With the coronavirus pandemic continuing to rage, government leaders have been preoccupied with how to get vaccines out quickly. Given that everyone can’t be vaccinated at once, who should get shots first? Who should get them second? How should the limited resources (shots, medical personnel, etc.) be managed to end the pandemic as soon as possible? While the Centers for Disease Control and Prevention has been efficient at quickly delivering vaccines to the states, the states are failing at implementing the widespread scheduling effort needed to get these vaccines into people’s arms, with many having administered less than 20 percent of the vaccines in their possession.
Medical ethics and public health concerns obviously are priorities in administering vaccines, but when we think about scheduling, it helps to turn to a domain of mathematics known as queueing theory. Queueing theory is the study of how to schedule waiting in lines. When queueing theorists design a scheduling policy, their first step is understanding their objective. For example, when the objective is to minimize average delay, queueing theorists have shown that one should schedule customers in shortest-first order; that is, favor those customers who can be served most quickly. This is why there are express lanes in supermarkets, and it’s also why data centers try to run short jobs ahead of long ones.
When scheduling vaccinations, the CDC and most states have adopted an oldest-first scheduling policy, which says that vaccines should be doled out largely in decreasing order of age. The multistage prioritization scheme begins with health care professionals and residents of long-term care facilities, then moves on to the oldest age group (75+), then the next oldest age group (65+), and so on. Coming from a medical ethics and moral philosophy standpoint, these agencies are developing policies on the assumption that older folks are more susceptible to death and hence should be prioritized.
There are two objectives for vaccine distribution:
▪ To maximize efficiency and throughput by getting the vaccine into as many arms as possible as quickly as possible.
▪ To minimize transmission by stopping COVID-19 from spreading further before the vaccine can do its work.
From the perspective of these two objectives, what is the right scheduling policy? We argue that youngest-first beats oldest-first for meeting the objectives. That is, vaccines should be given to the youngest eligible population first. The Moderna vaccine has been approved for people as young as 18, while the Food and Drug Administration authorized the Pfizer vaccine for those 16 and older.
From the efficiency perspective, it’s easier and faster to inoculate younger people. They are already co-located in a vast network of some 140,000 K-12 schools and colleges across the nation, many containing thousands of students on a single campus. Co-location alone greatly increases efficiency, since people don’t have to drive to a vaccination distribution center and wait in cars for their turn to be called. Second, young people are quicker to vaccinate. Furthermore, young people don’t require seating and bathrooms as they wait in line.
From the perspective of minimizing spread, contact tracing has shown that it is younger people who are doing the spreading. True, older people are more susceptible to death from COVID-19, but that’s only if they are exposed. COVID-19 has surged among millennials and Gen Z-ers, many of whom are asymptomatic and spread the disease to older people.
By vaccinating the youngest eligible first, both objectives are met: Efficiency and throughput are maximized through vaccinating more people more quickly, and spread is minimized while waiting for the vaccine to take effect. Of course, it’s still important to start by vaccinating health care professionals, because that is both efficient (since they’re often co-located) and fits with the goal of minimizing spread. Simultaneously, states should follow a youngest-first policy rather than the current oldest-first policy. States should bring schools and colleges into the vaccination business and take advantage of the infrastructure and networks that are already in place to maximize both vaccination rates and benefits.
Mor Harchol-Balter is a professor at Carnegie Mellon’s computer science department. Justin B. Hollander is a professor in the department of urban and environmental policy and planning at Tufts University.