One constantly hears the phrase “big data” these days. But what does it really mean? John McInally, a partner in data services of the Big Data Group at the Falls Church, Va.-based business solution provider CSC, says big data, by definition, is the ability to aggregate internal data with external data. The trend now, he says, is, “I have my data—now what analytics can I run against the data that I have internally, but also, how do I compare that data and those trends with external data sources?’” At CSC, McInally says, “we are going to be focusing on the uses of that data to provide analytic service back to you.”
It is important to remember that big data is more than just a sea of information; it is an opportunity to find insights in new and emerging types of data and content. Most healthcare organizations have been focusing on quality reporting and performance reporting that has been required (primarily by Medicare), but what they are embarking on now is a period of refinement, says Jordan Battani, the managing director of CSC’s Global Institute for Emerging Healthcare Practices. By refinement, Battani means that organizations know how to do the reporting, as there are techniques and tools at their disposal. There are some places that do it really well and that will spread throughout the industry, she says.
NEXT STEPS
The next real frontier of big data is around surveillance and predictive analytics. “Here, you do more than just look at what happened and serve it up as information. You now look at patterns over time so you could predict what will happen based on past events,” says Battani.
Battani notes solutions are beginning to emerge that will give providers the capability to do real time risk assessment to predict the onset of diseases such as sepsis. CSC is one that has a program that then serves up that information to care coordinators and caregivers so that interventions can be made before the disease manifests. “To be able to spot that before it fully manifests and make an intervention has a direct impact on everything we are worried about—the patient’s experience, the quality of care received, the safety of his/her experience, and it brings down costs. You want to be able to take that little illustration and write it largely across the population.”
As organizations are now starting to move all of their patient chart information from paper to the electronic health record (EHR), the time has come where they can study it and do analytics like never before, says Steven Shapiro, M.D., executive vice president and chief medical and scientific officer at the 20-plus-hospital University of Pittsburgh Medical Center (UPMC) health system. “What we’re doing with our data warehouse is getting our provider data, payer data, financial data, even phonetic and genomic information together, and when that happens, I think the opportunities are going to be great.”
Steven Shapiro
Shapiro points to UPMC’s five-year, $100 million deal—announced on Oct. 1, 2012—with prominent vendors dbMotion, IBM, Informatica and Oracle that will incorporate their sophisticated enterprise analytics. “Right now, we have over 200 different data sources, but we are going to aggregate it all and harmonize them. This will make that digital information usable, and will allow us to test and optimize our new models of care, which will really lead to improved care.”
But it’s more than just technology and money, says Dawn Mitchell, principal at Pittsburgh-based Aspen Advisors. “It all starts with leadership. It will require strong leadership to take control in areas of IT adoption, but almost more importantly, a change in management and workflow. There needs to be a much broader business strategy looking at cultural change rather than trying to acquire the latest innovative technologies to help with population health management. Technology is an enabler, it’s not the answer.”
Dawn Mitchell
She adds that while implementing EHRs are one piece of the puzzle, organizations should also be looking at data governance as a whole. Indeed, UPMC’s Shapiro says, data governance is critical. “If it’s garbage in, it’s going to be garbage out. Paying attention to the definitions of things and having everyone enter data in a similar way is key to allowing these larger things to occur.”
A WAYS TO GO
There is a broad consensus that healthcare is still behind other industries with big data, and one reason for this is lack of resources. UPMC may have more resources than others to take on that kind of big effort, but certainly, even within their EHR, a number of organizations can start to take much better advantage of their clinical information, says Shapiro.
There are also challenges just getting the clinical data all integrated in the EHR, says Mitchell. Organizations have still have their data siloed, and they need to bring all of that data together through some kind of enterprise warehouse so that they have access to the total data required for the future. The direct answer is one step at a time, and sometime organizations are trying to run before they can walk, Mitchell says.
But implementing the EHR isn’t enough; that alone will not get you to meaningful use, Mitchell warns. Even with the EHR, she says, physicians are still complaining, saying their productivity is down, and all of these things that are supposed to make their lives easier are not because they can’t trust the data. “So now there is a big data issue, in terms of quality of data, integration of data, and making sure there is only one true source of truth for reporting. This is the biggest challenge right now.”
According to CSC’s McInally, while many organizations are still in the preliminary stages of their big data initiatives, because of health information exchanges (HIEs), it is now possible to look at patterns and uses of data that were not really available until this meaningful use era, and exploiting the data in ways that have never been done before. “And when we show some success doing that, I think the predictive analytics market will take off,” he says.