What’s Lean or What Works? Centralizing Hospital Functions


Positive Steps

I had a chat with somebody from a hospital recently who asked about following one Lean principle and that leading to a conflict with another Lean principle.

Here was the gist of the discussion… she said that the hospital centralized its scheduling functions for different departments in a single location “because it was Lean.”

I asked, in the most polite tone I could muster, “In what way was that Lean?” (meaning “what Lean principle drove that decision?”).

I got sort of a blank stare in response.

She then said, “Well, we wanted to standardize the way scheduling was done.”

I asked, “Isn't standardization a separate issue from the location of the employees?”

If there is a general Lean principle that talks about centralization versus decentralization, I think it would be to decentralize functions like scheduling and registration. This is, like many Lean rules, a general rule of thumb that applies unless it doesn't make sense to do so (TPS as the “Thinking Production System“).

The old traditional mindset (seen especially in manufacturing) is to have “process centers” or a functional department layout — all of the grinding machines go in this department, all of the types of the turning machines go over there. This type of layout leads to big batches and poor flow.

One brilliant thing about Lean manufacturing is the creation of “production cells” that have all (or most) of the machines required to produce a particular product or product family. So we shift from a functional layout to a product-focused layout that will have better flow and shorter cycle times (and probably better quality as a result of the other two).

In the traditional mindset, the most important thing is the utilization of machines – keeping it high because it's thought that reduces the cost per part.

In a Lean approach, the most important thing is flow – which often leads to lower cost. You need “right sized” machines that fit into the production cells and these machines might be smaller than the big mega machines people put in the old process centers.

An example of putting capacity right in the flow in healthcare would be the hospital I visited in Sweden that put a small O.R. in a cardiac unit. The operating rooms (probably mostly for good reasons) are clustered together in a process center in the hospital.

But back to the scheduling situation… I think it matters less “what is Lean?” as it should be more about “what works?” and “how do we know what works?”

A centralized scheduling department would have some benefits:

  • Higher utilization of people?
  • Being easier to have people shift around as demand for scheduling phone calls was higher or lower for different departments?
  • Easier for the scheduling people to learn from each other, improving their practices?
  • Easier for the schedule function to be managed by a single functional manager?

But, having dedicated scheduling IN a department might have advantages:

  • Better integration of scheduling into the department itself?
  • More accurate scheduling based on learning how long certain appointments or procedures really take?
  • More convenient for the patients if they would otherwise have to walk to a “scheduling department” for some reason?

I think you could still work on standardizing the scheduling process (if that was important) with people working in distributed locations. This is one reason why you still tend to have vertical management silos even when employees are distributed as part of departments or value streams.

The hospital was facing some sort of problem caused by centralizing scheduling. I could guess what they are, but I don't know exactly what the problems are (this was a short discussion we had as part of a larger group). If something's not working, follow the Plan-Do-Study-Adjust process… if they tried something (Plan-Do) and found that it didn't work well (Study), then Adjust rather than being stubborn about what Lean supposedly told you to do.

Do you see similar situations in your workplace?

What do you think? Please scroll down (or click) to post a comment. Or please share the post with your thoughts on LinkedIn – and follow me or connect with me there.

Did you like this post? Make sure you don't miss a post or podcast — Subscribe to get notified about posts via email daily or weekly.

Check out my latest book, The Mistakes That Make Us: Cultivating a Culture of Learning and Innovation:

Get New Posts Sent To You

Select list(s):
Previous articlePodcast #182 – Lean Healthcare in Tanzania – Michael Grogan and Dr. Brenda Dmello
Next articleIs Lean Management a “Slow Idea”?
Mark Graban
Mark Graban is an internationally-recognized consultant, author, and professional speaker, and podcaster with experience in healthcare, manufacturing, and startups. Mark's new book is The Mistakes That Make Us: Cultivating a Culture of Learning and Innovation. He is also the author of Measures of Success: React Less, Lead Better, Improve More, the Shingo Award-winning books Lean Hospitals and Healthcare Kaizen, and the anthology Practicing Lean. Mark is also a Senior Advisor to the technology company KaiNexus.


  1. Yes, absolutely. These discussions about centralization vs. decentralization occur all the time in hospitals. I think your analysis of the pro’s and con’s of the scheduling department question is spot on. I also think you’re advice to use PDSA if something isn’t working is great. This would ensure that over time the hospital in question moves toward the right balance of centralization/decentralization based on what works, not what is specified by a lean heuristic and especially not a misguided interpretation of a lean principle (LAME).

    However, we would like to at least be headed in the right direction when we begin our PDSA tests. That’s where understanding the ideal condition/true north comes into play. If the hospital in question had a vision of being organized about patient flows (i.e. product-centered production cells) as the ideal condition, then they would pursue decentralization, addressing barriers with PDSA. From your analysis, we can extrapolate what some of those barriers might be in a decentralized scenario: How to balance resources when departmental workloads fluctuate? How to standardize in a wide array of settings? How to learn from each other when we’re separated? How to manage this specialty function?

    I think through repeated use of PDSA, all these barriers could be overcome. We dealt with these same issues on a similar project at a pediatric hospital, and were able to find countermeasures to these issues over time.

  2. I’m drawn as well to Michael Balle comment about visual and how without visualisation of pull and takt (admittedly here he is talking about manufacturing, but point is still relevant) you will be likely to be looking at the wrong problem and therefore committing waste by implementing the wrong solution.

    So I guess in a healthcare environment we need visual depiction of work so we can know where problems eg lack of resources on hand or excessive transportation (decentralization) or excessive motion or excessive inventory (centralization) needs remedying first.

  3. Thank you very much for sharing your thoughts, Mark.

    We had a similar question on table in our organization about centralized vs. decentralized (existing structure) phlebotomy services provided by different roles in the different outpatient clinics in the hospital. Analysis of current state indicated some of the major issues as inconsistencies in the pactice, various roles doing the job, work priority conflicts leading to higher wait times for the patients in the clinic (nurse being occupied with their primary tasks) and lot of discrepacies in the data collection leading to flawed information gathering system.

    3 future state options were proposed and one of the them included a complete centralization of these services. As you mentioned earlier there are clearly some advantages to the cetralized hospital fuctions however in this case, it also neglects the requirements of the patients with mobility issues, difficulty in predicting the patient volumes (also due to lack of/incorrect data that we have), potentially higher wait times for the patients and not to mention, union conflicts due to redefined roles.

    If we keep the existing decetralized structure then we might have to do the continuous runs of PDSA cycles to move towards more standardized work pratices while also keeping it patient centered.

    In my opinion, the answer might lie somewhere in between as we try to achieve the right balance of “patient centered care” and “standardized work processes”. In this case for example, we might be able to keep 2 or 3 Specimen Collection Centers for phlebotomy services (based on patient volume), and the patient volume going in to these centers can be regulated on a regular basis (or real time in an ideal scenario) to maintain the one piece flow, have fairly standardized processes, and to accommodate the needs of most (or all) of the patients.

    The decision is yet to be made in this case however I would love to hear your thoughts on this.

    Thank you

    • I think the best answer is often somewhere in the middle of two extremes, as you point out – possibly having a mix of centralized and decentralized.

      Looking at “inconsistent practice” as one problem statement, it seems like there are a few root causes and potential countermeasures that could be used to address that… whether that involves centralizing as a countermeasure or not.

      It’s important to understand why an expected countermeasure… what is the hypothesis about why centralizing would lead to a more consistent practice? Many people get caught up in solution thinking (such as “we heard centralizing was good”) and they don’t understand why it would work and what problem it really solves.

      It’s also important to ask why a more consistent process would be better – less pain to patients, better or faster lab results, etc?


      • We have a slightly different care model here since the hospital is a huge cluster of ambulatory clinics with limited number of inpatients and the OR is limited to only a day surgery. Due to the ambulatory clinics, patient volume is very high.

        Currently we have one Lab Test Center(LTC) in addition to the individual clinics drawing blood and often times, the patients in the clinic end up being sent to LTC due to various reasons anyway. Also decentralized services pose some challenges to maintain the best practices and to keep the roles up to date with the current standards in practice and training. Some additional benefits may also include faster portering service since we are down to only one location, and in turn faster lab results.

        Can most of these be achieved while staying decentralized as it is perceived as more “patient centered” approach? It can be to a certain extent. However technically speaking, even current model is not a completely “decentralized” model since we have one central lab test center. And future model can also not be completely “centralized” since there will be some clinics drawing blood in house due to the type of patient population (Urgent Care, OR or Infection Control).

        So the question is what is the optimum hybrid model which will address most of the issues from the root and how to identify that? I would wonder if there is any lean tool that would help organize our thought process better (not necessarily give the solution).

  4. Krishna, we had a similar situation at a large pediatric hospital, and we also ended up somewhere in the middle between pure centralization and decentralization. I think in most cases, this is where the practical sweet spot lies.

    However, what’s practical is not ideal. Over the long-term we need to be moving in the direction of the ideal. That might mean starting in the middle/sweet spot and moving toward pure decentralization over time, along the way figuring out how to maintain the benefits of centralization in an increasingly decentralized environment. This is where an organization having a strong habit of using PDSA to test ideas/countermeasures is invaluable.

  5. My experience is that most managers lack skills and abilities to manage non-core functions well. One of the best arguments for lean is to improve manager skills. I’ve also seen bad functions centralized with bad management and resulting in the worst of both worlds.


Please enter your comment!
Please enter your name here

This site uses Akismet to reduce spam. Learn how your comment data is processed.