HEALTHCARE QUALITY PAGE

Getting sampling right for locally conducted clinical audits or quality improvement projects

There are several misconceptions about sampling in local clinical audits and quality improvement (QI) projects. One is that a small sample — 30 or 50 cases, or even fewer — is sufficient because poor care in a few cases suggests the process needs improvement. Another is that a valid clinical audit must include every eligible patient on an ongoing basis.

These beliefs are myths. When designing a clinical audit or QI project, consider the sample in context, including how many relevant cases occur within a time period, such as a week. Depending on the objective, consider if the sample needs to represent all eligible patients or events. Research studies do not include every patient in a population, yet their findings still inform best clinical practice. The same statistical principles can — and probably should — be applied to clinical audits and QI projects.

You also need to decide what type of sample is appropriate. Should it represent all patients or events, or can it focus on problem areas or cases selected for convenience? Guidance is available on choosing the sampling approach, sample type, and number of cases to include.