The idea that new diseases come from human invasions of pristine environments probably stems from HIV, the virus causing AIDS. HIV originated as a simian virus and most likely crossed the species line around 1930. Yet AIDS did not turn into a human pandemic until sexual activity transformed a sluggish disease into an explosive and probably more virulent one, contends evolutionary biologist Paul Ewald of the University of Louisville. It was not just the spark of infection, but what happened to that infection when it entered human society, that converted a simian retrovirus into HIV and then into the deadly pandemic that is AIDS.

This is why pandemics are necessarily social phenomena. Each human pandemic exists because social conditions have allowed it to evolve. Some diseases move across species lines and yet harm us not at all. An example is Ebola-Reston, a virus deadly to monkeys. It infected four people at an animal facility in Reston, Virginia, in November 1989, but no one fell ill; the infections were detectable only through blood tests. At the same time, other, related Ebola viruses ravage the human body. Some particularly notorious strains almost always kill.

What most new diseases, including Ebola, don’t do very well is spread from person to person. Without such spread there is no pandemic. There isn’t even a new, self-sustaining human disease. If we are looking for the real engine for infection—the driver of the disease factory, if you like—we have to understand what human-to-human transmission is, and how it turns an animal disease into a human one.

To be transmitted, a germ has to be shed from one human host and picked up by another. Some germs, like cholera, make their hosts produce copious diarrhea. If there is poor sanitation or people are crowded together, those germs are likely to infect other hosts. Respiratory infections are shed into the air or onto surfaces. If someone sneezes into her hand and touches a doorknob that you touch afterwards, you may catch her infection. Some germs float through the air and you breathe them in. Measles and tuberculosis spread this way. These germs have to keep their hosts mobile. If you are knocked off your feet right away by an infection and you don’t walk around sneezing or coughing, it is harder for the germ to find another person to infect.

But manipulating the host into producing diarrhea or a germ-filled sneeze is not a trivial task. Highly virulent H5N1 avian flu, which many experts thought would trigger a deadly pandemic between 2004 and 2007, has so far been unable to convert from a chicken flu to a human one. On the vast chicken farms of Asia, where millions of birds are crowded under disease-factory conditions, H5N1 gained tremendous virulence, becoming a sophisticated chicken-killing machine. When people are exposed to high doses of this virus, they can succumb as well. But H5N1, well adapted to the higher temperatures and particular cell receptors found in chickens, does not spread from person to person. In its present form, therefore, it has shown no capacity to turn from a well-adapted chicken disease to a human one.

Experts have warned that H5N1 could recombine with a human virus at any moment, mutating so it becomes transmissible. But transmission doesn’t just happen; it evolves. It would take a whole series of mutations, and they would have to happen in the bodies of a chain of human beings, not chickens. Nor would a single recombination event be likely to spawn a fully transmissible human disease. Instead, you would have to start with a huge dose of virus, says flu virologist Earl Brown of the University of Ottawa. A huge dose would be more likely to contain many diverse strains. Then you would have to expose those strains to a chain of human hosts. As the virus moved from person to person (assuming it could), the strains that are most easily transmitted would be selected over the others. Eventually you would get a human-adapted virus, though what that virus would look like is anybody’s guess. To produce a truly lethal human flu or other virus, you would need the same sort of conditions that produced virulent chicken flu. You would need a disease factory for people.

Yet in 2005 scientists were so terrified of a lethal avian flu pandemic that they devised all sorts of preparations, which surely helped inspire the CDC’s Zombie Apocalypse campaign. Tamiflu, an antiviral drug, was stockpiled in enormous quantities, even though the late influenza virologist Edwin Kilbourne noted that it is useless for prophylaxis. The drug also has dangerous side effects—and there are practical limits for how long you could keep people on it, anyway. Governments and drug companies developed candidates for anti-pandemic vaccines, and some experts called for mass prophylactic vaccination, even though no human H5N1 virus existed. Government agencies urged people to stockpile water and peanut butter and tuna fish in preparation for what Michael Osterholm, director of the Center for Infectious Disease Research and Policy at the University of Minnesota, told the press would be “three years of a given hell.... I can’t think of any other risk, terrorism or Mother Nature included, that could potentially pose any greater risk to society than this.”

We never got three years of hell, but we did get a new pandemic influenza. While experts were fixated on birds in Asia, this strain came from pigs in Mexico. Packing pigs into cramped quarters produces another kind of disease factory. In the winter of 2009, people living near a giant industrial pig farm in La Gloria, Mexico, reportedly became ill with a severe respiratory virus. No one except local reporters and a few health officials paid attention.

Then the virus began to spread. By the time the supposed “index case”—a 5-year-old La Gloria child named Edgar Hernandez—was identified, the so-called swine flu had been transmitted widely across Mexico. Within days it landed across the border, and in just weeks it spread across the world. Pigs and people, both mammals, pass flu back and forth quite easily; a pig is more like a person than a chicken when it comes to temperature and cell receptors. Pigs were always more likely to produce a pandemic flu than chickens. But most people weren’t looking.

Enormous animal farms remain a source of danger, a far greater danger to people than new viruses emerging straight out of nature. Virologist Brown warns that we have to find new ways to raise animals that don’t pack them so closely together in huge industrial farms. Under such crowded conditions, even the most innocuous germs can rapidly evolve virulence. The deadly Escherichia coli outbreak in Germany last year, which sickened more than 4,000 people and killed more than 50, may be traceable to modern factory farming, which uses massive doses of antibiotics to curb animal infections, likely converting a normally benign microbe into an antibiotic-resistant killer.

Hospitals also encourage the emergence of lethal human microbes. Paul Ewald points out that any conditions allowing germs to be transported from a person immobilized by serious illness to new human hosts can produce deadly infections. Ironically, hospitals turn out to be highly efficient disease factories. They allow the proliferation and spread of dangerous germs among patients, and the evolution of those germs to extreme levels of virulence. The last decades have seen the evolution of virulent, antibiotic-resistant staphylococcus germs (MRSA), now spread worldwide, along with potentially lethal strains of Clostridium difficile, a usually benign intestinal infection. And the normally innocuous soil bacterium Acinetobacter baumannii has plagued veterans’ hospitals, evolving intense virulence and antibiotic resistance and afflicting wounded soldiers returning from Afghanistan and Iraq.

The CDC’s Zombie Apocalypse program is well intentioned, and for natural disasters it gives great advice. But preparing for a pandemic has almost nothing to do with personal supplies of crackers and duct tape. It lies in understanding the real engines of infection: the giant industrial farms and crowded hospitals with insufficient hygiene.

Many European hospitals have reduced their rates of MRSA transmission by screening patients with nasal swabs prior to admission, and some American hospitals are following suit. Shutting down industrial farming should probably come next. These germ incubators need to be replaced with less intensive farms that raise fewer animals on more land. Switching to grass-fed cattle, which live on the open range, would help as well. The price, including the end of cheap chicken, pork, or beef, might be more than we are willing to pay, even at risk to our health. But as long as there are industrial farms, the possibility of pandemics arising from them persists. Closing disease factories, wherever they occur, may not make for compelling cinema, but it is the most effective form of pandemic preparation we can devise.

Outbreaks around the world
When new germs break loose, humans—not nature—are often to blame.

In the modern world, humans wield considerable control over their environment—even over the diseases they face, as several recent outbreaks attest.

CHOLERA IN HAITI: Absent from Haiti for at least a century, cholera swept in after the powerful earthquake of January 2010. The timing was no coincidence: Genome sequencing reveals that the cholera strains that struck were identical to strains from South Asia. The microbes arrived from Nepal with a group of United Nations peacekeepers. The U.N. team disposed of raw sewage directly into streams and rivers.

PNEUMONIA AROUND THE GLOBE: A hospital-acquired, virulent strain of a common, gram-negative intestinal bacterium caused several severe outbreaks in 2011. This strain of Klebsiella pneumoniae is resistant to most antibiotics, including carbapenems, a class of drugs long used to treat this infection. The strain sparked outbreaks in Panama, France, the Netherlands, and Canada, preying on people with compromised immunity.

E. COLI  IN GERMANY: A deadly strain of Escherichia coli known as EHEC O104:H4 broke out in Germany last May, sickening more than 3,800 people there and killing 53. The disease also killed an Arizona man who had traveled to Germany. The strain, which can cause kidney failure, most likely originated in fenugreek seeds imported from Egypt. Researchers think the gene for the toxin causing kidney damage originally entered E. coli, a normally harmless intestinal bacterium, from Shigella, another intestinal germ. If fenugreek sprouts were the source, they were probably contaminated by runoff from an animal feedlot or from fresh manure or sewage, since E. coli grows only in human or animal intestines.

FOOD POISONING IN JAPAN: Last April and May, another potentially lethal E. coli strain infected diners at several Japanese restaurants serving raw meat. This strain, called O111, killed four people and infected at least 54, 19 of them critically. It can cause bloody diarrhea and sometimes kidney failure. E. coli O111 sickened 341 people in Oklahoma in 2008.

CLOSTRIDIUM IN CANADA: Virulent Clostridium difficile erupted in 25 outbreaks in Ontario hospitals in 2011, with dozens of potentially related deaths. Although Canadian hospitals promote vigorous hand-washing, the facilities themselves can be contaminated with this resistant germ, which forms spores that survive on surfaces. Anyone on antibiotics, which kill normal intestinal flora, is vulnerable to Clostridium infection, but the elderly and those with weak immune systems are at greatest risk.

Wendy Orent is the author of Plague: The Mysterious Past and Terrifying Future of the World's Most Dangerous Disease.