The Curse of the Empty ED Waiting Room
By Phillip M. Stephens, DHSc, PA-C
We measure things in medicine and pretend the numbers have meaning. One problem is a lack of measurement sophistication. Another is a lack of basic understanding of biological modeling. Let’s apply this hypothetically.
Walking through the Fast Track area of the Emergency Department one day, we see a nursing assistant pick up six charts in order to retrieve the six Fast Track patients waiting in the lobby. A nurse reminds her that Dr. Discrete Time (his name reflecting his modeling worldview) is working. Without further discussion, the nursing assistant puts half the charts back, retrieving only three patients from the lobby instead of six.
Asked what just happened, the nurse explains that Dr. Discrete Time typically accuses Fast Track of not working hard enough if he sees the lobby empty or very few patients in Fast Track rooms. He will complain or pull their staff to help him in other parts of the department if he momentarily sees this on the computer screen, slowing their flow even further. She explains that when he works, they have to work slower so he will think they are working faster.
But she goes on to explain that when Dr. Continuous Time (again named, cleverly, for his modeling methodology) is working, they do the opposite. They try to empty the lobby and turn the rooms over as fast as possible because Dr. Continuous Time thinks Fast Track is not doing their job unless the opposite occurs. He views flow being optimal if they can keep the lobby and rooms empty – the opposite perception of Dr. Discrete Time.
As a result, she says, when Dr. Continuous Time is on duty, they work faster. He never pulls staff and sometimes sends help if they get backed up, which they try not to let happen as things seem to run smoother when Dr. Continuous Time works.
She then provides more than anecdotal evidence by producing a patient list of a day when Dr. Continuous Time was working. The lobby and rooms stayed empty all day, yet they saw 30% more patients than when Dr. Discrete Time worked with the same daily volume.
Two separate processes are occurring in this scenario. A dichotomy between modeling and human behavioral economics is at work. Both must be considered. Discrete time modeling, as you have guessed, examines snapshots of time. Continuous time modeling examines a broader swath of experience. Discrete time modeling is a photo while Continuous time modeling is a video.
But neither is perfect. Neither accounts for acuity or variation, which are additional complicating variables. Sometimes this means seeing fewer is seeing more, making Dr. Discrete Time right on occasion–but that’s another story.
Nevertheless, how departments decide to model has an effect on human behavior as demonstrated by Dr. Discrete Time and Dr. Continuous Time. Unintended consequences occur when suboptimal models are used, as human behavior has a way of working around traditional modeling methods. Models are static constructs. Human behavior is dynamic. This is the very problem.
We also must remember that we call it a “model” because it isn’t reality. If it were reality we wouldn’t call it a model. We’d call it…reality.
Although we assume that modeling emergency department flow and behavior is a simple process, in reality, pioneering this field of measurement is quite complex. So for now we simply measure what is easiest to measure.
I discussed this dichotomy in greater detail in a recent article in Emergency Medicine News entitled, “Moneyball in the ED.” The article provides a glimpse into new paradigms we need to explore regarding efficiency measures.
As for the differences between discrete time and continuous time modeling, we need to be careful what we ask for, as we just might get it.
Phillip Stephens, DHSc, PA-C, is the associate practitioner site director for Emergency Medical Associates at Southeastern Regional Medical Center, Lumberton, N.C. He is adjunct faculty at A.T. Still University in Mesa, Ariz., and has practiced as an emergency medicine physician assistant for more than 20 years.