Last week, news broke that an influenza outbreak at Lackland Air Force Base in Texas had caused around 160 people to become sick.
The timing was both bizarre and not. On one hand, the U.S. armed forces ended its flu vaccine mandate in April. On the other, a flu outbreak in June is unusual.
Indeed, after Secretary of Defense Pete Hegseth announced the policy, an influenza outbreak in military barracks seemed inevitable. But so soon? That may have surprised some. But, in retrospect, it shouldn’t have.
Military officials in the Army, Navy, and Air Force reportedly took action this week to once again require flu shots for basic trainees, but it’s worth taking a closer look at what happened. The math on this is beautiful. The results are not. In short, Hegseth did not break epidemiology. He merely stumbled into the mathematically inevitable “find out” phase of his “FAFO” approach to vaccinating our nation’s newest influx of Air Force recruits.
The Epidemiology of an Outbreak
Outbreaks occur when the conditions permit. Remember when everyone was obsessed with the “reproduction number” of COVID-19 — the R₀, or “R-naught”? The R₀ is the average number of “secondary” infections one infected person will cause. (Note: R₀ refers to the reproduction number before mitigation measures are put in place. “Rt” — or sometimes just “R” — refers to the number once efforts to slow the spread down are in place. For now, we’ll mostly just use R₀.)
This is key: R₀ is a summation of biological, behavioral, and environmental factors. Many people assume that R₀ is purely biological. That is, how contagious a pathogen is. But the other features matter greatly.
In fact, whether an outbreak can even occur boils down to one mathematical statement: Is R₀ greater than 1.0? If, on average, an infected individual spreads a pathogen to more than one person (R₀≥1), an outbreak occurs. If R₀ is less than 1.0 (R₀<1), the outbreak dies off. That’s really it.
SIR, This Is an Outbreak
Here’s a simple schematic of the life cycle of an outbreak, using what’s called a “SIR” model.¹ (In fact, R₀ is derived from equations that feed into SIR models.) In any situation, there are susceptible (i.e., not immune and “available” to be infected), infected, and recovered people.²
Let’s imagine a virus enters a military barrack. If nobody is immune (i.e., no prior infection, no vaccination), everyone is susceptible. Within a short time, everyone will be exposed. With enough exposures, everyone will become infected. Eventually, everyone either recovers or does not (if they don’t, see footnote 2).
For this experiment, let’s also assume that each person in the barrack has contact with 15 people that are close enough for transmission. Of those 15, only three get infected (an attack rate of 20%). That would make the R₀ of this experiment 3.
Eventually, outbreaks die out because there are not enough susceptible people left to become infected and continue the chain of spread. It’s like a fire running out of wood. That also means that if an entire barrack gets infected in June, a new recruit carrying that same flu bug in July won’t cause an outbreak because there’s nobody left to infect. No “susceptibles”? No outbreak.
What Did Secretary Hegseth Do? He Increased ‘Susceptibles’
By ending the influenza vaccine mandate for U.S. military recruits, Hegseth massively increased the number of susceptibles at Lackland Air Force Base. That, indeed, raised the R₀ of influenza in that particular environment. In our experiment above, Patient Zero in the June outbreak had adequate close contact to infect 15 people and infected three of them. Those three spread it to nine people, and so forth.
But what if the virus was vaccine-preventable, and vaccination rates were 100%? (That’s what would have happened at Lackland before the policy change in April.) The answer: No outbreak. And now you know why: no susceptibles.³
According to reports, after the U.S. military ended its flu vaccine requirement, only around 40% of recruits have opted to receive it. That means that instead of nearly zero susceptibles during training, up to 60% of the new recruits might be susceptible (though it could be less, owing to 2026 infections or previous vaccinations whose effects have waned).
Applying this to our experiment above, the R₀ (“R_Pete,” if you will) is only 60% of three, or 1.8 (0.6 × 3.0 = 1.8). That’s still well above the R₀>1 needed to support an outbreak. That’s in contrast to the pre-April circumstances, in which the R₀ would effectively be 0 (or at least well under 1).
The Conditions Needed for an Outbreak
Having enough susceptibles is necessary but not sufficient to drive an outbreak. You also need the right conditions. When considering the likelihood of transmission (a major part of R₀ and SIR models), there are two major categories to consider, outside of biology: climate and human behavior.
First, climate matters greatly and is a major reason why there’s a cold and flu “season.” Temperature and humidity affect respiratory particles’ ability to spread through the air. In colder, drier air (i.e., winter), smaller particles can more easily remain suspended in the air. This means that in winter settings, people don’t need to be as close for transmission to happen. This allows the virus to travel farther and transmit more easily. In hot and humid air, particles travel shorter distances before falling to the ground where they can’t be inhaled or directly deposited into the eyes and mouth (sidebar: except toddlers).
But summer spread is possible. It’s just less likely. In fact, wastewater data shows us that many pathogens decrease in the summer — but do persist at lower levels.
Influenza A vanishes in the summer, while Influenza B persists at very low levels. That means it’s possible for someone to enter a military barrack in June with Influenza B (or other viruses like RSV, COVID-19, or human metapneumovirus) and start an outbreak. (Note: an Influenza A outbreak in June in the U.S. would be a surprise, but it could happen if, say, someone who had just arrived from the southern hemisphere, where Influenza A season is roughly the opposite of ours, showed up.)
But one person with an infection is not enough to drive an outbreak. People have to mix and come in contact with each other for spread to occur. In flu dynamics (and many other infectious diseases), the “contact rate” matters greatly. That is, how many people are interacting, how often, for how long, and in what kinds of settings?
In fact, one of the biggest differences between summer and fall or winter in terms of transmission risk and contact rates is that schools are out, removing a major driver of flu transmission. Studies on school holidays and breaks have found meaningful reductions in flu transmission, often in the range of about 20% to 30%, because kids, who can be very susceptible to flu, are no longer mixing every day. Some studies have shown that longer breaks lead to less flu transmission.
So, places like schools, college dorms, nursing homes, prisons, and yes, military barracks, are perfect hotbeds for outbreaks.
Leadership Matters: Preventable Outbreaks Leave Us Vulnerable to Adversaries
When Secretary Hegseth announced that U.S. military personnel would no longer be required to get flu shots, his tone was striking. His announcement began with an attack on the “disastrous” Biden administration, rather than a frank discussion of the science. He referred to vaccine mandates as an “era of betrayal.”
This was not a level-headed discussion of the risks and benefits of vaccines for a low-risk population (i.e., young military recruits are unlikely to have severe disease, though it can happen). The message was more in line with anti-vaccine rhetoric, describing the prior requirement as too broad and emphasizing individual choice. His tone read as discouraging vaccination as an act of medical freedom.
This is concerning. More than that, in military settings, vaccination policy is not just an individual medical preference. It is also a force-readiness decision. A sick training cohort is not effective, and leaves our recruits, and our country, more vulnerable to bad actors. Thankfully, some military officials have taken note and are once again requiring flu shots. The country should heed this warning.
Special thanks to Marisa Donnelly, PhD, who writes “Your Local Epidemiologist in New York,” for co-authoring this piece!
A version of this post originally appeared in “Inside Medicine.”
Footnotes:
1We often use SEIR, which has an extra compartment that describes the delay between being “exposed” to a pathogen and being infectious to others. But for the sake of this argument, we’ll use the simpler SIR model.
2This last compartment can also be split into those who survive (“recovered”) and those who do not (“removed”). Those who die get “removed” from the model.
3Here we are assuming the vaccine is 100% effective, but the concepts hold true for any vaccine that lowers infection rates sufficiently to drive R under 1.0.
