From Optional to Foundational: RWE in the Era of One Trial

I. The Persistent Execution Gap in RWE Generation in Clinical Development

We have previously articulated the case for systematic, integrated RWE generation beginning early in drug development — not as an afterthought, but as a foundational component of strategy.

This includes using representative real-world point-of-care data to build a deep, comprehensive, and current understanding of the characteristics, routine care, and outcomes of the indicated population. When developed early and intentionally, this evidence informs regulatory opportunities, shapes regulatory strategy, optimizes trial design, and ensures interpretability of trial results. It reduces uncertainty before it becomes an asset program risk.

Yet across biopharma, adoption remains uneven. RWE efforts are frequently reactive — mobilized to supplement existing published literature, address a regulatory question late in development, contextualize unexpected results, or compensate for design limitations. The industry does not lack awareness of RWE’s value. It struggles with early systematic execution.

Under evolving regulatory expectations, this execution gap will become increasingly consequential.

II. Why 2026 Raises the Stakes

In the first months of 2026, FDA has signaled a meaningful evolution in evidentiary expectations.

First, a recent NEJM publication by FDA leaders Drs. Vinay Prasad and Marty Makary describe a framework in which a single pivotal trial, accompanied by confirmatory evidence, is the new default to satisfy the statutory standard of “substantial evidence.” This moves away from the traditional expectation of two adequate and well-controlled trials and places new emphasis on the totality of evidence.

Second, FDA issued new draft guidance on the use of Bayesian statistical approaches to support inference on clinical effectiveness and safety. The guidance highlights the potential to incorporate prior information, borrow from external data, and apply adaptive methods — particularly for conditions with unmet need or affecting small populations.

Together, these developments point toward a regulatory environment that values efficiency and flexibility. But flexibility does not lower the evidentiary bar. It changes where and how evidence must be generated.

III. One Pivotal Trial Means No Margin for Design Error

Drs. Prasad and Makary outline a new default standard in which one robust, well-designed pivotal trial, combined with confirmatory evidence (e.g., independent clinical data, real-world evidence, mechanistic understanding, etc.), may meet the legal requirements for marketing authorization. They advocate for this new standard as a replacement for the historical default expectation of two pivotal trials to improve efficiency and reduce development costs, enabling faster patient access to therapies at lower cost.

Under a two-trial paradigm, design limitations in one study may be contextualized or balanced by findings from the other. Under a single pivotal trial framework, there is no opportunity to “average out” weaknesses. Design decisions carry amplified consequences. Trial design must therefore be grounded in robust, population-specific evidence. Sample size calculations, entry criteria, and endpoint selection depend on an accurate and current understanding of the indicated population’s baseline characteristics, standard of care, and natural history – this understanding can only come from robust real-world evidence.

Moreover, real-world evidence may contribute to the robust and compelling confirmatory evidence that is critical to enabling trust in the findings of a single pivotal trial. Based on our experience in supporting use of RWE in sponsor applications, and review of other FDA approval use cases and guidance (e.g., Demonstrating Substantial Evidence of Effectiveness With One Adequate and Well-Controlled Clinical Investigation and Confirmatory Evidence), there are many valuable uses of RWD in applications – natural history to support sample size assumptions or corroborate the outcome rate in the trial comparator arm, to name a few. 

This is not an academic consideration. Weak, misaligned, or limited confirmatory evidence will undermine the efficiency gained by conducting one trial.

IV. Bayesian Methods: Flexibility Requires Credibility

FDA’s draft Bayesian methods guidance recognizes how these approaches, when based on high-quality historical data, offer solutions in settings where traditional RCT approaches struggle and often, there is unmet need — small populations, rare diseases, ethical constraints on control arms. These approaches use external data to augment randomized concurrent controls, extrapolate to additional populations (e.g. from adults to pediatric patients), enable adaptive decision-making (e.g. timing of interim analyses), and improve the reliability of subgroup analysis. Successful application is supported by robust RWE of disease outcomes under standard of care, overall and within subgroups. This evidence is key to understanding how a fully randomized controlled trial is not feasible and may also inform the relevance of external information to another population and pre-specified success criteria for determining whether primary objectives are met. 

In this context, systematic RWE becomes even more foundational infrastructure. Without early, integrated evidence generation that carefully characterizes populations and treatment landscapes, sponsors may unintentionally introduce risk when responding to FDA’s flexibility around Bayesian methods.

V. Bringing It Together: Positioning for Efficiency 

The regulatory environment is evolving toward increased efficiency. FDA’s new pivotal trial framework and Bayesian guidance increase opportunities to accelerate development and address unmet need, but do not change the circumstances in which a single RCT, a non-randomized study (single arm trial + external control arm), or use of Bayesian methods may be acceptable. In this evolving environment, systematic, integrated RWE is no longer a supplementary exercise. It is a prerequisite for regulatory acceptability. 

At this point, Makary and Prasad’s articulation of a single pivotal investigation standard represents a new regulatory stance rather than formal guidance. As with many policy positions, it may evolve over time or shift under future leadership. The organizations that treat RWE generation as infrastructure — initiated early, maintained continuously, and integrated strategically — will be positioned to navigate the inevitable shifts. Those that continue to generate evidence reactively will find it difficult to respond effectively.