Author: Edward Rothberg, PhD
AI technologies have proved to be powerful weapons in the fight against Covid-19, enabling governments and companies to improve their decision-making, resilience and response to the pandemic.
Mathematical optimization is one of the many AI weapons that we have in our arsenal, and many industry professionals and academics today are discussing, developing and deploying applications of this problem-solving technology to help address Covid-19 challenges like determining the location of healthcare facilities to meet anticipated resource demand, calculating the timing of shelter-in-place orders to avoid overwhelming hospitals and causing unnecessary economic disruption, and optimizing stadium seating configurations while social distancing.
The Covid-19 challenge that is top of mind for everyone at the moment is vaccine allocation. Now that vaccines have been successfully developed and approved, the key question is how to get these vaccines from “the factory to the frontlines” and distribute and administer them to the masses in the most efficient and equitable manner possible. All around the globe, organizations from the public and private sectors are wrestling with this question and striving to create and implement effective vaccine distribution strategies.
One U.S.-based team of leading professionals from the operations research and public health fields has developed a mathematical optimization application — which was funded by the Centers for Disease Control and Prevention (CDC) and the Texas Department of State Health Services — to solve this vaccine allocation problem.
In this article, I will discuss what makes this vaccine allocation problem so challenging and how this team of researchers is employing mathematical optimization to tackle it.
Understanding The Vaccine Allocation Problem
In the midst of a pandemic such as the one we are facing today, public health officials must make rapid vaccine allocation decisions — choosing who will receive the vaccines, where and when.
Generally speaking, there are two different methods of distributing vaccines to the general population: pull-based allocation (which is driven by requests from healthcare providers for specific quantities that they need) and push-based allocation (which is focused on distributing quantities to achieve equity among priority groups). Typically, the majority of vaccine doses are distributed via a pull-based approach, while a smaller percentage of vaccine doses are held back by public health officials who must make decisions on how to allocate these doses to address and correct the inequities that may arise in a strictly pull-based process.
These decisions are highly complex from both a logistical and ethical perspective, and in order to achieve the ultimate goal of allocation — which is to distribute and administer vaccines equitably within priority groups such as healthcare personnel, the elderly and people with certain health risks — public health officials must take numerous factors into account, including:
- Fairness: A priority group in one geographic location should have the same access to vaccines as the same priority group in another geographic location.
- Simplicity: The path from vaccine allocation to distribution to administration should be as simple and smooth as possible for healthcare providers.
- Variety: There are multiple vaccines with different properties (e.g., one shot or two shots and unique refrigeration requirements), and it is imperative to ensure that the right vaccines get to the right people at the right locations and times.
- Quantity: Vaccines are distributed in batches and not in single doses, and numerous aspects such as temperature controls and defrosting need to be considered.
It’s simply not possible for the human brain to take all these factors into account and make fast, accurate decisions on how to allocate vaccines in the most equitable manner possible. To do this, you need an automated software tool like mathematical optimization.
Using Mathematical Optimization To Solve The Vaccine Allocation Problem
The team of researchers mentioned earlier in this article — from the University of Texas, Northwestern University, the Texas Department of State Health Services and the Santa Fe Institute — has built a mathematical optimization application (powered by my company’s mathematical optimization solver, the Gurobi Optimizer) that is capable of handling the Covid-19 vaccine allocation problem.
This application — the objective of which is to optimize the allocation of “multiple vaccine types to multiple priority groups, maximizing equal access” — can provide public health officials with a web-based decision-support tool that they can use to rapidly determine:
- What percentage of available vaccines to keep in reserve for push-based allocation.
- The most efficient and equitable way to distribute vaccines, taking into account all the factors highlighted in the previous section.
As distribution ramps up in the U.S. and around the world in the coming months, this application could be an essential tool for governments, enabling them to optimize the rollout of vaccines to achieve equity across their populations. With this goal in mind, the team of researchers is planning to make the application available to U.S. state governments in the near future.
Technology Can Help In The Fight Against Covid-19
To win the fight against Covid-19, we must enlist our brightest minds and employ our most powerful technologies. The development of the Covid-19 vaccines — which was executed over the past year with such amazing speed and skill — is a shining example of the incredible synergy between humanity and technology.
As the fight against the virus continues, the public and private sectors must continue to look for new and innovative ways to leverage technology — including AI tools — to help us address and overcome the challenges we are facing.
Now that safe and effective vaccines are available, the most pressing, critical challenge is how to distribute and administer these medicines as quickly and fairly as possible to the global population. As I have explained, this is a very complicated question involving numerous ethical and logistical considerations, but technologies like mathematical optimization can help us answer it.
This article was originally published on Forbes.com here.