Real-world Evidence and External Control Arms in Oncology: Progress, Friction, and the Path Forward

‍Introduction

On April 7, 2026, the Friends of Cancer Research (FOCR) convened a public meeting: Application of External Control Arms (ECAs) in Oncology Drug Development. The meeting presented preliminary results from FOCR’s ECA Pilot Project and explored the broader role of real-world evidence (RWE) and ECAs in regulatory decision-making through a keynote session with Dr. Amy Abernathy and 3 panel discussions.

Progress is undeniable. The FDA has now cited 13 approvals across CDER and CBER where RWE was used in regulatory decision-making. But despite this momentum, ECAs, and RWE more broadly, remain a work in progress: scientifically, operationally, and regulatorily.

The meeting itself reflected that tension: optimism about what is possible, paired with a clear acknowledgment of what still needs to be solved.

The Expanding Maturity of RWE

Since the 21st Century Cures Act (2016) and FOCR’s early pilot efforts (2017), the RWE ecosystem has evolved significantly. Data sources have expanded beyond claims to include electronic health records, linked datasets, and patient-reported outcomes. Analytical methods have matured, and the FDA has become more comfortable engaging with these approaches and has published over a dozen guidance documents addressing the use of RWE.

Dr. Amy Abernathy highlighted the progress made with data: “we are now moving toward longitudinal, multimodal data that captures the full patient journey across systems and time.”

However, she noted six key gaps:

• Data curation and transparent transformation

• AI governance and continuous validation

• Scientific rigor and clear statistical plans

• Documentation and traceability

• Systematic handling of missing data

• Convergence toward best practices

In summary, we have more data, better tools, and an understanding of the gaps, but we are still building the collective experience needed for scaling and standardization.

The FOCR ECA Pilot: Defining “Fit for Purpose”

FOCR’s ECA Pilot Project is the fourth in a series of initiatives aimed at developing and establishing methodology for using RWD in regulatory contexts. This pilot aimed to test whether multiple data partners could independently construct an external control arm that replicates a Phase 3 randomized controlled trial (RCT) in metastatic pancreatic cancer.

Selected data partners were given the trial eligibility criteria translated for RWD and asked to construct a cohort independently using their own data source. These cohorts were then compared to the original trial control arm.

The results were encouraging and showed that independent construction of ECAs across heterogeneous data sources is feasible under a shared statistical analysis plan.

At the same time, variability across partners was expected and observed. Differences from the randomized trial were attributed in part to known issues, such as trials enrolling healthier patients than those seen in real-world settings. Reasons for differences will be explored further in the pilot.

The takeaway is nuanced:
Feasibility has been demonstrated, but consistency and reproducibility remain open questions.

FDA Perspective: Fitness-for-Use Is Multidimensional

In the second panel, Dr. Marie Bradley, Senior Advisor for RWE in the Office of Medical Policy at CDER, was asked directly about how sponsors should think about factors such as relevance and completeness when assessing real-world data sources for constructing an ECA.

Dr. Bradley emphasized that fitness-for-use is inherently multidimensional.

At its core is data relevance:

• Does the dataset reflect the appropriate patient population?

• Does it capture the clinical context and sufficient follow-up?

But relevance alone is not sufficient. Sponsors must also demonstrate:

• Clinically meaningful endpoints

• Comparability to the trial population

• Reliable and consistent measurements for key variables

Notably, the FDA has not defined fixed thresholds for acceptable data quality. Expectations remain context-dependent and evolving.

Another important signal: the FDA continues to require patient-level data submissions for CDER and CBER, even when RWE is used. This is significant because late last year the Agency issued a statement that it was considering allowing RWE submissions to CBER and CDER without requiring identifiable patient-level data in all cases. If implemented, this represents a significant departure from standard FDA submission requirements. Dr. Bradley’s specific mention of this may indicate that the FDA has reconsidered this and will continue with the current requirements.

The Role of Data Standards

One notable gap in the discussion was the role of data standards.

Unlike Sentinel, which relies on a common data model and standardized, centralized query construction, the ECA pilot allowed each partner to independently implement eligibility criteria across heterogeneous data sources.

This raises critical questions:

• How much variability is introduced through independent interpretation and what contributes to the variability?

• What is gained or lost when data are standardized versus left in native persistent formats?

• In cases where cohorts are being created from large data sources, where in the pipeline is the optimal place for standardization to occur?

As ECAs scale and use cases widen, these questions will become increasingly important. As noted in one of the panels, very large datasets are often reduced dramatically when eligibility criteria are applied to create cohorts. I’ve seen patient counts go from the tens of thousands to below 100 after the application of inclusion/exclusion criteria. Standardization to facilitate characterizing the data, for example, the reason for a missing data element, a major issue with RWE, could have significant benefits when analyzing fitness for use.

Standardization may not be required to demonstrate feasibility, but it may be essential for reproducibility, comparability, and operational efficiency.

The Tension: Flexibility vs Certainty

A central theme throughout the meeting was a familiar tension:

Industry wants:

• Clear thresholds

• Defined benchmarks

• Predictability

FDA provides:

• Principles

• Contextual evaluation

• Flexibility as RWE data and methods evolve

As noted during the discussion, stakeholders often want both certainty and flexibility, but the two are inherently in tension.

In an evolving field, this may be unavoidable. Premature standardization could constrain innovation, while insufficient guidance can slow adoption. However, it was underscored that flexibility does not mean the FDA will reduce its evidentiary standards.

Toward a New Evidence Paradigm

Perhaps the most forward-looking insight was this:
We are still framing the conversation as RWD versus randomized controlled trials.

That framing is increasingly outdated.

The future lies in integrated evidence generation. Dr. Abernathy emphasized in her keynote that industry will be able to move faster when we bring RCT and RWE data together and start thinking about the complete evidence package and multiple ways to get there.

One particularly compelling idea raised:
When considering using an ECA, don’t just optimize real-world data to match trials, but optimize trials to better reflect the real world.

This shift could reduce the structural mismatch between trial populations and real-world patients, improving both generalizability and the viability of ECAs.

Conclusion: Progress and the Goal

The promise of RWE is clear:
It offers a pathway to faster, more efficient drug development and the potential to bring therapies to patients sooner. ECAs are one of those pathways.

But the path forward requires balance.

• Industry needs to invest in rigorous, transparent, and reproducible methods

• Regulators should continue to provide guidance while preserving flexibility in an evolving landscape

• Both must resist prematurely regulating what is not yet fully understood in the name consistency for predictability

The goal is not simply to make ECAs work. It is to make them credible, scalable, and decision-grade so we can effectively help bring safe and effective drugs to patients faster.

Implications for Data Strategy and Implementation

At Orizaba Solutions, this is exactly where we focus, helping federal agencies and sponsor companies bridge the gap between data availability and decision-grade evidence. Our work centers on modernizing, integrating, and operationalizing real-world data through data standards, governance, and AI-ready pipelines, with deep experience supporting FDA submissions, understanding RWE submission requirements, and large-scale healthcare datasets. As the use of RWE continues to evolve, we are particularly interested in helping organizations design data strategies and architectures that are not only fit-for-purpose but also scalable, transparent, and submission-ready.