Reliable Design - IHI
Transcript of Reliable Design - IHI
5/7/2014
1
Reliable DesignFrank Federico
May 2014
Why Are Processes Not Reliable?
Individual Autonomy
Focus on benchmark performance
Over-reliance on training, vigilance and hard work
Expecting that having a policy will result in improved
reliability
2
5/7/2014
2
3Example of a Run Chart Showing Implementing
the VAP Bundle
Integrate daily
goals with MDR
to identify
defects as a
Step 1 change
concept
Education
Baseline
Feedback on
compliance
Redundancy in the
form of a check by
RT built into 1 hour
scheduled vent
checks as a Step 2
change concept
(Baptist Memorial, Memphis)
What Makes Processes Reliable?
Deliberate design
All employees focus on reliability
Ideas are tested-tested-tested
Measure-measure-measure
4
5/7/2014
3
5
The Reliability Design Strategy
Prevent initial failure using intent, simplification and standardization
Identify defects (using redundancy/backup plans) and mitigate
Measure and then communicate learning from defects back into the design process
IHI Reliable Design MethodologyStart small- subset/segment of the population
Visualize the steps in the process using a high-level flow
diagram
Identify the defects in each step
Lead with the change concepts of simplification and
standardization
Develop a ‘back-up plan’ or redundancy
Test-test-test
Measure-measure-measure
6
5/7/2014
4
Start Small
Experience tells us that not all situations are the same
One standardized process will not work for all
Design a process to deliver reliable care for a group that
is easiest to work with
Learn from that group and spread to others
7
Selecting a Subset/Segment
Characteristics:
Easy to identify
Willing participants
Can learn how to design for other subsets
High enough volume to be able to test daily or every other day
“If we cannot make our process reliable for this group, what
are our chances with other groups?”
8
5/7/2014
5
Subset for the Ventilator Care Bundle
One ward
One doctor’s patients
9
Population
Segment
Small Test
of Change
10
5/7/2014
6
Visualize the Steps
High level flow diagram
3 to 5 Steps-keep it simple
Identify defects
– Is there a cascade?
– Where is the biggest defect?
– Identify what you will fix
11
Example of High Level Diagram
Patient
Admitted
Assessment
Completed
Prophylaxis
OrderedProphylaxis
Administered
Patient
Discharged to
Home
Fail to complete assessment
12
At least 95% of the patients will be assessed
VTE Prophylaxis
5/7/2014
7
13Example: Ventilator Care Bundle “Set Up”
Patient placed
on ventilator
Elements of
the bundle
ordered
Elements of
the bundle
accomplished
Patient
removed from
ventilator
Segment: Medical ICU With 2 Willing Doctors
Of the elements of the bundle, the
head of the bed elevation is most
commonly not accomplished
Our aim is to achieve 95% or better reliability at
keeping the head of the bed elevated
What is your goal?
At least 95% reliability for the process
Why not 100%?
Process must be capable and reliable
Process must be linked to outcome
14
5/7/2014
8
Will Patients Be Harmed?
Focus on non–catastrophic processes
If process fails, patient will not be harmed within next few
hours
15
Problem Solving
First order Problem Solving
– Fix the problem you see
Second order Problem Solving
– Solve the root of the problem
16
5/7/2014
9
Step 1
17
Getting Started- Let’s Work on Your
Identified Problem
Select subset or segment of population
Select process you want to make more reliable
Develop a high level flow diagram
Identify defects in each step
Select which defect you will fix first
State your reliability goal
– At least 95% reliable
18
5/7/2014
10
Step 2
19
Change Concepts
Simplification
– Are there steps in your processes that can be eliminated?
Standardization
– Best known process to achieve desired results
– But known today-may change with new knowledge or new
context
20
5/7/2014
11
Definition of standard work
“…current best practice, standardized work forms the
baseline for kaizen or continuous improvement. As the
standard is improved, the new standard becomes the
baseline for further improvements, and so on. Improving
standardized work is a never-ending process.”
http://www.lean.org/workshops/WorkshopDescription.cfm?WorkshopId=20
21
How to Standardize
Determine what you want to standardize and why
Ask people who do the work to contribute ideas
Ideas from others
Test-test-test
Measure-measure measure
Do not be afraid of giving up an idea if it does not work
22
5/7/2014
12
Standardize the Process
Who
What
When
Where
How
With what
23
Step 3
24
5/7/2014
13
Why Develop a Back Up Plan?
Very difficult to reach 100% every time with only one
step
Allowing for 80% reliability in first step gives opportunity
to design more freely
You may achieve 95% or better with only the first step-
but is it sustainable?
You should have a safeguard/backup plan in place
25
Examples of Backup Plan
Checkout after primary care visit
Visual marker for head of bed elevation
26
5/7/2014
14
Standardize the Backup Plan
Who
What
When
Where
How
With what
27
When to move to this step?
Standardization step must be at least 80%
reliable over time
Do not introduce backup plan if the first step is
not reliable
28
5/7/2014
15
When will the backup plan fail?
When it becomes the primary process
When it never finds a defect in the primary
process
When it becomes another function in our busy
days
29
Measurement
Needed to determine reliability of processes
Sampling is useful
Measurement alone does not result in
improvement
30
5/7/2014
16
Measurement
Select 5 to 10 events daily
Plot over time
Display charts
Easy to do
Collected by people who do the work
31
5 Charts/Day Run Chart 32
5/7/2014
17
5 Charts/Day Run Chart 33
How to Help Your Teams
Start with segmentation, flow diagram and
standardization
Standardization must achieve at least 80% before
moving on
If reliability not improving consider testing some other
steps
Use the following grids to coach teams
34
5/7/2014
18
Key Questions To Analyze Testing and
Implementation
Key Question Your Evaluation
Is the connection between goals and process clear?
Is the design strategy primarily vigilance and hard
work?
Has some degree of segmentation been used to test
the design?
Has standard work with testing been part of the
design?
Is a design methodology being used?
Are small tests of change being used in a rapid cycle?
Is data collection rapid enough?
35
Key Questions To Analyze Spread
Key Question Your Evaluation
Have you repeated the small test cycles as you
spread from the initial site?
Is the process of spread dependent on one person?
Has some degree of segmentation been used to
spread?
Has customization been allowed or encouraged?
Is the same team who developed the pilot now
responsible for spread?
Have you shifted your focus from process reliability
to outcomes too early?
36
5/7/2014
19
Take a moment to reflect on the work you
will be doing
What will you incorporate from
this session?
37