Online library of Quality,

Service Improvement
and Redesign tools

Statistical
process control

collaboration trust respect innovation courage compassion

Statistical process control
What is it?
There are two methods to support the robust statistical interpretation of measures
presented over time and to understand if your process has special cause and/or common
cause variation. These are run charts and statistical process control (SPC) charts. SPC can
help you understand the scale of any problem, gather information and identify possible
causes when used in conjunction with other investigative tools, eg process mapping and
spaghetti diagrams. You will then be able to measure the impact of any change and
evaluate its worth.
When to use it
SPC should be used throughout the life cycle of the project to help you identify a project,
get a baseline and evaluate how you are currently operating. SPC will also help you to
assess whether your project has made a sustainable difference.
How to use it
An SPC chart has an average line (mean or median – the mean is most often used in SPC
charts) and two control lines above and below the average line, both of which allow more
statistical interpretation.
Figure 1: Example of an SPC chart

40

35
Unit of measurement

30

25

20

15

10

0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30

Consecutive points

Data Average (26) Lower limit (16) Upper limit (37)

SPC tells us about the variation that exists in the systems that we are looking to improve.
S – statistical, because we use some statistical concepts to help us understand processes.
P – process, because we deliver our work through processes ie how we do things.
C – control, by this we mean predictable.

Quality, Service Improvement and Redesign Tools: Statistical process control

SPC charts are constructed by plotting data (we suggest a minimum of 25 data points or
more ideally) in time order, calculating and displaying the average (the mean) and some
data comparisons known as the upper and lower control limits as lines. These limits, which
are a function of the data, give an indication by means of chart interpretation rules as to
whether the process exhibits common cause (predictable) variation or whether there are
special causes. After plotting your chart, the next stage is therefore analysing the chart by
looking at how the values fall around the average and between the control limits.
Generally we use specialist software to create SPC charts, but charts can also be easily
created using MS Excel. There are four rules to interpret SPC charts and if you use
specialist software, these rules will be flagged so you don’t need to remember them. If
one of the rules has been broken, this means that special cause variation exists in the
system and once identified, can be removed. It is also perfectly normal for a process to
show no signs of special cause variation. This means that only common cause variation is
present and therefore by reducing variation you can improve the process further to deliver
standard results – ie it is necessary to ensure that only common cause variation is present
before you make a change.
Figure 2: Rule 1 – any single point outside the control limits:

Point above UCL

UCL UCL

MEAN MEAN

LCL LCL

Point below LCL

Quality, Service Improvement and Redesign Tools: Statistical process control

Figure 3: Rule 2 – a run of seven points all above or all below the centre line (a shift),
or a run of seven points all consecutively ascending or descending (a drift):

Seven points above centre line Seven points consecutively ascending

UCL UCL

MEAN MEAN

LCL LCL

Figure 4: Rule 3 – any unusual pattern or trends within the control limits:

Cyclic pattern Trend pattern

UCL UCL

LCL LCL

Quality, Service Improvement and Redesign Tools: Statistical process control

Figure 5: Rule 4 – the number of points within the middle third of the region between
the control limits differs markedly from two thirds of the total number of points:

Considerably less than 2/3 of Considerably more than 2/3 of

all the points fall in this zone all the points fall in this zone

UCL UCL

LCL LCL

If you want a more efficient system, you need to reduce variation. Common and special
causes of variation indicate the need for two different types of improvement. If
controlled variation (common cause) is displayed in the SPC chart, the process is stable
and predictable, which means the variation is inherent in the system.
If you want to improve the process, you will have to change the whole system. If
uncontrolled variation (special cause) is displayed in the SPC chart, the process is
unstable and unpredictable.
Variation may be caused by factors outside the process. In this case, you need to identify
these sources and resolve them, rather than change the system itself.
There are three issues that you should be aware of when using SPC charts to improve a
process:
1. You should not react to special cause variation by changing the process, as it may
not be the system at fault
2. You should not ignore special cause variation by assuming that it is part of the
process. It is usually caused by outside factors, which you need to understand in
order to remove
3. You should ensure that the chart is not comparing more than one process and
displaying false signals. An example of this would be data covering two hospital
sites, or two procedures that are very different.

Quality, Service Improvement and Redesign Tools: Statistical process control

TIPS

• You may need to collect the data for analysis as it may not be available. To be
statistically rigorous, the more frequently you record your observations the
better. Weekly or daily is better than monthly.
• Aggregate data is discouraged (ie the use of percentages) as this often hides the
pattern of the data).
• The problem you are observing may be the means by which you are measuring,
not what is really happening. Sometimes it is better not to act if you are unsure;
investigate further instead.
• Sometimes the process that you are looking to improve will display seasonal
variation and if this is the case you can split the chart to understand what is
normal for the different time periods (for example winter and summer) based on
where the data runs start and end.
• Remember that when you change something in the process, the data points after
the change will be from a new system. When you have a run of points that break
a rule, you will need to recalculate the SPC control limits to show an improvement
(showing the control limits of the new system).

What next?
Common cause variation is the desirable state of a system prior to undertaking a PDSA
cycle. If the system is unstable and unpredictable then we are not confident about
assigning improvement to the PDSA, it could be external factors affecting the result.
Undertaking analysis before improvement allows you to develop a strong baseline and
using weekly or daily data is a good starting point for understanding the baseline. The
following other tools will be helpful:
• Managing variation – essential reading for using SPC charts.
• Process mapping – useful for understanding variation revealed by SPC charts.
• Root cause analysis using five whys – useful for understanding variation revealed by
SPC charts.

Additional resources
Berwick, DM (1991) Controlling Variation in Healthcare: A Consultation from Walter
Shewhart, Med Care: 29: p1212–25
Bicheno, J and Catherwood, P (2005) Six Sigma and the Quality Toolbox: For Service and
Manufacturing, Picsie Books
Deming, WE (1986) Out of the Crisis, MIT: Massachusetts
Esain, A (2006) Problem Solving, TQM and Six Sigma in Rich, N, Bateman, N, Esain, A,
Massey, L and Samuel, D (Eds) Lean Evolution: Lessons from the Workplace

Quality, Service Improvement and Redesign Tools: Statistical process control

Rich, N et al (2006) Lean Evolution: Lessons from the Workplace, Cambridge University
Press
Shewhart, WA (1980) Economic Control of Quality of Manufactured Product: 50th
Anniversary Commemorative Reprint, SAQC/Quality Press, Reissue Edition
Wheeler, D (2000) Understanding Variation: The Key to Managing Chaos, SPC Press, 2nd
edition
Wheeler, D (2003) Making Sense of Data: SPC for the Service Sector, SPC Press: Knoxville
Thor et al (2007) Application of statistical process control in healthcare improvement:
systematic review. Quality Safety in Health Care (2007)16: 387–399.
Background
The method of SPC was created by Walter Shewhart during the 1920s when he was
investigating the production of faulty telephones for Bell Telephones. He found that the
company continually changed the production line when faulty handsets were produced.
This resulted in the production of different formats of handset – telephones that wouldn’t
talk to each other!
Through Shewhart’s work, Bell Telephones understood the importance of reducing the
variation in the manufacturing process to ensure Lean production. Shewhart framed
problems as ‘assignable cause’ and ‘chance cause’ and introduced the SPC chart to
differentiate between the two.
1920s: First control charts developed by the communications industry.
1930s: Adopted by other industries.
POST WAR: Shewhart tried to introduce the principles of SPC in American manufacturing
but was unsuccessful and so took his ideas to Japan.
1950s: Japan began to widely use the principles of SPC in industry.
1980s: Ford Motor Company adopted SPC principles due to pressure on the US market
from Japanese imports.
2000s: SPC adopted in the NHS.
Data should always be presented in a way that preserves the evidence. Shewhart (1931)
suggested that displaying data using averages and aggregates loses the richness of the
individual data points. SPC displays the individual data points (in the NHS these are often
individual patients), then provides analysis to interpret what the user sees.
SPC and PDSAs were created by the same person, intended to be used together. PDSA
improvement cycles should only be implemented in a system that displays common cause
variation. If special cause variation is present, then the system is not stable enough to
ensure the information is a result of the PDSA change.

Written by the ACT Academy for their Quality, Service Improvement and Redesign suite of programmes.
Contact: [email protected]