Is my work considered research?
Not everything that involves collecting data from people counts as research in the formal sense. The distinction matters because the obligations that come with formal research (ethics review, regulatory oversight, participant protections) do not apply to all data collection activities equally, and misclassifying a project in either direction has real consequences: treat genuine research as a service evaluation and you risk running an unapproved study; treat a legitimate audit as full research and you add months of unnecessary ethics review to something that never needed it.
This isn't a rare edge case either. Anyone who has worked across a hospital, university, or research-active organisation for long enough will have seen a project stall for weeks while someone tries to work out which category it actually falls into, often well after data collection has already started informally.
In the UK, the Health Research Authority (HRA) provides a decision-making tool aligned with the UK Policy Framework for Health and Social Care Research that helps you work out where your project sits. The framework centres on three key questions.
1. Are participants randomised to different groups?
Randomisation is a core feature of investigatory research. If people are being randomly allocated to receive (or not receive) some intervention, that is a strong indicator that the work is research. The intervention does not have to be a treatment versus placebo: any systematic allocation to different conditions counts. A project comparing two existing, already-approved care pathways by randomly assigning patients to one or the other is research, even though neither pathway is new or experimental in itself. The randomisation is what triggers the classification, not the novelty of what's being compared.
2. Does the study protocol change treatment, care, or services from accepted standards?
This question separates research from two other common activities: clinical audit (which reviews existing practice to improve quality) and service evaluation (which examines how standard care is delivered). If your work involves departing from accepted standards of care or service for any participants, it is likely to be classified as research. A project that simply measures how well an existing clinic is meeting its own targets, without changing anything about what patients receive, is much more likely to sit in audit or service evaluation territory instead.
3. Is the work designed to produce generalisable or transferable findings?
Research produces findings that can be applied beyond the specific participants involved. If your study is designed to generate conclusions that could be relevant to a broader population or transferred to a similar context, that is a defining characteristic of research. This is often the most subtle of the three questions in practice, because intent matters as much as design. Two projects can look almost identical operationally, but one is explicitly designed from the outset to produce publishable, generalisable conclusions, while the other is a one-off local check that happens to generate similar-looking data.
The three questions at a glance
| Question | Points toward research | Points away from research |
|---|---|---|
| Randomised allocation to groups? | Yes | No, or allocation follows existing clinical decision-making |
| Departs from accepted care or service standards? | Yes | No, work only reviews or evaluates existing standard practice |
| Designed to produce generalisable findings? | Yes, intended for wider or future application | No, findings apply only to this specific local context |
If any of these three criteria apply to your work, you are likely conducting formal research and will need to understand what that means in terms of permissions and oversight before you begin.
What happens next?
If your work is research, you may need a number of approvals in place before approaching participants. What those are depends on your study type, population, and setting. The HRA's own decision tool, along with the relevant ethics body for your institution, are the right starting points for working that out, and it's worth running the actual project through the tool rather than relying on judgement calls from the three questions alone, since the tool accounts for edge cases and combinations these summaries necessarily simplify.
Getting the classification right early saves significant time and avoids the risk of discovering a compliance gap mid-study, which is a considerably more disruptive place to discover it than during initial planning, when a wrong classification is still just a decision to revisit rather than a study already partway through data collection.