Syntegra recently worked with Janssen to generate high-fidelity synthetic data for their internal use. We’re excited about this initial work and our broader partnership. We recently spoke with Janssen’s RWE Method Lead, Sebastian Kloss, about the difficulties of data access in the EU and how synthetic data helps address these challenges to improve health economics and outcomes research (HEOR) and feasibility studies.
In the EU, patient-level data is notoriously difficult to access. What challenges do you and your team face when it comes to accessing the data you need for your real-world evidence (RWE) research?
First and foremost, I basically see it quite positively that patient-level data is not easily accessible in Europe. It ensures data privacy, integrity, security, ownership and ultimately trust. However, RWE research stays and falls with data quality and methodological know-how. As the data source landscape is continuously evolving and growing, it is crucial to identify, select and collaborate with those having the highest potential of finding answers to our research questions.
There are three key challenges for Pharma when it comes to data access: 1. Assessing a dataset’s value, quality and strategic value is difficult and often costly when there is no direct access. 2. Quick feasibility analyses, which can inform strategic decision making and be more competitive is often not possible. 3. Data ownership and methodological expertise often do not go hand in hand, so more complex analytical approaches or pilots can take a long time if direct data access is not possible.
How do these challenges impact your ability to conduct HEOR and feasibility studies?
This is a crucial aspect. Often, time is essential for getting a quick overview of available patients with specific conditions and/or variables. These queries can sometimes be time and resource intensive, which makes it very difficult to plan accordingly. Data owners usually get overwhelmed by these requests, and budget and sometimes even contracting are needed upfront.
How do you see the role synthetic data can play in RWE and broader pharmaceutical research?
Challenges with data access, competitiveness through quick feasibility studies and the assessment of external data quality are key goals since RWE became a standard piece in any evidence package, and they can be addressed with synthetic data. Another core use case of synthetic which is now evolving has been the anonymization of patient-level data to allow data sharing in an GDPR compliant manner. As more complex analytical approaches in the era of GPT-3 are now arising, synthetic data approaches can now be used to synthesize a patient journey more adequately, increasing analytical opportunities and more efficient collaborations with data owners.
Diving deeper into the use of synthetic data for feasibility, study design, and development of methods. How do you see this use case benefiting RWE or other research from a time or cost saving perspective?
RWE is constantly evolving and so is the external data and policy landscape. Especially for EMEA regions, with quite fragmented healthcare frameworks, synthetic data can bring more flexibility into research and feasibility. This includes the development of new, more targeted approaches of patient identification and evaluation. Working on these concepts in a data driven and not purely theoretical way can help identify early roadblocks and accelerate solutions.
How should patients think about synthetic data and their privacy?
This should be an area of more active communication and translation. Synthetic data per se needs to be explained and discussed on a lot of different levels, less so technically speaking but more from the privacy perspective. Through a more complex synthesization process, we are no longer talking about real patients and this needs to be explained while driving the debate on this topic.
Building off of the need for better external communication, the concept of synthetic data is still relatively new to healthcare and is not always well understood. What interested you about using synthetic data to solve these challenges? How has the use of synthetic data impacted your research?
A lot around synthetic data needs to happen within the industry. As data itself is considered the backbone of any research activity, the concept of garbage in, garbage out is on everyone’s mind. Convincing researchers and others that synthetic data can help maintain a high data quality without significant loss of information is very motivating. As a person with a scientific background, finding and internally working with reliable data within the industry (and especially focusing on European data) was always preferred. The concept of synthetic data could become a game changer with this regard.
Syntegra allowed us to work with a partner data source in Germany without any privacy restrictions, accelerating the RWE research needed to bring better treatments to patients.