, at a amazing institutions, and see all the disruptive things they’re doing, and understanding their challenges, and thinking again about how technology can solve them.
I find it incredibly rewarding and intellectually stimulating, and a tremendous privilege. Of course the hope is that we’re able to help folks take better care of their patients, and that’s a goal that we’re very excited to help achieve.