In 2026, the conversation around EHR-linked registries for FDA label expansion has moved well beyond fitness trackers and consumer-grade wearables. Sponsors, academic consortia, and even mid-sized biotechs are now wiring structured electronic health record (EHR) data directly into disease registries, turning routine clinical encounters into a continuously updating evidence base. The result is a faster, more defensible path to label expansion, and a regulatory environment where real-world evidence (RWE) is no longer a supplementary footnote but a central pillar of decision-making.
This shift matters because traditional label expansion has historically relied on randomized controlled trials (RCTs) that cost millions per indication, often enrolling narrow populations over multiple years. EHR-linked registries offer something different: longitudinal, deeply phenotyped data drawn from the very patients clinicians already see. When designed well, these registries can support indication-specific efficacy claims, safety labeling updates, and even comparative effectiveness rationales without spinning up a de novo trial.
Why Registries Outpaced Standalone Wearables in Regulatory Acceptance
Consumer wearables captured the public imagination, but they struggled in regulatory submissions for three reasons: data provenance was opaque, device interoperability was inconsistent, and patient adherence dropped off after the first 90 days. EHR-linked registries, by contrast, inherit data that is already captured as part of clinical care, including diagnoses, lab values, imaging reports, medication orders, and procedure codes. The information has a clinical context attached, which means reviewers can trace a specific outcome back to a documented encounter rather than a self-reported app entry.
This provenance advantage became especially visible in 2024 and 2025, when the FDA accepted EHR-derived registry data as primary supportive evidence in several supplemental biologics license applications (sBLAs). The pattern has accelerated in 2026, with reviewers increasingly comfortable with hybrid submissions where registry data fills gaps left by smaller pivotal trials.
The Architecture of an EHR-Linked Registry
Modern EHR-linked registries are not simple data dumps. They are engineered systems with several distinct layers:
- Data ingestion layer: Connects to hospital systems via FHIR APIs, HL7v2 pipelines, or vendor-specific extractors. Data is normalized into a common data model such as OMOP, PCORnet, or a sponsor-specific schema.
- Cohort definition engine: Uses computable phenotypes, often a combination of ICD-10 codes, SNOMED CT concepts, lab thresholds, and natural language processing (NLP) of clinical notes, to identify eligible patients.
- Outcome ascertainment module: Pulls structured outcomes where available (e.g., hospitalization, mortality) and uses validated NLP pipelines for unstructured outcomes like disease progression or adverse event descriptions.
- Longitudinal linkage: Connects records across visits, providers, and even health systems via deterministic or probabilistic matching, preserving patient privacy through tokenization.
- Governance and audit layer: Tracks who accessed what, when, and why, satisfying both HIPAA requirements and FDA audit expectations.
The result is a living dataset that updates as patients receive care, rather than a frozen snapshot collected at trial enrollment.
Real-World Evidence Submissions: What Reviewers Now Expect
Submitting EHR-derived RWE is not the same as submitting an RCT dossier. The FDA’s real-world evidence framework, refined through multiple draft guidances and pilot programs, has crystallized around several expectations that sponsors must meet.
1. Fitness-for-Use Documentation
Sponsors must demonstrate that each data element used in the analysis is fit for the specific regulatory question. This includes assessing missingness, accuracy, and timeliness. A registry claiming to support a cardiovascular outcomes claim, for example, needs to show that lipid panel results are captured consistently across participating sites, not just present in aggregate form.
2. Reproducible Analytics
The study protocol, statistical analysis plan, and analytic code are treated as part of the submission. Reviewers increasingly want executable notebooks or version-controlled pipelines they can re-run, not just PDF tables. This has prompted several registry sponsors to adopt open-source tooling like OHDSI’s cohort definitions and the CURE ID platform.
3. Pre-Specified Sensitivity Analyses
Because EHR data reflects routine care rather than protocolized follow-up, confounding is unavoidable. Regulators now expect pre-specified analyses using propensity score matching, target trial emulation, or active comparators to demonstrate that observed effects are not artifacts of prescribing patterns.
4. Patient Privacy and Data Governance
Beyond HIPAA, reviewers pay attention to how patient consent was obtained, whether data use agreements permit regulatory submission, and whether re-identification risks have been assessed. Sponsors who cannot answer these questions cleanly often see their submissions delayed.
Where Label Expansion Is Moving Fastest
Not every therapeutic area is equally suited to EHR-linked registries. The strongest traction in the past 18 months has been in:
- Oncology: Particularly for tumor-agnostic therapies and biomarker-defined subpopulations where trial enrollment is difficult.
- Rare disease: Where every additional confirmed responder is regulatory gold, registries can aggregate global experience efficiently.
- Chronic inflammatory conditions: Rheumatology, gastroenterology, and dermatology all benefit from long-term outcome capture in EHRs.
- Pediatric label updates: Where ethical and enrollment constraints limit traditional trials, EHR-derived evidence on off-label use is increasingly informing label revisions.
Challenges That Still Slow Things Down
Despite the momentum, EHR-linked registries are not a silver bullet. Three persistent challenges deserve attention.
Data Quality Variability
EHR systems were designed for billing and clinical workflow, not research. Lab values may be recorded in different units, diagnoses may be coded inconsistently, and notes may use idiosyncratic language. Without rigorous data curation, registries can produce misleading signals. Sponsors increasingly invest in centralized data quality teams that run monthly audits and feed results back to participating sites.
Site Engagement and Sustainability
Registries live or die on whether clinical sites continue to contribute high-quality data. Sites burdened by manual chart abstraction churn quickly. The most successful 2026-era registries embed passive data collection directly into clinical workflow, requiring abstractor effort only for complex validation cases.
Regulatory Skepticism on Subgroup Findings
Even when a registry analysis is methodologically sound, reviewers remain cautious about subgroup-driven label claims. A positive signal in a 200-patient subset of a 10,000-patient registry will not, on its own, support a label change. Sponsors are learning to either power their registries for specific subgroups prospectively or treat registry findings as supportive rather than pivotal evidence.
The Competitive Landscape: Who Is Building What
A handful of players have emerged as infrastructure providers for EHR-linked registries. Large academic consortia offer pre-built cohorts in specific disease areas. EHR vendors themselves are launching registry modules that allow sponsors to stand up custom datasets within existing hospital contracts. Specialty CROs have built hybrid offerings that combine registry operations with traditional clinical trial management, allowing sponsors to run a pivotal trial and a parallel registry under a single governance structure.
What unites these efforts is a recognition that the future of label expansion is hybrid. A small, well-designed RCT can establish efficacy, while a complementary EHR-linked registry can extend the label into populations, durations, and combinations that a trial cannot practically cover.
What This Means for Sponsors Planning 2026 Submissions
Sponsors considering an EHR-linked registry strategy should treat it as a long-term infrastructure investment, not a one-off data collection project. The most successful programs begin registry development 24 to 36 months before a planned supplemental filing, allowing time for site onboarding, data curation, and analytic validation. They also invest early in a target product profile that explicitly identifies which label claims will be registry-supported, which will be trial-supported, and which will need both.
Engagement with the FDA’s Office of Translational Sciences and relevant review divisions early in the process has become near-mandatory. Pre-submission meetings now routinely include dedicated RWE discussions, and reviewers will often suggest specific sensitivity analyses or external comparator choices before the formal submission arrives.
In short, EHR-linked registries have quietly become one of the most consequential tools in modern regulatory strategy. Wearables captured the headlines, but the real revolution is happening in the structured, governed, continuously updating datasets sitting inside hospital systems, and 2026 is shaping up to be the year that revolution moves firmly into the regulatory mainstream.
