HLTH Kicked Off by Unpacking Health Data Buzzwords
Healthcare is an information industry. Medical decisions, research advances, and health optimization apps — nearly everything that happens in healthcare is fundamentally built on health data. But as health data technologies have proliferated in the past decade, the buzzwords surrounding the industry can sometimes obfuscate the importance of what’s underneath.
So, at the 2022 HLTH conference, health data and technology innovators gathered to discuss what all these buzzwords really mean. What’s behind the buzz, and why does it matter? How do any of these tech trends impact patients?
As Janssen R&D’s chief data science officer Najat Khan, PhD, highlighted on the panel, we need to remain thoughtful and clear in our approach to building the future of healthcare: “Our level of rigor around health data needs to be extremely high — because the stakes are extremely high.”
Here are some key takeaways from the conversation.
Healthcare can be a massively inefficient industry, one that is resistant to change and stuck in legacy processes and ways of doing things. Over the past few years, between harnessing data, the emergence of powerful digital health technology, and the pandemic upending everyone’s sense of normalcy — we’ve finally built up momentum in envisioning what a better healthcare system could look like. It’s up to all of us to take advantage of that momentum, continue pushing the envelope on what’s possible, so that we can actually bring new approaches, technologies, and therapies to life — and deliver that benefit to patients.
As it stands, there is ongoing conflict between insurance companies, providers, and big pharma, who are each fighting for a different piece of the pie — and patients are caught in the middle. Data and technology have the power to change that, but only if we’re willing to think differently about how we bring that data together, and the new opportunities to make that data even more powerful to answer healthcare’s most challenging questions.
For instance, think about the value of seeing a data point that represents a patient’s diagnosis with MS. You can see the HCP who diagnosed them, where they live, and, maybe, some information about their course of treatment. Compare that to seeing that patient’s entire longitudinal medical history, both leading up to and following that diagnosis. Suddenly, you can understand the symptoms and patterns that may have presented long before diagnosis, specific trends in comorbidities, responses to therapy, changes in treatment regimen, and so much more. The latter approach provides the details necessary to actually drive deeper understanding of MS in the real world, address challenges and unmet needs across the entire patient journey, and support better patient outcomes.
In the simplest of terms, data science is about turning data into insights. Of course, it’s a bit more complicated than that, requiring an interdisciplinary approach to gathering and linking data together, cleaning and organizing that data into a usable format, and creating a secure and flexible data infrastructure that allows a user to extract as much value as possible. It also means gathering and sorting data with an eye toward eliminating or minimizing bias, so that as we build solutions and research on top of that data, we can illuminate an accurate, reliable picture of care patterns — rather than perpetuating and exacerbating existing disparities in care.
By providing the data, the infrastructure, the analytics, and the development tools, we can reduce the complexities of building new solutions and unearthing new patient insights.
Turquoise Health is a great example: They are building a platform to unlock healthcare cost insights and simplify payer-provider contracts. Using real-world data together with healthcare pricing information, they are building a net-new solution on Komodo’s platform that will illuminate cost and care patterns for different healthcare scenarios. The eventual technology will be both provider- and payer facing (breaking down pricing data, utilization, and reimbursement for different services) and patient facing (providing visibility into the full costs of a healthcare encounter).
The term “interoperability” often elicits eyerolls, but it is essential for different information systems to speak to one another. This is usually discussed in the context of two different EHR systems being able to communicate, seamlessly transferring patient records from one health system EHR to another. At Komodo, we’re actually much more interested in the synchronization of data sources. We know how to gather and curate data from multiple sources; the hard part is stitching that data to create an asset that is broad (contains a massive amount of patient records), deep (provides as much of each patient’s experience as possible), and diverse (includes multiple types of insight, from diagnoses and procedures to labs and demographic insight). It’s also important to ensure the data is carefully reviewed for redundancy and duplication.
Historically, the industry’s approach to buying, cleaning, and analyzing data has been clunky and inefficient — and lacks the level of nuance needed to make smart decisions or offer effective services. Over time, we’ve seen increasing data liquidity, which has unlocked the ability for companies like ours, which are focused on insight generation and model development, to really productize insights and develop models and AI that can be applied to large datasets at scale. The key to realizing the benefit of big data is in collecting that information accurately, in real time, processing it so that we can see the entire patient journey, and — critically — linking it with other datasets to create even more nuanced insights.
With linkability, we now have the power to bring new datasets into our platform and create net-new insight into patient behavior, symptom patterns, treatment efficacy, and more. For instance, Invitae is a genetic testing company that is linking its trove of genomics data with our platform to drive powerful new research initiatives. Alone, genetics data is limited in its value. However, when we integrate it with longitudinal patient journey data, we create the resources to uncover fascinating, unprecedented insight about the role of genetics in conditions such as epilepsy.Identified/De-Identified:
De-identified data refers to data that has been stripped of key personal information, like name, date of birth, and other details that could allow a data point to be traced back to an individual. De-identified data can provide incredibly valuable and nuanced insights. Even when stripped of identifying information, a specific patient’s data can still be linked together on a single token, unlocking the ability to make connections and draw insights about a patient’s full healthcare journey.
Currently, de-identified patient data is the backbone of real-world data research. However, identified patient data will be critical to future innovations and research. More and more, we’re seeing patients taking custody of their data and wanting to contribute that data to research. Linking identified data to a platform built on de-identified data creates massive opportunities for new insights. Ultimately, as we expand our thinking around identified and patient-owned data, there is benefit for both those patients (who gain unprecedented clarity and ease of access into their own medical records), pharma (who can use that data to drive new research and medical innovation), and then patients again (who ultimately benefit from that research and innovation).
Healthcare has historically been an industry that’s slow to evolve and resistant to change. Pandemic effects have forced a shift in perspective about how we were approaching health, disease, and treatment — no matter where in the industry you sit. For instance, If you’re in pharma, you realize you need to take digital approaches to HCP relationships, you need to rethink the standard, centralized approach to clinical trials, you need to critically consider race and ethnicity as factors in health and outcomes, and so on.
Recently, the industry has been specifically discussing the need for more modernization in clinical research, as we are innovating new treatments faster than we can prove their value and safety. Modernization, therefore, must go beyond the development of new therapies, and expand across all stages of the drug lifecycle, including clinical research and delivery of those therapies to patients. Part of that requires massive structural change, but data can also play a key role, whether that’s using data to inform site selection and inclusion/exclusion criteria or building an entire trial arm on real-world data.
We work with a company called AppliedVR, which is building immersive VR therapies for patients with chronic low back pain, accelerating an alternative to traditional pharmacological, opioid, or surgical treatment. That itself is incredibly innovative work — but they are also taking an innovative approach to researching and proving out the value of their technology. AppliedVR is using Komodo’s real-world data to form synthetic control arms to better understand the impact of their treatment on patient outcomes, sidestepping many of the legacy challenges of clinical trial design, like long recruitment timelines or limited metrics. To me, this is such a clear example of the power we have to modernize nearly every aspect of healthcare. Of course it all can’t happen at once, but it’s amazing to see players taking those strides towards the future of healthcare.
So What’s All the Buzz About?
Rather than continuing to say that the “future of healthcare” is data, we need to recognize that data is already firmly ingrained in every single aspect of healthcare. Instead, the true “future of healthcare” is about democratizing access to data-driven insights and using those insights for powerful, innovative new approaches to the industry’s biggest challenges.
Learn more about how Komodo is powering the future of healthcare data with our platform approach.