Making the leap from big data to better care - 4 myths to overcome

How to harness the power of real-time intelligence to enable more precise and personalized care
Written by Jean Drouin, MD, CEO of Clarify Health Solutions

"If a tree falls in a forest and no one is around to hear it, does it make a sound?" When it comes to big data in healthcare, clinicians and managers face a similar challenge. If an insight is generated and never makes it into the workflow to impact a decision, is it relevant?

There is a lot of hype centered on machine learning and artificial intelligence (AI), with many technology vendors touting the ability to transform care. Clinical insights, whether uncovered by sophisticated analytics or more traditional means, are most valuable when they provide new information that can positively influence care, and are delivered to the individual with the power to make the right decision, at the right time.

In keeping with this theme, here are four common misperceptions about the role of advanced analytics in delivering better care.

Myth 1: Variation is the enemy

Variation is a symptom of a problem rather than its root cause—and not all variation is negative. Some forms of variation are manifestations of new and better ways of doing things. The challenge is that most of the data that we have to work with is retrospective. We look back in the rearview mirror and do our best to determine what went well and what didn't go as well. Regardless of what we find, it's often too late to do anything about it.

Admittedly, retrospective analysis of variation is helpful to understand patterns and behaviors that can on average improve care for a population of patients. However, to truly personalize and deliver precise care for a specific individual, the ability to spot markers of emergent issues in order to avoid them would be transformative. The real enemy is our inability to spot and avoid problems before they occur.

Myth 2: Claims data is clinically useless

Many clinicians believe that claims data has limited value. However, machine learning enables analyses of multiple relationships in claims data to surface insights, including ones that clinicians previously may not have considered or been able to distill.

For example, using typical data found in an electronic health record (EHR), one physician may be lauded for his patients' low average acute length of stay (LOS). Meanwhile, his colleague in the same service line has a different practice pattern with a longer acute LOS, including a higher proportion of referrals to skilled nursing facilities (SNFs). However, incorporating the analysis of claims data reveals that the first doctor's readmission rate is much higher, while his counterpart ultimately generates similar total costs of care due to a lower readmission rate.

Such insights have been hidden in claims data for years, in part because it was difficult to obtain large enough data sets for meaningful analysis. This is beginning to change, making it possible for sophisticated analytical techniques used in the financial, logistics and consumer industries to be applied to patient care.

Myth 3: My patients are different

When presented with data about their practices, physicians often respond with: "But my patients are different." That's most likely because much of the data shared with clinicians in the past failed to provide true apples-to-apples comparisons that fully account for differences in patient populations. One of the major advantages of claims data is its sheer vastness. With millions of records, it is now possible to create benchmark panels that much more closely match a physician's own patients. Comparing physicians based on their average performance, while instructive at one level, hides a far more important question: "For that specific patient, did we as a care team do as good of a job as we could?" Ultimately, results at a population level are the sum of what transpired with each individual patient.

What's exciting about big data and analytics is their ability to enable more precise, personalized patient care. In that case, the question clinicians can ask themselves is "how precise am I at delivering the required care?" Yes, each physician's patients are different; the key to delivering the best care journey possible is to understand those differences and create a care plan specific to each patient's needs.

Myth 4: The solution is insights

Insights are only one part of the solution. What's needed is real-time care guidance that informs personalized care journeys. Imagine a world where patients could be guided, turn-by-turn, through their care journey just as driving apps such as Google Maps and Waze direct travelers today. Technology now exists to provide that personalized and precise, real-time care guidance. To fully enable such a system requires pulling together individual-level information across clinical, claims, demographic and patient-reported data to create a rich, real-time view of a patient's care journey.

Real-time intelligence can use models based on historical data to compare what's happening with a patient right now and determine whether it falls within the bounds of the expected outcome. For example, if a patient is three days post-op and spikes a fever, it may be an indication that the patient may have an infected wound that needs to be cared for before it turns into an emergent clinical issue. This timely intervention could save the patient and the provider a readmission. This proactive approach increases efficiency as well as improves patient outcomes and satisfaction.

By monitoring patient care journeys in real-time and comparing them to both historical and claims data, providers are able to make more timely interventions and focus on the patients with the highest risk. These tools quickly and efficiently improve outcomes, increase efficiency, and drive down cost.

New technologies bring new opportunities

We stand at an exciting moment. The recent availability and linking of large data sets; the advent of sophisticated analytical methods enabled by computing power; and the spread of smartphones that enable two-way communication with patients provide the pillars for real-time care guidance. The time has come to seize this opportunity to fix broken workflows and processes, and re-imagine the way care is delivered.

Jean Drouin, MD, is CEO of Clarify Health, a health IT company that enables more precise and personalized care through trusted real-time intelligence and care guidance. He brings 20 years of experience in healthcare management, strategy, operations, finance and cultural change. Prior to founding Clarify Health, Jean led McKinsey's Healthcare Digital and IT practice. He also built and served as the founding Head of McKinsey Advanced Healthcare Analytics (MAHA). Jean holds an MD and MBA from Stanford University and an AB in molecular biology from Princeton.