The illness struck the little baby suddenly.

It was a hot, sticky day late in the summer of 2017. Only 5 months old at the time, her little boy was a peaceful infant, his mother recalls. "He didn't make much of a fuss." The family lives in a small fishing town near the South China Sea in Sarawak, Malaysia, at the mouth of the Rajang River. Their tidy home sits atop stilts, above a maze of canals and families' rowboats tied to piers.

She has six children now; the baby was her fifth. We aren't using their names to protect the family from stigma around the son's illness.

On that humid August day, something was terribly wrong with her child. First, he became feverish. The mother thought he might have the flu or a cold. "The fever went away quickly," she says. But by evening, the child began coughing and struggled to catch his breath. "He was breathing very fast," she remembers.

She took the baby to the nearest clinic, but his condition deteriorated. Doctors rushed them to the nearest city, Sibu. It's three hours away by ambulance, depending on how the ferries are running.

At the hospital, doctors admitted the infant to the intensive care unit. By then, the baby's lungs had begun to fail. They were filled with mucus. He couldn't absorb enough oxygen, his mother says, and doctors connected him to a machine to help him breathe.

For three long days, the child didn't get better. His mother worried for his life. "I was so concerned," she says.

He had pneumonia. "But doctors didn't know why," she says. They ran tests looking for a cause — a bacterium or virus. All the tests for the usual culprits came back negative.

But one pediatrician at the hospital had the foresight to know that scientists might one day have the tools to figure out the cause of the child's life-threatening pneumonia and that perhaps he had a pathogen that no one had detected before. "We are looking for novel infections, even types of viruses that we might not be aware of," says Dr. Teck-Hock Toh, who teaches at SEGi University and heads the Clinical Research Centre at Sibu Hospital.

 Toh's team took a little white swab, like the ones in COVID-19 testing kits, and scraped inside the infant's nose. They took the sample to the laboratory, extracted the genetic material from the possible pathogens present and stored the sample in a freezer. In 2016 and 2017, Toh and his team collected about 600 samples like this one.

What doctors eventually found inside the sample — inside the baby's respiratory tract — has fueled a shift in scientists' understanding of how pandemics begin and made them reconsider the way they search for new threatening viruses. It has made them realize there could be an easier, more efficient way to find viruses like SARS-CoV-2 before they evolve into a global nightmare.

Spillover theories, old and revamped

For decades, scientists pretty much thought they understood how pandemics, such as COVID-19, began. It centers on this idea of what's called spillover.

Most new pathogens, up to 75%, come from animals. They're often viruses that have been circulating in animals for decades, even centuries. At some point, they jump — or "spill over" — into people.

For the past 10 years, I've been a global health reporter at NPR. That whole time, I've heard the same idea repeated over and over again about spillovers: They are extremely rare. Animal viruses tend to stay in their animal host. One way scientists have described it is that a virus spilling over is, in a way, winning the lottery: The virus is in the right place at the right time, and on top of that, it has special, rare characteristics that allow it to infect people. For all these events to coincide is remarkably rare, the thinking went.

This theory has shaped how scientists look for new deadly pathogens — or try to predict which ones could cause future pandemics. In particular, it led scientists to focus on searching for new viruses in wild animals. Since 2009, the U.S. government has spent hundreds of millions of dollars trapping wild animals, such as bats and rodents, cataloging all the viruses circulating in their bodies and then trying to predict which of these viruses will most likely spill over into people and cause a costly outbreak or pandemic. Unfortunately, this effort failed to detect SARS-CoV-2 before the virus could spread to multiple continents.

Over the past few years, a growing number of virologists and epidemiologists have begun to question whether this approach is feasible. Some have blatantly said it won't work.

"I think like projects cataloging viruses, doing virus discovery [in wild animals] is interesting from a scientific standpoint," says evolutionary biologist Stephen Goldstein at the University of Utah. "But from the standpoint of predicting pandemics, I think it's a ridiculous concept." The numbers just don't make sense, Goldstein says. Animals contain more than a million viruses, and only a tiny, tiny fraction of those will ever be able to infect people.

But what if the tiny fraction of animal viruses that do infect people actually jump into people way more frequently than scientists thought? What if spillovers aren't extremely rare but are common enough that scientists can actually detect them inside people?