It was bound to happen again. For the second time in two months, the United States has confirmed a case of bird flu in a dairy worker employed by a farm with H5N1-infected cows. “The only thing I’m surprised about is that it’s taken this long to get another confirmed case,” Steve Valeika, a veterinarian and an epidemiologist based in North Carolina, told me.The true case count is almost certainly higher. For weeks, anecdotal reports of sick farmworkers have been trickling in from around the nation, where H5N1 has been detected in dozens of herds in nine states, according to federal counts. Testing among humans and animals remains limited, and buy-in from farms is still spotty. The gap between reality and what the government can measure is hindering the world from realizing the full scope of the outbreak. And it may hamper experts’ ability to detect human-to-human spread, should that someday occur. “I wouldn’t be surprised if there have been dozens of cases at this point,” Valeika said.The risk to most of the public is still low, as federal guidelines continue to emphasize. But that assurance feels tenuous when “the threat to farmworkers remains high,” Jennifer Nuzzo, the director of the pandemic center at the Brown University School of Public Health, told me. Too often, infectious disease most affects a society’s most vulnerable people; now the future of this virus depends on America’s ability to protect a community whose health and safety are routinely discounted.[Read: America’s infectious-disease barometer is off]Like the first case of a dairy worker contracting avian flu, this second one has at least one reassuring element: Exposure in both cases seems to have involved heavy, repeated contact with infected, lactating animals and resulted in a mild illness that involved only eye symptoms. (In another U.S. case, from 2022, in which a man contracted the virus from poultry, fatigue was the only reported symptom.) Cow udders and human eyes both contain receptors for H5N1 that resemble the ones primarily found in birds, and experts suspect that those receptors are an easy entry point for the virus, which still seems to be very much an avian pathogen. To spread in earnest among people, the virus would still probably need to make a few more evolutionary leaps. For most of the public, “I’m not worried about H5 right now,” Seema Lakdawala, a virologist at Emory University, told me.People who work on dairy farms, though, have reason to worry, Lakdawala added. In the so-called parlors where dairy cows are milked, animals are strapped into machines that latch on to their udders, pump until the rate of flow slows, then release, swinging “off the animal at eye height,” Lakdawala told me, and blasting bystanders with frothy liquid. The machines aren’t necessarily sanitized between each animal—and what cleaning does occur often involves a high-pressure hose-down that also mists up milk. The entire process involves a lot of direct maneuvering of udders, as workers load machinery onto each cow and prime their initial milk flow manually. If workers aren’t directly getting milk on their hands—which will, at some point, touch their face—they’re “constantly being bombarded with aerosols, droplets, and spray,” Lakdawala said.When infected cows are present, that can mean a lot of virus exposure. Lakdawala’s lab has been studying how long H5N1 can persist on milky surfaces, and the initial results, not yet published in a scientific journal, suggest that the virus may linger for at least one to three hours on the same sorts of plastic and metal commonly used in milking equipment. That creates a clear conduit for the virus to move among animals, Lakdawala said—and a very easy path for a human to pick it up, too. Improper disposal of milk could also pose some transmission risk, especially milk from infected farm cows, which still have to be milked if they’re lactating. (Several farm cats appear to have caught the virus from drinking raw milk.) The USDA recommends heat-treating all milk before it’s discarded, but some farms, especially smaller ones, may not have consistent access to the necessary equipment or human power, Lakdawala told me.The CDC has urged farmworkers to don goggles, gloves, high-quality respirators, and other protective equipment in these environments. But those recommendations can’t really be enforced, and it’s unclear how many farms have been following them, or how many workers on those farms are complying. In the rising spring and summer heat, wearing that gear may get even less palatable, Lakdawala pointed out, especially in the steamy, cramped environments in which the people with the most exposure do the brunt of their work. Goggles and other tight-fitting eye protection, in particular, are tricky: “They get dirty very quickly,” Lakdawala said. Workers can’t see what they’re doing through milk-spattered lenses.Enthusiasm for testing cows and people has also been low on farms, as business owners and employees alike weigh the economic and personal risks they face if one of their herd is reported as sick. And although asymptomatic cows are likely responsible for a good degree of spread, the USDA requires testing of only a subset of the cows being moved between states. That basically ensures that “we won’t find a virus before a farmworker is exposed,” Nuzzo told me. Similarly, the CDC maintains that “testing of asymptomatic persons” for H5N1 “is not routinely recommended,” and close contacts of infected people aren’t guaranteed a screen for the virus. Those sorts of delays could allow infections to simmer—potentially past the window in which intervention with treatments such as Tamiflu or forestalling transmission to close contacts is possible. The fact that this second case was caught doesn’t mean that testing is anywhere near sufficient: The diagnosis was made for a farmworker in Michigan, which has more aggressively tested its dairy herds, Nuzzo said. Nuzzo and Lakdawala both argue that stockpiled vaccines should be offered en masse to farmworkers while their risk remains so high—but federal officials haven’t yet made the injections available. (The USDA and the CDC did not respond to requests for comment.)[Read: The bird-flu host we should worry about]These shortfalls would be concerning for any population contending with under-the-radar infections. But among farmworkers especially—a group that includes many migrants and uninsured individuals living in rural regions—H5N1 could play on existing health disparities, Anne Sosin, a public-health researcher at Dartmouth, told me. If protecting farmworkers is a priority, Valeika said, “I think we’re kind of failing.”Researchers are also unsure just how much risk infected farmworkers may pose to their close contacts. Other forms of pink eye are pretty transmissible—and someone who has recently rubbed their eye, Lakdawala said, could presumably pass H5N1 by touching someone else’s hand, which could then touch their face. Experts also remain worried that an infection in the eye might find a way to travel to other parts of the body, including the respiratory tract, especially if the virus were to pick up the sorts of mutations that could adapt it to the receptors in our lungs. (The Michigan dairy worker’s nose swab, thankfully, turned up negative for an H5 virus.)The virus doesn’t yet seem poised for such a jump. But these flu infections are still a problem for everyone. “If we fail to stop it in the highest-risk groups,” Sosin told me, the threat to the rest of the public will only grow. H5N1 may never spread human-to-human. If it does, though, it will almost certainly have been helped along by transmission in a community of people that American society has failed to properly protect.