Research ethics
A useful clinical question still needs a defensible design and the correct permissions.
Quick revision summary
Classify the project and secure the right review
Distinguish research designed to produce generalisable knowledge from audit against an explicit standard and evaluation of a current service. Use the HRA decision tools and local research or governance team when classification is uncertain.
Match design, participation and permissions
Write the question and proposed methods before seeking a label for the project. Research, audit and service evaluation can use some of the same data but serve different purposes.
Protect meaningful choice and honest reporting
Explain burdens, possible benefits, uncertainty and what withdrawal means for participation and already collected information. Avoid implying that refusal will affect ordinary care.
Common mistake
An intention to publish does not decide project classification, and a non-research label does not remove every approval requirement.
Research ethics, consent and project classification
A useful clinical question still needs a defensible design and the correct permissions.
Classify the project and secure the right review
Distinguish research designed to produce generalisable knowledge from audit against an explicit standard and evaluation of a current service. Use the HRA decision tools and local research or governance team when classification is uncertain. The project’s title, small sample or intention to publish does not decide its classification. Projects that are not research may still require local registration, information-governance assessment or university review.
Research planning includes a scientifically sound question, proportionate risk, appropriate sponsorship and relevant approvals. Consent should address participation, withdrawal, data use and foreseeable burdens without exaggerating individual benefit. Recruitment through a clinician relationship can create pressure, so make refusal possible without affecting ordinary care. Protect confidentiality, minimise data and specify access, storage and sharing. Patient involvement can improve relevance and design, but it does not replace participant consent or formal approvals.
Key distinctions in practice
| Finding or option | What it means | Clinical consequence |
|---|---|---|
| Research | Aims to generate generalisable knowledge | Use the relevant research governance and approval pathway. |
| Clinical audit | Tests care against an explicit standard | Register locally and plan action and remeasurement. |
| Service evaluation | Judges the existing service | Check local review and data requirements. |
| Consent and data | Participation and information require protection | Use proportionate, clear processes and authorised access. |
Match design, participation and permissions
Write the question and proposed methods before seeking a label for the project. Research, audit and service evaluation can use some of the same data but serve different purposes. Use official decision tools and local research or governance advice when unsure. A project does not escape review because it is small, retrospective or led by a student.
For research, identify the responsible sponsor and the required approvals before recruitment or data use begins. Distinguish agreement to take part from the lawful arrangements governing personal data. Patient involvement in designing a study can improve relevance, but those advisers do not give consent on behalf of future participants. Changes to an approved design may need further review before implementation.
Protect meaningful choice and honest reporting
Explain burdens, possible benefits, uncertainty and what withdrawal means for participation and already collected information. Avoid implying that refusal will affect ordinary care. Recruitment by a treating clinician needs care because patients may feel obliged to help. Provide accessible information and enough opportunity for questions.
Collect only the information required by the approved purpose and control access and storage appropriately. Report methods, limitations and results honestly, including findings that do not support the original hypothesis. Selective outcome reporting can mislead future care and wastes participants’ contribution. A scientifically weak design is also an ethical problem when it exposes people to burden without a useful prospect of answering the question.
A distinction to remember
An intention to publish does not decide project classification, and a non-research label does not remove every approval requirement.
Does an intention to publish automatically make a service evaluation research?
Reveal the reasoning
No. Classification depends on purpose and design. Use the HRA tools and local advice; publication and a small sample do not determine the category.
Guidance to read alongside this chapter
Test your recall
Why is ethical approval distinct from individual consent?
Show answer
Both oversight and a participant's own informed decision may be needed.
