Convergence: Thinking beyond the registry for real-world device evidence

Regulatory NewsRegulatory News | 17 September 2020 |  By 

Medical device registries can provide rich fields for data mining when real-world evidence is sought. But most medical devices will never land on a registry. Does this mean that these manufacturers will not be able to find a source of real-world data?
Not according to Heather Colvin, who spoke during a real world evidence-focused session at RAPS Convergence 2020. Colvin is the director of evidence and outcomes policy for MD regulatory affairs at Johnson & Johnson. She took a look beyond the registry, to see how real-world evidence (RWE) policy and practice might be advanced by using other data sources.
Registries work in some cases, but “What about the rest of us? What about the devices that are unlikely to be captured in registries?” asked Colvin.
Real-world data can come from many sources
Sharing a “cheat sheet” of all sources of device RWE, Colvin said, “Traditionally, some people have thought about complaints as not real-world data sources, but [they] can be a source of real-world data if compiled with other information.”
The Eudramed database, The US Food and Drug Administration’s (FDA’s) Manufacturer and User Facility Device Experience (MAUDE) database, and even internal company databases can be RWE sources, she said.  Increasingly, patient-generated information is being used as RWE. These can include survey results, patient preferences, social media posts, and data generated by wearables and sensors. Devices themselves can also now capture and store data about realtime use, Colvin pointed out.
EHR and claims data have strengths – especially when linked
However, electronic health records (EHRs), claims data, and registries linked to claims data all have potential to be fertile grounds for harvesting real-world data, said Colvin, turning her focus to these areas. “Times are changing, and more and more granular information is being captured in hospital systems and supply chain management tools, and in claims administrative data.”  Registries that can link claims and EHR data, she said, represent “the perfect universe” for RWE from these sources.
Each data source, Colvin pointed out, comes with its own pros and cons. Insurance claims are abundant and representative of the population, and give longitudinal data, but such details as the device identifier and patient-reported outcomes or scores are lacking. With the EHR, “you have clinically rich information,” but you may lack the volume and longitudinal tracking that insurance claims provide, she said. Billing databases may mesh well with insurance claims and are also very large datasets.
“Just because the data is there doesn’t mean that it’s appropriate for real-world evidence generation,” said Colvin. Also, “You need to manage expectations” so that your team does not expect RWE for every product. If a product is newly on the market, she said, or used in a very small population, there just has not been enough time for sufficient accrual of real-world data.
Both the increasing ubiquity of unique device identifier (UDI) codes and the “increasing granularity” of International Classification of Diseases-10 (ICD-10) codes have aided generation of real-world device information. In ICD-10, she said, has over 300 fracture codes that specify location, type and severity. “So you get increasingly rich information” from such sources.
Colvin suggested that device manufacturers start to think of the full map of possibilities for RWE generation, while keeping the product in mind. “Try to make that pros and cons list about what makes sense, given the product that you have.”
NEST is paving the way
“Collaborations pave the way to advances” in gathering and analyzing quality RWE, said Colvin. A multi-layered team of collaborators can includes trade associations, clinical societies, patient organizations and think tanks, but should also have a public-private partnership approach that includes regulators and payers. “Many groups have been working on this for a long time,” said Colvin, so device manufacturers should not feel they have to go it alone or start from scratch.
An example of such a public-private collaboration involves FDA’s consultation with stakeholders that resulted in the National Evaluation System for health Technology (NEST). “The idea was to have multiple people…participate in that system,” including not just FDA and manufacturers, but private insurers, the Centers for Medicare and Medicaid and even clinicians, said Colvin.
The seed money for NEST came from FDA’s Office of Technology Assessment, but medical device manufacturers agreed to contribute funding as part of the fourth Medical Device User Fee Agreement (MDUFA IV). “People are putting some skin in the game,” she said.
NEST, she said, is tackling two big tasks: first, gathering a network of diverse data sources that can work together to generate RWE. The network currently has data from over 141 million patients across over 3,000 outpatient sites and 162 hospitals and medical centers.
Currently, 14 research collaborators are working through 21 real-life scenarios provided by device manufacturers; these test cases are being used to stress test NEST network capabilities.
With the NEST test cases, researchers are using a workflow where they first develop the protocol and define data elements and identify the specific device or devices they wish to study. The method for abstracting device data must be set, and researchers must identify the patient populations to be studies and the particular outcomes and covariates to consider. Researchers must also develop the codes and algorithms that will be used to identify the data elements of interest. All of this, Colvin said, is in the service of making sure the RWE gathered is relevant and reliable, meaning it will be fit for its intended purpose.
Globally, efforts to gather RWE from diverse sources that go beyond the registry are increasing, as are attempts at harmonization and coordination. This means that policy challenges are also cropping up. These center around such issues as data quality and methodology, including how data sources will be characterized and whether data linkage can be accomplished. Governance issues around access, data sharing and transparency are being worked through as well, said Colvin. Those seeking to use RWE also must address patient informed consent, cybersecurity issues, and other ethical concerns.


© 2023 Regulatory Affairs Professionals Society.

Tags: devices, FDA, medical, US

Discover more of what matters to you

No taxonomy