An introduction to Statistical Process Control (SPC) and associated analysis with data for: Demonstration only

1
An introduction to Statistical Process Control
(SPC) and associated analysis with data for
Demonstration only

• Malcolm Boyes
• Health Outcomes Consultant
• GSK

2
This presentation aims to provide

• An understanding of the basic principles of
  funnel plots
• Examples of data for Demonstration Only

• Reference Source Data for cases
• 2. Reference Source Data for Population

3
Background to Statistical Process Control (SPC)

• Introduced by Walter Shewhart (Bell Telephone
  Laboratories 1924)
• The method was exported to Japan in the 1950s,
  where it was successfully applied in industry.
• SPC techniques demonstrate the simplicity and
  power of control charts at guiding their users
  towards appropriate action for improvement. 1

1. Mohammed MA, Cheng KK, Rouse A, Marshall T.
Bristol, Shipman, and clinical governance
Shewhart's forgotten lessons. Lancet 2001
357(9254)463-467
4
What is Statistical Process Control (SPC)

• Statistical Process Control SPC is defined as
• a philosophy, a strategy and a set of methods for
  ongoing improvement of systems, processes and
  outcomes 1
• Simple graphical way to display data and outcomes
• It is a method which identifies unusual variation
• Aims to improve quality

1. Evidence based practice Definition of SPC.
Available at http//www.evidencebasedpractice.or
g.uk/spc.htm Accessed 31/03/2009
5
Traditional approach in NHS

• It is common for performance data to be presented
  in the form of League tables or ranked data, for
  example DoH Hospital Episode Statistics (HES)
  data in Disease Management Information Toolkit 1

1. DOH. Disease management information toolkit.
Long Term Conditions. 2008 July. Accessed
16/07/09 Available fromhttp//www.dh.gov.uk/en/
Healthcare/Longtermconditions/DH_074772?IdcService
GET_FILEdID169229RenditionWeb
6
The Use of League Tables in Decision Making1

• League tables with only common cause variation
  may encourage unwarranted tampering
• League tables may lead to local special cause
  variation being ignored
• League tables may encourage the blame culture
  and are not linked directly to improvement
  activity

• 1. Roberts T. Understanding variation Online
  July 2005 cited July 2009 17 Pages Available
  from
• http//www.evidencebasedpractice.org.uk/documents/
  presentations/spc_TEBPCJune2005.ppt

7
How does this help?1

• Performance data should be used to guide quality
  improvement
• The purpose should be to find the assignable
  causes and understand their origin they should
  be prevented if bad and spread if good
• When only unassignable causes are present, the
  process can only be improved by changing things
  that affect the process all of the time

• 1. Roberts T. Understanding variation Online
  July 2005 cited July 2009 17 Pages Available
  from
• http//www.evidencebasedpractice.org.uk/documents/
  presentations/spc_TEBPCJune2005.ppt

8
Variation in a system is normal 1

• The variation is caused by factors that are
  inherent in the system over time
• They affect all outcomes
• This is common cause variation or
• The causes are unassignable
• Common cause variation can be reduced by tackling
  things that affect the process all the time

• 1. Roberts T. Understanding variation Online
  July 2005 cited July 2009 17 Pages Available
  from
• http//www.evidencebasedpractice.org.uk/documents/
  presentations/spc_TEBPCJune2005.ppt

9
Some variation may not be normal 1

• The factors are not present in the process all
  the time
• They do not affect everybody
• They arise because of specific circumstances
• This is special or assignable cause variation.

• 1. Roberts T. Understanding variation Online
  July 2005 cited July 2009 17 Pages Available
  from
• http//www.evidencebasedpractice.org.uk/documents/
  presentations/spc_TEBPCJune2005.ppt

10
Two types of SPC chart

• If you want to compare different individuals,
  units or hospitals etc over a single time period,
  a funnel chart may be helpful
• If you want to compare a single individual, unit
  or hospital over different time periods, atime
  chart may be helpful

11
Anatomy of an SPC Funnel Chart
Practices with higher or lower than average
admissions may be explained by a variety of
factors
Non Elective admissions / 100 Patients
Mean
List Size
Example data for illustrative purposes only
12
Anatomy of an SPC time chart1

• 1. Roberts T. Understanding variation Online
  July 2005 cited July 2009 17 Pages Available
  from
• http//www.evidencebasedpractice.org.uk/documents/
  presentations/spc_TEBPCJune2005.ppt

13
How ranking data may lead to misinterpretation
(1/2)
Hypothetical data showing how ranked data can
lead to misinterpretation
GP Practice Admissions List Size Admission Rate
Dr C 3 5 60.0
Dr G 20 35 57.1
Dr E 56 110 50.9
Dr F 23 54 42.6
Dr D 25 70 35.7
Dr H 28 123 22.8
Dr A 24 132 18.2
Dr B 11 333 3.3
Average 190 862 22.0
Hypothetical data developed by GSK for
illustrative purposes only
14
How ranking data may lead to misinterpretation
(2/2)
GP Practice Admissions List Size Admission Rate
Dr C 3 5 60.0
Dr G 20 35 57.1
Dr E 56 110 50.9
Dr F 23 54 42.6
Dr D 25 70 35.7
Dr H 28 123 22.8
Dr A 24 132 18.2
Dr B 11 333 3.3
Average 190 862 22.0
Hypothetical data developed by GSK for
illustrative purposes only
15
Presentation of Data for Demonstration purposes
only

• Presentation of Demonstration graphs, using mock
  up data.
• Unplanned COPD admissions per 100 COPD patients
  plotted against COPD list size

1. Reference Source Data for cases 2. Reference
Source Data for Population
16
Funnel Chart x versus y

Enter graph/graphs on this and following
slides
1. Reference Source Data for cases 2. Reference
Source Data for Population
17
Hypothetical data developed by GSK for
illustrative purposes only
1. Reference Source Data for cases 2. Reference
Source Data for Population
18
Hypothetical data developed by GSK for
illustrative purposes only
1. Reference Source Data for cases 2. Reference
Source Data for Population
19
Hypothetical data developed by GSK for
illustrative purposes only
1. Reference Source Data for cases 2. Reference
Source Data for Population
20
Hypothetical data developed by GSK for
illustrative purposes only
21

• For further information or to see the admissions
  data for COPD in your own Health Board contact
  Malcolm Boyes on
• Mob [phone]
• Email malcolm.2.boyes_at_gsk.com