What I learned at the ATP Course Spring 2011
[An email to my fellow classmates at the Advanced Training Program in Healthcare Quality and Delivery run by Brent James, MD at the Intermountain Healthcare Institute for Quality Improvement.]
I hope you all arrived home safely. It has been a real privilege
taking this course with you. After Tommy’s speech on Wednesday, my
husband asked me what we learned in the class that caused everyone to
nod their heads when Tommy said this course had changed his life. I
submit to you my list of 13. (I have triskadecaphilia and just
couldn’t narrow down to 10).
What I learned at the ATP course:
1. Human memory is limited and processes are complex.
This will change your life. We can only hold about 4-5 pieces of data
in our short-term memory. Most simple processes are far more complex
than that. This is probably why most patients with well-trained
physicians across the country have such dismal rates of aspirin
post-myocardial infarction recommendations in the absence of a system.
Imagine, for example, that you find yourself making pizza with your
husband, twin two-year olds and your baby on a weekend off. You put
flour on the pizza paddle, stretch the dough, add tomato sauce,
mozzarella, basil, roasted red peppers, sautéed mushrooms, and salami.
As your husband goes to put it in the oven, he asks if you added the
parmesan already. You did not. He does. Halfway through the ten
minute baking time, you realize you both forgot the olive oil. The
next pizza you stretch the dough, add tomato sauce, parmesan, and the
rest. As your husband tries to put the pizza in the oven, he’s unable
to get the dough to slide off the pizza paddle and onto the stone.
You forgot to put the flour on the bottom. The blob of pizza
ingredients and dough eventually is pushed onto the stone, and as you
watch the mozzarella sizzle on the stone there are plenty of words to
go around regarding whose fault this is. You both still forgot the
That evening, you educate your husband about process complexity and
system errors vs. bad apples and personal failure. You show him that
even if you do each of the ten steps required perfectly 95% of the
time, this leads to a perfect pizza less than 60% of the time.
The following weekend, your husband adds olive oil to the tomato sauce
and he is dedicated to flour management pre-pizza assembly because a
fault in that particular step has such severe consequences for the
rest of the pizza.
2. Taylor vs. Deming and human nature.
We all want the people below us to just do as they’re told (Taylor)
and the people above us to just leave us alone and let us do our job
(Deming). Even Brent James.
3. Appeal to values and motivation.
If you win people over, they will do anything for you. Make them be
part of your team and they will have your back. Assume everyone is
hard-working, smart, and wants to do what’s best for the patient/the
world (Deming). Then let them figure out how to improve the process.
4. Stories trump statistics. Relationships trump stories.
(A single death is a tragedy. A million deaths is a statistic. –Stalin)
We might think we are, but are not actually influenced by data. All
data is suspect and processed in ways we hardly understand. We trust
the data only in so far as we trust those showing us the data: the
author, the journal, the academic reputation of the university in the
resume. Everyone tells stories, you have to trust that they tell a
A version of this is perennial question: “Doc, what would you do if
this were you?” The patients ask someone they trust, someone they
perceive to care about them how to interpret the data.
This, in part, is probably why people, patients, resist computers and
electronic decision support tools. They often prefer an incompetent
human to a competent machine. We simply can’t adequately evaluate
whether to trust the machine as well as we believe we can evaluate
whether to trust a person.
5. I’m a husky. I need a statesman, a coordinator, and a technical
person to help make things happen.
a. The husky wants change to happen. Achievement is based on
completion. “Salmon is frozen? Just eat it.”
“Your feet hurt? Just keep going.”
“You’re tired after a week of work and life with three under three?
Just suck it up and let’s go ‘skiing’ with the kids.”
Someone, a golden retriever coordinator/facilitator, needs to check in
with the team.
“How many of you throw up after eating frozen salmon?”
“How many of you have bleeding feet?”
“How many of you fall asleep at 7pm from exhaustion? I mean parents,
not children, of course.”
b. You can be a husky, statesman, facilitator/coordinator or
technical advisor depending on the issue. Similarly, you can be an
early adopter, average adopter, or laggard depending on the specific
6. “In theory there is no difference between practice and theory. In
practice there is.”
7. Management is like the drunk looking for his keys where the light
is, not where he knows he dropped them.
And its corollary: “Beware of bling.” – Lee Pierce
Bling is fancy graphs and pseudoscience. If we flash numbers at you
with colorful graphs and fancy statistics calculated to the nth
significant figure, which you can’t possibly understand, they must be
right. Consultants show executives nice analyses with large bubbles
around cardiovascular disease and obstetrics.
“That’s where you’re spending your money. Focus on those, and ignore
the rest for decades.”
“But at some point we’re spending all our time improving heart failure
management for octogenarians and ignoring asthma management in
children,” observes the pediatrician.
“To take that into account, we should analyze care by episodes. This
analysis is too hard and too expensive, so nobody does it. The
biggest bubbles in our analysis are cardiovascular (percutaneous
coronary angioplasty!) and obstetrics. Hospitals/executives should
focus on those.”
That’s why it has taken/will take decades for some trivial request for
IT improvement from the pediatricians to be implemented. For some
things you need a system change.
8. No improvement is too small.
During the last week of the ATP course one of the assistants started
writing “OPEN” on the tops of milk bottles that were opened but still
sitting in the bucket of ice. We are not three-year olds. Most of us
only drank milk with our coffee. Now instead of having most of the
bottles opened and half full, there were only one or two bottles that
were opened. No matter how little influence I think I may have, I can
at least improve something small within my control.
80% of your processes should be in control so you can improve 20%.
Working on everything all the time is not improvement. It’s chaos.
10. Perspective: To improve you need front line knowledge and step
back perspective. Everyone thinks they have both when they usually
Look for evidence to identify the problem you think you understand.
You’ll be shocked at what you find.
11. Surrogate measures and goal displacement.
Patients care about hospital cleanliness, staff courtesy, and food
texture. These are surrogate measures for patient perception of
quality. I can’t possibly evaluate how much my doctor knows, but if
she’s successful (nice office, courteous staff, etc.) then she must be
competent. To alter patient perception of quality, you can improve
the process (hard) or change the surrogate measures (easier). Nobody
said doing the right thing was easy.
12. To make lifelong loyalty, first screw up, and then really make it
right. It’s hard to think someone’s exceptional unless you’ve been in
a challenging situation with them that allowed them to rise to the
occasion. As the maxim in politics, it’s not the sin, it’s the
cover up that gets you.
13. Stories trump statistics. Relationships trump stories. Worth repeating.