How to Conduct a
Clinical Audit

Step 1 — Pick a topic worth auditing

Not every process deserves a formal audit. A good topic is high-risk, high-volume, or has an evidence base you can measure against — and ideally, someone has already raised a concern about it (an incident, a near-miss, a recurring complaint). Vague topics like "improve patient care" are not auditable; "time from arrival to first antibiotic dose in sepsis cases" is.

Step 2 — Write a measurable standard and target

A standard has two parts: the criterion (what should happen, stated precisely) and the target (the percentage of the time it's considered acceptable). For anything safety-critical, the target is usually 100%. For everything else, pick a number your team can realistically hit and improve on — and agree, in writing, what counts as a fair exception before you start collecting data, not after.

Step 3 — Decide your sample and collect data

You rarely need to check every case. For most process audits, 20–50 recent cases give a reliable picture. Decide upfront whether you're looking back at existing records (faster, but only shows what was documented) or watching practice happen in real time (slower, but catches things records miss). Keep the data collection sheet anonymous — a case number, not a patient name.

Step 4 — Analyse the results against the standard

Every case sorts into one of three buckets: met the standard, didn't meet it but had a fair agreed exception, or didn't meet it with no good reason. That last bucket is where the real finding is. Report both the count and the percentage together — "45 out of 50 (90%)" is honest; "90%" alone can be misleading on a small sample.

Step 5 — Make changes and re-audit to confirm improvement

An audit that stops at "here's what we found" isn't finished — it's a measurement exercise. Identify the root cause, agree a corrective action with an owner and a date, then repeat the same data collection method after the change has had time to embed. The re-audit — not the first audit — is the actual evidence that something improved.

Where most hospitals get stuck

• Choosing a topic too broad to measure
• Skipping the sample-size/exception decisions until after data is already collected
• Reporting a percentage without the underlying count
• Never scheduling the re-audit, so the loop never closes

A guided tool that takes you through each stage in order — with worked examples alongside — closes most of these gaps without needing a separate quality-management course first.

Frequently Asked Questions

How long does a full clinical audit cycle take?

It varies by topic and how fast the change can embed, but most process-level audits run a planning-to-first-results cycle of 4–8 weeks, with the re-audit scheduled 3–6 months after the change is implemented to give it time to take hold.

Who should lead a clinical audit?

Ideally someone with both clinical credibility and enough authority to get a corrective action agreed and funded — commonly a CQI coordinator working with the relevant department's clinical lead.

What's the difference between a clinical audit and a quality indicator?

A quality indicator is ongoing routine monitoring of a metric. A clinical audit is a defined project: a specific standard, a specific sample, a specific action plan, and a specific re-audit to close the loop.