Pratap Khedkar co-wrote this blog post, the third in a series on ZS’s 2018 AI in Healthcare study, with Paul Darling.
Each time you visit the doctor, it’s the same scenario: You rattle off a bunch of information like medical history, lifestyle behaviors and choices, and current complaints, while she pounds away at a keyboard. You might have access to a portion of your medical record through a web-based patient portal, but otherwise that content is under the provider’s lock and key (unless you’re a New Hampshire resident). Don’t you wonder what your full electronic health record holds? Stakeholders throughout the healthcare ecosystem do, too.
Patient data is becoming one of healthcare’s most valuable assets, and data-rich sources like EHRs, clinical trials and digital health tools are fueling treatment development and clinical decision-making, informing population health management programs and boosting patient adherence to medications. If control over that data lands in patients’ hands, we’ll need to create a marketplace for data-seeking companies and stakeholders like providers and payers to bid on and exchange patient data. Will patients begin selling their data? Could this be a new side hustle for millennials? Will patient data continue to appear on the black market, or will this be a legitimate future enterprise? And what does all of this mean for the role that artificial intelligence could play in powering healthcare delivery down the line and who will have access to the patient data to fuel it?
It’s My Data, so Why Don’t I Own It?
In our 2018 AI in Healthcare study, for which we surveyed 400 patients and 400 doctors, we found that a majority of the doctors (72%) and patients (92%) who responded to our survey believe that patients should retain ownership and control of patient-generated data. Similarly, 81% of doctors and 78% of patients believe that patients also should control hospital-generated data, and that hospitals shouldn’t be able to use it without consent.
We also asked the patients how they feel about sharing different types of data and found that they’re more willing to share data when they can see the direct application to their own healthcare. About two-thirds (66%) of respondents said that they’d be willing to share traditional healthcare data including family history and prescription use, and 60% would be willing to share genetic, biometric and lifestyle data. In contrast, just 30% of patients said that they’d be willing to share more personal data like what can be collected from IoT connected devices, social media channels and devices’ location services. Interestingly, this last category of data is the most closely linked to innovation in digital and connected health interventions as well as wellness efforts from employers, payers and providers.
So patients want to control their own data (not surprising) and they’re more willing to share it if it could positively affect their own health and well-being (also seems logical). But in a healthcare landscape dotted with towers of vertical integration and sprawls of new entrants, all competing to develop new strategic partnerships and technologies, the question of data ownership and access is paramount. Despite their different business models and points of entry, each of these entities is trying to secure access to the data that it’ll need to fuel future innovation. This begs the question, what data ownership/sharing construct is better for innovation and wellness?
A recent Stanford University study on the economics of personal and medical data looked at the nonrival nature of data, or the notion that the data can be used by separate companies or organizations “simultaneously without being depleted.” It can be used and reused multiple times—and by many different stakeholders. The study’s authors looked at three ownership scenarios: government-owned and -regulated data, private corporation-owned data, and consumer-owned data, concluding that consumers are the best equipped to balance the economic gains vs. privacy challenges of sharing their own data. They found that the scenario in which consumers own the data also maximized financial and technological growth. When governments regulate data usage or when private corporations opt not to share data to maintain a competitive edge, innovation suffers.
Corporate players, of course, currently see the data itself as an exclusive asset, but the marketplace tells us that data access is the future asset. In other words, when patients control their own data, gaining access to those patients and their data will become an entirely new business. For example, Nebula Genomics and Longenesis have partnered to create a new blockchain-based marketplace that shifts health data back into the hands of those who generate it to create a global marketplace that would allow patients to license their data for profit. Essentially, data brokers would make cryptocurrency payments to buy or use patient data, including blood test results, medical histories, genetic profiles and other sensitive information. Nebula is banking on new, highly valued data assets, as the market for genomic data alone is expected to be worth billions in the next few years. And because genomic data is closely tied to the bets that the healthcare industry is making on personalized medicine, Nebula will also offer its own genome testing service to consumers and, to complete the service, will make the data available to the marketplace.
Once You Secure Access to Patient Data, How Do You Leverage It?
Let’s say that patients do assume ownership of their own data, which seems to be likely, given the U.S.’s growing interest in establishing new standards for data ownership and privacy along the lines of the EU’s General Data Protection Regulation. Data access and analysis could be hampered by interoperability issues, limiting the potential of an AI-powered healthcare system.
The average patient bounces between multiple sites of care including hospitals, urgent care centers and pharmacies, leaving behind a distinct data trail at each. That doesn’t even account for the slew of patient data being collected by personal fitness trackers and other wearable technologies. The healthcare system has to figure out how to capture these distinct data sets in a standardized way.
Apple may have already found the answer, or at least the first generation of an answer: The tech company was quick out of the data gates with its patient health records app. Dozens of health systems including Henry Ford Health System and Cleveland Clinic have begun offering the FHIR-enabled tool to patients, enabling them to store their own EHR data and beam it to providers as needed.
And other tech giants are trying their hand at a solution, too: Amazon, Google, IBM, Microsoft, Oracle and Salesforce are joining forces to address interoperability issues in healthcare. Banding together, these companies aim to refine data standards like FHIR to enable true interoperability and unlock healthcare data’s potential. And with the interoperability framework Carequality, the raw network exchange already exists, enabling EHRs to be shared via FHIR among hospitals and practices in different networks and health systems. CMS also is entering the data-sharing arena by launching Blue Button 2.0, which gives beneficiaries access to, and full control over, Medicare data including drug prescriptions, primary-care treatment and costs.
Meanwhile, Walmart’s investments in blockchain technology could bring the retail giant one step closer to solving the dilemma of who owns—and who has access to—a patient’s EHR. Walmart was awarded a patent on a system that stores a patient’s EHR in a healthcare blockchain-enabled database that’s accessible via a wearable device and provides limited access to first responders in emergency situations.
It almost seems like the roads to interoperability are nearing completion, but the traffic barriers still need to be removed. And once we get beyond data access, we have to tackle the challenge of strategically introducing artificial intelligence into the healthcare equation. As the answers to the industry’s most vexing data challenges continue to be scripted by tech giants and governments—paired with an evolving marketplace that creates incentives for exchanging data—we can only imagine that AI will continue to penetrate healthcare. And the deeper that AI bores into the fabric of healthcare, the more it will pull at its seams. Can we expect the technology’s presence to create a dichotomy in the type of healthcare available? AI could push healthcare away from its “traditional” (Norman Rockwell-esque) custom-care model in favor of an “industrial” (factory-like), highly efficient, data-driven model that prides itself on repeatable outcomes.
With a burgeoning pool of patient data and increasingly sophisticated tools to mine it, U.S. healthcare could indeed be on the precipice of a transformation. Increasingly, it seems that patients are taking a turn at the wheel, and AI is pressing on the accelerator.
We’ll explore AI’s potential impact on healthcare delivery models in our next blog post.
BLOG POST: The State of AI Adoption in Pharma