Rethinking Clinical Evidence Requirements in Rare Diseases

Key Takeaways

• Rare disease clinical trials often require alternative evidence strategies because small patient populations, complex biology, and limited natural history data can make conventional randomized studies difficult.

• The FDA’s rare disease evidence frameworks emphasize regulatory flexibility while preserving the statutory requirement to demonstrate safety and effectiveness.

• Natural history studies, registries, RWD, and external controls can help contextualize treatment effects in rare disease development when traditional control arms are impractical.

• Novel endpoints and surrogate endpoints may support rare disease drug development, but they require strong scientific rationale, regulatory engagement, and, in accelerated approval pathways, confirmatory evidence.

• Sponsors and development partners should integrate evidence strategy early across clinical design, endpoint development, biomarker planning, data quality, regulatory engagement, and operational execution.

Rare disease development begins with a paradox. Each individual condition affects a small population, but rare diseases collectively represent a major public health burden. In the United States, the Orphan Drug Act defines a rare disease as a condition affecting fewer than 200,000 people, yet more than 10,000 rare diseases affect more than 30 million people in the United States, or approximately one in 10 people, and about half of those affected are children.¹

That scale does not make rare disease development straightforward. The small population affected by any single rare disease can make clinical trials difficult, particularly when patients are geographically dispersed, clinical expertise is concentrated in a small number of centers, and disease progression is poorly understood. The U.S. Food and Drug Administration (FDA) identifies small patient populations, complex biology, and limited understanding of natural history as challenges in rare disease medical product development.¹

These features put pressure on the assumptions that underlie traditional clinical development. In many common diseases, sponsors can often design large randomized trials, enroll enough participants to power multiple endpoints, and rely on established measures of disease progression or clinical response. In rare diseases, the available population may be too small to support that model. The untreated disease course may be poorly characterized. Patients may have significant unmet need and few, if any, approved treatment options. In some cases, the population eligible for a targeted therapy may be a subset of an already rare condition.

The central regulatory question is therefore not whether rare disease therapies require evidence. They do. The question is what kind of evidence can credibly support regulatory decision-making when the conventional model is not feasible. Rare disease development demands rigor, but that rigor may need to be demonstrated through a different configuration of data: natural history information, real-world data (RWD), external controls, novel endpoints, surrogate markers, confirmatory evidence, and close regulatory engagement.

This shift is not a retreat from scientific standards. It is a recognition that a rigid insistence on conventional trial designs can be poorly matched to diseases with very small populations and urgent unmet need. The challenge for sponsors is to build evidence packages that are practical enough to execute, robust enough to interpret, and persuasive enough to support confidence in safety and effectiveness.

Regulatory flexibility in rare diseases is sometimes misunderstood as a willingness to accept less evidence. A more accurate interpretation is that regulators are clarifying how different forms of evidence may be assembled to meet the same underlying expectation: that therapies demonstrate safety and effectiveness.

The FDA’s Rare Disease Evidence Principles (RDEP) process offers one of the clearest examples of this thinking. RDEP is proposed for certain rare diseases with very small patient populations or subpopulations, generally fewer than 1,000 people in the United States, significant unmet medical need, and a known in-born genetic defect that is the major driver of pathophysiology.² Those criteria matter because they describe a setting in which conventional development approaches may be especially difficult. At the same time, the FDA states that drugs considered through this process would still be reviewed under the statutory marketing approval standard of safety and efficacy.²

That balance is the core of rare disease evidence reform. Regulators are acknowledging that the usual quantity of clinical studies and traditional placebo-controlled designs may be difficult to generate in very small rare disease populations, but that acknowledgment does not remove the requirement for a convincing evidentiary basis.² Instead, it shifts attention toward the scientific logic of the full development package.

For sponsors, this means the evidence strategy must be disease-specific from the beginning. Trial design cannot be separated from biology, natural history, endpoint development, patient identification, and the feasibility of generating comparative data. A program for a slowly progressive disease with established endpoints may allow a different approach than a program for an ultra-rare pediatric disease with limited longitudinal data. A therapy directed at a known genetic defect may raise different evidentiary questions than a therapy with a less direct mechanistic rationale.

The FDA’s rare disease development guidance is intended to assist sponsors in conducting efficient and successful rare disease drug development programs by addressing issues commonly encountered in rare disease development.³ That emphasis on program design is important. Rare disease evidence requirements are not determined only at the pivotal-trial stage. They are shaped by the earliest decisions about what data to collect, how to characterize the disease, which endpoints to develop, and when to seek regulatory input.

The practical implication is that flexibility must be earned through planning. A sponsor seeking to rely on a smaller trial, an external comparator, a novel endpoint, or a surrogate marker must be able to explain why that approach is appropriate for the disease, the population, the mechanism of action, the available data, and the unmet need. The smaller the patient population, the more important it becomes to ensure that every element of the evidence package is intentional, interpretable, and aligned with regulatory expectations.

For many rare diseases, evidence generation starts before a therapeutic trial begins. Sponsors first need to understand what happens to patients in the absence of intervention. The FDA describes natural history studies as studies that collect information about a disease in the absence of intervention, from onset through resolution or death.⁴ The agency also states that natural history information is usually unavailable or incomplete for most rare diseases and is therefore particularly needed in rare diseases.

Natural history evidence can help define the basic contours of a disease: how symptoms emerge, how quickly progression occurs, how variable the course may be, which subgroups differ, and which clinical changes are meaningful. It can also help inform endpoint selection and provide context for interpreting change in small interventional studies. Without that foundation, an observed effect in a single-arm or small controlled study may be difficult to distinguish from expected variability in disease course.

Natural history data and RWD can also help define and characterize disease progression, patient populations, novel biomarkers, genetic relationships, and treatment effects in rare diseases.⁵ This broader role is particularly important when trial populations are small and every participant contributes significantly to the interpretability of the evidence package.

The same foundation can support the use of external controls. An externally controlled trial is one in which outcomes in participants receiving the test treatment according to a protocol are compared with outcomes in a group of people external to the trial who did not receive the same treatment.⁶ The external control group may be historical or concurrent but in another setting.

External controls can be especially relevant when it is difficult to enroll enough patients into a randomized controlled trial, when a no-treatment or placebo arm raises ethical concerns, or when the disease is serious and there is high unmet need. A review of FDA approval decisions from 2000 to 2019 identified 45 approvals in which the FDA accepted external control data in the benefit–risk assessment, with reasons including disease rarity, ethical concerns about placebo or no-treatment arms, seriousness of the condition, and high unmet medical need.⁷

The use of external controls changes the nature of the evidentiary burden. It places significant weight on whether the external data are comparable to the treated population, whether outcomes were measured consistently, whether baseline differences can be accounted for, and whether missing data, bias, or measurement error could distort interpretation. In the same review, retrospective natural history data, including retrospective patient-record reviews, were the most common external-control source, accounting for 44% of the approvals in which the FDA accepted external control data.⁷ That finding reinforces the importance of building high-quality natural history resources before they are needed for regulatory interpretation.

Real-world evidence (RWE) is also moving from concept to regulatory application in rare diseases, although its use remains selective and context dependent. A systematic review of FDA approvals from January 2017 to October 2022 identified 20 non-oncologic rare disease applications with orphan drug designation that used RWD to support efficacy outcomes.⁸ In that review, 19 of the 20 applications used only retrospective RWD, and one used both retrospective and prospective RWD. The RWD study types included natural history– or registry-based retrospective historical controls, retrospective medical chart reviews, and external RWD controls from other studies.

The same review underscores why RWE must be handled carefully. The FDA generally accepts RWD studies that demonstrated large effect size, but the agency also raised concerns about data quality and comparability, including baseline differences, missing information, bias, and measurement error.⁸ Those concerns are not secondary details. They determine whether RWD can clarify treatment effect or introduce uncertainty.

Registries can help address some of these needs when they are designed with regulatory use in mind. A registry is an organized system that collects clinical and other data in a standardized format for a population defined by a disease, condition, or drug exposure. A 2024 workshop on natural history studies and registries in rare disease treatment development addressed registry and natural history data that are fit for regulatory purposes and the use of these resources to inform regulatory decision-making.⁹

The lesson for sponsors is direct. Natural history studies, registries, RWD, and external controls are not interchangeable labels for supplemental evidence. Each must be designed, curated, and analyzed in relation to the specific regulatory question it is meant to answer. When traditional controls are infeasible, the quality of the comparator data becomes central to whether the evidence package can support interpretation.

Brineura and the Scrutiny of Natural History Comparisons

Brineura offers a useful example of how natural history data can become part of rare disease regulatory interpretation while still receiving significant scrutiny. The FDA’s statistical review for Brineura describes a phase I/II, first-in-human, single-arm, open-label study and a treatment extension study, as well as the use of a natural history cohort from the DEM-CHILD database.¹⁰

The review also describes the dual purpose of the submission: evaluating Brineura and assessing the adequacy of the CLN2 rating scale and the comparability between the natural history cohort and treatment study. That point is important because it shows that the comparator was not simply accepted as background context. The FDA evaluated whether the outcome measure and the external comparison could support interpretation.

The FDA’s review described challenges involving external control data, including rating-scale comparability and data and analysis quality, while ultimately finding the overall data and analysis quality acceptable with documented limitations.¹⁰ The case illustrates the larger rare disease evidence principle: natural history comparisons can be critical, but their usefulness depends on whether the data, measures, and analyses are strong enough to bear regulatory weight.

Rare disease evidence challenges are not limited to trial size or comparator selection. In many programs, one of the most difficult questions is what to measure. A disease may progress slowly, symptoms may fluctuate, functional changes may be difficult to quantify, and the most meaningful outcomes for patients may not align easily with conventional clinical endpoints. When the eligible population is very small, an endpoint that is noisy, poorly characterized, or weakly connected to clinical benefit can make the entire development program harder to interpret.

The FDA’s Rare Disease Endpoint Advancement (RDEA) Pilot Program reflects the importance of this issue. The program supports novel efficacy endpoint development for drugs that treat rare diseases and provides a mechanism for sponsors to collaborate with FDA throughout endpoint development.¹¹ That focus is significant because endpoint development is not merely a statistical or operational detail in rare disease studies. It can determine whether a feasible trial can produce evidence that regulators can evaluate.

At the same time, the FDA cautions that RDEA advice should not be equated with a special protocol assessment or with agreement that a proposed approach will be sufficient to support approval.¹¹ This caution reinforces the distinction between regulatory engagement and regulatory assurance. Sponsors may receive input on endpoint development, but they remain responsible for demonstrating that the endpoint is meaningful, measurable, and appropriate for the disease context.

Surrogate endpoints raise a related set of questions. The FDA’s Accelerated Approval Program allows earlier approval of drugs for serious conditions that fill an unmet medical need based on a surrogate endpoint.¹² The FDA defines a surrogate endpoint as a marker thought to predict clinical benefit but not itself a measure of clinical benefit. For some rare diseases, such endpoints may offer a way to support earlier regulatory decisions when waiting for direct clinical outcomes would be difficult, slow, or impractical.

Accelerated approval, however, does not end the evidence-generation process. The FDA states that companies must conduct studies to confirm anticipated clinical benefit after accelerated approval; if confirmatory studies verify clinical benefit, FDA grants traditional approval, and if they do not, FDA has procedures that could lead to removal from the market.¹²

Surrogate endpoint acceptability is also highly context dependent. According to the FDA, acceptability is determined case by case and depends in part on the disease, patient population, mechanism of action, and available treatments.¹³ The agency also states that its surrogate endpoint table is not a replacement for discussions with the appropriate Center for Drug Evaluation and Research (CDER) or Center for Biologics Evaluation and Research (CBER) review division.

That context dependence should shape how sponsors think about endpoint strategy. A biomarker or functional measure that is compelling in one disease may be insufficient in another. A surrogate marker that aligns closely with disease biology and mechanism of action may still require evidence that it is reasonably likely to predict clinical benefit. In rare diseases, endpoint strategy must therefore be developed as part of the scientific and regulatory foundation of the program, not added late to a trial design that has already been fixed.

Duchenne Muscular Dystrophy and Surrogate Endpoint-Based Accelerated Approval

Duchenne muscular dystrophy (DMD) provides examples of surrogate endpoint-based accelerated approval in a serious rare disease with unmet need. The FDA’s accelerated approval of Exondys 51 was based on increased dystrophin in skeletal muscle as a surrogate endpoint, with the FDA concluding that the increase was reasonably likely to predict clinical benefit in some patients with DMD amenable to exon 51 skipping.¹⁴

The FDA also granted accelerated approval to Vyondys 53 for patients with confirmed DMD mutations amenable to exon 53 skipping.¹⁵ These examples are best understood within the accelerated approval framework: a surrogate endpoint can support earlier approval when it is reasonably likely to predict clinical benefit, but confirmatory evidence remains necessary to verify the anticipated benefit.¹²

Rare disease development increasingly depends on the totality of evidence. That phrase can be overused, but in this context it has practical meaning. A small clinical study may not be interpretable without natural history data. A natural history comparison may not be persuasive without comparable endpoints and standardized data collection. A surrogate endpoint may not be meaningful without mechanistic plausibility and regulatory agreement on its relevance. RWD may support efficacy interpretation in some applications, but only if the underlying data are reliable, relevant, and sufficiently comparable.

The FDA’s guidance on substantial evidence provides a useful broader framework. The agency states that its evidentiary standard has not changed since 1998 while explaining how one adequate and well-controlled clinical investigation plus confirmatory evidence can meet the substantial evidence requirement.¹⁶ This guidance is not rare disease-specific, but it is relevant because it reinforces that evidence strength can come from the relationship among multiple sources of data, not only from repeating one conventional trial format.

For rare diseases, this mindset changes how development programs should be designed. Sponsors need to determine early which data streams will be essential to interpretation. That may include untreated disease-course data, prospective registry information, external comparator data, endpoint-development evidence, biomarker data, pharmacodynamic measures, or confirmatory clinical evidence. The specific combination will vary by disease and product, but the logic should be coherent before pivotal decisions are made.

A totality-of-evidence approach also requires discipline. Combining multiple weak data sources does not automatically produce a strong evidence package. Each component must address a defined uncertainty. Natural history data may help explain expected progression. External controls may help contextualize outcomes when randomization is infeasible. RWD may supplement or contextualize trial findings. A surrogate endpoint may support earlier approval when it is reasonably likely to predict clinical benefit. Confirmatory evidence may strengthen the case that observed effects are meaningful and durable.

This is where rare disease development differs most sharply from a checklist-driven approach to clinical evidence. The strength of the package depends not only on the presence of multiple data types but on whether they converge on the same conclusion, whether their limitations are understood, and whether the remaining uncertainty is acceptable in light of the disease, available therapies, and unmet need.

Rare disease evidence strategy needs to begin early because many of the most important decisions cannot be corrected late in development. If natural history data are incomplete, a sponsor may not be able to build a credible external comparator when it becomes clear that a randomized control arm is infeasible. If an endpoint has not been characterized, a small pivotal study may produce results that are difficult to interpret. If RWD sources are inconsistent or incomplete, they may create more uncertainty than clarity.

Sponsors should therefore treat evidence generation as an integrated development strategy. The first question is not simply how to design the pivotal trial. It is what evidence will be needed to make the pivotal trial interpretable. That includes understanding the untreated disease course, identifying clinically meaningful endpoints, evaluating whether external controls or RWD could support interpretation, determining whether a surrogate endpoint may be appropriate, and planning for confirmatory evidence if accelerated approval is pursued.

Regulatory engagement is central to that process. The FDA’s RDEA Pilot offers a mechanism for collaboration throughout rare disease endpoint development, but the FDA also cautions that such advice does not guarantee that an endpoint strategy will be sufficient for approval.¹¹ The agency’s surrogate endpoint resources similarly emphasize that the acceptability of a surrogate endpoint is context dependent and that general surrogate endpoint information does not replace engagement with the relevant review division.¹³

For development partners, the implications are broader than clinical trial operations. Contract research organizations (CROs), contract development and manufacturing organizations (CDMOs), analytical laboratories, specialty logistics providers, data partners, and patient-identification vendors may all influence the quality of evidence in rare disease programs. Their contributions matter most where execution affects interpretability: biomarker assay performance, sample collection consistency, endpoint measurement, patient screening, data standardization, chain of custody, documentation quality, and the reliability of longitudinal follow-up.

This is particularly important because rare disease trials have little room for avoidable noise. In a large trial, variability can sometimes be managed through scale. In a rare disease study, every missing data point, inconsistent measurement, assay issue, or poorly documented deviation can have outsized consequences. Operational quality and evidence quality are closely linked.

The same principle applies to manufacturing and analytical strategy. For therapies targeting small populations, process changes, assay comparability, release testing, stability, and product characterization can intersect with clinical evidence strategy. While the regulatory sources discussed here focus primarily on clinical evidence, rare disease development programs often depend on the ability to connect product understanding with clinical interpretation, especially when biomarkers, pharmacodynamic measures, or surrogate endpoints are central to the evidence package.

The future of rare disease development will favor sponsors that design with this integration in mind. A credible evidence package will not emerge from isolated workstreams brought together at submission. It will come from coordinated decisions about the disease model, patient population, endpoint, comparator, data source, statistical approach, regulatory pathway, and confirmatory plan.

Rare disease evidence requirements are being rethought because rare disease development makes that rethinking necessary. The goal is not to make approval easier by making evidence weaker. The goal is to make rigorous evaluation possible when the standard tools of clinical development cannot be applied in the usual way. For patients with serious rare diseases and few treatment options, that distinction matters. For sponsors, it sets a high bar for planning, execution, and scientific clarity.

1. “Rare Diseases at FDA.” U.S. Food and Drug Administration. 20 Apr. 2026.

2. “CDER/CBER Rare Disease Evidence Principles (RDEP).” U.S. Food and Drug Administration. 3 Sep. 2025.

3. “Rare Diseases: Considerations for the Development of Drugs and Biological Products.” U.S. Food and Drug Administration. 21 Dec. 2023.

4. Rare Diseases: Natural History Studies for Drug Development: Draft Guidance for Industry. U.S. Food and Drug Administration. Mar. 2019.

5. Liu, Jing, et al. “Natural History and Real-World Data in Rare Diseases: Applications, Limitations, and Future Perspectives.” The Journal of Clinical Pharmacology. 62: S38–S55 (2022).

6. “Considerations for the Design and Conduct of Externally Controlled Trials for Drug and Biological Products.” U.S. Food and Drug Administration. Feb. 2023.

7. Jahanshahi, Mahta, et al. “The Use of External Controls in FDA Regulatory Decision Making.” Therapeutic Innovation & Regulatory Science. 55: 1019–1035 (2021).

8. Vaghela, Shailja, et al. “A Systematic Review of Real-World Evidence (RWE) Supportive of New Drug and Biologic License Application Approvals in Rare Diseases.” Orphanet Journal of Rare Diseases. 19: 117 (2024).

9. “Natural History Studies and Registries in the Development of Rare Disease Treatments.” Reagan-Udall Foundation for the FDA. 13 May 2024.

10. “BLA 761052 Statistical Review: Brineura (cerliponase alfa).” U.S. Food and Drug Administration. 18 Apr. 2017.

11. “Rare Disease Endpoint Advancement Pilot Program.” U.S. Food and Drug Administration. 3 Jan. 2024.

12. “Accelerated Approval Program.” U.S. Food and Drug Administration. 15 May 2026.

13. “Table of Surrogate Endpoints That Were the Basis of Drug Approval or Licensure.” U.S. Food and Drug Administration. 29 Apr. 2026.

14. FDA Grants Accelerated Approval to First Drug for Duchenne Muscular Dystrophy. U.S. Food and Drug Administration. 19 Sep. 2016.

15. FDA Grants Accelerated Approval to First Targeted Treatment for Rare Duchenne Muscular Dystrophy Mutation. U.S. Food and Drug Administration. 12 Dec. 2019.

16. “Demonstrating Substantial Evidence of Effectiveness With One Adequate and Well-Controlled Clinical Investigation and Confirmatory Evidence.” U.S. Food and Drug Administration. Sep. 2023.