10The design of any audit project will include early considerations of the following:

  • How long will it take to carry out the audit?
  • Do you have the time and enthusiasm to complete the audit? 
  • What information do I need and how will I collect it?

Decide what data is necessary to be able to answer your audit question. Each data item should link to specific criteria. Do not collect information just because it might be interesting or useful - it may well not add anything to your project and will mean the project takes you more time.

Different data collectors may interpret some entries in the same record in different ways. Establish consistencies from the start. Ensure the data collection systems are clear and agreed. In practice there will often be only one person performing the data collection anyway.

Should I pilot my data collection method?

Piloting your tool on a few cases is important before undertaking any audit. It will help you iron out any problems and ensures you are presented with the right information, rather than a quantity of unusable data’ (UBHT 2005). 

How many records / events should we look at?

In deciding whether you are going to look at every patient in your population, or just a sample of them, consider;

How critical is the aspect of care that you are investigating? (Is it to the extent that you need to look at every single patient?)

How much time do you have available to conduct the study? (if your population consists of 1000 patients and the data you require is only contained in their case notes, you will almost certainly want to audit a sample of the population)

The sample chosen should be small enough to allow for rapid data collection but large enough to be representative.

How should we select our sample / target population?

Depending on your audit topic you might be auditing an aspect of treatment / care which applies to all patients (e.g. obtaining an appointment). It is more likely however that you will be interested in a defined group of people who share certain characteristics: most typically the fact that they have the same medical condition, or have received the same form of treatment. For example, patients over 50 years of age with ischaemic heart disease - this forms your audit population.