How To Right the Train Wreck of COVID-19 Vaccine Distribution
by Susan Dentzer and Eugene Litvak
The rollout of vaccines against COVID-19 — a “dismal failure” to date, in President-elect Joe Biden’s words — has been nothing short of a train wreck. As of mid-January, the Centers for Disease Control and Prevention reports, only about 12.3 million of the 31 million doses of vaccine shipped to states have been stuck into individuals’ arms. Vaccines are in short supply in some areas, while those shipped to other states and hospitals are sitting in refrigerators and going unused. Senior citizens sometimes wait in line for hours to get vaccinated, but in other locations, breeze right through. Meanwhile, the nation is recording record new numbers of COVID cases daily, record hospitalizations, and record deaths.
Like many train wrecks, this one was avoidable — and course correction now will constitute “one of the most challenging operational efforts we have undertaken as a nation,” Biden has observed. His new vaccine plan incorporates important new elements: Potentially invoking the Defense Production Act to boost supplies of vaccine; helping states to stand up mass vaccination sites; mobilizing “thousands of clinical and nonclinical professionals” to staff them, and even paying for National Guard troops to administer vaccines. But all of these elements will still have to be woven together in carefully calibrated operational plans within each state. That’s where it will be important to rely on the principles of operations management — the same tools that systems engineers use, for example, to avoid train wrecks.
On the whole, these engineers must maintain an exquisite balance of supply and demand: The numbers of trains, train cars, and personnel available to operate on time and safely, and the volume of passengers who want to travel, and when. Hundreds of other factors influence these two sides of the equation and enable “throughput” — in the case of trains, the number of passengers, or the amount of freight, that can be transported around the country. If a conductor falls asleep at the controls, or a crippling snowstorm or electrical outage occurs, throughput is blocked, and a train wreck can occur.
How might the vaccine rollouts already under way have been different if systems engineering principles had been employed?
First, the federal government and states would have taken very different approaches to estimate and meet vaccine demand. As it was, the federal government initially allocated supplies to states based on their populations, aiming to send just enough to administer one round of a two-shot regimen to the highest-priority groups. But it should have gone further, asking states for a far deeper analysis of how they might bring vaccine supply and demand into optimal balance, given the many factors involved.
On the demand side, states should have been asked to estimate just how many of their residents fit into all the categories of people initially prioritized for vaccinations by a CDC panel in phase 1a: frontline health care workers and residents of long-term care facilities. These are knowable numbers, and these — not the overall state population — should have been the initial point of departure for the vaccine allocation process. States should also have been given a definite time limit on delivering the vaccines — for example, getting all people in phase 1-a vaccinated by the middle of January.
Next, on the supply side, states should have been asked to estimate how many people in the state could be mobilized to dispense shots over that time period, and what locations could be used for that purpose. (This should have been a rapid, “ballpark” analysis compiled over just a few days.) Such factors are every bit as important as the actual stocks of the vaccine on hand in determining what “supplies” of shots could actually be delivered into individuals’ arms within the deadline. That’s because any one of them could prove to be a bottleneck in the process — and the dimension of any single bottleneck will ultimately determine the speed of the process, or “throughput.” So if the bottleneck is not enough people to administer the vaccine, no matter how large the supplies of vaccine are, the speed of the process will be limited.
Forcing this exercise of estimating true supply and demand on states would have compelled them to develop realistic estimates of how many people could be vaccinated in the time allotted. The federal government should then have sent states only the amount that they could realistically deliver in the appropriate time frame. If individual states wanted more, they would have to take steps to change aspects of the “supply” equation — for example, by opening up new locations for mass vaccination, or mobilizing the National Guard or others to assist in administering vaccines.
Now that the Biden administration plans to help in some of these efforts, states have a better shot at bringing supply and demand into balance. But what is critically important now is synchronizing all aspects of the vaccine response: The number of people to be vaccinated; the number of vaccines available; the number of people who can administer them, and the number of places. Otherwise, there will still be bottlenecks, and some amount of vaccine will either go unused, or supplies will fall short.
To our knowledge, no state has yet included operations management experts in their state vaccine advisory groups. It’s not too late for them to acquire that expertise to help them formulate their plans.
After the current vaccine phases 1a and 1b (people 75 and older plus frontline essential workers such as grocery store workers, firefighters, and police officers) will come phase 1c — people ages 65 to 74, people ages 16–64 with underlying medical conditions, and other essential workers. These groups will easily number at least 150 million nationwide if all agree to be vaccinated. States should be working now to firm up the numbers — asking the Social Security Administration, for example, to tell them how many Medicare beneficiaries live in each zip code. They should also make realistic estimates of how many people will willingly forego the vaccine due to hesitancy and resistance, even as they develop educational and outreach efforts to drive the numbers down.
With all of these measures in place, states will have a firmer fix on true vaccine demand. Then they can address the supply side.
As vaccine supplies build in the coming weeks, states need to count and enlist all possible points of vaccine distribution — for example, all physicians’ offices, clinics, pharmacies, and health centers — as well as create the new mass vaccination sites proposed by Biden. They can also mobilize the entire corps of each state’s physicians, nurses, other health care providers, and National Guardsmen, and detail where they will administer vaccines and what their daily administration targets will be. But all of these pieces, on both the broader supply and demand sides, will need to match up. States should produce the equivalent of timed flow charts demonstrating how these newly synchronized systems will work.
These plans would constitute just the sort of skilled operations management that keeps trains running on time, and safely. It could still keep the effort to fight the deadliest pandemic in a century on track as well. ##
Susan Dentzer is Senior Policy Fellow at the Robert J. Margolis Center for Health Policy at Duke University. Eugene Litvak is adjunct professor of operations management at Harvard’s T.H.Chan School of Public Health