Payer-Provider Data Integration Not Without Pitfalls

Combining payers' claims data with providers' clinical data is an essential element of revolutionizing the delivery of healthcare. But there will be unintended and unfavorable consequences, says one data expert.

I gaze quickly at the watchcorder on my wrist. It's 4:05, my pulse rate is 74 beats per min and the blood sugar monitor is flashing at 81… I missed lunch, again.

I have a long drive home and it's always good to get ahead of the traffic coming out of Boston. It's Jan. 20, 2020, and tomorrow will be my sixth work anniversary at HealthLeaders. On the spur of the moment, I decide to pop into a nearby smoky bar to celebrate.

My belly has barely touched the lacquered mahogany expanse separating me and the bartenders at Phil's when the watchcorder lights up like a Christmas tree. The no smoking light. The no drinking light. The provider, payer, and employer wireless notification lights.

I'm in my 50s and my doctors have advised me to put my bad health habits behind me, especially smoky bars. My watchcorder's GPS, and its array of sensors embedded within the device and inside my collar have betrayed me.

I can count on a finger-wagging at my next primary care doctor appointment, an email from my insurance carrier announcing a $10 surcharge on my healthcare policy, and a letter from HealthLeaders' EAP urging me to seek smoking cessation and alcohol abuse counseling.

This scenario may be years away, but we're already seeing what can happen to personal health data. "There's a whole spectrum of desirable-to-undesirable consequences out there," David Burton, MD, former chairman and senior VP of the Salt Lake City-based data analytics firm Health Catalyst, told me last week.

"It's broader than healthcare—all the data mining that's going on. But it is amplified by the personal nature of the healthcare data. [Data mining for] marketing is one thing. It's another thing for someone to know where I am and what I'm doing."

Burton has practiced as an emergency room doctor and launched a health plan that now has about 600,000 beneficiaries in Utah and Idaho. He says the biggest pitfall providers and payers face in joint data integration projects is their troubled past. "The industry needs to mature to the point that this isn't a shell game where you get a piece of the other side's shell," Burton told me. "This is all about [reducing] cost."

The adversarial days of the past are haunting healthcare data integration projects, he says, noting that among providers, "only want to share the data that will get them paid." On the payer side, Burton says insurance carriers have hidden claims data behind locked doors. "There has been a historically adversarial relationship."

For beneficiaries, "there is always the risk of privacy," Burton says.

Combining the vast payer and provider data sets of consumers' highly sensitive medical history information creates a treasure trove for data analysts. "In the ideal, you have to have access to the claims data, but the insurer also needs to bear down on your care data to get down to the granular level," Burton says. Lax data governance and stewardship are prime risks for healthcare organizations. "It doesn't take much of a breach to ruin your reputation," he adds.

As the pace of data integration projects in healthcare quickens, there are stumbling blocks for payers and providers to avoid, says Jeff Cohen, co-founder of Denver-based Welltok, which describes itself as a "health optimization pioneer" on its website. The challenges, Cohen told me last week, include data access, data accuracy, "how fresh it is," and compliance with privacy protection laws such as HIPAA.

Challenges are also associated with "the technology [itself] and how you're going to mine that data," Cohen says. "It's the type of technology, the cost of the technology, and securing the data. Oftentimes, [picking the wrong technology] causes these efforts to fail."

Cohen says there are inherent challenges in the data for payer and providers. "The 'structure' of the data in healthcare makes these projects difficult," he told me, adding that there within the tremendous volumes of data, there is spotty consistency, lapses in accuracy, formatting variability, and a clinical coding morass.

Welltok is working with IBM's Watson supercomputer to overcome the structural obstacle to healthcare data integration projects and help consumers optimize their health. They are working together to build a new app that leverages Watson's unique ability to comprehend multiple types of data and provide consumers with personalized recommendations that evolve with changes in their activities, benefits or preferences.

With Watson, users can bring in 'unstructured' data, which means data from PDF and Word documents can be gobbled up, just like data stored in traditional databases. That data gets ingested into a massive data store where it can be analyzed.

IBM calls Watson a cognitive technology. "More than 100 different techniques are used to analyze natural language, identify sources, find and generate hypotheses, find and score evidence, and merge and rank hypotheses," according to Big Blue.

For payers and providers launching data integration projects, consumers are the key to avoiding risk, Cohen told me. "There are unintended consequences. You have to think through letting the consumer control the access and use of some of those data streams," he says. "You want people to feel comfortable and to share in the benefits of this sharing of data… Ultimately, the consumer has to be in control."

Data integration projects offer tremendous opportunities for healthcare organizations. But caution is advisable whenever a revolution is swirling around you. "There are going to be some bumps along the way," Cohen told me. "It's not going to be easy. There will be some failures."