How to Write a Clinical Audit

An audit measures practice against an agreed standard — which means the standard has to exist before you collect anything.

Clinical audit compares current practice against an explicit standard and acts on the difference. It is distinct from research: research asks what the right thing to do is, while audit asks whether the thing already known to be right is actually being done. Confusing the two is the most common conceptual error in student audit work, and it affects everything from ethics approval to how findings are reported.

Choose a topic worth auditing

Good audit topics share features: the practice matters to patient outcomes, there is a recognised standard, data is obtainable, and change is feasible if a gap is found. Auditing something nobody can act on wastes everyone’s effort.

High-volume, high-risk or known-problem areas are the conventional priorities, and local incident data frequently points to them.

Identify the standard first

Audit is meaningless without a standard to measure against. It should come from a recognised source — national guidance, professional body standards, local policy derived from evidence — and be cited explicitly.

The standard must then be converted into measurable criteria with a target. “Pain should be assessed regularly” cannot be audited; “pain score documented within one hour of admission in 95% of cases” can.

Set criteria and targets explicitly

Each criterion needs a definition precise enough that two people auditing the same record would reach the same judgement. Ambiguity at this stage produces unreliable data that cannot support any conclusion.

• What exactly counts as compliance?
• Where in the record would the evidence appear?
• What are the legitimate exceptions?
• What target applies, and where does it come from?

Targets below 100% need justification. Where the standard admits clinical exceptions, state how they are handled rather than treating them as failures.

Plan the data collection

Decide the sample, the period, the source and the tool before collecting anything. Retrospective record review is the most common approach in student audits; prospective collection gives better data and takes longer.

Design a simple collection form with one row per case and one column per criterion. Pilot it on a handful of records — ambiguities in the form surface immediately and are cheap to fix before the full collection.

Get the governance right

Audit does not usually require research ethics approval, but it does require local registration and information governance clearance. Assuming otherwise is a common and serious error.

Handle data accordingly: collect the minimum necessary, anonymise at the point of collection, store on approved systems, and never take identifiable data off site.

Analyse against the criteria

Analysis in audit is usually straightforward descriptive statistics: the proportion meeting each criterion, compared with the target. Resist the temptation to add inferential tests that the design does not support — audit samples are not designed to generalise.

Present each criterion separately. An overall compliance figure obscures which specific element of care is failing, which is the information needed to act.

Interpret the gap, do not just report it

Finding that 62% of cases met a 95% target is a result; explaining why is the analysis. Consider whether the standard was known, whether the process supports compliance, whether documentation rather than practice is the failure, and whether the exceptions were legitimate.

This distinction matters practically: a documentation problem and a practice problem need entirely different interventions.

Recommend changes that can be implemented

Recommendations should follow from the analysis and be specific enough to act on: what changes, who is responsible, by when, and how it will be supported. Recommendations aimed at individual effort — “staff should remember to” — rarely produce sustained change; those that alter the system usually do.

Close the loop

An audit that is never repeated has not completed its cycle. Re-audit after the change has had time to take effect, using the same criteria and method so the comparison is valid. Reporting the re-audit is what distinguishes audit from a one-off survey, and in assessed work the plan for it is usually expected even where the re-audit itself falls outside your timeframe.

Present the audit clearly

Written audit reports follow a recognisable structure: background and rationale, the standard and its source, criteria and targets, method, results against each criterion, discussion of the gap, recommendations, and the re-audit plan. Where your programme provides a template, follow it exactly, since audit reports are usually assessed against those sections.

Present results per criterion rather than as an overall figure, and use simple charts where they aid comparison against the target.

Involve the people whose practice you are auditing

Audits imposed on a team generate defensiveness and rarely change anything; audits that involve the team in interpreting results and designing the response usually do. Share findings without attributing failures to individuals, and ask the people doing the work what prevents compliance — they generally know, and the answer is frequently a process problem rather than a knowledge one.