Quality Improvement, Clinical Audit and Research in Healthcare
Choosing the right project: QI, clinical audit, service evaluation or research
A practical method for classifying a proposed project by purpose, question and method so that governance, approvals and evaluation match the work being undertaken.
Learning objectives
• Classify a proposed activity by its purpose rather than by the label preferred by the team. • Distinguish QI, clinical audit, service evaluation and research while recognising overlap. • Explain why classification matters for governance, consent, ethics, data use and the claims that can be made. • Use HRA/HQIP decision support and local governance when classification is uncertain. • Recognise “label shopping” and other unsafe shortcuts.
Learning objectives
- Classify a proposed activity by its purpose rather than by the label preferred by the team.
- Distinguish QI, clinical audit, service evaluation and research while recognising overlap.
- Explain why classification matters for governance, consent, ethics, data use and the claims that can be made.
- Use HRA/HQIP decision support and local governance when classification is uncertain.
- Recognise “label shopping” and other unsafe shortcuts.
Start with the question, not the label
The safest classification begins with purpose. “Are we meeting an explicit standard?” suggests clinical audit. “How is this existing service performing or experienced?” suggests service evaluation. “Can we improve this local process through iterative change?” suggests QI. “Can we generate new, generalisable knowledge by answering a defined research question?” may suggest research. Method alone does not settle the classification: interviews, surveys and record review can appear in several project types.
Clinical audit
Clinical audit compares practice with explicit criteria or standards, then uses the findings to improve care and re-measures to determine whether the gap has closed. A complete audit therefore requires a defensible standard, defined criteria, appropriate sampling, action and re-audit. Merely describing current practice is not a complete audit cycle.
Quality improvement
QI is local, systematic learning aimed at improving a process or outcome. It commonly uses repeated measurement, process understanding, change theory and iterative testing. QI may use evidence-based standards but is not limited to measuring compliance. It asks how a system can be changed and whether those changes produce improvement in context.
Service evaluation
Service evaluation asks how an existing service performs, whether it meets objectives, and how it is experienced. It may use routine outcomes, surveys, interviews, pathway analysis or mixed methods. It should still have a clear question and governance plan. Calling an activity “evaluation” does not remove obligations around confidentiality, data protection, accessibility or burden.
Research
Research is designed to generate knowledge beyond the immediate local service and uses methods appropriate to the research question. Formal research may require sponsorship, HRA/REC or other regulatory approvals depending on design and setting. Teams should use current HRA decision support and local R&D/research governance advice rather than deciding classification solely on intuition.
Why classification matters
Classification affects who must approve the work, what information/consent processes are required, how data may be accessed, whether intervention risk is acceptable, whether results can support generalisable claims, and what reporting standards apply. Misclassification can expose participants and organisations to risk. Conversely, over-classifying simple local improvement as research can create unnecessary delay and discourage useful learning.
Worked examples
Example A: compare VTE assessment completion with a national standard and re-audit after intervention - clinical audit. Example B: test a new bedside handover prompt on one ward using weekly measures - QI. Example C: evaluate patient experience of a new outpatient pathway introduced six months ago - service evaluation. Example D: randomise patients to two novel follow-up strategies to estimate comparative effectiveness for wider use - likely research and requires formal governance assessment.
Classification worksheet
Write four lines: (1) primary purpose, (2) who/what is being changed or observed, (3) intended audience for the conclusion, and (4) governance route proposed. Then test the classification using HRA/HQIP/local tools. If the answer changes when you describe the project honestly, revisit the design rather than choosing the easiest label.
