Delivering the COVID-19 vaccine to the population safely: A logistic challenge solved with the Theory of Constraints
For the last 10 months, the global population of the world is going through a pandemic sanitary crisis that forced people to lockdown and work from home, stopped social interactions, and put a toll on the economy of most countries. The inequalities are rising, and the toll on our healthcare systems is increasing.
The announcement of several vaccines coming to the population in the near future is a great relief. Nearly all governments have prepared a strategy of vaccination, determining which are the priority communities. Those priority communities are not always the same, but all those strategies have a common point: after vaccinating the high priority people, a mass vaccination campaign is organized to bring the vaccine to the population.
This mass campaign will be, in most of the cases, voluntary. This is a challenge for the logistic organization, and the theory of constraints can bring a solution to design a simple but highly efficient distribution system.
The Issue
The distribution of the vaccine to priority communities is simpler as the number of doses needed at a place shall be known, but for the mass vaccination, we don’t have this information available.
Most of the organizations will rely on forecast systems, like the ones used in retail industry to anticipate the demand, except that for mass vaccination, the historical data to build a model is not available. The most reliable estimate of the demand is that we should face a high initial demand, then the demand decreases and extends for a long period as some people will wait to have some feedback before getting vaccinated. An estimate of six months for the duration of the mass vaccination campaign seems reasonable both from the point of view of the needed logistics and from the point of view of the availability of doses of the vaccine. If the vaccine needs more than one dose, the duration of the campaign may have to extend way more than six months.
The curve of demand should look as the graph below. With such a distribution, it is very difficult to forecast the demand for the vaccine reliably over time.
Vaccines are products that need some specific conditions of storage. Some vaccine even need extra low temperatures to be stored until the last moment before injecting it. This makes the equipment needed at the vaccination center expensive if you want to store large quantities.
Furthermore, these conditions of storage and transportation makes endpoint to endpoint transportation more complicated and expensive.
The logistic system shall succeed to have the right number of doses at the right place at the right time. This translates into:
- Avoiding overstock to prevent storage issues that could void precious doses of vaccine
- Avoid local stock out that could prevent a part of the population from finding the dose of the vaccine in their vaccination center
- Adapt to variation of the demand as fast as possible
- Avoid direct transportation of vaccine dose between vaccination centers as this is complicated, slow to organize and therefore inefficient.
The solution
The first trend will be to rely on a forecast software and usual logistic system, but there is a simpler and more efficient solution: Synchronous Logistics.
The Usual Solution
The main objective being to have most of the population vaccinated, stock out of vaccine dose in a vaccination center shall be avoided as far as possible. The usual solution is to push as many doses as possible to the vaccination center. This would make sure that there is always a dose for any person presenting herself at the vaccination center.
To organize the most cost-effective transportation, a policy of minimum economical quantity to transport is usually applied that sets the minimum quantity of doses of the vaccine that should be delivered at once to a vaccination center.
This solution has collateral consequences of significant importance:
- The vaccine has a limited shelf life and may void before to be used;
- The vaccination centers have a high cost of installation due to the necessity of being equipped with cold storage devices that may find no use after the campaign.
- The minimal economical quantity to transport can result in the quantity of the vaccine in a center exceeding the storage capacity of that site, putting doses at risk of becoming voided;
- The bigger the quantity of dose in a vaccination center, the bigger the difficulty to apply the rule of “first in, first out.” This procedure ensures that the first doses that were delivered on the site are used first.
Synchronous Logistics
Synchronous Logistics is based on the principles of the theory of constraints as they are applied in distribution applications.
The principle is to protect the capacity of injecting a dose of the vaccine to the person who needs it and present herself in a vaccination center.
The main ideas are
- Keep the doses where they are safe and as high as possible in the supply chain to use it to mitigate better the variation of each center. This means keeping the dose in the distribution center until we know that they have high chances actually to be quickly used in the vaccination center.
- Use a simple buffer indicator at the level of the vaccination center and the distribution center to adapt the stock to the demand and to create alarms if the use of the doses is too high or too low compared with the expectations.
- Organize a daily or twice a day replenishment of only what has been consumed, no economic minimum quantity policy applying.
Initial stock and dynamic buffer management
The initial stock of the vaccine at the vaccination center is set to two or three times the expected consumption between two planned replenishments. The shorter the time between two replenishments is, the more efficient the adaptation of the system to the variation of demand will be. The initial value is not of importance, as the system is self healing and self-correcting, it will adapt in a matter of a few days to the actual demand as seen below.
A dynamic buffer management system shall be used to adapt the stock to the demand.
The stock of the vaccination center is divided in three quantities or zones: green, yellow and red of equal quantities. If there are 300 doses in the center, then there are 100 doses in every zone.
At the end of the replenishment interval, the zone, in which the remaining inventory of dose of the vaccine is, indicates what action shall be taken.
- If the inventory is in the green zone at the end of a replenishment period, the replenishment shall compensate for the exact number of doses that has been used in this period.
- If the inventory finish in the green zone for a set number of replenishment periods (for example 3) in a row, then the global inventory of dose shall be diminished by 20% (or another factor that depends on your risk perception).
- If the inventory is in the yellow zone at the end of a replenishment period, the replenishment shall compensate for the exact number of doses that has been used in this period. This is the zone in which we expect to finish at the end of each replenishment period when the inventory is adapted to the demand.
- If the inventory is in the red zone at the end of a replenishment period, the replenishment shall compensate for the exact number of doses that has been used in this period.
- If the inventory finish in the red zone for a set number of replenishment periods (by example 3, but it doesn’t need to be the same as for the green zone) in a row, then the global inventory of dose shall be increased by 30% (or another factor that depends on your risk perception. That factor doesn’t need to be the same as for the green zone).
This system can be even managed manually in the vaccination center. There is no need for any fancy software. The buffer zones can also be used to detect risk of getting short on doses and ask for emergency delivery between two normal replenishments.
For example, if the stock of a vaccination center enter the red zone early in the morning, it is the indication that an emergency replenishment should be requested to avoid running out of dose early during the day.
A Simulation
To demonstrate the efficiency of the Synchronous Logistic system for the delivery of the vaccine, a simulation has been prepared as described below.
This simulation covers a vaccination campaign of six months. The simulated vaccination center is designed to serve a maximum of 200 people per day, and no emergency replenishment has been considered.
The Demand
The demand is composed of the sum of two components. A first component with a slow variation corresponds to the evolution of the demand during the campaign. It begins with a peak for the first 30 days and a steady decrease after the 30th days until the end of the campaign.
On top of the first component, a second one is added with a random variation simulating the daily variation in the number of people presenting themselves at the vaccination center. The graph below shows the resulting demand.
This kind of demand is difficult to forecast because of the randomness of the simulation. The lack of historical data and the short duration of the campaign makes most forecast software useless or at least unreliable.
Usual Logistic System — Push System
We call it “push system” as it “push” the doses to the vaccination centers based on a forecast.
To cover the needs, the initial stock is set at 300 doses, i.e., 50% more than the expected capacity of the center. For simplicity, this is also the minimum economical quantity for transport.
The minimum inventory of the vaccination center is set at 75 doses. This means that every time the inventory is below 75 doses, 300 new doses are sent to the vaccination center the next day.
The inventory during the campaign is represented below. In a period of high demand for the vaccine, the replenishments are made more often. When the demand decrease, by the end of the campaign, the periods of replenishment become more distant.
Despite the stock level that far exceeds the peak demand, the vaccination centers lacks doses from time to time to fulfill the demand.
The graph below shows the missed vaccinations during the campaign.
The graph below shows the vaccinations executed at the vaccination center during the campaign
As a result, 13,580 people presented themselves at the vaccination center during the campaign. 11,984 were vaccinated, but 1,596 didn’t find a dose at the center. For this result, a total of 12,300 doses of the vaccine were mobilized with an average inventory of 301 doses in the vaccination center.
Applying Synchronous Logistic and Dynamic Buffer Management
This system is a “pull system.” This means that the demand will “pull” doses from the distribution network. Furthermore, the inventory at the vaccination center will be adapted dynamically to this demand by a Dynamic Buffer Management approach.
The initial inventory is set at three times the minimum inventory, i.e., 3 × 75 = 215 doses.
The replenishment is scheduled daily to compensate what has been consumed during the day.
The graph below shows how the maximum inventory is adapted to the decreasing demand over time. The second graph below the inventory curve shows how the signals (green, yellow and red) drive the adjustments of inventory levels over the duration of the campaign.
This method brings a significative reduction of missed vaccination due to lack of doses of the vaccine in the vaccination center. The graph below shows these missed vaccination over the duration of the campaign.
The graph below shows the vaccination executed during the campaign.
As a result, 13,580 people presented themselves at the vaccination center during the campaign. 13,566 were vaccinated, but 14 didn’t find a dose at the center. For this result, a total of 13,634 doses of the vaccine were mobilized, with an average inventory of 207 doses in the vaccination center.
Results
The simulation shows that with an average inventory of nearly 30% less than with the usual method, the Synchronous Logistic method allows vaccinating 13% more patients. Only 0.1% of the patients didn’t find a dose at the vaccination center against 11.8% with the usual method.
The Synchronous Logistic method allows vaccinating 99.9% of the patients coming to the vaccination center, while the usual method allows vaccinating only 88.25% of the patients that presented themselves at the vaccination center.
Conclusion
The synchronous logistics methodology allows a much more efficient delivery of the vaccine to the population without needing to maintain a high average inventory in the vaccination center.
The lower average inventory means that the installation of the vaccination centers are cheaper as they need less cold storage.
It delivers nearly a 100% availability of doses despite a very variable simulated demand.
The cost of a daily replenishment is largely compensated by the gain in efficiency and lower average inventory.
Furthermore, this method doesn’t require any special software or equipment and can be applied manually or in a simple spreadsheet at the level of the vaccination center and with a simple spreadsheet at the level of the distribution center.
If you want to learn more about this methodology, don’t hesitate to contact me. The simulation is available for audit if you need.
This article has been published also on LinkedIn
Didier Varlot
Senior consultant in Business Continuity and Theory of Constraints, Owner and CEO of SNTC.
Didier is a project manager with 35 years experience in project recovery and 25 years of application of the Theory of Constraints. He uses a mix of Theory of Constraints, Agile and Open organization (the TAO Way) to improve operation. His references go from railway industry to healthcare services, from chemical industries to green energy supply.
He is the author and moderator of the Thinking Logical: Synchronous Momentum and has his own publication on medium.
You can follow him also on Twitter, or LinkedIn
Contact Didier via LinkedIn, via email or via phone +40 744 501 044 (FaceTime and WhatsApp).
