The great majority of hospitals in the United States manage hospital supplies using what is called the “Par Level” method. One of the strengths of this method, it is claimed, is that it works well in the face of variable usage. The actual usage of a hospital supply item is unlike that of a manufacturing part, in that the rate of use is dependent on the patient need at the time, which is difficult to control or forecast. A factory can predict with good accuracy how many widgets it will need, based on a production plan and a Bill of Material. A hospital can certainly analyze historical usage data, but cannot control that pesky variable, the patient, very well. Defenders of the Par Level method therefore make the case that it is the only replenishment method for the variable usage environment. In this article I will describe how the Par Level method works, how the Par Level method could be replaced using a method called Kanban, and how a Kanban system could be used properly to respond to the admitted challenge of high variability of usage.
On the face of it the Par Level method makes good sense. The basic idea is to set a target quantity of a given item that you want to stock at the place where it is used. A rule of thumb of, say, a two-day supply based on average usage, would be a reasonable goal when setting these target or “par” quantities. As the supplies are consumed, based on actual need during the day, the quantity on hand will decrease. The heart of the Par Level method is to have a supplies handler review the on-hand quantity, and restock the items that are needed to bring the total quantity back “up to par”. If the item is consumed more quickly than usual, having a supplies handler review the inventory daily will help to ensure that the item is not depleted completely. The key to the success of the Par Level method in the face of uncertain usage is to check stocking levels frequently, usually once a day.
The Kanban system, mentioned above as an alternative, is the material replenishment method used by leading manufacturing companies around the world, including the Toyota Motor Company. It is similar to the Par Level method in that a target level of material is calculated, based on average usage and a certain number of days of usage. In this sense the methods are virtually identical. The difference is in the method for refilling items that have been used. Instead of physically looking and counting every day, the Kanban method sets up a signal (the literal meaning of the word Kanban) for replenishment that eliminates totally the need for counting or assessing inventory quantities. This is done in a variety of ways, with the most common one being to divide the item quantity into two, and signal for replenishment when the first quantity has been used up. Details of how a Kanban system works are readily available elsewhere, and the method is well-known, so I don’t need to repeat it here. The benefits compared to the Par Level method are substantial, including a reduction in the number of resupply trips, the elimination of counting, a potential reduction in overall inventory levels, and a reduction in supply shortages.
The focus of this article, however, is on one critical issue: what happens when the actual usage of an item exceeds the planned usage? How do the Par and Kanban systems respond to unplanned consumption? After all, if usage were known with total certainty ahead of time, there would be no excuse for ever running out of an item. Is daily checking and counting the only alternative available to hospitals? Let’s compare the two methods.
The Par Level Method. A supplies handler will visit a supplies stocking area daily, with a clip-board or hand-held device, and assess the current stocking level of each item. Items that need to be “brought up to par” will be noted, and these items will be refilled on a subsequent trip, usually later the same day. If during the day an item is consumed at a rate higher than planned, such that the entire supply is consumed in a day, the item balance will either go to zero, or someone will need to notify the material department and an expedited delivery will need to be done. The intention is that this will happen infrequently, and that a daily check will be sufficient to cover variability of usage.
The Kanban Method. Under a Kanban system the supplies handler does not count or assess item quantities at all. As the initial supply quantity is consumed and depleted (the “primary quantity”), the nurse or tech will create a signal by placing a reorder card in a collection box, raising a flag indicator, placing the empty container in a collection spot, or posting some other kind of simple signal. The supplies handler simply responds to the signal, and refills a standard quantity every time. While the supply is being refilled, nurses and techs will be consuming the second quantity of the item. No time is wasted eyeballing or counting supplies, and once a regular replenishment cycle is set up there is no need for a separate trip to the stocking area to evaluate what is needed. Kanban cards or bins can be picked up as a part of the routine replenishment trip, or what is called the “milk run”.
If supplies are used more quickly than planned, because of a temporary emergency or some other reason, they will need to be replenished more quickly than planned, as with the Par Level method. The difference is this: under a Kanban system you will set up a routine replenishment cycle (the milk run) that will be more frequent than the Par Level cycle. Whereas the Par Level frequency is typically one cycle per shift or one cycle per day, the Kanban system will plan to deliver twice per shift, or even more frequently. This becomes possible without adding to your supplies handler headcount because no counting is required. The milk run cycle involves picking up Kanban cards (or not even that if the Kanban system is automated), picking standard quantities from a “supermarket” or Central Stores, and delivering these full Kanban quantities to the correct stocking locations. This process is many times more efficient than the typical Par Level process, and will enable a supplies handler to double or triple the number of replenishment trips that he/she is able to complete.
How does this help in responding to variable demand? Let’s explain. If the milk run cycle is four hours, and you do three milk runs a day, and the Kanban quantity represents a two-day supply based on average usage, the actual usage would need to exceed the average by 600% before we would need to be concerned about running out completely. While the milk run method will not guarantee that you will never ever run out of a supply, the chances of doing so are remote. The manpower to support this system could be less than you have today, since you are doing no counting trips, no counting, and no picking of variable quantities.
Let’s go through this again, so you’re 100% clear about the example above. If you consume a given supply at an average two-day rate, you will receive a signal to refill a Kanban quantity every two days or twentyfour working hours. If you consume the supply at double the planned rate, you will receive the signal after only one day, or twelve working hours. If you consume the supply at four times the average usage, you will receive a replenishment signal in six hours. Finally, if you consume the supply at six times the normal rate, you will receive the replenishment signal in four hours, equal to your replenishment cycle. Your supplies handler and milk run cycle should be able to handle that without expediting or running out. Of course, in this example if you consume the supply faster than 600% of your average consumption rate, you do run the risk of running out. Kanban is not magic.
The winner in this comparison of Par Level versus Kanban on the issue of responsiveness to variable usage is: Kanban. As we have seen, the key to responsiveness to change is a short replenishment cycle, and the Par Level system is too labor intensive to permit a high number of replenishment cycles without blowing your material management budget through the roof. Kanban, on the other hand, requires no counting, standardizes the replenishment quantities, eliminates the very undesirable practice of “eyeballing” supplies quantities, and permits a higher number of replenishment cycles without driving up material handling cost.