Home > The Lean Post> One Idea for Improving Hospital Emergency Room Care
The Lean Post
Sharing how the world is making things better through lean.

One Idea for Improving Hospital Emergency Room Care

by Sami Bahri
June 12, 2014

One Idea for Improving Hospital Emergency Room Care

by Sami Bahri
June 12, 2014 | Comments (9)

I've been thinking a lot about how to improve patient care, in my own organization and in healthcare in general. How can we give patients better care, faster (when they need it most)?

Leveling the workload has been described as the foundation of the Toyota Production System. Much of my career has been dedicated to understanding how to apply leveling to dentistry. And the more I think about leveling in dentistry, the more I think about how it might help congestion in hospital emergency departments.

In Kanban: Just-In-Time at Toyota, Taiichi Ohno wrote: “What we have accomplished was to equalize the amount of work by 'leveling' the peak. What five people could not handle before is now done by two or three.”

By leveling the workload Ohno produced the same output with fewer people. In healthcare, however, leveling is not as simple as in manufacturing where you can produce a car in one location and sell it in another. The healthcare consumer (patient) has to be present when we perform the service (treatment).

In a normal day in my dental practice, we level our schedule to match our staff capacity, without exceeding the patient’s tolerance for delivery time. In emergencies, however, we need to deliver care faster, which causes unevenness in the workload — Mura. We know the antidote to Mura is flexibility in the delivery system.

To create flexibility, we spread the workload among staff and over time. So for us in dentistry (as it would be in any healthcare profession), workload leveling is internal; it deals with employees and with equalizing work distribution among employees. In manufacturing and other non-healthcare services leveled production deals with output—finished products received by external consumers.

What personnel are we talking about? Do we mean spread the assisting work among assistants, front desk work among front desk personnel, and hygiene work among hygienists? Certainly not. That would be siloed thinking. What we’ve done is spread the workload across every person in the office, within the limits of laws and regulations.

How have we achieved this? First, we cross-trained everyone. Second, we made sure people were willing to cross functional barriers. We made this optional at first; any front desk assistant, for example, had the right to refuse helping chair side because that was not in his/her job description. Over time, as people left, we included that flexibility requirement in the job description. Today, every employee moves between departments. Third, we standardized work in the each area to limit training requirements.

And finally, we created the position of Patient Care Flow Manager (“Flow Manager” for short) who redirects other staff members throughout the day to where they are needed most. She uses a paper kanban to fine tune the operation of matching capacity to demand. It’s common practice today to see assistants helping out by answering phone calls when the front desk is busy, or vice versa. What ten people could not handle before is now done by six, and everyone is less stressed.

Can we extrapolate to hospitals? 

In an interview with Chet Marchwinski, “Reflections on 25 Years of Lean,” Dan Jones has said, “Hospitals have two flows, the emergency flow and the planned flow—the elective procedures and so on. And they go through very common steps that were not recognized before as common steps. And the criticality of providing the right test results, the right support, the right therapies, the right drugs at the right time, at the right place, that was not there before and now it’s been raised to the surface.”

Dan’s comment about only “two flows” in a hospital is a revelation that, to me, seems fundamental to healthcare improvement. In today’s hospitals, the two flows belong to two different departments, with more complaints about congestion in the ED. Hospitals keep trying to resolve the problem with different degrees of success.

Suggestion: Merge two work flows by leveling the workload

Mark Graban, a lean coach who has worked with many healthcare organizations, has written: “Are we taking advantage of opportunities to level load workloads or processes where we can… or at least better match staffing to demand in areas where demand is not shapeable (like emergency departments)?” It’s an important question!

To match staffing to demand, hospitals might consider redeploying staff from other parts of the hospital. Our positive experience with leveling, with workload reduction, cross training, and a flexible workforce, suggests that the solution in emergency department (or any other department) congestion might be found outside the bounds of the department. Lean teaches us that we find effective countermeasures to problems when we think across the organization.

Resources, especially personnel, should be able to move between the hospital and the emergency department as needed. The load should be shared by the whole hospital. Leveling may work differently in healthcare than in manufacturing, but it doesn’t mean it’s not possible. To do it, cross-training between the emergency department and the rest of the hospital would be a requirement. Another requirement would be to standardize processes between floors and departments in order to minimize training needs and facilitate personnel transfer.

After reading Dan Jones’s work and running my own experiments in dentistry, most with great success for our organization, I’m curious as to which hospitals and healthcare organizations would consider this new way of working. Which healthcare organizations are already practicing this way?

The views expressed in this post do not necessarily represent the views or policies of The Lean Enterprise Institute.
Search Posts:
Was this post... Click all that apply
17 people say YES
26 people say YES
16 people say YES
8 people say YES
Related Posts
9 Comments | Post a Comment
Daniel Jones June 16, 2014
3 People AGREE with this comment
I just caught up with this Post. Against most peoples' expectations emergency demand turns out to be very stable and the variation highly predictable, so levelling work in the ED is about designing the staff rostas to match capacity with demand. As in any industry we create the experienced peaks and troughs in workload, not the customer or patient! But speeding up the flow of patients through the ED only results in frustration as they wait for a bed to be admitted into. Which is why a lot of lean initiatives in ED run into the sand. Only when the hospital addresses the delays in discharging patients when they are done do these frustrations ease and ED initiatives come into their own. It turns out that the key to levelling from a whole hospital perspective is levelling the workload to discharge patients in a timely fashion. All of which can be undone by bed management working to keep patients in a bit longer because they are paid by the bed day! Hospitals are complicated places with lots of hidden and often perverse incentives

Reply »

Sami Bahri June 18, 2014


Thank you for your valuable comments. Your last statement  about the bed day pay reminds me of Dr. John Toussaint’s call for a better compensation system when he wrote: “We must find a way to reward and encourage more efficient, better-quality health care, and that’s what we will get. “

Reply »

Mark Jaben June 17, 2014
1 Person AGREES with this comment
Sami and Dan,

As an ED doctor, I have found that the patient arrivals do seem predictable when viewed over spans of 24 hours and longer.  Unevenness, however, comes from  shorter inter hour variability and often follows a 3 hour pattern. This is what produces the 'experience' of ED staff that is not appreciated when lookiing over those longer time spans. If the ED cannot get caught up from one rush before the next ladders on, then it gets way behind leading to the gridlock we see. In-hospital discharges certainly apply for admitted patients, but this is only roughly 25% of ED volumes.

A focus on levelling the work during these inflows/ overload situations and restoring capacity before the next onslaught is the key.
And that takesautonomy and decentralized, real time decision making based on overall department and hospital guideliens.

Reply »

Sami Bahri June 18, 2014
1 Person AGREES with this reply

It is that autonomy and decentralized, real time decision making that I’m hoping to get. On the practical side, though, we will also need cross training and standardization—which can take a lot of time and efforts. In any case, I think (and you seem to agree) that the whole hospital should absorb the waves of variation in the ED, whether they come from the 25 percent of patients that are admitted, or the 75 percent that are not.

Reply »

Daniel Jones June 23, 2014
Yes the mid-morning peak always trips the team up and they feel they never catch up for the rest of the shift. But it happens every day so it is possible to think of various countermeasures to address it, depending on your situation.

Reply »

Mark Darvill June 17, 2014
1 Person AGREES with this comment
Most articles on healthcare tend not to address the issues that we experience so I was delighted to read Sami's piece on flow and levelling.

We are a combined General Practice and Accident & Medical Centre in New Zealand.  Approx half of our daily arrivals have booked appointments with the GPs, the other half arrive unannounced, simply walking in when it suits them.  Patient presentations are evenly balanced between chronic disease management and acute medicine/accident - traditionally different skill sets for clinicians.

As Mark explains above, the walk-in volumes are predictable on a macro level but impossible to predict on a shorter time frame.  All walk-ins require triage within ten minutes of arrival so we have trained our entire nursing team to be able to triage and use queuing theory to determine when additional nurses are needed to move into triage.  We also employ 'virtual rooms' which are briefly unoccupied clinical areas that can be put to use as triage areas when demand exists.

Knowing how the walk-in demand varies on a macro level is useful but you cannot roster a doctor on for two hours, they need (demand) a longer duty period, so we vary our appointment templates to give gaps between appointments in which GPs can 'pick up' walk-in patients.

We're not there yet, variation in demand can still catch us out, and I'd be delighted to read more on others experience in dealing with the issue.

Reply »

Sami Bahri June 18, 2014

I find your comment about leaving gaps between appointments very important. We leave some gaps as well, and we make our appointments slightly longer than expected. We base those gaps on the takt time of repetitive procedures.  The takt time is extracted from what you called the macro level of demand.

Reply »

MArk Jaben June 19, 2014
Well, this is very interesting. Sami, I grew up in Jacksonville and my parents reside there still. Mark, I practiced in Taupo for some time.

Sami, the rest of the hospital could pitch in, except they don't necessarily have the capacity to break free to assist either and are staffed just enough to keep up with their flow. Hospitals have yet to figure out what manufacturing has- that running your people at less than 110% capacity has benefits.

When I have seen succcess with this unpredictable variation, it is based on the realization of the 'real' cycles, and the support/guidance on how to prioritize the work that is present at the time combined with real time problem solving as to what the obstacle is at that moment- there are so many possible ones- all done by the available staff in the ED.

In other words, there is often enough capacity in the existing staff to get caught up if deployed to the real issues of the moment, but an over all structure to support this real time problem solving among the staff present is lacking.

Good and appropriate visual management to alert staff to those circumstances; joint purpose and problem solving; management/leadership support of this problem solving are all needed. What's really lacking is the credibility and trust among all the parties to focus and collaborate here.

Reply »

Mark Graban June 19, 2014
Sami - thanks for pushing people to think differently. The "way we've always done it" doesn't have to be the way we'll always do it.

I don't see this as a matter of "level loading" -- heijunka would mean trying to shape demand to change when people arrive (we can do this for a clinic or elective surgeries, but not an E.D. -- except maybe for the people who come for primary care).

Toyota and many other companies level production since it's harder to shape demand. They produce the same amount each day, basically, because it's a lot of work to change the cycle time of a truck line from 60 seconds to 55 seconds. It's easy to add an hour or two of overtime... and it's easy to (gasp) build inventory! There's inventory in the supply chain to level production and still sell a different amount each day.

I think the things to explore in an E.D. are:

1) Properly staffing to demand (setting shift patterns to match that predictable pattern of patient arrivals)

2) Utilize cross-training within the department, where it's right and makes sense, or "float" people from other parts of the hospital (again, when it makes sense).

I see more of the first than I do the second. I'm not sure why.

Reply »

Search Posts:
"Too Busy to Walk the Gemba"
A Reflection on Competition
Ask Art: How Should We Staff and Run Kaizen?
Please include links as plain text URLs only. Do not copy and paste directly from a web page or other document. Doing so may pick up additional HTML that will not function here.
URLs will be converted to functioning links when your comment is displayed on the site.
Here's an example:
See this article for more details: https://www.lean.org/whatslean