Diane Miller, from the Virginia Mason Institute, argues that one of the advantages of an industrial setting is that it’s easier to see some of the Lean techniques in action, because everything is there, right in front of you, rather than hidden away in multiple processes, spread over time, as is often the case in health and social care. I saw a good example of this, in the use of a Drum-Buffer-Rope system in part of a production line in a car factory I visited recently.
This term – Drum-Buffer-Rope – is associated with Eliyahu Goldratt and Theory of Constraints. Goldratt argues that any system has at least one constraint that limits the speed of the whole process. To reduce the impact of this constraint, the system needs to maximise the use of this constrained step. In the long-term, you would be likely to want to remove the constraint if you wanted to increase capacity. Until you can do that, it’s important to reduce the limiting effect of the constraint.
So, for example, if you have an eight step process and Step Six has a lower capacity than the other steps, then you need to maximise the use of Step Six. This means that, as far as possible, you want to keep this step supplied with work, as time lost at this function can never be recovered. This constraining step, which limits the process, is known as the Drum, because it sets the drumbeat for the rest of the process. There is no point in going faster than the Drum, because in this example you simply pile up work after Step Five, while it waits for Step Six to catch up – which it may never do.
In order to ensure that the constrained process works at its maximum, Goldratt argues for the creation of a Buffer – a time period in which the Drum can function, even if the flow in to it is disrupted. In practice, this requires the build up of a stock in front of the Drum. In this example, some work is held in a Buffer between Step Five and Step Six, to allow the Drum to keep functioning for a while, even if one of the earlier processes is disrupted. The time required depends on how likely it is that the flow will be interrupted, and for how long.
The Rope then comes in as an analogy for the control of the flow in to the process, before Step One. If the Buffer is full, there is no point in taking more material in to the system, as it will simply build up at some other point in the system. The Rope. linked to the Buffer, opens and closes the gateway in to the process.
At Nissan UK, I saw a production line that stopped three times while I was there, for between twenty seconds and two minutes. A sign flashed up over head, ‘Buffer Full’ – and the upstream processes stopped, until the Buffer dropped back in the required number. The people working on earlier stages of the production, on the part of the production line beside which I was standing, just – stopped. And waited.
The health care application of this is often seen in out-patient clinics, but can occur in other settings as well. I gave a previous example of an ophthalmology clinic where the constraint was the time with the doctor. In a process I observed, nurses and optometrists worked hard to push people through the system as fast as they safely could, but these people then waited for the doctor, and the number of people waiting for the doctors grew and grew as the clinic progressed.
Pushing people through the system faster than the constraint could cope, did not speed up the process. If anything, it made it more turbulent, because it became more difficult to work out exactly what was happening in the clinic, and who was waiting for what part of the process, because people were waiting anywhere in the clinic they could find a seat. Also, the number of people waiting seemed to produce a real sense of urgency in the nurses and optometrists, who worked even harder – but this made no difference at all to the Lead Time, because there the central problem was that the upstream process was working faster than the constraint could function.
In this system, there would be a logic to being a little ahead, so that the doctor never needed to wait to see the next patient, and so that their time is maximised. This would, in effect, be a buffer. To decrease the lead time for patients, however, patients needed to be brought in to the process matched to the Drum, the overall constraint, by means of careful scheduling. The only alternative would be to increase the capacity of the constraint. In this case, the service did both. They increased the doctor time available, by bringing in the first patients in sequence, so that they were ready for the doctor, the constraint, as soon as the doctor began their work. They also smoothed out demand, so that there was less variation over time, and so that the people waiting to see the doctor – the people in the Buffer – were limited to a smaller number. This maximised the doctor’s time, but also decreased the waiting for patients.
Moving to a Buffer does not get a service, and therefore its clients and patients, to one piece flow, but it can be a pragmatic option until a constraint is resolved. Making the source of the limitation clear, as Nissan did during my visit, makes the true constraint much more apparent, and therefore more of a focus for improvement.