Patient Safety with Six Sigma, Lean, or Theory of Constraints

Text-only Preview

Patient Safety with Six Sigma, Lean, or Theory of Constraints

Joseph L. Quetsch


Understanding between healthcare clinicians and
This presentation is a synthesis of well known
engineers is facilitated by presenting three improvement
practices and publications. There is ample anecdotal
methods compared with Patient Safety. The methods
evidence in the literature and amongst practitioners of
discussed are Six Sigma (6σ), Lean, and Theory of
process improvement that each of the systematic methods
Constraints (TOC). Patient Safety is a dominant paradigm
of improvement yields desirable results efficiently for
in healthcare today. Effective organizations will first adopt
correct healthcare opportunities. The presenter is unaware
a proven method for improvement, and then apply tools
of randomized controlled research comparing
correctly with understanding.
improvement systems in healthcare.
Improvement methods are contrasted along aspects

including 1. original purpose 2. implicit goals 3. focus of
attention and effort 4. social values 5. size and types of

problems approached 6. characteristics of the organization
Requesting a tool for a quick and easy repair of a
to use the method and 7. typical tools and artifacts.
process may be a symptom of misunderstanding. The
The handout is a table with columns {Six Sigma,
request may reveal an assumption that correct things are
Lean, Theory of Constraints, Patient Safety}, and rows
happening correctly and just need a little tweak. As
{each aspect enumerated 1-7 above}. The presentation will
healthcare practitioners are reluctant to treat only
highlight similarities and differences using healthcare
symptoms and eager to intervene at root cause or at
examples. The reader will have a framework useful for
intermediate mechanisms, so engineers are reluctant to
quickly informing a healthcare improvement team at the
treat apparent symptoms and eager to know the root source
point of selecting an approach or considering a tool.
and intermediate determinants of behavior for the relevant


Common ground
A typical request of the engineer from a healthcare
Each improvement method is a way of seeing the
client is to show a tool, perhaps without context or method.
world. Each method determines perception initially and
Healthcare clients may be inpatient with process
the pursuit of evidence. Each method has a style of
improvement engineers, or vice versa, while incompatible
hypothesis formation and a preference for hypothesis
paradigms are not well recognized.
testing (Kuhn, 1996). We have skipped the arguments that
Opportunities in problem form may be approached
“Results are determined by process” and only by process.
successfully by more than one method and proven methods
We here axiomatically know that to improve results we
are rarely mutually exclusive. In practice, professionals
must improve process. We agree on a sufficiently broad
“borrow” techniques from other methods or solve each
notion of process as some set of the determinants of the
part of a project with the appropriate approach for that
result. The methods discussed here are ways of
component. When the team comes down to applying a
understanding processes.
tool, the team may be more successful where the
Not all clinicians understand or agree with the
assumptions contained in the tool are understood. Tool
engineers’ notion of process. Some good faith
use is action, action occurs in context, thus proven tools
improvement efforts have been harmful. Clinicians have
were proven in context. Random use of tools without
learned that some bad faith attempts to take control or
regard to the method within which the tool evolved may be
money came wrapped as improvements. Our laws or
random behavior.
customs do not restrict who can pretend to be an
This presentation contrasts four paradigms for
improvement expert. Besides actually being competent at
improving healthcare. A detailed understanding of each
one or more proven methods, we must successfully
method or detailed understanding of American healthcare
communicate an understanding of our engineering
today is beyond the scope of this presentation. The
sciences. We may not desire an identical view of some
audience is assumed to be generally aware of the
phenomenon, but we desire to comprehend the other
approaches discussed and familiar with problems in
professionals’ perception.
Page 1 of 3

Myths about improvement sciences abound.
effort scientific and the Lean method reproducible
Improvers are often believed by other improvers to have
(Womack & Jones, 2003).
blinders or lack common sense. Few scientists abandon
Theory of Constraints (TOC) chooses to see the world
common sense. While the descriptions below discuss only
through accountants’ reports designed to show real
special strengths of a method, the reader should not infer
operations, before accounting games begin. (Corbett,
the absence of other views within a method. For example,
1996) Throughput accounting directly opposes creative
Six sigma people are perfectly capable of improvement
accounting. Traditional tax or finance or cost accounts
without statistics, and every type uses some statistics. All
remain necessary but only throughput accounting is used
methods include coaching and personal growth. All
to support operations. Throughput accounting judges local
methods value community and non-profit social values.
management decisions according to impact upon global
This paper intends to contrast the methods.
cash flow by following the logical consequences of a
Successful improvers are adept at defining the
management decision through each downstream process
problem or opportunity in helpful ways. Helpful for
all the way to the global cash flow (Goldratt, 2004). Then
developing understanding of the problem and achieving
the accountant logically backward-chains from global cash
the opportunity. The same problem, perhaps even
flow to assign a dollar value to the local management
equivalent notions of understanding, might be expressed in
decision. TOC attempts to maintain cause and effect
6-sigma as Y = F (X1, X2, X3 …), in Lean as 5 Whys, in
relationships throughout the enterprise, exposing the costs
TOC as drum-buffer-rope and in Patient Safety as FMEA
of local optimization. The name ‘theory of constraints’
(Failure Mode and Effects Analysis).
derives from the aspect that once the steps are know, only
Successful improvers are adept at discovery and
the bottlenecks receive management attention or
measure. Systematic methods provide styles for
investment. TOC’s narrow focus on throughput and
generating hypothesis and testing. Methods tend to define
bottleneck may make TOC less threatening compared to
problems such that the tools work there. Methods also
other improvement methods. Patient throughput or health
tend to define the project management within which tools
status can be improved using the same thinking TOC has
are used.
developed for global cash flow. Clinicians and line

managers may use throughput accounting to illustrate the
wisdom of decisions not supported by arbitrary or creative
The reader may wish to explore Table 1 before or after
the following remarks. Table 1 title: High-Level Side-by-
Patient safety chooses to see the world through errors.
Side Comparison of Dominant Improvement Styles.
(Institute of Medicine (U.S.). Committee on Quality of
Six sigma chooses to see the world through statistics.
Health Care in America., 2001) Viewing the world through
Six sigma is far more than statistics, but reject a statistical
errors presents several hazards or hurdles for process
view of your world and you must reject Six sigma.
engineers. If errors are defined as exceptions, then looking
Statistics are an amazingly powerful way to see what
at the errors is not looking at the ordinary routine
human eyes were not designed to detect. Much of 20th
processes. If errors are defined by a final end result, then
century science confirms the utility of statistical
there are many plausible processes and no method for
approaches. Few scientists reject statistics, but many
replacing the black box generating final results. If errors
prefer more tangible methods. To be content with Six
are defined as deviation from a professional standard, then
sigma requires faith as well as competence in statistics.
resources may be consumed validating the abstract
Six sigma may have a tendency to ignore what does not fit
standard or quibbling over the deviation. Viewing the
a solvable statistical function. Six sigma is associated with
world through errors may be inconsistent with viewing the
an inclusive calculation of return on investment, prescribed
processes that are routinely producing desired correct
team composition and prescribed behaviors. Viewing the
results. The view through an error may impair perception
world through mathematical functions is both rigorous and
of a need for process redesign if a patch remedy looks
curious. (Pyzdek, 2003)
good for that error. (U.S.). Committee on Quality of
Lean chooses to see the world with human eyes
Health Care in America., 2001)
directly on the unfiltered, possibly undistorted process.
Patient safety makes good use of a scientific body of
Lean observes directly rather than via functions or reports.
knowledge specializing in aspects of systems leading to
Humans unconstrained by functions or filters may have
failure or errors. Patient safety practitioners who use the
fewer blinders for the first looks, but the need to see and
safety science tools for more than diagnosis or discovery
communicate scientifically requires disciplined use of
are likely to apply Six sigma or Lean methods while
checklists and the inclusion of measure and category
engineering replacement systems. TOC arguments may
before the view is considered adequate. Statistics are used
serve to internally sell the new safer systems. Other
for hypothesis testing. Lean may have a tendency to ignore
patient safety practitioners may remain in a comfort zone
what is not apparent to direct observation. Disciplined
of professional traditions and keep original processes with
adherence to established Lean practices may keep the
a safety patch.
Page 2 of 3

A common belief in patient safety circles is that no
Zimmerman, B., Lindberg, C., & Plsek, P.
prescriptive method is necessary- improvement will
(2001). Edgeware: insights from
emerge from sharing ideas. Idea sharing as a method is
complexity science for health care
consistent with a view of healthcare as a complex adaptive
system that is best managed without prescriptive control
leaders (second ed.). Irving TX: VHA.
(Zimmerman, Lindberg, & Plsek, 2001). Patient safety

does not require a particular view of process or provide a

method for pursuing improvement. A systematic,
Biographical Sketch
coherent, guiding scientific paradigm for improvement

may someday emerge from the patient safety movement.
Joseph L. Quetsch MD, BS Computer Science, is

principal of Care Process Remodel Associates, LLC,
assisting healthcare organizations to change efficiently and

correctly using mainstream sciences.
Problems and opportunities should be addressed with

a scientific method, which incidentally uses tools. Use of
Next page is Table 1: High-Level Side-by-Side
tools out of context may be random behavior.
Comparison of Dominant Improvement Styles, which is
Each method is derived from its own view of the
the handout for the presentation.
world. Given that view, the tools and hypothesis and

behaviors make sense and yield consistent results.

Improvement methods were contrasted to facilitate

selection. High level characteristics of the methods were

presented as rough concepts to facilitate matching the

organization and the opportunity with a method. A table
compares side by side aspects of the methods.


Corbett, T. (1996). Throughput Accounting.
Goldratt, E. M. (2004). The goal : a process of
ongoing improvement (3rd ed.).
Burlington, VT: Gower.
Institute of Medicine (U.S.). Committee on
Quality of Health Care in America.
(2001). Crossing the quality chasm : a
new health system for the 21st century
Washington, D.C.: National Academy
Kuhn, T. S. (1996). The structure of scientific
revolutions (3rd ed.). Chicago, IL:
University of Chicago Press.
Pyzdek, T. (2003). The Six Sigma handbook : a
complete guide for green belts, black
belts, and managers at all levels
and expanded ed.). New York: McGraw-
Womack, J. P., & Jones, D. T. (2003). Lean
thinking : banish waste and create wealth
in your corporation
(1st Free Press ed.).
New York: Free Press.
Page 3 of 3

High-Level Side-by-Side Comparison of Dominant Improvement Styles
Quetsch JL, Society for Health Systems, Orlando, February 18 2007
Theory of Constraints
Patient Safety
1 Original Purpose
Maximize reliability
Maximize value
Maximize profit
Minimize bad results
2 Goal or Theme Eliminate variation
Eliminate waste
Reality accounting
Eliminate error
3 Focus of effort
Statistical correlation
Process visibility
Optimal pace
Peer review
4 Social Values
Uniformity in all things.
Adapt to do more with same
Maximize return on investment
Judgment, autonomy and
Statisticians drive
Grass roots driven
Throughput accountants drive
Care professionals drive
Adhere to the defined, decreed Agree to perform the current
Think through implications, followAdhere to professions, each
best process
best common process
to real cash flow consequences profession knows best practice
customer voice factored in,
customer is only judge, value is global cash flow shows value. no uniform measure of value.
statistician judges value
what will customer pay?
Measures are typically ratio
indicators, actuarial risk, quality
of life.
5 Types of Problem Statistically tractable problems Whatever is discovered, visibility Optimization. Each local unit
Incident reports, surveillance
(large counts of similar items)
enables common sense.
optimized for global result.
Large problems, prefer $500k
No contribution is too small.
Single minded focus on the
Politically correct problems
current constraint
6 Organizational fit
Large organizations, central
Any size organization, delegated Any size organization, central
Any size organization, healthcare6
control style, operations chiefs control, line mangers make
view for global information,
professionals control, decisions
make decisions
accountants make decisions
7 Typical artifacts
Statistical graphs, control loops Value stream maps, project
Stories, metaphor, chains of
Policy statements, inspection
reason, chains of producers
Theory of Constraints
Patient Safety

Document Outline

  • Patient Safety with Six Sigma 2006c18.doc
  • SHS_2007_handout_Quetsch_2006c18submitted.xls