We are all engaged in a hospital-wide a system of
patient flow or patient care. We are each part of the
whole. The emergency department is connected
to the ICU. The ICU is connected to the OR. The
discharge and discharge processes are connected
to our admission capabilities and capacity. It’s
like the “Dry Bones” song you learned as a child,
“The foot bone’s connected to the leg bone, the
leg bone’s connected to the knee bone, the knee
bone’s connected to the thigh bone” and so forth.
Overall flow, or “the system,” can only be improved
by applying several key strategic concepts to these
disparate but equal parts.
April 2024 ONCOLOGY CARTOON by DR KANHU CHARAN PATRO
Key Strategies for Improving Hospital Flow
1. WHITE PAPER
Key Strategies for Improving Hospital Flow
BY:
Kirk B. Jensen, MD, MBA, FACEP
Content
The ED is a part of a hospital-wide
system of patient flow
Every system produces the results
it is designed to produce
Key strategic concepts to improve
patient flow
You cannot optimize a system by
optimizing just one part of the system
3. KEY STRATEGIES FOR IMPROVING HOSPITAL FLOW
Introduction................................................................................................................................................. 2
Every System Is Designed to Produce the Results It Produces....................................................................... 3
Demand-Capacity Management................................................................................................................... 3
Real-Time Monitoring of Patient Flow......................................................................................................... 4
Forecasting................................................................................................................................................... 4
Queuing Theory........................................................................................................................................... 5
The Theory of Constraints............................................................................................................................ 6
Managing Variation...................................................................................................................................... 7
The Appreciation of a System....................................................................................................................... 8
In Summary................................................................................................................................................. 8
About the Author......................................................................................................................................... 9
Contact..................................................................................................................................................... 10
4. INTRODUCTION
We are all engaged in a hospital-wide a system of
patient flow or patient care. We are each part of the
whole. The emergency department is connected
to the ICU. The ICU is connected to the OR. The
discharge and discharge processes are connected
to our admission capabilities and capacity. It’s
like the “Dry Bones” song you learned as a child,
“The foot bone’s connected to the leg bone, the
leg bone’s connected to the knee bone, the knee
bone’s connected to the thigh bone” and so forth.
Overall flow, or “the system,” can only be improved
by applying several key strategic concepts to these
disparate but equal parts.
2 |
5. KEY STRATEGIES FOR IMPROVING HOSPITAL FLOW
Back at that time when I was training, I was an
idealistic emergency physician managing an inner
city ER in downtown Los Angeles serving the poor.
I stumbled across an article written by Don Berwick
that said “every system is perfectly designed to
produce the results it produces.” It changed my life.
(Berwick MD, Donald M, Continuous Improvement
as an Ideal in Health Care, NEJM, January 5th, 1989,
p50).
I started looking at the ER as a set of processes, as
a system, with inputs, throughputs, and outputs. I
had to change those processes if I wanted different
results or outputs, because they were often “perfectly
designed to produce” some of the very results that
neither I, my co-workers, nor my patients liked.
As Thomas Edison has pointed out, “Discontent is the
first necessity of progress.”
So the question became: Are waits and delays
inevitable? We think the answer is no.
There are several key strategic concepts which can be
applied to improve patient flow:
•
•
•
•
•
•
•
Demand-capacity management;
Real-time monitoring of patient flow;
Forecasting service and service demand;
A basic understanding of queuing theory;
The theory of constraints;
Managing variation; and
Appreciation of a system.
Demand-Capacity Management
What should capacity look like to guarantee quality
care? Here’s a profile of demand and capacity.
NUMBER OF PATIENTS
Every System Is Designed to
Produce the Results It Produces
TIME
The horizontal axis is time and the vertical axis
is the actual number of emergency department
patient by hour. In the ED this often seems totally
unpredictable, however, if you look closely, you will
notice there is lot of white space above and below the
graph and a line drawn through the middle of the
curve begins to approximate predicted demand by
hour of the day or day of the week.
How do we match demand and capacity? You need to
predict demand based upon your historical data. You
need to match your service capacity to your patient
demand. You need daily patient flow predictions by
hour of the day and plans to service those predictions.
You also need to implement a real-time dashboard
for key operational cycle times and then measure,
monitor, and service those cycle times. You need to
do this while respecting the desires, concerns, and
goals of your people, of your team. Demand-capacity
management is critical. We cannot plan on or even
hope for our capacity routinely meeting our demand
without planning, foresight, and action.
|3
6. Real-Time Monitoring
of Patient Flow
All by setting up a system where units could grade
how busy they were, put a temporary halt to new
admissions while they managed the patients they had,
McDonald’s does it. They have cameras they
nicknamed “Hyperactive Bob” that monitor incoming
customer traffic. They use software to predict what
the “soccer mom in the van with three kids” is going
to order. They can get five to ten minutes upstream
on predicting what their demand is going to be.
McDonald’s reduced their waiting time and cut waste
in half using this process.
monitor hospital-wide patient flow, and disseminate
this information in real-time to all of the key
departments and personnel.
Now imagine a busy freeway. It’s a dark Friday night
and you’re driving 75 miles an hour on this busy
freeway. Now, turn off the dashboard and headlights.
How long do you think you could stand it? How long
could you maintain your sense of safety and mastery?
This is what we do every day in our hospitals; we
show up, our real-time dashboard is not available, our
headlights are turned off, and then we brag about how
well we can manage the chaos. There is a better way.
You need an inpatient flow dashboard. It can be grease
board, pen, or electronic, but you need some way of
monitoring flow in real time.
Forecasting
Do you look at the weather forecast before you leave
home? Forecasting plays a significant role in your life
on a daily basis and ought to play a significant role in
your life within the hospital. You can predict patient
flow. You can predict unscheduled arrivals with about
80 to 85% accuracy. You don’t know their names but
you do or should know who’s coming, what’s coming,
why they are coming, and what set of resources they’ll
need.
How many Friday nights does it take before you
decide your next Friday night is going to be different?
How many flu seasons does it take before you decide
your next flu season’s going to be different? Patient
flow is predictable.
Here’s what happened at Luther Middlefort Hospital
once they introduced a hospital-wide intranet
dashboard and a system for capping patient flow to
units that were overwhelmed:
• Increased patient throughput, resulting in
increased revenue of about $200,000 per month;
• Increased the percentage of patients who were put
into bed within one hour from 23% to 40%;
• Reduced ED diversions from 12% to 1 to 2%; and
• Decreased the overall number of open nursing
positions from about 10% to 1%.
4 |
0:00 1 :00
2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 1 0:00 1 1 :00 1 2:00 1 3:00 1 4:00 1 5:00 1 6:00 1 7:00 1 8:00 1 9:00 20:00 21 :00 22:00 23:00
FY2004 Q-1
1 54
1 49
1 20
81
83
79
99
1 53
1 66
269
253
277
235
260
274
268
294
307
332
352
345
299
278
21 1
FY2005 Q-1
1 60
119
1 07
83
71
76
85
1 06
1 56
208
226
230
260
243
260
260
304
286
302
333
287
270
260
1 98
This emergency department patient flow graph
charts arrivals, patient volume, and arrival by hour
of the day.
7. KEY STRATEGIES FOR IMPROVING HOSPITAL FLOW
This is your emergency department. This could
actually be any emergency department in the United
States. The actual number of patients per hour may
vary but the curve is the same.
Here’s a queuing model in action in a MICU:
The key questions are:
•
•
•
•
How many patients are coming?
When are they coming?
What are they going to need?
Is our service capacity going to match
patient demand?
Queuing Theory
Queuing theory is the art and science of matching
fixed resources to unscheduled demand. You are
engaged in a queuing model any time you are in a
system with unscheduled or uncontrolled arrivals.
Whether you’re in a line waiting to register, at
Starbucks awaiting your coffee, at McDonald’s to
order a burger, or serving patients in the ER, that is a
queuing system.
Facts about queuing systems:
• Systems that are serving unscheduled or
uncontrolled arrivals behave in a characteristic
fashion.
• When patient inflow and service times are
random, the response to increasing utilization is
nonlinear.
• As utilization rises above 80 to 85%, waits and
rejections increase exponentially.
The great thing is that at high levels of utilization,
small changes can lead to big improvements.
The horizontal axis is the utilization percentage. At
0%, the MICU is totally empty. At 100%, every bed
is full.
The vertical axis is the rejection rate. If the MICU
is totally empty, the odds of you being rejected for a
requested bed are zero. If the MICU is totally full, the
odds of rejection are close to 100%.
Notice the shape of the curve. As you get to 60%
utilization, it starts to take off. You will note this is
not a linear curve. It rises steeply and the sweet spot is
probably about 80 to 85%.
Operating at 100% capacity or occupancy, in a
system with unscheduled arrivals and/or variation
in service times, is absolutely the wrong way to go.
There is simply no way you can run a queuing system
at 100% utilization. You want to be at about 80 to
85%. In unit after unit where they have been able
to drop down to 80 to 85, or even 90% utilization,
hospitals are better able to handle the inflows and the
variation. As a result, throughput goes up, profits go
up, and healthcare worker satisfaction goes up.
|5
8. The Theory of Constraints
The Theory of Constraints was articulated by Eli
Goldratt. His novel is called The Goal. It’s a quick
read. It’s about a guy who is trying to save his job and
his marriage. He is trying to save his relationship with
his son, and he does it through the understanding and
application of the theory of constraints. After reading
this, you will never ever walk into an ER or a hospital
without thinking about the theory of constraints.
satisfaction went up, but were disappointed to learn
that throughput cycle time, their most important
objective, did not change.
Does that mean that an admission nurse is a bad
idea? Absolutely not. What it means is that in this
particular hospital, the functions of the admission
nurse were not a critical bottleneck to the flow or
process of getting the admitted patient to an inpatient unit.
What the hospital should have done was map out
their process flow, look at the cycle times, and then
experiment with an admission nurse to see if it
reduced the process times. Not only does the theory
of constraints allow you to understand how a system
works, but it also helps you identify which bottlenecks
are irritants and which bottlenecks are essential to the
critical path.
The key principles:
• Patient care is a network of queues and service
transitions.
• An hour lost at a bottleneck is an hour lost for the
whole system.
• Time saved at a non-bottleneck is a mirage.
• Efforts spent improving a non-critical bottleneck
will not improve the overall performance of your
process or system.
Here’s one example: A hospital was trying to reduce
the cycle time for admitting patients from the ER
to upstairs. They thought hiring an admission nurse
would be the solution. They hired four people
and deployed a full-time admission nurse. They
pleasantly discovered that patient satisfaction,
nursing satisfaction, and emergency physician
6 |
Efforts spent improving a non-critical bottleneck
will not improve the overall performance of your
system.
To make matters even more complex, bottlenecks
can jump around. For a certain period of time, the
bottleneck may be a bed. For another period of time,
it may be lab. For another period of time, it may be
the discharge process.
Another example. One hospital found that getting
patients out of the PACU was a problem. Patients
were delayed in the PACU, potentially backing
up the OR. Through further investigation, they
discovered that all orthopedic patients leave on an
anticoagulation regimen. The true bottleneck was
actually getting medication delivered on the day of
discharge.
9. AVERAGE NUMBER OF PEOPLE ON HOLD
20
18
16
14
12
10
8
6
4
2
0
Calls / hr = 29
NOTE:
- Average Call lasts 2 minutes.
- Calls are answered by one person full time
Calls / hr = 28
Calls / hr = 27
Calls / hr = 25
LOW
MEDIUM
HIGH
VARIATION OF CALL LENGTH
Now you are confronted with several possibilities.
One – you can wait to try to speed things up the day
of discharge. Two – you can try to speed things up
the day before discharge. Or three – you can get the
pharmacy to deliver the meds on post-op day one.
They had a just-in-time delivery system that wasn’t
just in time. Patients need this medicine when they’re
discharged. This was prolonging discharge, sometimes
by a whole day. This was a key or critical constraint.
The solution was to move delivery up to day one.
After analysis, rapid-cycle testing and andingprototype
the improved outcome was 43 out of 43 patients
received medication on post-op day one with a simple
process change.
If we have 25 calls per hour with low variability,
almost nobody ends up on hold.
Managing Variation
There are multiple sources of variation in health care.
Are all our sick patients the same? Is every congestive
heart failure patient the same as every other congestive
heart failure patient? Is every AV malformation the
same as every other AV malformation? You know the
answer is a resounding no. We have clinical variability.
We have flow variability. We have variability in our
processes. We have variability in how people work.
Each of these sources of variation is additive. They do
not cancel out each other; they add up.
A telephone helpline is a queuing system. Here is an
interesting example of the challenges we face: Let’s
say, that on average, a call to this helpline lasts two
minutes. The calls are answered by one full-time
person, or a single-server queue. Can the system
handle 30 calls an hour without putting people on
hold?
With high variability and 25 calls per hour, we have
two to three people on hold.
But if there are 29 calls in an hour (high volume) and
high variability, we end up with 18 people on hold.
It is the variation that wreaks havoc with your
systems, your service, and your people.
That is one of several reasons why the sweet spot
for patient flow is not 100% utilization. Because of
variability, we need a slack to handle that variation.
|7
10. Look at the variation in health care patients, processes,
and service delivery and think about what that does
to your systems. Some of the key leverage points are:
focus on operational efficiency, prioritize available
resources, flexed responses, flexible scheduling, predict
demand, smooth demand, and improve the accuracy
of your predictions.
The Appreciation of a System
This is articulated in Deming’s system of profound
knowledge. There are four components to Deming’s
system of profound knowledge:
1. Appreciation of a system;
2. Knowledge of variation;
You can’t have every part of the system busy all of
the time and have a system that’s optimized and
performing perfectly. Some components of the system
need sufficient slack to take up some of the work and
variation of the other parts of the system. You cannot
have every part of the system functioning at 100%
utilization or capacity if you want flow through the
system to be maximized.
The larger the system, the harder it is to optimize but
the greater the total benefits. While it may be difficult
to improve emergency department patient flow, it
may be much harder to improve hospital-wide patient
flow. It is a challenge to improve flow through the
OR, but it may be even more difficult to improve flow
throughout the entire system.
3. Theory of knowledge; and
4. Knowledge of psychology.
In Summary
A system must have an aim. Without aim, there is
no system. The aim of the system must be clear to
everyone in the system. You have to decide what the
aim of your system is. Otherwise, you will have good
people working under significantly different priorities.
For some people, it’s patient safety. For others, it’s
patient satisfaction. For some, it’s patient flow. It
could also be workforce satisfaction or controlling
costs. You must look at patient flow and operations
through a global lens to optimize the entire system.
The key strategic concepts that you need to have a
working understanding and appreciation for are:
demand-capacity management, real-time monitoring
of patient flow, forecasting, queuing theory, the theory
of constraints, managing variation, and appreciation
of a system.
You cannot optimize a system by optimizing each
part of the system.
8 |
“Genius is 1 percent inspiration and 99%
perspiration.” – Thomas Edison
The work may be hard, but the results are gratifying.
The number one reason to get this right is that it is
good for your patients and it is good for your people
who take care of your patients.