Professional Documents
Culture Documents
Using Run Charts To Establish Special Cause Variation: Carol Haraden, PHD
Using Run Charts To Establish Special Cause Variation: Carol Haraden, PHD
March 3, 2017
Framework for Clinical Excellence
Patient Safety
Culture
Psychological Accountability
Safety
Leadership
Teamwork &
Communication
Engagement of
Patients & Family
Transparency
Negotiation
Reliability Continuous
Improvement Learning
5.9%
1.1%
Jan 13 Jan 14
Coronary Artery Bypass Graft
7
0
Apr
Aug
Jun
Jul
Jan-13
Jan-14
Feb
Oct
Mar
Nov
Sep
Dec
May
Coronary Artery Bypass Graft
7
0
Apr
Jun
Aug
Jul
Nov
May
Jan-13
Jan-14
Mar
Feb
Sep
Oct
Dec
Coronary Artery Bypass Graft
7
0
Apr
Aug
Jun
Jul
Jan-13
Jan-14
Feb
Oct
Mar
Nov
Sep
Dec
May
Understanding Data Variation
1.0
UCL = 0.88
0.6
Mean = 0.54
0.4
0.0
10
11
12
13
14
15
16
17
18
19
1
9
STATIC VIEW DYNAMIC VIEW
Descriptive Statistics Line Chart
Mean, Median & Mode
Run Chart
Minimum/Maximum/Range
Standard Deviation Control Chart
Bar graphs/Pie charts Statistical Process Control (SPC)
7
Kaiser Permanente Improvement Institute
© 2014 Kaiser Foundation Health Plan, Inc. For internal use only.
Improvement uses static and dynamic data
Static views are suited Dynamic views are
to assess variation at best for measuring
a point in time changes in data
variation
100%
1000
90%
80%
800
Significance Unusual
70%
Processing Time
600 60%
of Factors 50%
Observations
400 40%
30%
200 20%
Sudden
Shifts
10%
0 0%
ll q M ib a HP
De m
pa IB sh
Co To
System
Trends
Best Practice
Spread to entire Region!
New CHF
Protocol
Introduced
New CHF
Protocol
Introduced
10
Kaiser Permanente Improvement Institute
© 2014 Kaiser Foundation Health Plan, Inc. For internal use only.
How will we know that a change is an improvement?
1. By understanding the variation that lives
within your data
No
action
Reject Action taken
taken
defectives on all
here
occurrences
Better Quality Worse Better Quality Worse
Old Way New Way
(Quality Assurance) (Quality Improvement)
12
Source: Robert Lloyd, Ph.D.
Tabular Data Display
Frozen Section Turnaround Time
(minutes)
16 15 26
8 17 25
25 25 17
7 7 9
9 23
X=16.8 26
16 9 24
24 8 18
16 16 15
17 17 18
Graphical Data Display
Frozen Section Turnaround Time
Run Chart (minutes)
30
28
26 X=16.8
24
22
20
MINUTES
18
16
14
12
10
8
6
4
2
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
SEQUENCE
Graphical Data Display
Frozen Section Turnaround Time Histogram
(minutes)
5
0
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
Four Dimensions of Data
SHAPE
CENTER
SPREAD
SEQUENCE
Types of Variation
Common Cause Variation Special Cause Variation
• Is not ‘good variation’ • Is not ‘bad variation’
• Is stable and predictable • Unstable and
• Due to the design of the unpredictable
process • Due to irregular or
• Does not mean that the unnatural causes-
variation is acceptable intentional or
unintentional
• Does not mean that the
variation is acceptable
18
Your Drive to Work….
• How much time does it usually take at 7:30 AM on a
Monday morning?
70
60
50
40
30
20
10
0
08
08
08
08
08
8
8
00
00
00
00
00
00
00
00
00
00
20
20
20
20
20
/2
/2
/2
/2
/2
/2
/2
/2
/2
/2
1/
8/
5/
3/
7/
15
22
29
12
19
26
10
17
24
31
3/
3/
4/
5/
6/
3/
3/
3/
4/
4/
4/
5/
5/
5/
5/
Points equally likely above or below center line
There will be a high data point and a low, but this is expected
No trends or shifts or other patterns
Courtesy of Richard Scoville, PhD, IHI Improvement Advisor
Two Types of Special Causes
Unintentional
When the system
is out of control
and unstable
Minutes ED to OR per
1000
Patient
600
200
system 0
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64
Sequential Patients
Minutes ED to OR per
1000
800
Patient
600
400
200
1 4 7 10 13 16 19 22 25 28 31 34 37 40 43 46 49 52 55 58 61 64
Sequential Patients
21
Example of Data for Judgment
(Perfect Care Bundles – all aspects of a bundle must
be met in order to receive credit)
Does this tabular display of data help us understand how to
improve care?
Care Region TYD Average
Bundle Average Q1 Q2 Q3 Q4
AMI 79 79 79 81 80 79
CHF 61 56 58 63 62 60
Pneumonia 46 16 16 20 31 20
SSI 52 41 43 54 49 47
Legend
Better than or equal to the Region
70
60
50
Bundle Reliability
40
30
20
10
0
1 2 3 4
Quarters
SSI: Special Cause or Common Cause?
70
60
50
Bundle Reliability
40
30
20
10
0
1 2 3 4
Quarters
25
What is wrong with this chart?
Comparison is region average- is the color assigned based on
best practice or best performance by region even when not best
practice?
Is there enough data to make any decision?
No goal stated- is the goal green or best practice?
What is rewarded? Special cause or common cause?
Appropriate Management Response to Common & Special 26
Causes of Variation
Source: Carey, R. and Lloyd, R. Measuring Quality Improvement in Healthcare: A Guide to Statistical Process Control
Applications. ASQ Press, Milwaukee, WI, 2001, page 153.
Attributes of a Leader Who Understands Variation
Leaders understand the different ways that variation is viewed.
They use graphical methods to learn from data and expect others to consider
variation in their decisions and actions.
They understand the concept of stable and unstable processes and the potential
losses due to tampering.
5.25
Measure
Pounds of Red Bag Waste
5.00
4.75
Median=4.610
4.50
~
4.25
X (CL)
4.00
3.75
3.50
3.25
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
Four simple run rules are used to determine if special cause variation is present
Normal Distribution with Standard Deviations 31
“What is the variation in one system over time?”
Walter A. Shewhart - early 1920’s, Bell Laboratories
Static View
time
Mean = 2.0
2.2
Micrograms/ML
1.7
1.2
0.7
0.2
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
Blood Samples
Analysis of Run Charts
Special Cause Rule Number 2: Trends
Five or more consecutive points all going up or all going down. If the value of
two or more consecutive points is the same, only count the first point and
ignore the repeating values; like values do not make or break a trend.
ADVERSE DRUG REACTIONS
Number of Adverse Drug
Mean = 3.0
5
4
Reactions
3
2
1
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
Week Number
Analysis of Run Charts
Special Cause Rule Number 3: Patterns
Any non-random pattern may be an indication of a special cause variation. A general rule is to
investigate any non-random pattern that recurs eight or more consecutive times.
110
105
100
95
90
85
80
75
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
70
60 UCL
TIME IN DAYS
50
40
30
20
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
COLPOSCOPY PATIENTS
Medication Administration Process
45
SHIFT DOWN
Elapsed Time to Administer Medication in
40 Mean = 22.5
35
30
Minutes
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
Medication Sequence
Abnormal Pap Test Follow-up Process
60
PATTERN
Median = 35
50
40
Time in Days
30
20
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
Colposcopy Patients
Process for Obtaining a Stat Consult
6
SHIFT UP
5
Median = 3.75
4
Time in Hours
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
Consult Patients
Process for Admitting from Outpatient Clinic
6
TREND
PATTERN
5
Median = 3.0
4
Time in Hours
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
Patient: Admissions
P41
60
RANDOM VARIATION
Median = 35
50
40
Time in Days
30
20
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
Colposcopy Patients
Take a moment to reflect
on your own work.
What will you incorporate from
this session into your plans?