Why It Matters
Without using the right tool, it can be difficult to understand what your data is telling you.
Processing ...

Avoid Two Common Mistakes with a Shewhart Chart

By David Williams | Wednesday, May 30, 2018
Avoid Two Common Mistakes with a Shewhart Chart

Photo by NeONBRAND|Unsplash

Leaders are frequently faced with having to improve results in their organizations. Without a method, they can fall prey to two big mistakes: 1) acting like something is a unique event when it’s normal for the process, or 2) ignoring issues that are truly special, assuming they are normal. Making the right assessment can be difficult without the right tool: a Shewhart chart, also known as a control chart.

Shewhart charts, which display data over time with upper and lower control limits, are the only way to differentiate between common (predictable) and special cause variation. Common cause variation is inherent in the process, while special cause variation is due to an attributable cause. On the chart, common cause variation falls between the upper and lower control limit, and special cause variation is found above or below it or when one of several rules exist (example, a run of either eight or more points above or below the mean). A good example is your commute time. Some variation has to do with the process itself, such as hitting a red or a green light. If something out of the ordinary happens, like a car crash, that’s special cause variation.

Shewhart chart

Phil Monroe, a current hospital board member, explained why he uses Shewhart charts on the Deming Institute’s podcast: “I need to understand, ‘Is this process predictable?’ If it isn’t, I want it to be. If it is, then I can consider whether I’m happy with it and whether a process improvement project makes sense. The best way to answer this is to get the last 20 data points and plot it on a Shewhart statistical process control chart.”

People often question whether they need as many as 20 points, because they don’t have the data, or they track monthly or quarterly data. In most cases, you can start a Shewhart Chart with 12 data points and create trial limits. The reason for 20-30 data points is that’s when you have enough data to have confidence in the control limits used for determining special cause. If that means going back historically, do it. If that means increasing the frequency of measurement to daily or weekly, do that. If you have no data, start now with a run chart. You’d rarely want to measure less often than monthly.

Understanding whether you have common cause versus special cause variation helps guide your actions. If your process is unpredictable, you want to figure out what’s causing that special cause variation and remove it from the system — for example, you might address equipment or procedure issues that are leading some people to do the work differently than others. If your process is predictable but you don’t like the current performance, you must change the process producing it to get different results.

A key first role of a leader is to foster reliable, predictable processes, and then decide whether you like the performance and make a plan to improve it.

first last

Average Content Rating
(0 user)
Please login to rate or comment on this content.
User Comments


© 2020 Institute for Healthcare Improvement. All rights reserved.