Cost savings are always key drivers of new initiatives. And in today's healthcare industry, as priorities continue to shift and pressure is added to increase revenues and improve outcomes, one element could be a key player in making it all happen: big data.
"We think it's going to separate winners from losers in many markets over the next five years," said Russ Richmond, MD, CEO of healthcare solutions and consulting company Objective Health. "The institutions that are capable of first understanding where the market is going … are going to have tremendous advantages over the ones who can't or won't do this. We believe that over time, it's going to become a core competency for hospitals, and it won't be something seen as extra or nice to have – it's going to become a core part of how they operate going forward."
Richmond outlines four tips for leveraging big data at hospitals:
1. Understand what kind of data you have, and where it's stored. One of the biggest challenges for hospitals, Richmond said, is recognizing what data is available and retrieving it from where it's stored. "Hospitals are largely unaware of how data can be deployed to help them improve the value of what they're delivering and the care they're giving, and how it can help them grow," he said. "What we find almost universally is hospitals discover all the nooks and crannies in their institutions where they have valuable data assets." These assets could be in the form of labor data bases containing employee information, patient level transactions or other encounter data, or even time-stamped data telling how patients are moving through the OR or ED. "It's about identifying where these assets are and [combining] them," Richmond said. "That's why we call it 'connecting the dots,' to give them views not just within these data streams, but also across them in a combinatorial way."
[See also: How to harness Big Data for improving public health.]
2. Get data into the right format and perform some plumbing. The second step, Richmond said, is for organizations to do a little plumbing to combine their data. "This is the barrier for most hospitals," he said. "First, they don't know what data they have or how valuable it is; they just know they have it. But the second is the actual plumbing work of getting [data] into a useable format." Once you get it into a useable format, he continued, you can more easily leverage technology to get even more data out of the data streams. "It's literally extracting this data from their core systems and working them into a format that is common across institutions," he said. That allows hospitals "to compare and benchmark data sets, cleaning the data to remove and correct [it] for the intricacies of the hospital's operating environment. And once it's there … and you've created a reputable process to get data in the cloud, then you can apply all the types of technology, some of which is open source, etc. to predict the future with that data and generate value for hospitals."
3. Run analytics on datasets. Running analytics on big data, Richmond said, is tricky, since based on what he's seen, most hospitals and hospital systems don't have analysts who are capable of big data analytics. "Most hospitals, they have data analysts, but these are analysts who use Excel or Access on their lap top to manipulate data sets, and the scale and size of what we're talking about is quite different," he said. "It requires industrial grade technology, and it's not going to run on your lap top." Choosing a partner who's able to run these analytics has benefits, Richmond said, and it allows hospitals to break away from the traditional types of analytics they perform. "What's happened in the past, [hospitals] want to do retrospective analysis," he said. "But we're finding some of the most powerful things is prospective predictive modeling, [or] predicting the future based on predicting demand curves for the hospital, based on who's likely to come next week, next month, or next year." He added this function also allows hospitals to predict which of their service lines will grow, "like cardiology or orthopedics," he said. "These are predictions where you can't use the past because the market is evolving so quickly that the past isn't a good predictor of the future."
4. Get the data into the hands of the decision-makers. "That's the real trick, and it's probably the hardest step because what we find is, very few organizations are capable of generating executive-level insight out of this information," Richmond said. It's key to deliver data in a way that each exec can understands, he added. "The results need to get to senior executives in a format that they recognize and is in a way that meets the cadence if their management." So, for example, data refreshes at the right rate and is able to be used in operational meetings. "In our view, it's probably the least appreciated but also the most important step because you can have great analytics and great stuff, but until you get them into the hands of a decision maker, it's literally going to behave differently."
Follow Michelle McNickle on Twitter, @Michelle_writes