The purpose of our project was to evaluate St. Luke's Emergency Room facility design in concert with operational design and track the impact of such designs on patient care.
Using MedModel, we simulated the combination of three different physical
designs with three different operational approaches to make sure there
was a fit between the two. In fact, the hospital actually ended up spending
more money in the redesign than they had planned but clinical capacity
and nursing reach was increased upwards of 40%.
The Emergency Room under study had been redesigned, redirected and added
to over many years. Unfortunately, no master plan was ever developed
that dictated exactly what was to be achieved by each change. Simply
put, new equipment was purchased and new nursing processes and staffing
limitations were implemented without the benefit of a single coherent
management vision. Simulation helped us combine operational redesign
with physical redesign so that the two, often evaluated separately,
were not only compatible but also mutually supporting when brought together.
The first thing we did was identify and document the various problems. Through observation we found that St. Luke's was losing over 35% of their available nursing time, simply because of the way the unit was laid out. In essence, there was no ability to share nursing hours between patients; the problem was lost nursing ability. As a result, waiting times were incredibly long, often averaging in excess of 3.5 hours!
Initially, the emergency room was divided into five nursing units, most
of which were physically separated and out-of-sight from the others.
Recognizing the fact that by reducing the amount of time nursing personnel
were isolated with specific patients, we would be able to increase available
nursing time to do other tasks. We asked the following questions: "What
if we removed the limitations? What if we allowed nurses located in
different pods to share patient supervision and care time?" In response,
the hospital suggested several designs ranging from five nursing units
to two or three. Our idea was why stop there? Let's try to achieve a
design that depends on a single nursing group. In other words, let's
achieve as much sharing as possible. Once we did that, we started seeing
some major decreases in patient length of stay. With the model, we could
quickly test different scenarios very easily. More importantly, if the
model showed improvement in nurse utilization or a reduction in patient
waiting time, management was willing to consider the option…
and some of the ideas were fairly radical.
Finally, we designed a new system for triage which focused on getting
a lot of testing and specimen collection done up front. We called this
redesign, "Triage Plus". Triage Plus depended on using nurse practitioners
as screening agents with the skill and experience to identify and request
more advanced testing in a more timely manner. This advanced classification
system kept 20% of the patients out of the ER. In fact many of them
never left Triage.
What we ended up with was a system of totally de-specialized treatment
rooms, the elimination of all nursing pods, and a greatly increased
and far more efficient triage area characterized by a new way of collecting
and processing samples as well as getting timely, case-based x-rays.
What the design achieved, according to the model, was a decrease in
length of stay in the unit by as much as 40%! The quality of care increased
along with nursing satisfaction as well. The staff was able to do what
they needed to do more efficiently and with greater safety.
A Lesson in Complexity
As an architect, I approach modeling differently than most analysts.
For example, I've found through experience that most design questions
don't require models of extreme complexity and overwhelming detail.
What they do require, however, is an initial vision of what needs to
be done and what members of the staff have experience with respect to
possible improvements. Combining each option with its parallel design
is the real task and the one that produces the greatest benefit. In
short, keeping things simple makes modeling easier and far more productive
and generally doesn't sacrifice one bit of validity!
-Victor Lange, ProModel Corporation