Beyond Clinical Trials: Real-World Evidence in Medical Devices
Article Summary
Real-world evidence, electronic health records and AI are expanding the options for medical device manufacturers to generate clinical evidence, address regulatory requirements and manage research costs.Article Contents
Introduction
Traditional clinical trials remain an important part of medical device development, but they are not the only way manufacturers can generate meaningful clinical evidence. Real-world evidence, electronic health records and emerging technologies such as artificial intelligence are creating additional opportunities to understand device performance and address questions that conventional studies may not answer.
For manufacturers, particularly those working with limited budgets and resources, understanding these alternatives can make a significant difference to their evidence-generation strategy.
John McDonald, Senior Director of Clinical Evidence Generation at Teleflex Incorporated, explains how manufacturers can use different evidence sources, what makes real-world data reliable and where AI could accelerate the process.
What Can Real-World Evidence Tell Us That Clinical Trials Cannot?
Traditional clinical studies can require multi-million-pound budgets and take several years to complete. They also typically involve carefully selected patient populations, which may not always reflect the diversity of patients encountered in routine clinical practice.
Real-world evidence offers a complementary approach, helping manufacturers address evidence gaps that may be difficult or impractical to investigate through traditional clinical trials.
“Real-world evidence can address gaps that might not be achievable or even feasible through traditional clinical studies,” explains McDonald.
One important opportunity is improving the representation of patient populations in clinical evidence.
Certain groups, including ethnic minorities and paediatric patients, have historically been underrepresented in clinical research. By accessing larger healthcare datasets, manufacturers may be able to investigate outcomes across a broader range of patients and geographical locations.
For example, electronic health records can provide access to information from different healthcare settings and patient populations, potentially offering a more representative picture of how devices are used in everyday practice.
Real-world evidence can also help manufacturers understand how a device performs outside the controlled environment of a clinical investigation. This may reveal practical insights into its use, identify unmet needs or highlight opportunities to investigate additional indications and product claims.
For McDonald, this is one of the important benefits of post-market evidence generation: it provides manufacturers with intelligence about how their products are actually being used.

What Makes Real-World Data Reliable Enough for Regulatory Acceptance?
Having access to large datasets does not automatically mean the resulting evidence will be suitable for regulatory decision-making.
Manufacturers need to demonstrate that their data sources are relevant to the research question and sufficiently reliable to support the conclusions being drawn.
McDonald highlights device identification as one of the first practical considerations.
Can the manufacturer reliably identify its specific device within the dataset? Is it possible to distinguish the product using a device identifier or brand name?
If a device cannot be confidently identified, the value of the resulting evidence may be limited.
Data quality is equally important. Manufacturers need to understand where the information comes from, how it has been collected and whether it is sufficiently complete and relevant for the intended purpose.
In the United States, the FDA has published guidance setting out expectations for the use of real-world data and evidence in medical device regulatory decision-making.
For manufacturers, understanding these expectations early is important. The objective should not simply be to collect as much data as possible, but to establish whether the available information can support the specific regulatory question.
Generating Clinical Evidence When Budgets Are Limited
For smaller medical device manufacturers, the cost of clinical evidence generation can be a significant barrier.
However, a randomised controlled trial is not necessarily the only way to generate useful evidence. The appropriate methodology depends on the device, the intended claims, the available data and the regulatory requirements.
McDonald points to European guidance on clinical evidence, including MDCG 2020-6, which discusses different types and levels of clinical evidence in the context of legacy devices.
Alternative approaches may include retrospective data analysis, registry information and structured clinician surveys.
For example, a manufacturer could consider asking clinicians to assess patient outcomes using information available in electronic health records. Depending on the research question and the quality of the information collected, this could provide a more cost-effective way of generating clinical data than a prospective investigation.
However, manufacturers need to assess whether the proposed methodology is appropriate for their intended purpose and whether the resulting evidence will meet the expectations of the relevant regulatory authorities.
There are also practical considerations beyond regulatory acceptance. If the manufacturer intends to publish the findings, it needs to consider whether the chosen methodology and resulting evidence will meet the requirements of the intended journal.
The key is to explore the available options rather than automatically assuming that a large prospective clinical study is required.

Can AI Make Clinical Evidence Generation Faster?
AI is creating new opportunities to improve the speed and efficiency of clinical evidence generation.
Its applications range from searching documents and supporting the development of standard operating procedures (SOPs) to analysing information and helping researchers investigate specific clinical questions.
For McDonald, speed is likely to be one of AI’s biggest advantages.
However, he distinguishes between using AI to improve internal processes and relying on it to generate evidence supporting a regulatory claim.
Using AI to search documents or assist with administrative tasks is very different from using an AI model to support an expanded medical device indication or a new product claim.
In the latter situation, manufacturers need to demonstrate appropriate governance, oversight and model validation. They must be able to show how models have been developed and validated, and that their outputs are sufficiently reliable and repeatable.
This also requires organisations to have the appropriate skills and expertise to manage the technology.
AI could help manufacturers reach research outcomes more quickly, support publications and accelerate evidence generation for potential indication expansions. But its value will depend on whether it is implemented within a robust governance framework.
Choosing the Right Evidence-Generation Approach
Real-world evidence, traditional clinical investigations and emerging technologies all offer opportunities for medical device manufacturers. However, none represents a universal solution.
The appropriate approach depends on what the manufacturer needs to demonstrate, which patient populations it needs to understand and what data are available.
A new device with an ambitious clinical claim may require a prospective clinical investigation. An established product may offer opportunities to generate further evidence through existing healthcare datasets. An organisation looking to investigate economic value may need to consider a different type of data altogether.
Manufacturers should therefore consider the full range of methodologies available before deciding how to invest their resources.
As McDonald’s experience highlights, the most useful evidence strategy is one that connects the research question with an achievable methodology and a clear understanding of regulatory expectations.
Disclaimer. The views and opinions expressed in this article are solely those of the author and do not necessarily reflect the official policy or position of Test Labs Limited. The content provided is for informational purposes only and is not intended to constitute legal or professional advice. Test Labs assumes no responsibility for any errors or omissions in the content of this article, nor for any actions taken in reliance thereon.
Get It Done, With Certainty.
Contact us about your testing requirements, we aim to respond the same day.
Get resources & industry updates direct to your inbox
We’ll email you 1-2 times a week at the maximum and never share your information