FDA Issues Revised Draft Guidance on Demonstrating Substantial Evidence of Effectiveness for Human Drugs and Biological Products

On June 22, 2026, the U.S. Department of Health and Human Services (HHS) unveiled Operation TrialBlazer, a department-wide effort to accelerate clinical research and development centered in the United States.¹ As part of that initiative, the U.S. Food and Drug Administration (FDA or the Agency) is taking multi-pronged actions to help facilitate early- and late-stage clinical development, including:

1. proposing a pilot program intended to reduce the start-up time for first-in-human (FIH) clinical trials (see Arnold & Porter’s Advisory);
2. clarifying the chemistry, manufacturing, and controls (CMC) expectations for Investigational New Drug (IND) submissions to help sponsors efficiently generate and submit the phase-appropriate data needed to support phase 1 clinical trials; and
3. issuing a significantly revised draft of FDA’s critical Substantial Evidence Guidance, which sets out regulatory expectations for sponsors regarding the type and quantity of data and information necessary to meet the statutory standard for “substantial evidence” of effectiveness in support of a drug or biological product application, as discussed herein.²

The HHS Office of Inspector General is also evaluating whether to update the safe harbor regulations under the federal Anti-Kickback Statute, or the exceptions to the civil monetary penalty provision prohibiting inducements to beneficiaries to address remuneration provided to individuals in connection with their participation in clinical trials.³

FDA’s revised draft guidance, Demonstrating Substantial Evidence of Effectiveness for Human Drug and Biological Products (the “2026 Draft Substantial Evidence Guidance”), substantially revises FDA’s December 2019 draft guidance of the same name, which itself was an update to FDA’s 1998 guidance on the same topic.⁴ The 2026 Draft Substantial Evidence Guidance reframes how sponsors should plan to meet the “substantial evidence” standard for demonstrating effectiveness.⁵ In particular, it shifts focus from the previous default posture of two “adequate and well-controlled investigations”⁶ onto what had been considered the exception to that general requirement: a single adequate and well-controlled clinical investigation with confirmatory evidence.⁷

FDA’s rewrite is intended to “clarify circumstances in which drug developers may be able to rely on data from one adequate and well-controlled pivotal clinical investigation with confirmatory evidence, to demonstrate substantial evidence of effectiveness for drug approval.”⁸ In so doing, the 2026 Draft Substantial Evidence Guidance appears to: 

• Broaden the circumstances in which a single pivotal trial strategy with confirmatory evidence would be expected to meet the standard for substantial evidence of effectiveness in support of a marketing application — and thus potentially shift the burden onto FDA to justify requiring a second trial;
• Clarify that confirmatory evidence can come from several sources, including related adequate and well-controlled trial data, data supporting a related indication for the same drug, evidence from other approved drugs in the same pharmacologic class, mechanistic and biological information, early-phase clinical data, natural history or registry data, and other external information;
• Explain that confirmatory evidence must be strong enough — and from a reliable enough source — to yield the confirmatory evidence that a second adequate and well-controlled trial otherwise would have done; 
• Focus FDA’s confirmatory evidence analysis on the strength of the evidence as determined by trial design, conduct, prespecified analysis, endpoint selection, results, data missingness, consistency across endpoints and subgroups, and the overall development program; and
• Describe FDA’s revised policy regarding flexibility in statistical analysis, including recognition that a p-value greater than a one-sided 0.025 threshold may be acceptable in some circumstances, while also warning that a one-sided 0.025 threshold may be insufficient where the pretrial probability of effectiveness is low.

The 2026 Draft Substantial Evidence Guidance also signals a preference for clinical endpoints when possible, stating that “[u]se of a clinical endpoint is preferred when feasible”; use of a surrogate endpoint is relegated to “[a]n alternative approach.”⁹ If maintained in final guidance, this could signal a more general shift in FDA’s thinking on accelerated approval.

Background: The Substantial Evidence of Effectiveness Standard

As detailed in Arnold & Porter’s prior Advisory, section 505(d) of the Federal Food, Drug, and Cosmetic Act (FD&C Act) requires “substantial evidence” of effectiveness for approval of a drug. The statutory definition refers to adequate and well-controlled investigations by qualified experts. It provides that, if FDA determines based on relevant science that “data from one adequate and well-controlled clinical investigation and confirmatory evidence” are sufficient to establish effectiveness, FDA may consider that data and evidence to constitute substantial evidence.¹⁰

FDA’s implementing regulations explain that investigations should be adequate and well-controlled to be able to distinguish the effect of a drug from other influences, such as spontaneous change in the disease, placebo effect, or biased observation.¹¹ The regulations also identify study design features that ordinarily will be considered adequate and well-controlled: a clear protocol, appropriate controls, methods of patient selection and assignment that minimize bias, methods to minimize subject and observer bias, well-defined and reliable response measures, and adequate analytical methods.¹²

The 2026 Draft Substantial Evidence Guidance does not purport to alter the statutory or regulatory standard. But it is significant insofar as it describes how, when finalized, FDA intends to interpret and apply the statutory standard in a way that differs from historical practice (even if a large swath of drugs already have been approved in reliance on a single trial with confirmatory evidence). This is particularly telling when read against FDA’s broader, recent theme of emphasizing the quality, relevance, and biological coherence of the evidence package rather than insisting on multiple pivotal trials in every application.¹³

A Single Trial with Confirmatory Evidence as the New Default

The 2026 Draft Substantial Evidence Guidance puts the FD&C Act’s single-trial-with-confirmatory-evidence language at the center of development planning. According to the 2026 Draft Substantial Evidence Guidance, FDA will consider the design, conduct, analysis, and persuasiveness of the single trial; the source and strength of the confirmatory evidence; disease-specific considerations such as seriousness, unmet need, and prevalence; and whether more than one adequate and well-controlled trial would be ethical and practicable.  Thus, sponsors should be prepared to explain, before initiating the single pivotal trial, why the proposed clinical investigation is adequate and well-controlled and why the proposed confirmatory evidence is sufficient.  Additional trials may be required when a single trial is not sufficiently representative, or where more evidence is needed to support the safety or benefit-risk calculus — though FDA appears poised to consider those situations the exception rather than the rule, and the draft guidance notes that in some situations a convincingly positive, well-conducted trial showing a substantial decrease in mortality may make a second trial impractical or unethical.¹⁴

FDA recommends sponsors discuss their proposed approach to demonstrating substantial evidence early in development, including at a pre-IND meeting and no later than at the end of phase 2. Sponsors should be prepared to present the proposed trial design, the confirmatory evidence package, the legal basis for any reliance on external or third-party evidence, and the statistical rationale.¹⁵

Strength of the Single Trial

The 2026 Draft Substantial Evidence Guidance sets forth an expectation that to support approval, the single pivotal trial will be particularly persuasive, such that it can carry most of the evidentiary weight of “substantial evidence.”¹⁶ This means that the trial must be generalizable to U.S. clinical practice, reflective of a broad and representative population across multiple sites, include a control arm and supportive therapies that reflect the current standard of care, and utilize a clinically meaningful primary endpoint. FDA identifies multiple factors that affect whether the evidence can support a fair and responsible expert conclusion about effectiveness. These factors include:

• Trial design. FDA focuses on the control group, randomization, blinding, endpoint selection, eligibility criteria, site selection, representativeness, the standard of care, and whether the design is appropriate for the clinical question. The single trial must also be sufficiently powered to convincingly demonstrate an effect.  
• Trial conduct. FDA emphasizes data quality, adherence to treatment and protocol, completeness of follow-up, minimization of bias, and the effect of missing data.
• Analysis plan. FDA expects prespecification, control of type I error where applicable, appropriate estimands, sensitivity analyses, and well-justified frequentist or Bayesian methods.
• Trial results. FDA will consider statistical persuasiveness, clinical meaningfulness, magnitude of effect, uncertainty, consistency across endpoints and subgroups, and robustness to analysis assumptions.
• Overall development program. FDA will evaluate the pivotal trial in light of early-phase data, external information, dose and mechanism information, any inconsistent data, and all relevant adequate and well-controlled trials.

To ensure sufficient strength, sponsors should endeavor to also build into the single trial (and the related statistical analysis plan) prespecified supportive secondary endpoints, as well as supportive results across important subsets, high-quality conduct, comprehensive follow-up, minimal data missingness, and robustness of assumptions. Sponsors also should identify any potentially inconsistent evidence early; an adequate and well-controlled trial showing no effect, or even harm, with confidence intervals that rule out meaningful effects, could call into question positive results from other trials unless there is a clear and compelling explanation for the difference.¹⁷

Options for Confirmatory Evidence 

The 2026 Draft Substantial Evidence Guidance also lays out several options that FDA may find persuasive as confirmatory evidence to support data from a single trial: 

1. Related adequate and well-controlled trial data. FDA generally expects strong confirmatory evidence to come from related trials in related diseases or conditions or for related products. For an already-approved drug, FDA explains that a single pivotal trial may be supported by the adequate and well-controlled trials that supported approval for a different but closely related indication. FDA identifies the degree of similarity in disease pathophysiology, mechanism of action, and efficacy endpoints as critical considerations.¹⁸
2. Adequate and well-controlled trials demonstrating the effectiveness of other approved drugs in the same pharmacologic class. Here, the strength and relevance of the confirmatory evidence will depend on factors such as the similarity of the mechanism of action, whether similar endpoints were measured, the consistency of effects across the class, whether the new drug has similar effects, and the number of approved drugs in the class.¹⁹ This could be particularly beneficial for sponsors in competitive drug classes where FDA is more likely to accept existing data as persuasive confirmatory evidence.
3. Natural history and other external information. Natural history or registry data may serve as confirmatory evidence in appropriate circumstances, particularly where the clinical course without treatment is well characterized, and the treated trial result is difficult to attribute to bias or random variation. Early-phase information supporting the drug’s mechanism and dose, external information about disease pathophysiology and natural history, effects of the drug in related diseases, and effects of drugs with similar mechanisms of action can all help to support FDA’s expectation that the drug will be effective.²⁰

FDA appears to recognize that its recommendation of reliance on other approved NDAs and BLAs raises important legal questions, and the Agency notes that reliance on data concerning a different drug may raise legal and regulatory concerns.²¹ If an NDA applicant does not own or have a right of reference to the other drug’s data, reliance on FDA’s prior finding of safety and effectiveness will convert the application into a 505(b)(2) application and trigger associated patent-certification and exclusivity considerations.²² And importantly, for biological products (for which there is no pathway comparable to a 505(b)(2)), FDA reiterates its long-standing policy that a section 351(a) applicant must include all of the data and information necessary for approval in the Biologics License Application (BLA). In the alternative, FDA explains that to rely on a prior determination of safety, purity, and potency for another biological product to support approval, the applicant would need to submit a biosimilar BLA and otherwise meet the requirements to demonstrate biosimilarity.²³ (As an aside, we cannot help but note the tension between FDA’s position here — that sponsors may not rely on prior knowledge to support a BLA without a right of reference — and FDA’s draft guidance, Leveraging Prior Knowledge in the Development of Human Gene Therapy Products Incorporating Genome Editing, which appears to encourage sponsors to use relevant prior knowledge, including knowledge from previous clinically studied products, to support development of genome-editing gene therapy biological products.)²⁴ Stakeholders may wish to ask FDA to distinguish clearly between reliance on published or public scientific knowledge to support a scientific inference and legal reliance on another sponsor’s proprietary data or FDA’s prior approval finding.

The Revised Draft Guidance Offers Statistical Flexibility

The 2026 Draft Substantial Evidence Guidance also incorporates a more flexible approach to statistical analysis, explaining that the overall persuasiveness of a trial’s results may be influenced by the magnitude of the p-value, or by alternative measures such as the posterior probability of effectiveness in a Bayesian analysis, as well as by the magnitude and clinical meaningfulness of the effect on the primary endpoint. FDA also notes that a statistically significant result in a large trial may not be clinically meaningful. At the same time, even small effects may be clinically meaningful when the effect is on survival or irreversible morbidity. To that end, the 2026 Draft Substantial Evidence Guidance explains that the appropriate significance level will depend on the prior, or pretrial, probability that the drug is effective. Where the prior probability is low, a single trial using the common one-sided 0.025 significance level may not adequately limit the probability of false-positive effectiveness conclusions. Conversely, in other circumstances, including where prior knowledge or strong confirmatory evidence increases confidence in effectiveness, FDA indicates that a p-value greater than one-sided 0.025 may be acceptable as part of the totality of evidence.²⁵

Sponsors considering statistical flexibility should seek alignment with FDA early, build the rationale into the protocol and statistical analysis plan, and explain how the proposed approach satisfies the statutory function of substantial evidence. Commenters may wish to ask FDA to provide examples to illustrate when FDA would accept a less stringent or more stringent threshold and how Bayesian operating characteristics should be calibrated in single-trial-with-confirmatory-evidence programs.

Other Points of Interest

1. Animal Rule is beyond the scope of the Revised Guidance. FDA expressly excludes Animal Rule approvals from the scope of the 2026 Draft Substantial Evidence Guidance.²⁶ The Animal Rule regulations permit reliance on animal studies to establish effectiveness for certain products when human efficacy studies are not ethical or feasible. Still, FDA states that those regulations and related considerations are beyond the scope of this draft guidance.
2. Accelerated approval and surrogate endpoints. As noted above, the 2026 Draft Substantial Evidence Guidance states that the use of a clinical endpoint is preferred when feasible and that the use of a surrogate endpoint requires appropriate scientific justification linking the drug's effects on the surrogate endpoint to its effects on a relevant clinical endpoint. It also recognizes that surrogate endpoints reasonably likely to predict clinical benefit can support accelerated approval when statutory criteria are met, while validated surrogate endpoints can support traditional approval.²⁷ Sponsors should consider whether to comment on how the single-trial-with-confirmatory-evidence framework applies to accelerated approval programs, particularly where the pivotal evidence rests on a surrogate or intermediate clinical endpoint.
3. Relationship to plausible mechanism pathway. The 2026 Draft Substantial Evidence Guidance’s connection to FDA’s Considerations for the Use of the Plausible Mechanism Framework to Develop Individualized Therapies that Target Specific Genetic Conditions with Known Biological Cause Draft Guidance (“Plausible Mechanism Draft Guidance”) is also notable.²⁸ In that separate draft guidance, FDA proposed recommendations for individualized therapies targeting specific genetic conditions with known biological causes, including genome editing and RNA-based therapies for very small patient populations.²⁹ The revised substantial evidence guidance places that framework within the broader concept of regulatory flexibility, though sponsors may wish to consider commenting on how mechanistic evidence, biological plausibility, and early clinical evidence can be used outside the narrow individualized-therapy context.

Conclusion

The 2026 Draft Substantial Evidence Guidance gives direction on how FDA intends to operationalize the default use of a single trial with confirmatory evidence to meet “substantial evidence.”  It also leaves a number of open questions and room for commenters to influence the final guidance. In particular, stakeholders may wish to comment on the following topics, among others. 

• When a single adequate and well-controlled trial will be considered “highly persuasive,” including whether FDA will provide therapeutic-area-specific examples.
• The boundary between permissible scientific reliance on public or platform knowledge and legal reliance on another sponsor’s data or FDA’s prior findings, particularly for 351(a) BLAs.
• Whether FDA’s preference for clinical endpoints foreshadows disfavored use of accelerated approval, which has been an important approach to achieving patient access to new therapies for several decades.
• Examples of statistical flexibility, including when a p-value greater than one-sided 0.025 may be acceptable and when a more stringent threshold may be expected.
• Expectations for multiregional or largely ex-U.S. trials, including what FDA views as a sufficient number of U.S. patients and how sponsors can bridge representativeness gaps.

If you have any questions, would like more information, or would like to discuss submitting a comment to FDA’s 2026 Draft Substantial Evidence Guidance (by September 22, 2026), FDA’s Request for Information (RFI) on the Expedited Investigational New Drug Pilot program (by July 22, 2026), or HHS Office of Inspector General’s RFI (by August 24, 2026), please reach out to one of the authors of this Advisory or your existing Arnold & Porter contacts.

© Arnold & Porter Kaye Scholer LLP 2026 All Rights Reserved. This Advisory is intended to be a general summary of the law and does not constitute legal advice. You should consult with counsel to determine applicable legal requirements in a specific fact situation.