Forget Pull or Push, Focus on Improving Patient FLOW Instead


Back in the manufacturing world, the distinction between “push” and “pull” is often much more obvious than it is in healthcare. Push production occurs when products are built based on the forecast were something other than actual customer needs. In the Lean approach, pull systems are the ideal to prevent overproduction, to reduce inventory costs, and to make sure the customer gets what they need when they need.

In healthcare, pull systems can be very effective for the replenishment of supplies and medications. But the conversation often gets muddled when we are talking about patients. It “pull” or “push” better for patient flow? Is that even the right question?

I was recently at a conference where a team presented about emergency patient flow (or the lack thereof) at their hospital. They described the old condition as one where the ED “pushed” admitted patients to the floor, making a call to say “hey, we need a bed.” The problem was that the inpatient unit was often not ready to take the patient or they didn't have the space available – so the patient wasn't moved.

This describes poor patient flow, not having the resources available or not being able to move the patient when it's medically necessary. This handoff of patients from emergency to the inpatient unit needs to be a two way agreement. I agree it would be bad to push a patient to a floor if they are going to just sit in the hallway without the proper service or medical attention.

In the presentation, they talked about moving to a pull system that “has roots in the lean thinking methodology and is far more effective than a push system.”

They described a pull system as one where the inpatient floor triggered the movement of a patient's by informing emergency department that they are ready to receive a patient.

If the patient is waiting until the unit is ready, is that really improving patient flow? Is it just an academic quibble to call that a pull instead of a push? Is it really “far more effective” to wait until the unit “pulls” the patient? Either way, in the push or pull approaches, it seems like the patient didn't move until the inpatient unit was ready – so what's the difference then?? It might be less frustrating to the inpatient unit team, but is that really the primary goal?

I'd argue that the true “pull” that's necessary is a patient “pulling” upon the resources that are needed and having those resources made available or delivered immediately based on that demand. If a patient needs a room, that patient should, in effect, “pull” the room. The language around this is a bit silly because were not physically pulling the room to the patient, we are moving the patient up to the room.

So again, who cares about push or pull? Are rooms available when they are needed, for the sake of the patient and for the sake of the emergency department not getting overcrowded? Do we have the right capacity available, when and where it's needed?  I think those are the right questions to be asking so we can actually improve patient flow.

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 articleCollected Quality & Cost Results from Healthcare Value Network Members
Next articleBuild, Measure, Learn – in Startups, Healthcare, and Careers
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. Hear, hear.

    People get these things terribly mixed up. I think that you pulled it apart nicely.

    Another way of thinking about it is to note that it is the customer that pulls. Always. In manufacturing that customer is at the end of the assembly line pulling their chosen product toward them. On admission to hospital the patient is the customer and they pull the care and treatment they need toward them. So they pull the need for an inpatient bed. But in fact, we know that we only pull when we can’t flow, so if a patient flows to the next stage of their care, then there is no need for any pulling and the confusion melts away.

    Excellent post.

  2. The customer always, always, always pulls.
    In fact, I would argue that hospitals have been using pull much longer than factories!
    It’s almost inherent in what they do, it being difficult to operate on a condition that a patient doesn’t have that doesn’t cause them to visit the hospital!

    • But like I said, it doesn’t matter if you have pull with terrible flow.

      Factories have done a better job of figuring out flow, generally, even if they were using a “push” model.

  3. I was the data collector in a large teaching hospital’s ED for a flow improvement project. Having lots of time to sit around and think between observations I debated this “push vs. pull” concept.

    The unique situation hospitals seem to be in is that the “customer” is also in a way the material sent through the system and the wants and needs of the customer aren’t known until they show up at the ED making it hard to be in a position for smooth flow.

    My role in the aforementioned project was to measure various steps between the point where the patient is deemed admit-able by the ED (push) to their actual arrival on the med/surg floor. The main bottleneck was waiting for the intern to come down to assess the patient prior to writing admit orders (pull).

    Having just learned “pull” is the ideal in lean systems I remember thinking a system set up to focus on a competition between medical teams to be the next one to “pull” a patient up to the floor would be ideal. The patient gets what they want (quick admission / less time in the ED), the medical teams would get some sort of financial incentive for admitting more patients, and the hospital gets more efficient use of its resources (less wasted time/space).

    This might require pairing nurses/floors with specific medical teams in order to align goals of all involved. It would also require nursing and medical staff to have simple/transparent communication with the ED to gauge when there is a patient ready to be pulled.

    I’m new to this stuff and would appreciate any comments or suggestions any of you have towards my line of thinking.

    • Andy – thanks for your comment. I often hear in healthcare about “how competitive people are” and how posting metrics or giving incentives can make a big difference. I guess physicians are conditioned by “piece work” payment systems that reward them for volume above all else.

      I would have a lot of concerns about the possible dysfunctions that would occur with a “contest” like you describe. You’d definitely want to make sure all stakeholders (nurses, doctors, etc.) are aligned. What could be possible dysfunctions if the patient is getting pulled to the unit that’s not ideal for their condition (telemetry vs. med/surg, for example)?

      Rather than trying to artificially inject extrinsic motivation, I like to think through why the natural intrinsic motivation to care for patients doesn’t seem to be presenting itself, if units are not wanting more patients… why is that motivation not there?

      I wonder if questions like that get closer to the root cause of the problem than would contests and more extrinsic motivation?

      I’m curious what others think about this…


      • Thanks for the input Mark!

        Perhaps competition was the wrong word to describe what I envisioned. What I meant was that rather than making the medical teams admit patients in a set order why not let them take patient’s when they are able?

        I feel (feel = no data to support), the root cause of floors not wanting to admit patients in a timely manner is that each new patient represents an unknown risk to be undertaken, and at the same time also represents less time to spend with existing patients.

        A system where the MDs and nurses pull patients when they are ready for them hopefully offsets both of these issues. Pulling a patient “when ready” means everyone has researched the patient’s ED records to understand the risk involved with caring for that patient. It also means existing patients are stabilized with their needs met. A “competition” (still not the right word) between teams to admit patients focuses the team on improving their processes for researching new patients and stabilizing existing ones.

        The issue I see arising is that there is a point where doctors and nurses have enough patients where their motivation to learn/help is being met and an additional patient just means less time to learn/help existing patients properly, not to mention increased risk towards their license. That is why I suggested the financial incentive.

        Perhaps the answer is to continue lean innovation so that the tipping point continues to rise so that MDs and nurses can comfortably care for more and more patients?

        • rather than making the medical teams admit patients in a set order why not let them take patient’s when they are able?

          Because we are not here for the staff. We are here to care for patients. We do not set up the system for the convenience of the doctors, nurses or anyone else except for the patients.

          Teams should take patients when the patients need it, not when the teams are able.

          Improving the system for the patient will increase capacity and the quality of care so that teams can take more patients, more quickly and care for them better. But we must, must, must design the system around the patient. It is the only way to get genuine improvement.

          I think that you agree with some of that by your last paragraph.



    • Andy,

      I have to agree with Mark on this one. Setting up artificial competition is a bad idea. You want to find a way to spark the intrinsic motivation to care for patients to help them to work together to improve the flow. Competition will have them work less well together.

      Also the ideal state in Lean is not pull, it is flow.

      Thirdly, having financial incentives to admit patients will encourage teams to admit patients, but not necessarily to give them the right care.

      Think about the system from the patient’s point of view. That is the key. If there is a bottleneck waiting for an intern, then think about that from the point of view of the patient and their care.

      If you can’t yet achieve flow don’t think about how the team pulls a patient to be admitted. Think about how the patient pulls the intern to do the assessment in the shortest time.

      Good luck. Let us know how it goes.

      • Appreciate the input Rob!

        Wanted to play devil’s advocate with the concept of flow…

        I agree the ideal end state in lean is flow, but to focus on flow when improving processes seems more geared towards the conceptual.

        Water and processes can either be pushed or pulled, but they can’t be “flowed.”

        Isn’t it easier, cheaper, and more efficient to achieve flow by “pulling” water using gravity than it is to “push” water using a pump?

        Is this not also true for processes and that is why visual cues to “pull” material through the system are so important in Lean?

        • Andy,

          Why do you think that focusing on flow when improving processes is conceptual? Flow can be a reality, but only if you make it happen.

          Let’s imagine a patient that comes into an emergency room. He has a broken arm. When he walks in he doesn’t see a triage nurse he walks straight into be seen by a doctor, that doctor does a quick assessment and sends the patient to x-ray where he goes straight in. As the patient is walking out of the scanning room, the doctor already has the x-ray on his tablet and is instructing a nurse practitioner in how they are together going to set the man’s arm. The patient walks into a cubicle where the doctor and the nurse do what is necessary and while the nurse finishes of the cast the doctor is writing a prescription that is already being fulfilled by the ER pharmacist. The patient walks out with a small supply of pain killers, a set arm, a prescription for more drugs should he need them and a booking in outpatients or his local doctor to remove the cast or do another scan in a few weeks. It all took 45 minutes. At no time was the patient waiting, filling forms, going through triage or any other batching or sorting. That is an example of flow. (Forgive me if the medical details are wrong, I am not a clinician. I’m sure you get the drift.)

          Flow just means that the work never is delayed, has errors or requires rework. It just keeps having value added, without any breaks, until the work is finished. That is certainly ideal, but it is not simply conceptual and it is very possible.

          Also, at the risk of starting a pointless semantic argument about metaphors, water isn’t pulled downhill, it flows downhill.

          So if we are talking about water moving downhill under gravity then that is flow and that is a similar to what we want to see for our work. A patient (a drop of water) presents at our clinic (the top of a hill) and he flows though the system of care (down the hillside) until he is discharged (reaches the sea).



          • Rob,

            Loving this conversation!

            But I would counter your description of ED flow is exactly what frustrates those of us providing the care. It is far too conceptual and geared toward an end product that is ideal, but extremely complicated to get to from the current state.

            The wheels come off the flow wagon the instant the MD has to set a broken arm and a patient bleeding and a patient with chest pain and assess a patient detoxing from alcohol withdrawal. The number (and sometimes skill of the physician) becomes the bottle neck and patients must then be “pulled” to the MD based on need/urgency.

            Clinicians dislike nonsensical consultant buzz words like “flow,” and instead need step one answers to get to “flow” from the current state. Focusing on flow just glazes the eyes of the care giver because we know how to push or pull something, but how the heck do you “flow” something?!?

            • I don’t think “flow” is that esoteric of a concept. Do you have the capacity to meet patient needs at a given time? If yes, you have flow, if not, things get backed up.

              We can, at best, predict demand and workloads. Andy, you’re right to point out that there’s variation and things that aren’t well predicted. So, at times, flow will suffer.

              The Crane and Noon book on Lean E.D.’s makes this point clear, I believe – that we can plan for normal variation in demand and workloads. To say “we can’t plan, we never know how many patients are coming” is an old conventional wisdom that’s been pretty well proven false, I’d say. I’d suggest their book as the go-to guide on E.D. lean and flow.

              The Definitive Guide to Emergency Department Operational Improvement: Employing Le… by Joseph T. Crane


              • Thanks for the great resource Mark. I will check this out.

                Want to assure you and Rob that I am not trying to be a blog comment instigator to you or Rob, just wanted to play devils advocate from the point of view of the novice clinician. I appreciate your comments! Thanks for letting me participate.

                • Andy,

                  In a way you are right that water is like work in that you can’t flow work any more than you can flow water. But you can create the right conditions for flow. For water a good incline and a smooth, unobstructed channel. For work processes and systems that are designed from the patient’s point of view with no error, delay or rework.

                  I think Mark’s comments about variation in demand are very important. I would also like to suggest that the closer you get to work that flows, the more work does flow. The shorter the time it takes to treat patients (more effectively of course) the freer the MD will be for the next patient. It is the opposite of diminishing returns. The better you do, the better you will do.

                  If I remember the story correctly, someone once asked the golfer Gary Player if he believed in luck. Player replied that, “It’s funny, the more I practise, the luckier I get.”

                  And never apologise for asking good questions!

  4. Great Topic! We just kicked off a project on this very topic at my hospital.

    I agree that pull systems differ in manufacturing versus a hospital.

    In manufacturing we envision the customer at the end of the line consuming a given product. The pull from the final process and thats starts the chain into preceeding processes.

    In a hospital the patient is part of the process so it adds a different dynamic. I still envision it as the last step is the patient going home, so similiarly we can pull from the last process step going forward.

    Mark’s point about about flow being more important that push/pull is spot on. I manufacturing flow comes more natural, but there is a huge advantage in consistency in demand which allows for capacity planning etc. I doubt you ever see demand shift at any hour of a given day in manufacturing like you do everyday in a hospital.

    To that point we are working to adjust our discharge times to the morning, so that beds are available come afternoon, when the ED traffic starts to pick up.

    As for motivation/incentives, that has a lot to do with it. We have one unit that excels due to motivated leadership. Their motto is “patients equal paychecks” and when a bed is empty there’s a sense of urgency to get it filled. Other units will not report empty beds, some say it due to lack of incentive to fill the bed (whats in it for the nurses to keep their beds full?) or not wanting to overburden nurses who can be dealilng with patient that require special/extra attention and care.

    Ultimately, if there’s no capacity in downstream processes then push/pull is negligable.

  5. I think Rob has a key insight. The larger focus is on flow. A singular view just on pull leads to contortions in how to think about this topic because the patient is both the “material” that moves and the customer. Pull requires signaling to other processes–something the patient is not always well equipped to do. In contrast, the staff of the hospital has a multitude of methods and techniques to manage flow. First seek flow, and then pull where you can, then push where you currently have no other options.

  6. Interesting! Many people in our industry (we grow plants in greenhouses) get hung up thinking they need to design pull systems across their value stream or they are not “truly lean”. I think people take too literally Womack’s rubric (including “Pull”) and think they are not lean if they haven’t checked off each of the bullets. As with the situation you describe, we have a variety of pull systems for replenishment of supplies and internal processes, but there is no meaningful approach to one-point scheduling based on customer withdrawal from a finished goods supermarket!

    Plants (like people) don’t care if the next process is ready for them or not, they are experiencing time in a way that a widget or a Camry does not – growing and changing, needing sunlight and water (or feeling pain, fatigue, delay, etc., in the case of sick people). The next process MUST be ready and have capacity.

    Healthcare is maybe like a build-to-order/engineer-to-order environment? Patients present themselves as a work order and it is “flow” from there. “Customer pull from finished goods supermarket” of a repetitive manufacturing model just doesn’t make any sense in this context. The challenge is to make the flow balanced and seamless so there is no waiting.

    It is another lesson teaching us that lean is all about knowing the real work and solving the real problems, not copying or applying tools, even with top level concepts like “pull”.

    Art Smalley made the same point (in exacting detail) in his April 19 “Lean Edge” post.

  7. Thanks, everybody, for their comments.

    I’d make one more, possibly provocative point:

    If a healthcare professional NEEDS a financial incentive to “take one more patient,” as in “I can’t be bothered with these damn patients” (which seems to be the attitude once in a while), then maybe that person should consider finding a new profession.

    Again, though, if somebody’s attitude has gotten to be that bad, we should look for the organizational and managerial root causes of such attitudes. People are very rarely burned out on their first day of work. Why do they get burned out? How much of that is preventable and avoidable?

    • A provocative point indeed!

      We don’t fault Lawyers, Accountants, or Consultants for taking on one more client to reap additional financial benefit. Why is it frowned upon for healthcare professionals to explore a similar system to serve patients?

      Could we consider increased specialization among physicians, declines in the number of the primary care physicians, projected nursing shortages, and nursing surveys showing high percentages seeking to leave the bedside as proof that people are considering new/different professions to avoid being overwhelmed with more patients?

      Are the organizational and managerial causes of poor attitudes related to large numbers of quality providers leaving the bedside to specialize and increase their personal financial return and control of their own destiny while still helping SELECT patients?

      Is Lean the tool to empower bedside providers to improve their environment enough to keep them satisfied financially and personally so that they can focus on treating EVERY patient who needs care?

      • Andy – I didn’t make my point clearly. I’m not faulting doctors for getting paid for treating patients. I’m not frowning upon that.

        What I’m frowning upon is the desire to create games and additional incentives to do what they’re already being paid to do – treat patients!

        I’m not advocating anybody be “overwhelmed” – as with the primary flow discussion, we have to do our best to balance patient demands with the supply of beds and caregivers. Many of the “we need to give people incentives” discussions arise when a unit has capacity but people “feel” overwhelmed.

        Lean reduces that overwhelmed feeling by reducing waste and the resulting frustration that causes burnout.

        From my experience, if we focus on allowing healthcare professionals to feel less overwhelmed and more personally satisfied, as your last point says, then we can worry less about financial games and competitions.

        • Mark,

          Couldn’t agree more that Lean reduces the feelings of being “overwhelmed” and that is why I am so interested in your blog and introducing Lean as a tool for frontline providers.

          Sorry if I’ve been a difficult commenter but I wanted to get expert responses to the kind of issues/complaints my co-workers raise with me when I try to explain Lean to them. Thanks for the help!

  8. Mark, it is my belief, along the lines of Alfie Kohn’s work, that highly trained professionals do not need financial incentives to grab and sustain their interests, that these incentives can actually disengage or distract folks, if not simply turning off the “bottom 90%” of performers. As burnout…lots of reasons…most preventable, internal and external.


Please enter your comment!
Please enter your name here

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