The FDA is changing the rules. Marking a fundamental shift in how we define acceptable clinical proof, it has indicated that one well-controlled trial, supported by confirmatory evidence––in particular, real-world evidence––will become the default for drug approvals. Many are applauding this as a sign of more regulatory flexibility. But there is a more significant transition at play.

Following decades of regulatory requirements for two independent trials to demonstrate efficacy and safety, elevating real-world data (RWD) to the status of regulatory-grade evidence will bring both opportunity and challenges to the industry. Sponsors, regulators and researchers must now reassess the definition of acceptable evidence in drug development.

An Evolutionary Shift

The FDA’s openness to new evidence standards, specifically real-world data, did not happen overnight. Over the past decade, the regulator has taken incremental steps to incorporate real-world evidence into its decision-making. This began with encouraging the use of RWD to supplement traditional randomized trials, before progressing to support its use in building external control arms.

We are at the next stage in that evolution. The FDA is not simply relaxing standards. Rather, it has recognized the advances of AI capabilities to analyze vast amounts of unstructured and previously inaccessible data at a scale of analysis that would require an impossible amount of human labor. These new tools can harmonize fragmented data sources and clinical protocols, unifying them in one cohesive dataset. This provides a level of interoperability that dramatically accelerates data-intensive workflows, from endpoint validation to continuous safety monitoring.

In that sense, the regulators are moving faster than some parts of the industry in acknowledging that new analytical tools can improve the rigor of evidence generation. It is the pharmaceutical organizations that must be willing to change the processes and evidence frameworks they’ve been using for decades.

Addressing the Strain of the Traditional Trial Model

The move from the two-trial stipulation to a one-trial requirement and increased acceptance of real-world evidence is also a result of inherent pressures on the traditional system. Patient recruitment has been a huge pain point across indications. Fewer patients are available and willing to participate in the growing number of trials. Control groups have posed an even greater challenge.

This created a paradox where clinical trials remain the standard for regulatory approval, yet it is increasingly difficult to adequately populate the studies to fully represent patient populations. As a result, there are cases where drugs have been approved based on trials involving small and fragmented cohorts, regulators having been forced to accept them due to a lack of alternatives.

The FDA’s shift towards real-world evidence reflects a genuine effort to address this longstanding problem. Rather than lowering standards, it is raising them back to previous levels, while expanding the availability of tools that can meet them. The FDA is not making approvals easier. It is removing the excuses. Drug developers can now access better data and so are expected to use it.

The Missing Link from Data to Evidence

While the use of RWD has rapidly expanded, this is not what regulators are evaluating. Their priority is evidence, and it is this gap between data and evidence generation where efforts tend to fall short. Data access is no longer the primary constraint––large amounts of datasets are now readily available to organizations.

The challenge now is to make the data usable and convert it into regulatory-grade evidence. This requires unifying data from multiple disparate sources and converting that raw data into a clean, structured format that reflects real-world clinical settings and patient journeys, before applying rigorous study design that addresses biases and confounding variables. This process is complex and resource-intensive, requiring a level of epidemiological expertise and observational research capabilities that the industry has traditionally lacked. Meanwhile, the analytical tools traditionally employed to do this were overly cumbersome and complicated to use.

AI is shifting this dynamic - not just in granting access to data, but directly arming researchers with the advanced analytics needed to understand it. Advances in AI-driven analytics can help close this gap and accelerate the transition from RWD to regulator-ready evidence by enabling more consistent data structuring and preemptive bias detection. Natural language processing (NLP) models can now extract specific clinical outcomes from thousands of unstructured, messy physician notes, reports, clinical summaries and research documents. By standardizing this fragmented data from disparate sources, AI can create a unified, quantifiable dataset that regulators can actually evaluate.

However, the industry is still in the early stages of building these capabilities. Until it does, the primary challenge will be to guarantee that evidence drawn from real-world data is regulatorily sound.

The Strategic Implications for Pharma and CROs

As real-world evidence plays an increasingly central role in regulatory decision-making, integrating RWD into development programs can no longer be a late-stage consideration. For biopharma companies, this should begin prior to even designing the study, grounding the development plan with real-world patient journeys.

This early use of RWD can help identify treatment patterns, comorbidities and drug combinations that theoretical models cannot anticipate. It can also enable clinical teams to build - and stick to - highly informed, pragmatic studies from the start. Incorporating RWD into the foundation of a trial can mitigate structural biases and identify strategic opportunities that might otherwise be missed.

Large CROs that were early to establish RWD capabilities are already seeing increased demand, even where confidence in their output is inconsistent, as pharmaceutical companies seek partners who can navigate these evolving requirements. Smaller organizations, recognizing the opportunities, are also entering the space, incorporating real-world data and AI into their offerings. While their capabilities are still uneven, this is likely to improve through investment and acquisitions, as demand for integrated evidence generation increases.

Infrastructure and Technology Implications

Traditional approaches to real-world evidence, built around small, controlled datasets cannot scale to meet the volume and complexity of real-world evidence. As the sizes and number of datasets continue to grow, these manual processes cannot provide the necessary speed and accuracy. Meanwhile, automated platforms capable of handling large amounts of data are still viewed with skepticism, due to a lack of transparency or difficulties in validating their outputs.

Trust is becoming the central requirement, and the development of systems that can handle the data while meeting the necessary scientific and regulatory standards. These platforms must provide reasoning, traceability, validation of underlying data points and rigorous analysis. The ability to analyze regulatory-grade data without compromising scientific rigor or privacy will resolve much of the skepticism around platforms, which should be semi-automated rather than automated. As these platforms mature to incorporate stronger safeguards around the use of AI, we will see more consistent and scalable evidence generation.

The Future of Evidence Generation

Drug development is set to achieve the holy grail of continuous real-world evidence generation. But reaching that will require a comprehensive foundation built on advanced AI and true data interoperability. The companies best positioned to succeed in this new era may not be massive legacy organizations. Instead, small and medium pharma and biotech companies hold an advantage. Not being restricted by decades of rigid infrastructure, these agile players can quickly adopt and leverage rapidly evolving technologies and modern RWD networks. As the FDA shift comes to fruition, the winners will be those who stop treating evidence as a one-time regulatory hurdle and instead harness it as a continuous, dynamic lifecycle.