Non-Interventional Study

A non-interventional study is a clinical study of a medicinal product in which participants are treated according to normal clinical practice, rather than being assigned to a treatment strategy by a study protocol. Under EU Regulation (EU) No 536/2014, Article 2(2)(4) defines it directly and negatively: “a non-interventional study is a clinical study other than a clinical trial.” The classification matters far beyond terminology — whether a study is “interventional” or “non-interventional” determines whether it needs prior competent-authority/ethics-committee approval under the EU Clinical Trials Regulation, whether it triggers an FDA Investigational New Drug (IND) application, and how it must be registered and reported.

This entry covers the classification criteria themselves — the operational test used to sort a proposed study into one bucket or the other. For the EU regulatory definition of the product being studied once a study is classified as an interventional clinical trial, see CASRAI’s entry on the Investigational Medicinal Product (IMP). For the broader taxonomy of observational vs. experimental study design as a research-methodology question (not a regulatory-reporting one), see Research Study Types — that entry deliberately does not cover the regulatory classification test this one does.

The operational classification test

Under the EU framework, a study is a clinical trial (and therefore not non-interventional) if it has any one of three characteristics, originally set out in Directive 2001/20/EC and carried forward into Regulation (EU) No 536/2014’s definition of “clinical trial” at Article 2(2)(2):

• The assignment of the participant to a particular therapeutic strategy is decided in advance by a trial protocol and is not part of normal clinical practice of the member state concerned;
• The decision to prescribe the medicinal product is taken together with the decision to include the participant in the study (i.e., enrollment and prescribing aren’t independent decisions); or
• Diagnostic or monitoring procedures are applied to participants that go beyond normal clinical practice.

If none of these apply — the medicine is prescribed in the ordinary way a treating clinician would prescribe it, the prescribing decision is clearly separate from the decision to enroll the patient in the study, and no protocol-driven extra tests or visits are imposed — the study is non-interventional, regardless of how rigorous its data collection or analysis otherwise is. A large, carefully designed multi-country registry study can be non-interventional; a small single-site study can still be an interventional clinical trial if it fixes the treatment assignment or adds a protocol-mandated scan.

Why the line is drawn here, not at “observational vs. experimental”

It’s tempting to treat “non-interventional” as a synonym for “observational,” and in most everyday usage they overlap heavily — non-interventional study designs are typically observational cohort, case-control, cross-sectional, or self-controlled designs. But the EU/FDA regulatory test is narrower and more mechanical than the general research-methodology distinction between observational and experimental design (covered on CASRAI’s Research Study Types entry). The regulatory test asks three specific procedural questions about how treatment assignment, prescribing, and monitoring actually happen in the study — not whether the researcher is, in some general sense, “manipulating” an exposure. A study can be non-experimental in spirit but still trip one of the three criteria (for example, by bundling a protocol-mandated extra blood draw into an otherwise routine-care cohort) and be reclassified as an interventional clinical trial as a result.

FDA draft guidance issued in 2024 (Considerations Regarding Non-Interventional Studies for Real-World Evidence) uses closely parallel language for the US context: a non-interventional study is one where patients receive the marketed drug during routine medical practice and are not assigned an intervention determined by a protocol, and such studies are not “clinical investigations” under 21 CFR 312.3 and don’t require an IND. The EU and US framings converge on the same functional test — protocol-driven assignment vs. routine practice — even though the two regimes don’t use identical statutory language.

What the classification determines

• Approval pathway. An EU clinical trial requires authorization through the Clinical Trials Information System (CTIS) under Regulation (EU) No 536/2014, including ethics committee review, before it can start. A non-interventional study falls outside the Regulation’s scope entirely and is instead governed by national law and (for products under safety monitoring) EMA/national pharmacovigilance guidance — the approval burden and timeline are materially different.
• IND requirement (US). A study that meets FDA’s clinical-investigation criteria generally requires an IND; a genuinely non-interventional study of an already-marketed drug typically does not, though sponsors should not assume this without checking the current FDA guidance against their specific design.
• Registration and reporting. Interventional clinical trials are the category subject to prospective registration and results-reporting obligations (e.g., ClinicalTrials.gov under FDAAA 801, or EU CTIS publication requirements) — see CASRAI’s Clinical Trial Registration entry. Non-interventional studies are frequently still registered (many journals and pharmacovigilance frameworks encourage or require it), but they sit outside the specific statutory clinical-trial registries and reporting rules that apply to interventional trials.

Worked examples

Example — non-interventional: A pharmaceutical company runs a multi-site prospective cohort study following patients who have already been prescribed a marketed diabetes drug by their own physicians, in the course of ordinary care. Enrollment happens after the prescribing decision, the drug and dose are whatever the treating physician already chose, and the only study-specific activity is an additional patient-reported quality-of-life questionnaire at existing routine visits. Because assignment isn’t protocol-driven, prescribing and enrollment decisions are separate, and no extra diagnostic/monitoring procedure beyond normal practice is imposed, this is non-interventional.

Counter-example — interventional despite an “observational” label: A study is designed to compare two already-approved drugs head-to-head, and the protocol randomly assigns which of the two drugs each newly enrolled patient receives, even though both are standard-of-care options a clinician might reasonably choose between. Because the treatment strategy is decided in advance by the protocol rather than left to ordinary clinical judgment, this is an interventional clinical trial — regardless of the fact that both arms use marketed, already-approved products and the study team may informally describe it as “pragmatic” or “real-world.”

Related terms

• Research Study Types — the general observational-vs-experimental design taxonomy this entry’s regulatory test sits alongside.
• Investigational Medicinal Product (IMP) — the EU/UK term for the product under test once a study is classified as an interventional clinical trial.
• Clinical Trial Registration — the registration/reporting obligations that attach specifically to interventional trials.
• ICH GCP (Good Clinical Practice) and ICH E6(R3) — the conduct standard that governs interventional clinical trials.