In this episode of the McKinsey on Healthcare podcast, McKinsey senior partner Prashanth Reddy speaks to Datavant CEO Pete McCabe to discuss current issues, such as data fragmentation and access, the benefits of tokenization, value-based care, and the need for neutrality.
Today, Datavant is set to become one of the world’s largest healthcare data ecosystems, enabling patients, providers, payers, healthcare data analytics companies, patient-facing applications, government agencies, and life-science institutions to exchange patient-level data securely. McCabe believes that the need for data connectivity “is overwhelmingly compelling, and it’s going to be fueled by a ubiquitous network that is privacy centric and that’s neutral.” An edited and condensed transcript follows.
Prashanth Reddy: Could you talk us through why data connectivity is so important?
Pete McCabe: If you think about one of the biggest advancements in healthcare over the last hundred years, it’s powering every health decision with data. And when you’re able to do that, first and foremost you put an informed patient at the center of his or her care. Second, you get to practice personalized healthcare. This would mean that when anyone enters a hospital, the medical team would already have their historical data.
Third, there is an opportunity to advance the speed and lower the costs for life science companies and therapeutic companies developing new drugs and treatments. As a McKinsey study says, 20 to 25 percent of US healthcare spending is wasted. That’s about a trillion dollars, and 50 to 75 percent of that waste could be eliminated by a better utilization of data. This would mean $500 billion to $750 billion in reduced costs for the US healthcare system alone. Therefore, solving fragmentation is about enabling every health decision to be powered by data.
Prashanth Reddy: What are the leading causes of healthcare data fragmentation?
Pete McCabe: Fragmentation is a daunting problem. In the US healthcare system, there are tens of thousands of different organizations using thousands of different IT platforms, with an almost infinite number of different standards and different privacy controls. Let me give you a personal example. I’m relatively healthy and young. I have probably visited over 20 health providers in my lifetime. Therefore, my personal data sits in over a hundred different systems. Nobody has a full picture of my health data, not even me. There is a complexity and fragmentation in the US that you don’t see anywhere else in the world.
Prashanth Reddy: How can we unlock the power of connected data and ensure that fragmentation doesn’t harm the process?
Pete McCabe: At Datavant, we have a strong view that we need to solve this problem from the bottom up and the inside out. Fragmentation is exponentially greater than any of us. I’ve worked in lots of different industries, and it’s greater than our instincts would tell us. Our view is that you start at the bottom. On behalf of our hospitals, we release data compliantly. The process started manually; then we had to build in more and more technology solutions to automate those flows.
An important consideration when talking about data is that there’s identifiable health data and there’s de-identifiable health data. You need to deal with each type of data differently. Talking about de-identifiable data, in the US you have to follow HIPAA1 standards. If you have a data set and you don’t have authorization, you must de-identify that data. For this we use tokenization technology, which can remove the patient’s name and personal details.
For example, health providers need to access and compare data—pharmacy data and clinical data or genomics data and diagnostic-testing data. It is possible to implant a linkable token that says patient X at CVS Pharmacy is the same as patient Z at Emory Healthcare system. However, the patient would only be referred to by a number, and a central technology would decode that data to identify these two as the same patient, without any personal details being shared.
Identifiable data needs to start with authorization, and we need sophisticated processes and software to make sure that we are never putting data into the health data ecosystem without consent.
Prashanth Reddy: What is required to make this happen?
Pete McCabe: Covered entities have TPO2 authorization for the data. A health plan has TPO rights to its clients’ patient data because this will help drive value-based care. Therefore, mechanisms such as TPO allow organizations—in a limited, controlled, and defined way—to access or retrieve data for purposes of advancing care.
Prashanth Reddy: Does this also apply to research?
Pete McCabe: Yes, absolutely. COVID-19 is a great example of 20 or more leading institutions de-identifying data, using tokenization, and sharing that data for research purposes. This way, data can be made available to researchers to determine anything from the effectiveness of a new drug to correlations on demographics and preexisting conditions.
The FDA3 is quite supportive of reducing the amount of time and cost involved in developing new drugs.4 They have introduced a concept called real-world evidence, which sits on top of real-world data. Connecting these data can rapidly reduce the cycle times and cost of developing new drugs and speed up getting treatments into the right hands.
Prashanth Reddy: What is tokenization and what impact has it had?
Pete McCabe: Tokenization, for Datavant, means cutting-edge, patent-pending de-identification technology that replaces private patient information with an encrypted token that can’t be reverse-engineered to reveal the original information. The technology can create patient-specific tokens in any data set, which means that now two different data sets can be combined using the patient tokens to match the corresponding records, without ever sharing the underlying patient information.5
When you create that liquidity and data linking, you empower researchers, doctors, and entrepreneurs to advance care with incredible information. This switchboard effect allows you to connect, control, and apply flows of data into an institution and out of an institution. The question is, how do we get this switchboard into every health institution? When you start to link data, big data could reach its full potential in healthcare.
Prashanth Reddy: Is ensuring complete neutrality critical for this?
Pete McCabe: The idea of neutrality is important. At Datavant, we are very clear, from a collaborative viewpoint, that we’re neutral. We don’t sell data and or analyze it. We create data liquidity in a ubiquitous, privacy-centric, neutral way, with all the data stewardship control to the sources.
A lot of companies do want to analyze data or sell data, for obvious business reasons. However, this would simply create more fragmentation, due to the nature of market competition. For this reason, at Datavant we’re specific and purposeful about neutrality.
Prashanth Reddy: Do you believe this should be regulated at industry level?
Pete McCabe: In theory, I would recommend regulation. However, due to data fragmentation and the complexity of our health system, in reality regulation is nearly impossible. I think an ongoing conversation and collaboration between the public and private sectors is more probable. Innovation technologies such as tokenization can close gaps in data connectivity. This has happened organically by way of the market, and we feel confident about the continuation of that innovation. Regulation may be part of it, but it is not the whole solution.
Prashanth Reddy: How should patients think about this in the context of their consent?
Pete McCabe: This is an important question. Whenever any of us enter a healthcare institution, we engage in an active consent process. This allows a patient to control and influence how their data is used. This also allows the institution to de-identify and use the data under its governorship for the advancement of care.
Prashanth Reddy: What role do you see data connectivity playing in a healthcare transformation?
Pete McCabe: The foundation of value-based care implicitly requires data. Value-based care can only be as good as the underlying data that supports it. At Datavant, we are huge supporters of value-based care. Today there’s a clear drive toward value-based care. We need to speed up the underlying data to support value-based care. It’s one of the things we see the payers and providers and self-insurers driving hard, and we’re very anxious to support them in advancing their objectives.
Amazon can tell me what books I’ll like, and Netflix can recommend movies. The Internet of Things takes real-time feedback and makes personalized recommendations in people’s lives. We need to be doing this in healthcare. Whether at an individual or population health level, the transformational power is massive. I think value-based care is just the beginning of driving this potential forward. If the internet can recommend a piece of clothing based on a consumer’s buying history, why can’t a doctor recommend a drug treatment based on a holistic data view of that individual’s life?
Prashanth Reddy: What trends will be shaping the healthcare data ecosystem in the future?
Pete McCabe: I think you’re going to see things we are already aware of, such as artificial hearts, mapping of the genome, and antibiotics, biopharmaceuticals, diagnostic imaging, et cetera. I think you’re also going to see efforts to power every healthcare decision with data. I’m convinced we are going to solve this problem. There is already innovation driving change, and the need is overwhelmingly compelling. Furthermore, it will be fueled by a ubiquitous network that is privacy centric and neutral.