We are excited today to announce Datavant for Clinical Trials, a solution that makes it easy for sponsors, CROs, and regulators to confidently bridge traditional clinical trial data with real-world data. This morning, we announced a multi-year strategic partnership with Parexel to roll out this connectivity solution across all of their clinical trials, and we believe all clinical trials should be linked to real-world evidence 5 years from now.
The explosion of healthcare data and analytics over the last decade has made possible a new paradigm for clinical research that incorporates information from beyond the clinical trial context — although many frictions have made this difficult to put into practice. At scale, real-world data promises to make clinical research cheaper, faster, and more accurate.
Datavant’s vision is to connect the world’s health data to improve patient outcomes, and we see our launch today as a step towards eliminating some of those frictions by easily linking the clinical trial and non-clinical worlds.
This post will walk through what this launch means, and how we aim to remove friction holding back the use of real world data.
Background: The Promise and Peril of “Real-world data”
In the US, millions of healthcare events occur each day. Patients visit their healthcare providers for annual check-ups, fill and refill prescriptions for medication, undergo complex treatment for a chronic condition, or use wearable tracking devices to understand their activity and sleep. Each of these interactions can be thought of as an experiment, which collectively can unveil insights on a disease and how to best treat it.
This information about normal doctors interacting with normal patients, outside the context of a tightly-controlled clinical trial, is called “real-world data.”
There are numerous advantages to using real-world data to inform the research and development of new therapies, including the amount and types of data available, the large sample size empowering the ability to understand the effects of interventions in small segments of the population, the significantly lower cost of running a RWD-supported trial, and the ability to study more representative populations (women and minorities are under-represented in medical research). The FDA is increasingly recognizing the value of real-world data, incorporating it for post-marketing surveillance and other long-term studies, and defining guidelines for the use of RWD in studies that could be used to submit new therapies for FDA approval. The effective use of real-world data could revolutionize the way clinical trials take place.
Yet for all of the momentum, we still have a ways to go. One of the reasons that so few sponsors incorporate RWD into their study design is the challenge of sourcing the right data on the right patients to answer the research question at hand. This is hard for a few reasons:
- Lack of randomization in pure real-world studies. Clinical trials offer the rigor of randomization to control for biases in populations who receive treatment.
- Data fragmentation. The relevant data is often spread across various data sources; even something as simple as measuring 10-year mortality rates of patients with a particular treatment requires researchers to collect data from multiple institutions that often don’t share data (the institutions that know what treatment was provided generally differ from the institutions that may know outcomes 10 years later).
- Data messiness. Data collected in the real-world is often more messy and error-prone than in tightly-controlled clinical trials.
- Approach is not apparent at study start. The research questions and data sources that are relevant for a study may not become clear until after the trial has already started or been completed.
- Patient consent & privacy. Informed consent must sufficiently support RWD utilization and the aggregate RWD must be handled in a way that protects patient privacy.
All of these factors make it difficult to use real-world data to its fullest potential in clinical development. To overcome them, what is needed is infrastructure that enables data sharing and linkage across a variety of data types while preserving patient privacy.
Datavant’s Product Offering:
Our belief is that there are ways to combine the power of clinical trial randomization with the scale and representativeness of real-world data, and accomplish the best of both worlds. That’s why we’re thrilled to introduce Datavant for Clinical Trials to help companies use real-world data to recruit patients for trials, measure long-term outcomes, run hybrid studies, and ultimately develop safer and more effective treatments.
When a patient enrolls in a Datavant-supported clinical trial, they are assigned an encrypted, irreversible Datavant Patient Key. This Key can be used to link that patient’s records across data types, while protecting each patient’s privacy. For instance, if a patient studied in a trial also appears in both a disease registry and a separate hospital’s electronic health records, researchers could use the Key to link both of those records with the trial data without actually needing to know the patient’s identity. This creates the flexibility to connect RWD at any point in time without identifying the trial participants, even if the need was not contemplated at study start.
Flexible linking of RWD to clinical trial data can help sponsors more easily answer questions that may arise after study completion:
- What are the factors that drive outliers in the data — either patients who are “super responders” to therapy or those with a challenging adverse event profile?
- What are the health outcomes of the studied patients after the study window closes?
- What happened to patients that were lost to follow up or have missing values?
The Datavant Patient Key can be used to link patient records across Datavant’s vast ecosystem, which includes hundreds of data partners across all data types. With upfront patient consent, sponsors will be able to work with other ecosystem partners to pull in medical and pharmacy claims, electronic health records, labs, disease registries, wearable device data, social determinants data and much more to supplement their trial data. Data will remain de-identified throughout the process, so analytics based on the data are possible without compromising patient privacy.
By introducing de-identified Datavant Patient Keys into clinical trials, we can ensure that each clinical trial does not become a new data silo, and that sponsors can continue to learn from those trials long after database lock. And for patients that choose to participate in clinical trials in part to help others, introducing the ability to learn more from their participation allows them to play an even more significant role in medical research. In short, they become part of the process of innovation in healthcare.
In five years, our goal is for every clinical trial to contain Datavant Patient Keys that can be used to link that clinical trial data to real-world data. Running a large, multi-site clinical trial is far too costly and risky of an endeavor to create one new data silo. By making clinical and real-world data linkable, we will be able to learn significantly more from each clinical trial that is run regardless of whether the trial’s specific endpoints are met. This could accelerate the pace of innovation and treatment delivery for patients, and take us one step closer to realizing the vision of a learning healthcare system.