Here’s a conversation I’ve had more times than I can count. A clinical development director at a rare disease biotech — sharp, experienced, under real pressure — sits across from me and says some version of the same thing: “We know we can’t run a traditional RCT. We know regulators expect one. We don’t know how to bridge that gap.”
That gap is the entire problem of rare disease drug development in three sentences.
Rare diseases collectively affect an estimated 300 million people globally — and yet most of those conditions have no approved treatment. The development programmes that do exist are operating under constraints that make the standard pharmaceutical evidence playbook not just expensive, but often literally impossible. Patient populations measured in hundreds. Ethics committees that won’t approve placebo arms when no alternative exists. Recruitment timelines that could outlast the company’s runway. And regulatory agencies that still — quite reasonably — want credible evidence before they approve anything.
Real-world evidence isn’t a workaround for this problem. It’s the answer to it. And in 2025, synthetic control arms, natural history studies, and externally controlled trial designs have moved decisively from experimental methodology to established regulatory practice. The question is no longer whether these approaches work. The question is whether your evidence generation strategy is designed rigorously enough to survive FDA and EMA scrutiny — and that question has a very specific technical answer.
This post is the honest version of that answer.
Why Standard RCT Design Breaks Down — and Why Everyone Already Knows It
Let’s not spend too long on the problem statement, because if you’re reading this post, you’ve probably already lived it. But it’s worth being precise about exactly where the randomised controlled trial fails in rare disease contexts — because the failure mode shapes the solution.
The most obvious constraint is patient population size. When the global diagnosed population for your indication is measured in the low thousands — or lower — powering a traditional two-arm randomised trial requires recruiting a share of all eligible patients worldwide that is simply not achievable. Griggs et al. documented this structural problem comprehensively over a decade ago, and it hasn’t improved. If anything, as orphan drug designation criteria have broadened, the challenge has deepened.
Then there’s the ethics problem. Randomising patients with no treatment alternatives to placebo is increasingly unacceptable — to ethics committees, to patient advocacy organisations, and to regulators themselves. The FDA’s own rare disease framework acknowledges this explicitly. When there is no standard of care and the natural history of the disease is severe, a placebo-controlled design is often not just scientifically suboptimal — it’s ethically indefensible.
And then there’s the problem nobody talks about as much as they should: natural history variability. Many rare diseases have highly heterogeneous progression — the same condition manifesting very differently across patients. Without adequate baseline characterisation of that natural history, designing a trial with credible endpoints and sufficient statistical power is genuinely difficult. You can’t know what a clinically meaningful treatment effect looks like if you don’t know what the untreated course of disease looks like.
The FDA’s orphan drug programme, breakthrough therapy designation, and accelerated approval pathways all exist precisely because regulators understand these constraints. They have, over decades, built flexibility into the regulatory system for rare disease contexts. The question is how to use that flexibility rigorously — which is where most sponsors run into trouble.
The Real-World Evidence Toolkit: What It Actually Includes
There are three primary RWE approaches that matter in rare disease development. They’re often discussed as alternatives to each other. They’re better understood as complementary tools that serve different evidentiary functions — and the most successful rare disease programmes use all three strategically.
Synthetic Control Arms
A synthetic control arm is a comparator cohort constructed from real-world data — typically a combination of patient registries, claims data, and EHR records — that is matched to the trial’s treatment arm using statistical methods designed to approximate the balance you’d achieve through randomisation. Propensity score matching, inverse probability weighting, and covariate adjustment are the methodological backbone. The output is a comparator group that didn’t undergo randomisation but is designed to be as comparable as possible to the treated patients.
The regulatory track record for synthetic control arms is now substantial enough to take seriously. Thorlund et al. provided the foundational primer on the methodology, and the approval of treatments across oncology and rare metabolic diseases using external control designs has since created a meaningful body of precedent. A 2022 Nature Communications analysis of Pralsetinib in non-small cell lung cancer demonstrated how real-world synthetic control arms can address comparative effectiveness questions in contexts where randomisation isn’t viable — and that methodology has direct application to rare disease settings.
But here’s what the published literature doesn’t quite capture: the documentation burden. The methodological standards for a regulatory-grade synthetic control arm submission are considerably more extensive than most internal teams anticipate. Data source justification, covariate balance reporting, sensitivity analyses across alternative matching specifications, bias assessment — the regulatory expectation is comprehensive transparency. Getting the methodology right and getting the documentation right are two separate challenges. For more on synthetic patient cohort construction methodology, our earlier post on synthetic patient cohorts for clinical trial design covers the foundational concepts in detail.
Natural History Studies
If synthetic control arms are the most visible tool in the rare disease RWE toolkit, natural history studies are the most foundational — and the most consistently underinvested by sponsors who start thinking about their evidence strategy too late in development.
A natural history study characterises the progression of a disease in the absence of the treatment being developed. It establishes baseline outcome trajectories, defines the endpoints that will be clinically meaningful in a trial setting, and — crucially — provides the evidentiary backbone that regulators need to interpret a single-arm trial or external control comparison. Without a credible natural history characterisation, a single-arm trial produces results that regulators cannot confidently contextualise. With one, those same results can support a submission.
Real-world data sources are transforming how natural history studies are constructed. Patient registries — where they exist and are maintained to adequate data quality standards — remain the most valuable source. But claims data, EHR records, and linked multi-source datasets are increasingly used to construct natural history cohorts that would take decades to assemble through prospective data collection. The FDA’s 2019 guidance on observational studies using real-world data provides the methodological framework — but applying it well in rare disease contexts requires disease-specific expertise that goes considerably beyond the guidance document itself.
One honest caution: regulatory scrutiny of natural history study methodology has intensified since the early precedent-setting rare disease approvals. Approaches that sailed through review five years ago are facing harder questions today. The bar is rising, and study designs need to account for that.
External Control Arms
The broader category encompassing both synthetic controls and historical controls — and the one that has the most clearly articulated regulatory framework thanks to the FDA’s 2023 guidance on externally controlled trials. This document is essential reading for any rare disease clinical development or regulatory affairs team. It sets out what the FDA expects in terms of data source quality, matching methodology, bias assessment, and submission documentation — and it creates a clearer pathway than existed even three years ago.
The distinction between external controls as formal comparators versus real-world benchmarks matters here. A synthetic control arm designed to serve as the primary comparator in a pivotal submission faces a higher methodological bar than one used as a supportive benchmark alongside a single-arm trial. Understanding exactly what regulatory function your external control is meant to serve before designing the study is the difference between a submission that works and one that gets a complete response letter.
The Regulatory Landscape: What FDA and EMA Will Actually Accept
This is the section regulatory affairs directors will read most carefully — and rightly so. Because the regulatory landscape for RWE in rare disease is genuinely more nuanced than most summaries suggest. It’s also moving faster than many internal teams are tracking.
On the FDA side, the evolution has been substantial. The 2018 RWE Framework, the 2021 RWD/RWE guidance, and the 2023 externally controlled trials guidance form a coherent framework that, taken together, is genuinely permissive of well-designed external control approaches in rare disease contexts. A recent analysis by Innes et al. documented the real-world use of RWE in new drug and biologics license application approvals — and the trend is clear. The FDA is approving drugs based on real-world evidence, with increasing confidence in the methodology when the submissions are done properly.
Breakthrough therapy designation and accelerated approval interact with RWE strategy in ways that aren’t always well understood internally. Accelerated approval, in particular, creates a two-stage evidentiary burden: demonstrating effect on a surrogate endpoint at approval, then confirming clinical benefit through post-market studies. How those post-market study requirements are structured — and whether RWE can satisfy them — varies by indication and is subject to negotiation with FDA. Getting that negotiation right requires regulatory strategy expertise that most rare disease sponsors are not carrying internally.
On the EMA side, the parallel framework is equally substantive. The EMA’s rare disease regulatory framework and the 2025 EMA Reflection Paper on real-world data in non-interventional studies set out the European standard. PRIME designation and adaptive pathways create additional flexibility for rare disease sponsors who engage early with EMA scientific advice. The practical challenge for sponsors pursuing simultaneous global approvals is that FDA and EMA expectations don’t perfectly align — and designing a study that satisfies both frameworks simultaneously requires navigating that divergence deliberately.
The international consensus framework sits in the CIOMS Working Group XIII report on RWE in regulatory decision-making — published in 2025 and now the authoritative synthesis of where international regulators have landed on RWE methodology standards. If you haven’t read it, you should.
Honest assessment: regulatory acceptance for synthetic control arms is established in oncology and a growing number of metabolic rare diseases. It’s still evolving for conditions where the natural history is less well characterised and where the data infrastructure is thinner. The methodology is accepted in principle; the question is whether any specific submission meets the methodological bar that regulators will apply to that specific disease and data context.
Data Sources: What’s Available and What’s Actually Fit for Purpose
The rare disease data landscape is fundamentally different from the broader real-world data market — and treating it as equivalent is one of the most common and costly mistakes sponsors make.
Patient registries are the most valuable data source for many rare diseases, but they come with serious caveats. Many are maintained by patient advocacy organisations without the data quality standards required for regulatory submissions. Many are geographically fragmented — a European registry and a US registry for the same condition, with incompatible data structures and different data governance frameworks. And many simply don’t exist yet for conditions where organised patient communities haven’t yet emerged. RECORD statement compliance for registry-based studies is the minimum standard — but applying it well requires understanding exactly what data elements are present, how they were collected, and where the gaps are.
Claims data presents a different problem in rare disease contexts. Standard health insurance claims databases offer broad population coverage — but when the diagnosed patient population for your condition numbers in the hundreds globally, standard insurance databases contain too few patients to construct an adequately matched external control cohort. This doesn’t make claims data useless; it makes it a supporting source rather than a primary one for ultra-rare conditions. For conditions at the rarer end of the orphan spectrum, you’re often dealing with tens to low hundreds of identifiable patients in even the largest national claims databases. That reality needs to inform study design from the beginning, not as a surprise at the analysis stage.
EHR data is rich in clinical detail but fragmented. The rare disease specialist centres that treat your condition — often ten to twenty academic medical centres globally — each hold a piece of the patient picture. Accessing and linking that data requires negotiating separate data use agreements with each institution, navigating different national data protection frameworks, and resolving the interoperability challenges that make linked EHR data genuinely difficult to work with at scale. International data collaboration is often necessary; it is never simple. Across all rare disease data sources, phenotype validation in claims data is especially critical — misclassification in small patient populations has a disproportionate impact on analytical validity that larger dataset sizes cannot correct for.
Synthetic data augmentation is increasingly being considered to address the patient volume constraint — and for good reason. Murphy et al. (2025) demonstrated how synthetic data generation can accelerate rare disease research while preserving privacy. Perkonoja et al. (2025) provided the systematic review of longitudinal synthetic data methods that is now the reference for anyone evaluating generation methodology. The critical question isn’t whether synthetic data can be generated — it can. The question is whether the generation methodology is robust enough for the specific regulatory purpose the data is meant to serve. That assessment requires expertise that most internal teams, honestly, don’t have. For a deeper look at the broader clinical and preclinical dataset landscape, including historical borrowing and Bayesian approaches, see our Clinical and Preclinical Datasets pillar resource.
Building the Evidence Generation Strategy: A Stage-by-Stage Framework
The rare disease evidence generation strategy needs to be built from the evidence question backwards — not from the data source forwards. Here’s what that looks like at each development stage.
Pre-IND and Early Development
The foundational work at this stage is disease characterisation. What is the natural history of the condition? What is the diagnosed prevalence across the markets you intend to pursue approval in? What endpoints are clinically meaningful — and what does real-world data tell you about how those endpoints manifest in untreated patients?
This is where natural history data and real-world prevalence analysis do their most important work. Getting this characterisation right doesn’t just inform trial design — it shapes the regulatory strategy, the patient population sizing for the commercial model, and the HTA evidence requirements that market access teams will need to address years later. For pharma and biotech teams at this stage, our resource on diagnosed prevalence and patient data solutions covers the analytical approaches in detail.
Phase II and Proof of Concept
Single-arm trials with pre-specified external controls are the workhorse of rare disease Phase II. The external control cohort — whether constructed from natural history data, matched real-world claims, or registry records — needs to be defined and validated before the trial reads out, not retrofitted afterwards. Regulators are rightly sceptical of external control designs that appear to have been constructed specifically to make the treatment look good. Pre-specification and transparency are the methodological standards that distinguish credible from convenient.
Pivotal Programme
This is where the synthetic control arm or externally controlled trial design needs to meet the full FDA and EMA methodological standard — including propensity matching, covariate balance documentation, sensitivity analyses, and bias assessment. Selukar et al. (2025) provide the current state of the art for synthetic control arm methodology in haematological malignancies — which is instructive for rare disease programmes more broadly in terms of what the methodological expectations look like in practice. The Viele et al. framework for historical control data remains foundational for understanding the statistical properties of external comparator designs.
Post-Approval and Market Access
Post-market evidence commitments triggered by accelerated approval or conditional marketing authorisation are, in my experience, consistently underplanned. Sponsors focus on getting to approval and don’t fully model the evidence generation obligations that approval conditions create. And then there’s the payer evidence challenge: HTA bodies across Continental Europe are increasingly demanding real-world effectiveness evidence for orphan drugs even in conditions with no treatment alternatives. The humanitarian urgency argument is facing real resistance from payers managing constrained budgets. The real-world data generated during development can — with the right design choices — do double duty as the foundation for HEOR modelling and payer submissions. For teams at this stage, our resource on HEOR simulation and cost modelling with aggregated insurance data is directly relevant.
Why Expert Advisory Support Is Not Optional in Rare Disease RWE
I’m going to be direct about this, because I think the advisory value proposition is sometimes stated too gently in content like this.
The rare disease RWE landscape is more specialised, more fragmented, and more rapidly evolving than the general real-world data market. Navigating it without expert guidance creates specific, quantifiable risks. A failed external control arm submission — one that gets a complete response letter because the data source didn’t meet regulatory standards or the covariate matching methodology wasn’t documented adequately — doesn’t just cost the direct investment in the study. It costs one to three years of development timeline at the point of maximum capital exposure. For a rare disease biotech with limited runway, that can be an existential outcome.
The methodological expertise required for a regulatory-grade synthetic control arm — propensity matching, covariate selection informed by disease biology, sensitivity analysis design, bias assessment documentation — is genuinely specialist. Most internal biostatistics teams at rare disease companies have not designed this type of submission before. That’s not a criticism; it’s a structural reality of working in an area where every programme is rare by definition.
The ISPOR-ISPE Good Practices framework sets out the methodological standards for real-world evidence studies — but applying those standards to a specific rare disease context, with specific data source options and specific regulatory expectations, requires deep familiarity with both the methodology and the landscape that most companies are not carrying internally.
And the vendor landscape for rare disease data is genuinely complex. Multiple providers claim rare disease data solutions. Those claims vary enormously in what they actually deliver — in terms of coverage for specific conditions, data quality standards, and the methodological rigour required to use the data in a regulatory submission. Understanding which data sources match which specific use cases, for which regulatory purposes, in which therapeutic areas — that requires market knowledge that goes beyond what any single company can reasonably develop from first principles.
Getting Started: The Four Questions to Answer Before Any Data Decision
The single most valuable thing an expert advisor can do for a rare disease sponsor at the beginning of an evidence generation engagement is help them answer four questions before any data partnership or study protocol decision is made. In our experience, most teams skip directly to question three — and that’s where expensive misalignments happen.
Question one: What is the evidence question? Is this a natural history characterisation for regulatory submission? A synthetic control arm for a pivotal trial? A post-approval effectiveness study for HTA purposes? The answer determines everything downstream — data sources, study design, methodological standards, documentation requirements.
Question two: What does the regulatory bar actually look like? Not in the abstract, but for your specific indication, your specific development stage, and the specific agencies you’ll be submitting to. The FDA’s Advancing Real-World Evidence Program has generated substantial clarity on what acceptable looks like — but that clarity needs to be interpreted for your specific context. The 2024 FDA guidance on electronic health records and medical claims data is the most recent statement of the standard.
Question three: What data infrastructure actually exists for this condition? Not what vendors claim is available — what has been validated, at what quality standard, and for what methodological purposes. That assessment requires independent expertise.
Question four: What combination of data sources, study design, and analytical methodology creates a submission-ready evidence package? The answer is almost never a single data source. It’s almost always a considered combination — and getting the combination right before committing to any single partnership is where expert advisory support compresses timelines and prevents waste.
This is exactly the conversation that a MEDDDICAL advisory engagement opens with. Not which vendor to use — that comes later, once the evidence question and the regulatory bar are clearly defined. Starting with requirements definition rather than vendor selection is the structural difference between evidence strategies that work and ones that don’t.
If your rare disease programme is at a decision point on evidence generation strategy — whether you’re at pre-IND, approaching a pivotal design decision, or managing post-approval commitments — reach out to discuss your specific context. The conversation that defines requirements before any commitment is made is the most valuable one we have.
Sources & References
- Haendel, M., et al. (2020). How many rare diseases are there? Nature Reviews Drug Discovery, 19(2), 77–78. https://www.nature.com/articles/d41573-019-00180-y
- Kaufmann, P., et al. (2010). Rare diseases: a global look at efforts and opportunities. Health Research Policy and Systems, 8(1), 1–5. https://health-policy-systems.biomedcentral.com/articles/10.1186/1478-4505-8-22
- Griggs, R. C., et al. (2009). Clinical research for rare disease: opportunities, challenges, and solutions. Molecular Genetics and Metabolism, 96(1), 20–26. https://www.sciencedirect.com/science/article/pii/S1096719208003272
- Kesselheim, A. S., & Doherty, M. (2021). Real-world evidence in drug regulation — challenges and opportunities. Health Affairs, 40(2). https://www.healthaffairs.org/doi/10.1377/hlthaff.2020.01635
- U.S. Food and Drug Administration. Developing products for rare diseases and conditions. https://www.fda.gov/patients/rare-diseases-fda
- U.S. Food and Drug Administration. Orphan Drug Act — relevant excerpts and implementation. https://www.fda.gov/industry/developing-products-rare-diseases-conditions/orphan-drug-act-relevant-excerpts
- Thorlund, K., et al. (2020). Synthetic and external controls in clinical trials — a primer for researchers. Clinical Epidemiology. https://pubmed.ncbi.nlm.nih.gov/32440224/
- Selukar, S., et al. (2025). Synthetic control arms and other uses of external data in clinical trials for hematological malignancies. Blood Reviews. https://pubmed.ncbi.nlm.nih.gov/40753019/
- Goel, S., et al. (2022). Addressing challenges with real-world synthetic control arms to demonstrate the comparative effectiveness of Pralsetinib in non-small cell lung cancer. Nature Communications. https://www.nature.com/articles/s41467-022-30908-1
- U.S. Food and Drug Administration (2019). Considerations for the design, conduct, and analysis of observational studies using real-world data. https://www.fda.gov
- U.S. Food and Drug Administration (2023). Considerations for the design and conduct of externally controlled trials for drug and biological products. https://www.fda.gov/media/164960/download
- Viele, K., et al. (2014). Use of historical control data for assessing treatment effects in clinical trials. Pharmaceutical Statistics, 13(1), 41–54. https://onlinelibrary.wiley.com/doi/10.1002/pst.1589
- U.S. Food and Drug Administration (2018). Framework for FDA’s Real-World Evidence Program. https://www.fda.gov/media/120060/download
- U.S. Food and Drug Administration (2021). Considerations for the use of real-world data and real-world evidence to support regulatory decision-making. https://www.fda.gov
- Innes, et al. (2025). Real-world evidence in new drug and biologics license application approvals during fiscal years 2020–2022. Clinical Pharmacology & Therapeutics. https://ascpt.onlinelibrary.wiley.com/doi/10.1002/cpt.3688
- European Medicines Agency. Rare diseases: overview and regulatory framework. https://www.ema.europa.eu/en/human-regulatory-overview/research-development/medicines-rare-diseases
- European Medicines Agency (2025). Reflection paper on the use of real-world data in non-interventional studies to generate real-world evidence. https://www.ema.europa.eu
- Hennessy, et al. (2025). Real-world data and real-world evidence in regulatory decision making: report summary from the CIOMS Working Group XIII. Pharmacoepidemiology and Drug Safety. https://onlinelibrary.wiley.com/doi/10.1002/pds.70117
- Hey, S. P., & Kimmelman, J. (2021). The use of external controls in FDA regulatory decision making. Clinical Pharmacology & Therapeutics. https://pmc.ncbi.nlm.nih.gov/articles/PMC8332598/
- Murphy, M. L., et al. (2025). Synthetic data generation: a privacy-preserving approach to accelerate rare disease research. Frontiers in Digital Health. https://www.frontiersin.org
- Perkonoja, K., et al. (2025). Methods for generating and evaluating synthetic longitudinal patient data: a systematic review. Journal of Healthcare Informatics Research. https://link.springer.com/article/10.1007/s41666-025-00223-7
- Jordon, J., et al. (2021). Synthetic data in machine learning for medicine and healthcare. Nature Biomedical Engineering, 5, 493–497. https://www.nature.com/articles/s41551-021-00751-8
- Benchimol, E. I., et al. (2015). The RECORD statement. PLOS Medicine, 12(10). https://journals.plos.org/plosmedicine
- Dagenais, S., et al. (2022). Use of real-world evidence to drive drug development strategy and inform clinical trial design. Clinical Pharmacology & Therapeutics, 111(1), 77–89. https://pubmed.ncbi.nlm.nih.gov/34839524/
- Makady, A., et al. (2017). Policies for use of real-world data in health technology assessment. Value in Health, 20(3), 520–532. https://www.valueinhealthjournal.com
- ISPOR-ISPE Special Task Force. Good practices for real-world data studies of treatment and/or comparative effectiveness. https://www.ispor.org
- Dreyer, N. A., et al. (2018). GRACE principles: recognizing high-quality observational studies of comparative effectiveness. American Journal of Managed Care, 16(6), 467–471. https://www.ajmc.com
- Smith, et al. (2025). US Food and Drug Administration’s Advancing Real-World Evidence Program: initial experience. Clinical Pharmacology & Therapeutics. https://ascpt.onlinelibrary.wiley.com/doi/10.1002/cpt.3700
- ISPOR (2022). Real-world evidence transparency initiative: guiding principles. https://www.ispor.org/strategic-initiatives/real-world-evidence/real-world-evidence-transparency-initiative
- U.S. Food and Drug Administration (2024). Real-world data: assessing electronic health records and medical claims data to support regulatory decision-making. https://www.fda.gov
