Aug 262014
 

All of our processes contain variation. Understanding variation is vital when deciding how to improve our processes and services.

David Williams

David Williams

Variation is the natural fluctuation that we see in our processes. For example, the number of times the phone rings at work a day is never constant. The differing number of times it rings a day is the variation we see in this process.There are two different types of variation, “common cause” variation and “special cause” variation.

An example of how variation works…

Every day I drive to work. It normally takes me about 55 minutes, if there are no unusual occurrences, but this does vary. It rarely takes exactly the same time to drive to work due to levels of traffic, weather, or the timing of traffic signals. These time differences are expected. It is common cause variation.

One day, there was an accident on the Motorway. My journey to work took 94 minutes. This is special cause variation. If this happened to you, would you change your route to work every other day just because of this single occurrence?

A few days later road works began on the motorway. The journey to work now took 75 minutes. For the next few days it took around 75 minutes as well. It looks like this new drive time would continue, so I decided to look for a different route.

This data can be quickly articulated in a run chart, which helps identify the type of variation occurring.

David Williams Data blog graph

The data is plotted in the ordered it occurs, a line is then added showing the median value.

You can see the drive to work is usually around 55 minutes and the accident leads to the journey of 94 minutes. Using a run chart like this quickly shows that this is special cause variation by how different that data point is to the rest. The journey times then revert back to having common cause variation around 55 minutes for a further six data points, i.e. six days.

Then the journey time increases due to the road works. The data in the run chart shows that this has caused a ‘process shift’ to a resulting drive time of around 75 minutes. At this point it makes sense to look for a different route to work.

Applying this to the workplace…

In the workplace, when we have a process that is producing unsatisfactory results, understanding whether we are seeing common or special cause variation is vital when deciding our method on addressing the issue.

When we see special cause variation, we may be tempted to redesign the system around these one-off events. These one-off events may require escalation processes, but rarely need a complete process redesign.

Looking at the data over time allows us to understand the common cause variation within a process. When a process consistently gives us a result that is undesired, then changing the process can be considered. Continuing to measure the process will allow us to see if the changes we implement improve the process.

David Williams is the Information & Improvement Lead in 1000 Lives Improvement.

Do you know how your service varies over time? Do you react to special cause variation? Tell us in the comments below, or tweet us using @1000LivesPlus.

This blog is a summary of a section of the Silver Improving Quality Together course. For more information, please visit the Improving Quality Together website

  2 Responses to “Making the right decision based on data”

  1. […] last time I blogged, I was talking about improving your journey to work by making decisions based on data, but I also wanted to look at how to prioritise which changes are of the biggest benefit and which […]

  2. […] last time I blogged, I was talking about improving your journey to work by making decisions based on data, but I also wanted to look at how to prioritise which changes are of the biggest benefit and which […]

 Leave a Reply

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

(required)

(required)