How can small- to mid-sized biopharma generate insights from RWD? Q&A with Michael Sanky of Optum Life Sciences
Optum, a new Aetion data collaborator, offers a robust range of real-world data (RWD) assets to support biopharma in generating real-world evidence (RWE) on clinical interventions across a range of therapeutic areas.
For nearly eight years, Michael Sanky, Senior Vice President of Optum Life Sciences, has worked for Optum, guiding the clients through data strategy development and analytics queries. In a recent conversation, Michael, who leads the Optum Life Sciences data and analytics team, addressed the opportunity he sees for RWE in the small and mid-sized biopharma space, including the most common RWE organizational models, the challenges unique to smaller organizations, and key characteristics that drive success in RWE groups.
Read on as Michael shares more, including how technologies like natural language processing (NLP) can enhance RWD analysis capabilities.
Responses have been edited for clarity and length.
Q: How is Optum’s data well suited for RWE research?
A: Optum’s best-in-class data assets are used to generate RWE about therapeutic interventions. Our data sits at the intersection of clinical health care delivery and claims reimbursement in the United States.
I’d like to highlight three of our data sets. First is our claims data asset, derived from a database of administrative health claims for members of commercial and Medicare Advantage health plans. There are nearly 20 million members on an annual basis with both pharmacy and medical coverage, creating an eligibility-controlled view of health care utilization and cost.
The second database is our electronic health record (EHR) data asset, derived from dozens of large U.S. health care systems. This represents more than 100 million deidentified patients with deep, clinically specific data, including lab and microbiology results, vital signs and observable measurements, inpatient data, and immunization records. Our natural language processing technology extracts information from provider notes, including signs and symptoms, family history, cancer stage and histology, disease severity scores, and biomarker results. The clinical specificity in the EHR data is essential to create comparable cohorts and measure outcomes—all critical to RWE.
Diagnosis codes, procedures, and medications alone from claims data are not enough to generate meaningful RWE. Consequently, a third data set to highlight is our integrated data asset, which combines administrative claims with EHR data on an individual person level. This asset is particularly powerful in identifying factors like total cost of care for specific clinical cohorts. For example, we can stratify a diabetes population by body mass index, which we calculate from height and weight observations in the EHR data, and then quantify costs for specific clinical cohorts. There are now more than 60 million lives in this asset.
Q: How do your small and mid-sized biopharma clients typically work with your data? And in which stage of the drug life cycle or therapeutic area are you seeing the most RWE adoption?
A: Essentially, there are three RWE consumption models. First, our clients may host the data in their infrastructure and analyze it themselves. Second, clients may access the data via tools like the Aetion Evidence Platform®, which helps them generate RWE. Third, our clients may outsource the analytics to consultants. These three models are not necessarily mutually exclusive; some clients host and analyze the data internally, perhaps using a platform, but they may also outsource some work.
Historically, smaller companies tend to outsource more analytics. In recent years, we’ve seen wider adoption in small and mid-sized biopharma clients, who access data either through platform providers or, in some cases, directly hosting the data, particularly as the cloud has enabled more flexible solutions.
In terms of RWE adoption trends, smaller biopharma companies tend to focus on specialty disease areas, whether that’s rare disease, oncology, immunology, or others. And while smaller companies may not have the same dedicated RWE centers of excellence as larger companies, we still see data used across different functional areas, ranging from R&D to commercial forecasting, outcomes research, and safety and surveillance functions.
Q: How do you work with biopharma partners to help them maximize the impact of the RWD, and of the evidence they generate?
A: At Optum, we adopted a high-touch client model fairly early, one of the first research-ready EHR data assets. These data require close collaboration with the client to align study objectives with disease variables, and to ensure the data methods and functions established by researchers and data scientists are appropriate based on the clinical documentation of electronic medical record (EMR) systems.
Remember, real-world data typically is not created for research. These data are created either for claims billing and reimbursement on the administrative side, or clinical documentation of patient care. Of course, what makes these data so powerful and compelling is that they hold the key to quantifying therapeutic benefit in real-world populations, as opposed to the relatively narrow populations that we see in randomized controlled trials (RCTs).
Q: What challenges in using or adopting RWE are unique to the small and mid-sized biopharma space, and how have you helped clients troubleshoot?
A: Going back to the consumption model framework, many challenges relate to their data infrastructure, and whether they have teams of dedicated data scientists as we see in larger organizations. That said, the buyers and data consumers in smaller companies today are not any less sophisticated than those in larger companies. In many cases, we see a more thorough collaboration across their organization, to the extent that, for example, an outcomes researcher deeply understands the brand’s commercial strategy.
This typically creates a more connected strategy for RWD to plug into. The major challenges remain in data consumption, in terms of the infrastructure and resources to analyze it. But in terms of the strategic approach and questions asked, we see similarities across the industry, regardless of client size.
Q: Do you see opportunity for an enterprise-wide adoption of RWE in the small and mid-sized biopharma market?
A: Definitely. As the industry acquires more experience generating evidence from RWD, we’re seeing more employees with data science and epidemiology skills and experience migrating throughout the industry. From the regulatory perspective, we also see agencies continuing to advance in their willingness to consider RWE in their decision-making.
The demand for RWE has been growing steadily for more than five years. As we continue to see lower cost and more flexible infrastructure solutions in the cloud environment, combined with platform tools like Aetion’s, which are designed to translate real-world data into evidence, we don’t see demand slowing down.
Q: What qualities have you observed in the most effective RWE teams?
A: Curiosity is a key trait, and with that, both comfort with and understanding of the methodological approaches needed to address some RWD complexities. Again, these data are not generated for research and they are not RCTs telling a very tight story—which, by the way, don’t always square with the reality of clinical practice.
The teams having the most success are those that understand the data being captured, and how to design a study and apply that data to address key research questions.
Q: In partnering with Aetion, what benefits do you envision for the small to mid-sized biopharma market?
A: The two biggest benefits are ease of adoption and time to value. Clients can access the data, generate evidence, and begin deriving value from day one, rather than the time typically required to build out a data infrastructure, load the data, and then create processes to manipulate and query the data.
Q: What are the most exciting future opportunities for RWD? What advances have you seen with NLP or other technologies?
A: Certainly NLP has been tremendous. Much of the provider-patient interaction is documented in narrative, unstructured text, rather than structured EMR inputs. Integrating this technology has been a major focus for us, and it has helped to illuminate the value of RWD. Coupled with that, object character recognition allows us to extract information from scanned documents. This is incredibly powerful, particularly as we spend more time in specialty disease areas.
In terms of future opportunities, the next frontier is moving beyond the traditional health care delivery system, and the world of claims and clinical data, to begin incorporating more patient-powered data. These include elements obtained from wearable medical devices. Another key future opportunity is to combine genomic data with phenotypic claims and EHR data, which will accelerate new discoveries in drug development.
Q: Any final thoughts?
A: I’d like to acknowledge COVID-19 and this historically challenging time. It’s disrupting many of our processes in the industry; for example, patient recruitment in clinical trials has substantially decreased, as patients aren’t physically visiting their providers as frequently.
These disruptions will further elevate the need and importance for RWD and RWE to generate daily insights from claims and clinical documentation. Indeed, we are seeing increased interest from regulatory agencies, including the FDA, for incorporating such data into surveillance and drug approval programs.
We’re very much looking forward to continuing to partner with our clients on this journey.
The Evidence Digest
Subscribe to our newsletter for insights and updates on research and applications for real-world evidence in life sciences and value-based health care.