Big Data Yields Big Potential for Managed Care

Payers and providers now have access to more data than ever before, but simply having the access doesn't guarantee the success. For providers and payers to find substantial ways to improve quality while reducing the cost of health care, experts in the industry say "Big Data" must be used in a targeted, incremental way to identify actionable changes.

"There's a misunderstanding of so called ‘big data’—that somehow if we just have more data, that there's more truth to be found and that is not necessarily so," says Scott Zeger, PhD, professor of biostatistics at Johns Hopkins Bloomberg School of Public Health. "More data doesn't mean more knowledge. More data just means more data." 

Big data is a term that's growing traction across the health care industry. It's used to describe large sets of data, whether it's patient claims data, genomic data, clinical outcomes data, or data from consumer devices like a FitBit or Apple Watch, which are now increasingly available to providers and payers as data storage capabilities have significantly expanded.

"Data is being produced constantly around all sorts of events, patient encounters, and quality performance, so there's more data than providers actually know what to do with," says Peter Strack, CHFP, CPC senior advisor for strategy and development, Altarum Institute, a nonprofit health systems research and consulting organization.

The Role Big Data Can Play

The potential for big data in the managed care industry is significant and many payers and providers are already using it to identify opportunities within their system to improve costs or quality. 

Dorrie Guest, director, Deloitte Services, says her clients have seen great results in driving alternate care patterns that produce higher quality and lower cost through the use of big data. She says both payers and providers have used data around clinical encounters and care pathways to determine what will deliver the highest quality outcomes among clusters of clinicians and services at the lowest cost, also integrating patient preferences and patient engagement. 

For instance, she says if you are able to examine congestive heart failure patients who have been treated in naturally occurring clusters of clinicians and hospitals to collectively treat patients, you can examine the services rendered along with the acute or chronic disease management given to determine what outcomes have been produced and what resulted in a better encounter-based quality indicator at a lower cost.

"You can use that, in multiple ways, to translate that to other naturally occurring clusters of clinicians in facilities so that they learn and can also be tracking toward that more value-oriented care delivery model or you can use it to steer patients using the results of the data to make better decisions about their pathway through the care process," she says. 

Guest says as the industry moves toward more value-based care, she's also seen big data used to determine the demographics, service needs, and characteristics of the populations being served to determine the health needs of the patients.

"Maybe you don't need so many pediatric endocrinologists but you need more adult cardiologists,” Guest said of using data to determine needs. She stressed the importance of figuring out the balance of the right compliment of physicians, and making sure those physicians have the right quality and cost-performance while working together, as well as making sure they are in the locations where you have density of the population you are servicing.

Mark Bethke, an actuary and a director at Deloitte Services, says providers who take on shared risk agreements can also use big data they receive from payers as part of the arrangement to determine what actionable steps can be made to meet the quality standards established in the agreement.

But while some payers and providers are further along in using big data to make better data-driven decisions, others are still struggling with its applications.

"In the industry as a whole, there's a recognition that this is a good thing to do. The challenge is applying the resources internally and the investments to ultimately drive these types of capabilities in house. Some are just faster than others," Strack says.

Obstacles in the Path

Experts agree that one of the biggest challenges for payers and providers is finding ways to use big data to make actionable changes within their organizations.

Dr Zeger, who also serves as the director of Johns Hopkins' Individualized Health Initiatives program known as Hopkins inHealth, says that one of the thoughts that seems to pervade medicine today is the idea that the best way to use data is to capture it in the electronic medical record for a doctor to view while treating a patient. However, with the explosion of data available, he says doctors are now bombarded with increasingly complex data sets that include images, DNA sequences, clinical outcomes data, and more that they are not optimally able to synthesize on their own.

For this reason, the individualized health initiative is focused on how to best use data to make it relevant in practice.

"The individualized health initiative, this so-called Hopkins inHealth, is really trying to build tools, both data and science tools but also biomedical tools, that enable us to not only capture the data but synthesize it and capture its biological or its medical meaning and provide that data to the clinician in a way that she can understand what it means and make better decisions with her patient," he says. 

They plan to develop 20 to 30 of these tools in the next 3 years. 

One of the tools they've created is for the treatment of prostate cancer. Dr Zeger says if prostate cancer is caught early, research has shown that about half of the tumors discovered are indolent. The best treatment strategy for these patients may be to observe the patient rather than do a radical prostatectomy or radiation treatment, but since it's difficult to determine which patients may have indolent tumors, Dr Zeger says Hopkins inHealth has studied biopsy data, prostate specific antigen (PSA) measurements over time, prostatectomy reports, and pathology reports from a large sample of patients to identify who is most at-risk for having aggressive tumors.

Their tool will allow physicians to use data from the patient in front of them to predict the chance his tumor is indolent.

"The point then, is that the clinician and the patient can take this probability and make a decision about what do I want to do in light of my circumstance, in my circumstance in life, and it's really a choice in the end, but it's a choice that's informed by a scientific evaluation of the data about that man in comparison to what we know about other men," he says. 

In addition to the specific tools developed by Hopkins inHealth, they hope to create a platform where others are able to develop their own tools as well. 

Guest says to use big data to improve clinical outcomes and reduce costs, providers must present the data in a way that resonates with clinicians. Data must also be accurate or physicians may be likely to discount it and continue to work in the way they've seen results produced in the past.

Experts agree that another potential pitfall, particularly for those providers who aren't used to working with large sets of data, can be trying to take on too much too soon. 

"There's baby steps that need to occur with a lot of these providers," Bethke says.

Strack agrees, saying incorporating big data often works best when providers or health plans focus on a particular question or problem that has caused concern and trying to use data in a targeted way to address that specific problem. 

"These initiatives can fall down when they are too global or too broad," he says. "You need to demonstrate some quick successes, show how it can be applied, improve something, improve a condition or improve an aspect of care, indentify a particular care bundle for a service that is more cost-effective or produces a better outcome, and then use that to gain momentum." 

As genomic data becomes more available and data sets are growing in complexity, experts say that those who are likely to find the greatest success using big data are those who've already done well using existing small data, such as patient age, when making treatment decisions. 

"I am all for big data when the ‘big’ gives you more information, more knowledge, but I also know that in medical centers and in managed care organizations all over the country we don't use little data very well," Dr Zeger says.

Overall, big data itself isn't the sole answer to providing high-quality, low-cost population-based care. It's just one piece of the puzzle; however, if data is used effectively it could help drive industry cost-savings.

"Whether it's small data or big data, complicated algorithms or straightforward math, you can utilize data to drive really interesting analytics but if you're not putting the results of those analytics into the hands of folks who are going to do something different operationally, whether that's through clinical operations or health plan network operations or what have you, you're not going to have moved the needle," Guest says.