Regrettably, Robert F. Kennedy, Jr. is now the Secretary of the Department of Health and Human Services (HHS). In his characteristic dystonic growl, he has called for increased research and policy focus on addressing the root causes of chronic disease. President Trump’s February executive order establishing a presidential Commission to Make America Healthy Again puts Kennedy, as Chair, in charge of a crackpot list of objectives; some of the least eyebrow-raising examples include producing assessments of chronic disease prevalence and contributors to childhood chronic disease in the US.
But Kennedy has, at the same time, endorsed an eight-year “break” in federal infectious disease research, consonant with his (false) belief, advanced in his recent Senate confirmation hearings, that infectious diseases receive more (increasingly threatened) federal research dollars than chronic ones. While this is far from the most alarming development around Kennedy’s tenure as HHS Secretary (that honor might go to the ongoing measles outbreak radiating from West Texas), it is instructive as a window on his anti-vaccine politics. Chronic diseases, at least until the slashing began, were indeed the major focus of federally-funded health research; and that imbalance between chronic and infectious research efforts may not be ideal.
The fact that, contra the guy whose job it is to know this, we already devote outsize resources to the root causes of chronic diseases reflects a conceptual paradigm particular to public health and epidemiology. This paradigm, the concept of “epidemiologic transition,” is taught in public health schools as an axiom: as nations “develop” (get wealthier), they undergo a parallel quantitative and qualitative transformation in population health. Quantitatively, in that wealthier countries have lower mortality generally, and qualitatively in that the causes of mortality shift from infectious diseases (like measles or tuberculosis) to chronic conditions and afflictions of longevity, like cardiovascular disease and cancer.
This epidemiologic transition paradigm dates to the early 1970s, and in fact had a long first life prior to its revival, about 20 years later, as received wisdom in public health. The original epidemiologic transition idea was the result of efforts to integrate family planning (also known as population control) with health services and epidemiology under the aegis of the World Health Organization (WHO), wresting some of this authority away from demographers. We don’t need, for our purposes, to fully comb out the tangled roots of this concept or sketch the totality of its strange history. (For a thorough treatment, see Weisz and Olszynko-Gryn, 2010).
The concept of the epidemiologic transition—infectious diseases afflict poor countries and chronic diseases affect rich ones—universalizes very specific historical, technological, political, and geographic circumstances as innate processes of “development,” which unfold in each country according to its particular conditions. It gravely misrepresents the way that countries are related to one another in a global capitalist economy (i.e., countries are not independently “developing,” but are rather associated with each other through relationships of power and exploitation, such that some countries are “developed” because other countries are poor). Most importantly, as this theory of population morbidity thrives in its second life as public health ideology, it inculcates a very specific way of thinking about infectious diseases.
Per the epidemiologic transition framework, infectious diseases can be “solved”—even if they aren’t solved in some places, they are at least theoretically solvable with the right level and application of technical investment. The tangible consequence of reliance on this framework is that biomedical scientists struggle to think about infectious diseases in social terms. This conceptual under-resourcing is set to merge with and amplify the real, material defunding of science currently underway, compliments of Trump and Kennedy. The timing could not be worse, as we are confronted now with an infectious disease threat rife with devilish complexity, in both ecological and social realms.
Avian flu, or HPAI (highly pathogenic avian influenza) as it is sometimes called, is not new. Since its emergence in Asia in the late 1990s, there have been several sporadic HPAI outbreaks, mainly occurring among farmed poultry in Asian countries with occasional zoonotic (animal-to-human) transmission. Per Jahid and Nolting (2025), there were 878 human cases between January 2003 and July 2023, of which around 52% were fatal. (It should be noted here that this is the “apparent” case-fatality rate, likely to be overstated because milder zoonotic cases go undetected; a recent review estimates the actual case-fatality rate for human H5N1 cases at around 2%. It’s also important to keep in mind that the scale of human infections is miniscule relative to the scale at which birds and other animals contract it.)
Previous outbreaks have generally been successfully curtailed with tried-and-true methods of “stamping out” or “culling” (a term of art for, basically, the ruthless liquidation of infected flocks)—including the 2014-2015 U.S. outbreak of avian flu H5N8 that was ultimately contained by culling nearly 50 million birds. Previous experiences suggest a template for action: find the virus, vaccinate or cull birds, maybe think about compensating poultry producers for lost inventory. But the current outbreak of avian flu H5N1—which began in 2020 as the world was preoccupied with COVID—is different.
This outbreak is caused by a genotype of avian flu H5N1 called 2.3.4.4b. It has been able to spread explosively in wild birds in addition to commercial poultry flocks. This, in turn, has driven the continuous intercontinental spread of the virus over the last four years, seeding outbreaks among (to name a few) farmed mink in Europe, sea mammals all along the South American coast, and both farmed poultry and dairy cattle in the United States. The persistence of the virus in wild birds, season after season, year after year, and its ability to establish extensive reservoirs in mammalian hosts (mink, sea lions, cows) make this outbreak qualitatively different than previous ones. It also indicates, as Peacock et al. (2024) note in a comprehensive review article, that containment strategies used in prior outbreaks are not working. What is different? H5N1 2.3.4.4b is one of several descendants of a specific lineage of avian influenza that has shown an incredible propensity to thrive and to reassort.
But let’s back up. What does this mean? To understand the social import of technical descriptions like “a propensity to reassort,” it actually helps to start small. Avian flu is a type of influenza virus. A virus, at its most simple and schematic, is just a bit of genetic material inside a molecular envelope. At the most basic level, a virus works something like this: attach to a host cell (say, a lovely pink one in the soft mucosal lining of your nasal passages), get into the host cell, and use the host cell’s “machinery” (in the industrial metaphor favored by writers of pop-sci and textbooks) to make more copies of itself. These copies then burst out of the host cell, like baby spiders from an egg sac, and go on to find new cells to infect, repeating the cycle.
There are four broad categories of influenza, grouped according to the range of hosts they tend to infect: A, B, C, and D. We will focus on influenza A, of which avian flu is a subtype, and which is mostly found in birds and waterfowl, although some strains of influenza A also—as this year’s winter flu season can attest—readily infect humans. Seasonal influenza epidemics in the winter months are usually characterized by some mixture of influenza A and influenza B strains, although the influenza A strains causing seasonal flu outbreaks are not avian strains, but rather are human-adapted. This means they can easily pass between humans via coughing, sneezing, breathing, and so forth.
The so-called “species barrier” between human and avian flu is in many respects a literal barrier at the molecular level. Influenza has two important macromolecules on the surface of its viral envelope called hemagglutinin and neuraminidase, denoted H and N (hence the HxNx naming convention). Both are important in the species-specificity, infectivity, and spread of influenza. The specific characteristics of the H molecule, in particular, constitutes the main species barrier. To borrow an analogy from Mike Davis’s book The Monster At Our Door, H is the “key” that recognizes and binds to receptors on “locked” host cells, cracking them open and allowing the virus in.
Different types of H molecules—for influenza A, there are 18—open different receptor “locks” on the host cell’s surface. These different receptors are distributed differently among the tissues in the bodies of the various potential host species for influenza A. The H5 molecule of H5N1, for example, can engage a type of receptor that is found in great abundance in the lower gastrointestinal tracts of birds, particularly waterfowl like ducks, geese, and swans. In the current ongoing outbreak in U.S. cattle, a viral tropism, a sort of preference, for the mammary or milk-producing tissue of cows has been documented.
This is because mammary tissues express this same receptor in abundance as well. (This is also why raw milk is likely to be oozing with H5N1, and means that milking practices have important implications for the spread of H5N1 among cattle, and from cattle to humans.) Humans do have some cells with this H5 receptor, but these tend to be found in the lower respiratory tract (deep in the lungs) and in the conjunctiva (whites) of the eye. Which is to say, not in the tissues that make for pandemic potential: those of the upper respiratory tract, nose and throat. While animal-to-human transmission of H5N1 is currently possible, it is much more difficult than human-to-human transmission of this kind would be.
In order for H5N1 to jump the species barrier and acquire pandemic potential, it would have to somehow develop the capability to bind to the types of receptors that are expressed in abundance in the upper respiratory tract of humans. There are two ways that an influenza A virus (henceforth IAV) like H5N1 could develop this capability. The first is mutation, or evolution. Influenza viruses mutate quickly and constantly because the enzyme, RNA polymerase, that replicates their RNA genome is sloppy and error-prone. More mutations means more variability among individual virus particles for natural selection to work on.
Since mutation is random, most mutations tend to compromise the viability of the virus particle somehow, dying out as quickly as they arise; others, though, might confer some kind of survival advantage that allows the viral particles carrying the mutation to outcompete other variants. It is possible that a viral strain could acquire the several mutations it would need to adapt to human-to-human spread this way, and it becomes more likely the more that the virus circulates, and the more opportunities it has to mutate.
Even in a widespread panzootic (a pandemic in animals), this type of evolutionary change is still stochastic, undirected, and improbable. What’s more, some parts of the flu virus, in particular the critical H molecule, appear to be somewhat “conserved” from an evolutionary standpoint—even when the virus has a lot of opportunity to mutate, the portion of the genome encoding the H molecule remains relatively stable. This could be because of the critical role of H in the fitness and survival of a given viral strain, which could disfavor mutants in a kind of evolutionary “don’t fix it if it ain’t broke” arrangement.
While, perhaps unsurprisingly, news coverage of viral surveillance tends to play up concern over the discovery of mammalian adaptations, many of these mutations take place in other components of the viral genome, like those encoding the virus’s rather minimal structural proteins, that mutate more readily. While steps should obviously be taken to reduce the spread and host range of avian influenza in order to deprive the virus of opportunities to mutate, this isn’t the scarier of the two possible routes of change.
The scarier possibility is that a flu virus takes a sort of shortcut, called reassortment. Reassortment is a little idiosyncrasy of influenza virology (some other viruses can do it, too). The viral genome of influenza is “segmented,” meaning that the genes on it are arranged in neat little packets along the chromosome. The sequences encoding the H and N molecules, for example, are on different segments. IAVs can thus perform a sinister party trick. If a single host cell is infected with two different flu strains at the same time, the strains can “reassort,” meaning they can swap genome segments with one another.
While the possibility of H5N1 evolving pandemic potential via mutation and natural selection is a real threat, the possibility that H5N1 simply acquires it via a one-time reassortment event is a threat that exists at a different level of possibility, and consequences. This latter possibility is what keeps virologists and epidemiologists up at night. The aphorism, in this part of the public health world, that “reassortments make pandemics” reflects this anxiety—and the historical record substantiates this concern. The flu strain causing the 1918 pandemic is thought to have been a reassortant strain (though for various practical reasons, the exact lineage of that viral strain is hard to reconstruct). So too with the 1968-70 (H3N2) and 2009 (H1N1 or “Swine ‘09”) pandemics.
A short note about pigs. In their review, Peacock et al. (2024) describe the “continued absence” of H5N1 from pigs during this outbreak as “highly fortunate.” This is because pigs are ideal vessels for reassortment events on two levels: the biological and the social. Biologically, pigs are susceptible to infection by a wide range of influenza viruses, both avian and human-adapted—many of their cell types express the receptors for both avian and human-adapted H molecules. And socially, the way that pigs are raised as commodities encourages infection and reassortment. As Li et al. (2025) describe, with characteristic scientific understatement, “the complex ecosystem and herd management practices taking place in swine position swine herds as ideal locations for maintaining influenza infections endemic, facilitating the emergence of new strains.”
In the current outbreak, H5N1 has not, for now, been detected in pigs—highly fortunate. But in identifying reassortment as a major threat, the focus of the analysis suddenly telescopes out from the nanometer-size viral particles and their behavior in cells to the macro-scale patterns of interactions between animals and people. Reassortment events happen when different types of animals and people get together under different conditions—in fact, the strains circulating in wild birds right now are themselves reassortant strains (recall that H5 viruses of this lineage love to reassort). Favorable conditions for IAV reassortment include, infamously, an extensive interface between wild and domestic birds, such as exists in the Poyang Lake region of southern China, where commercial chickens and ducks are raised in intimate proximity to a number of wild species of waterfowl.
Reassortment, though a molecular event, represents a qualitatively different phenomenon than the paradigm of the epidemiologic transition can accommodate. While reassortment is an action of viral particles in host cells, its scientific and social relevance derive from the macrosocial patterns that make it more likely for certain viral strains to find their way into the same host cell at the same time.
These patterns are deeply structured by the unfolding historical process in which we find ourselves, capitalism. Capitalism compels people to do certain things under certain conditions, enables certain ecological arrangements and forecloses others, and incentivizes certain kinds of alterations and perturbations in the pattern of ongoing contact between humans, animals, and environments, both built and “natural.”
Market reforms in China in the 1970s and 1980s, for example, contributed to transformations in land use, animal husbandry (in both scale and density), and influenza ecology. Removal of limits on the number of animals that a given enterprise was allowed to raise and the vertical integration of small-scale poultry farms into larger supply chains led by so-called “dragon-head” corporations drove an explosion in the number of domestic ducks and chickens raised in certain areas of China, Poyang Lake among them. There was also an intensification in the density of domestic animals raised right up against the wildlife interface—along the same waterways frequented by the waterfowl that quietly play host to influenza A viruses. Because vertically integrated producers bear all the financial risk for flocks that must be culled due to avian flu outbreaks, there is an incentive for farmers to also raise free-grazing “sideline” ducks alongside their commodity flocks, creating what anthropologist Lyle Fearnley calls a “positive feedback loop” in the spread, evolution, and reassortment of bird flu.
This is not the only place where this type of interfacing occurs. And I’m not trying to say anything special or particularly abstruse about this example vis-à-vis capitalism—capitalism is just what we happen to live under—nor do I aim to overattribute any special causative property to capitalism here, as leftist analyses of public health can unfortunately tend to do. Sometimes the machinations of capitalist accumulation are directly causative of population health patterns, but perhaps just as often, they aren’t, working instead in mysterious ways that are more mediated, distributed, and indirect.
What I’m trying to do here is integrate something of a systems-level understanding of social patterns with a bit of basic biology. The paradigm of the epidemiologic transition, and the current state of public health thinking more generally, make this a difficult task. Too often the basic biology is siloed off from the social realm in specialist journals or mystified by technical jargon and impenetrable complexity. My attempt here is to lay out some boundaries of probability and uncertainty amidst a very complex situation. The uncertainty bounds remain wide, surely, but, in formulating such subjects to help make the public aware of the real threats that we do and do not face, I think we need all the help we can get.
What would make a reassortment with pandemic potential more likely during our present outbreak of H5N1? To begin to bracket this question within uncertainty bounds, consider some of the pathways by which a regular human seasonal flu virus and H5N1 could come to coinfect the same host cell. In more influenza-ecological terms, what are the contours and surface areas of the interfaces between wild animals, domestic animals, and humans? Focusing on humans for now, some basic epidemiology can, in its awkward way, help narrow the uncertainty a bit more.
It is certain that some human “spillover” (when a pathogen specific to one type of organism infects another type of organism) cases of H5N1 go undetected, but looking at documented cases does provide a sense of where spillovers are taking place. The current outbreak of H5N1 in the U.S. began in early 2021. Since then, there have been 70 documented domestic cases, almost all of which (67) involved some known exposure to animals. Additionally, the vast majority of those 67 cases represent occupational exposures to animals. Aha—so now we can imagine one possible scenario that could make a coinfection, and by extension a reassortment event, more likely.
Let’s imagine a record flu season in the U.S. Let’s imagine a worker on a dairy farm, sick with flu—say they have a kid in day care—who has to come in to work because they don’t get paid sick days. We can further imagine that this worker seeks no treatment because, as an agricultural worker, they’re likely to be uninsured, undocumented, or both—structural violences that strongly discourage engagement with the healthcare system.
Even if H5N1 cases have been detected in the herds, even if the farm owner (individual or corporation) is participating in whatever reporting or regulatory requirements obtain in their jurisdiction, we can easily imagine that this worker is not provided with personal equipment (PPE). Or maybe they do get some PPE, but it’s too hot or uncomfortable to wear consistently. Let’s envision this worker attaching and reattaching milking machines to the udders of different cows as part of their routine job activity. Doing this can not only spread H5N1 between cattle, but the resulting “spatters and aerosols,” often happening at the eye level of the worker (recall the susceptibility of the conjunctiva), can result in human spillovers, too.
This little thought experiment is just one of a huge number of potential circumstances we could elaborate. (We haven’t even considered pigs in this example.) That number, while huge, is—importantly—finite. From the perspective of the epidemiologic transition, this combinatorial complexity in viral and host dynamics across multiple scales (cellular, organismal, social, ecological) makes the landscape of avian flu into a hopeless slough of technical despond. From a more socially-oriented perspective on infectious disease, though, each notch in this fractal represents a possible opportunity for intervention and mitigation.
There are levers we can pull to reduce the risk of a bad reassortment event by quite a lot. These points of general intervention have nothing to do with H5N1 in particular. The recent record flu season, for example, is related to a number of factors: poor uptake of the seasonal flu vaccine (under 50% this winter, well below the recommended threshold of 70%), a poor match between the vaccine strains selected for the vaccines and the strains circulating, and unsupportive social, labor, and education policies that effectively force people to work while sick.
The Trump administration has pulled the United States out of the World Health Organization (WHO), which means the loss of access to the ongoing influenza surveillance that’s coordinated through the organization’s Collaborating Center for Surveillance, Epidemiology, and Control of Influenza. The abrupt cancellation of a Food and Drug Administration advisory panel on strain selection for the 2025-2026 flu season, while not likely to limit access to flu vaccines next season on its own, portends more bad consequences of the WHO withdrawal. The point is that, the dimming likelihood of success under this administration notwithstanding, there are specific technical and policy steps at the level of seasonal flu response that could serve as pressure points for advocacy.
Things are murkier and more vexing on the side of the issue that does have to do specifically with avian flu. This outbreak is only the latest of several since the emergence of HPAI, of which H5N1 is but one strain. Since 2021, migratory birds have carried H5N1 all over the globe; some papers have documented multiple independent reintroductions of the virus into North America via bird migration. The spillover into cattle that occurred in late 2023 is a likely result of contact with wild birds, as is a more recent, independent spillover event. The critical role of wild birds in this outbreak challenges conventional thinking on the epidemiologic transition model and the technical conquerability of infectious diseases; it also raises questions about the most appropriate control strategies.
Elon Musk and RFK Jr. have both recently posted their addled thoughts on bird vaccination, asserting that avian flu should just be allowed to rip through flocks, killing the “vulnerable” birds and leaving the strong, a kind of ketamine-warped and brain worm-inspired Great Birdington Declaration.
These men do not know what they’re talking about, but it’s true that vaccinating birds can have some adverse consequences, including selecting for vaccine-resistant strains of influenza. Prior vaccination campaigns, like China’s ambitious programs to vaccinate all domestic birds, and their records of successes (reducing zoonotic transmission) and failures (potential for cryptic spread, since the shots do not prevent infection) can and should be used to inform the design of a vaccination strategy appropriate to this outbreak. Given that the other options are, per RFK Jr. and Musk, just allowing avian influenza to run its course, or intensive programs of stamping-out, destroying large numbers of commercial flocks, vaccination strategies should be seriously and carefully considered. Unfortunately, in the prevailing political climate, that the issue will receive due consideration is dubious at best.
There are further interventions that are technically possible—giving limited doses of H5N1 vaccine to people at a high risk of exposure, making more H5N1 vaccine, investing in platforms and programs to scale up development and distribution of new influenza vaccines generally, increasing uptake of seasonal flu vaccines. The FDA did eventually make recommendations for the 2025-2026 flu shots: the composition will include H1N1, H3N2, and an influenza B strain. There are also preparatory and mitigation steps that could be taken, given the fundamental uncertainty and the novel technical difficulty involved in this particular outbreak of avian flu.
There is a world, just parallel to our own, where, in the event of a pandemic scenario, workplaces, schools, and federal and state governments could use the lessons of our recent experience with COVID-19 to minimize spread, and, ultimately, complications and death. These steps—technically indicated, technically possible—grow less and less likely with each passing day, as the horizon of political possibility around public health constricts further and further around the singular goal of building a patronage network to support TikTok influencers and the supplements they sell. The tragedy of public health is on full display here, as it has always been: it is technically very simple and socially very complex.
Avian flu right now exists in a twinned condition: it is a salient, present threat, a novel problem we must deal with—and simultaneously, it is very much not inevitable that it will play out like the worst-case scenarios that are casually cast around by doomers and feared by embattled public health professionals. As COVID made clear, social policies, above and beyond scientific ones, have tremendous impact on population health and disease spread. Occupational health and safety measures, like distributing PPE that is both protective and wearable to agricultural workers, free (and private) testing, and social policies like those that make it easier for workers to organize, that ensure paid time off for illness, that reduce financial or other barriers to accessing health care—or that protect people from deportation sweeps—would each shift the risk landscape around a potential avian flu pandemic. Taken together, the mitigating effects could be substantial. But to say that these sorts of policies are deeply disfavored at this moment of ascendant fascism is an understatement. So what—the only thing people ever really want to know—do we do?
Frameworks like the epidemiologic transition are important because they structure, at the subconscious, ideological level, how we think about what we should do now, what is possible to do, and what we can make possible. From the point of view of the epidemiologic transition, and mainstream thinking in public health and in science, all we need are the right technical tweaks to put right the problem of avian flu. In this outlook, the fact that American scientific research is being slashed and burned—raided, plundered, vandalized, the earth salted, the spoils distributed to beef tallow content creators and the richest man on the planet—makes it less likely that our good, dedicated scientists will be able to devise those technical tweaks.
So, in this view, the horizon of solutions to this political problem extend only to “Stand[ing] Up for Science,” to rally around narrow parochial demands related to the technicalities of federal grant funding for research. To insist that the federal government just put it all back the way it was. There is a ripening danger of disorganization and demobilization here, related to the epistemic commitments of narrow, technical thinking in the sciences. Science is a social process embedded in higher-order social processes; no part moves in isolation.
The attacks on science are not only malicious and blundering, though they are that too. As I wrote earlier this year, they are part of the right’s hegemonic project to take control of the definition of reality itself, established and licensed through the empirical sciences in what we call “facts.” The attacks—like this week’s bone-deep cuts to the Centers for Disease Control and Prevention, the National Institutes of Health, the Food and Drug Administration, and the National Institute for Occupational Safety and Health, to name a few—have to be understood in this specific context.
It is not accidental that the programs and agencies most pertinent to the risk of avian flu and other outbreaks are the most threatened. It is being done in specific retribution for the COVID pandemic, and how COVID undermined Trump’s ability to define reality, to control the narrative around himself. There were the horse dewormer diehards in the first Trump term, of course, but it ultimately was COVID, more than anything else, that tanked his reelection prospects. Former Trump campaign adviser Brad Parscale said as much in a recent PBS documentary. According to him, COVID was a “scientific fact,” not amenable to media manipulation, and you can’t message your way around a scientific fact. The wholesale liquidation of American science underway right now seems to say—the hell I can’t.
It’s a grim picture, I know. Public health is both conceptually and materially under-resourced for the threat of avian flu, and, worse, is now deliberately being destroyed. The conceptual limitations of public health and the structural defunding of science will have synergistic destructive effects on our ability to prepare for and respond to avian flu.
Even so, in these most disfavorable of structural and political conditions, the worst is not inevitable. If we view science as a technical specialty sealed off from the rest of society, we’ll understand avian flu as a technical issue, and the fixing of NIH grant indirect rates as simply flagrant attacks on science for science’s sake. In this sense, the right’s assaults are fatally wounding, distressing, incomprehensible—and demobilizing. They undermine what we take for granted as a social good to which everyone aspires: public health.
If, however, we can situate these attacks within the larger crisis (of capitalism, of class struggle against its degrading effects, and for the dignity and freedom of regular people) and link them together in a broad, narrative understanding of how science functions in society, we can see them for what they are: concerted attempts by a fascistic political formation of grifters and billionaires to consolidate power, self-deal, and destroy whatever institutions of civil society dare oppose them.
Each point of complexity, as with avian influenza, is a possible site of struggle, and each moment of disintegration a possible point of recomposition. The challenge, as ever, is organizational and social. As for the sciences, they will be antifascist, or they will simply cease to be at all. ♦
