Guide to using run charts

Data and Measurement - How we put run chart theory into practice for improvement projects

A run chart is a simple analytical tool that helps us understand changes in data over time. We describe here, within parameters typical of a quality improvement project, how Healthcare Improvement Scotland constructs, uses and interprets run charts to understand change in a system or process.

The purpose of this guide is to provide some simple, practical advice on how we deal with data in run charts.

For an introduction to run charts and rules for interpretation, we recommend the paper published by Perla, Provost and Murray, The Run Chart: a Simple Analytical Tool for Learning from Variation in Healthcare Processes1.

Although, their paper provides clear guidance on how to construct a basic run chart and detect a signal of non-random variation, the next steps you would take in your analysis are not necessarily covered. This is largely due to the subjectivity of some of these decisions.

We hope this guide on our approach will help with your journey of data interpretation. Provost & Murray also go into more detail in The Health Care Data Guide: Learning from Data for Improvement2.

Data sets for quality improvement projects take many shapes and sizes; we cannot provide generic guidance that will suit every eventuality, but for the purpose of this guide, we will work within a basic set of data criteria:

  • data are displayed monthlyat a minimum, weekly process data are preferable
  • if using a sample,sample size is 20 or more for each data point, and
  • time-between run charts are used when more than half the data points areat the extreme values on the scale (for example0 or 100 on a percentage scale).

We think that the approach taken will suit most applications, but if your data set falls out with the parameters above then some more thought may be required.

Published Date: 30 June 2017