What is a run chart?
A run chart plots a measure over time with its median as a reference line. It is the workhorse of improvement work (PDSA cycles, Lean and healthcare quality) because it answers one question with very little data: is the process changing, or is it just normal variation? Unlike a control chart, it needs no control limits and makes no assumption about the distribution of the data.
Run chart rules for non-random signals
- Shift: 6 or more consecutive points all above or all below the median. Points on the median are skipped.
- Trend: 5 or more consecutive points all going up or all going down. Repeated values do not count.
- Runs: too few or too many runs for the number of points not on the median. The tool calculates the limits from the exact runs distribution (Swed-Eisenhart), about 5% risk of a false signal.
- Astronomical point: a value that is obviously different from all others. Judge this one visually.
Run chart example
The example tracks the weekly median door-to-needle time in a stroke unit over 26 weeks. A pre-shift huddle was introduced in week 15. The baseline median is 41 minutes; after the change the median is 30.5 minutes. All 12 points after the change sit below the baseline median, a clear shift (along with the two weeks just before it), and the data has only 7 runs where 8 to 17 are expected. Because lower is better, the shift is in the desired direction: evidence that the huddle worked.
Run chart or control chart?
Use a run chart early, with little data, or when the data is skewed or made of rates and counts that change in size. Move to a control chart once you have 20 or more stable points and want to monitor the process, and to a hypothesis test when you need a formal before-and-after comparison.
How to use this Run Chart tool
- Enter your measure. Name the measure, its unit and whether higher or lower is better.
- Paste the data in time order. Paste labels (dates, weeks) and values from Excel into the table.
- Mark the change. Click the point where you introduced a change, or type its number, to compare baseline and after.
- Read the signals. Shifts and trends are marked in red, and the runs test compares the number of runs with the expected range.
- Annotate and export. Add notes for changes and events, then export the chart to PDF, PNG or Excel.
Frequently asked questions
What is a run chart?
A run chart is a line graph of data in time order with the median drawn as a centre line. It is the simplest way to see whether a process is changing over time and whether a change you made led to an improvement.
What are the run chart rules?
The common rules (Provost and Murray) are: a shift of 6 or more consecutive points above or below the median; a trend of 5 or more points all going up or all going down; too few or too many runs for the number of points; and an astronomical data point that is obviously different from the rest. Points exactly on the median are skipped when counting shifts and runs.
What is a run?
A run is a series of one or more consecutive points on the same side of the median. Count runs by counting how many times the line crosses the median and adding one. Too few runs means the data stays on one side for too long; too many means it alternates more than chance would.
How many data points do I need?
Start with at least 10 to 12 points; the runs test needs at least 10 points that are not on the median. For a baseline before a change, 12 to 20 points is a good target.
What is the difference between a run chart and a control chart?
A run chart uses the median and probability-based run rules and needs no assumptions about the distribution. A control chart adds control limits calculated from the process variation and is more sensitive to individual unusual points. Start with a run chart and move to a control chart once you have 20 or more points.
How do I show a before-and-after change?
Mark the point where the change was introduced. The tool then calculates the median from the baseline points only and extends it, so the run rules test whether the new points differ from the old process. It also shows the median after the change.