But as global health leaders struggle to ensure pandemic vaccines won't just be a tool for wealthy countries, influenza scientists admit they face an enormous conundrum, one that could stand in the way of efforts to transform vaccine for the few into vaccine for the many.
Simply put, scientists can't be certain how much vaccine is needed to protect people against novel influenza viruses such as H5N1 avian flu, because they don't know what the immune system of a person protected against a new flu strain would look like.
Sure, they can observe whether immunization with H5N1 vaccine produces certain antibodies and to what levels the antibodies rise, but they have no way of gauging how much protection those antibodies will provide if the person is exposed to the virus.
"We can't get the answer to that until the pandemic comes. There's just no way," admits Dr. Robert Webster, the renown flu researcher from St. Jude Children's Hospital in Memphis, Tenn.
The uncertainty about what protection against H5N1 actually looks like is bedevilling ongoing debates over whether people should be pre-vaccinated against the virus or whether a program of ultra-low or single dose vaccines could be safely used to stretch limited supplies and protect millions more people.
"There are two very bad scenarios. And it's not that anybody's wrong or right, it's just that the science isn't there," explains Dr. Jesse Goodman, director of the branch of the U.S. Food and Drug administration that assesses and licenses vaccines.
"So one bad scenario is that if you could have used less (vaccine per person) and you didn't have enough doses and you didn't use less. Less people are protected. But an equally bad scenario is you cut a dose to a certain thing and it doesn't protect anyone."
"The truth is probably somewhere in between. And the science . . . to get there is not going to be trivial."
In the vernacular of the vaccine world, scientists don't know what the so-called correlates of protection are for H5N1 vaccine -- the immune system markers found in the blood that are the signs that a person has been effectively protected against infection.
Scientists think they know what levels of which antibodies translate into protection when people are vaccinated against human flu viruses -- H3N2 and H1N1. It's thought that certain levels of antibodies which block the hemagglutinin protein on the exterior of flu viruses are critical to protection against those strains.
The measurements used by industry and government regulatory agencies were devised years back based on what are called challenge studies, where people were vaccinated and researchers then deliberately tried to infect them with flu viruses. The researchers then charted the antibody responses triggered by exposure to the viruses.
But it's neither ethical nor safe to deliberately infect people with H5N1 avian flu, which has killed roughly 60 per cent of those known to have become infected with it. So scientists are using the tests and scales worked out for H3N2 and H1N1 vaccines when trying to assess whether H5N1 vaccine will be effective and at what doses.
Regulatory agencies like the FDA are using this best-guess assumption and have told vaccine manufacturers that this is the bar H5N1 vaccines will have to hurdle in order to gain approval.
But each flu subtype has its own characteristics and quirks, a fact that makes assuming something about H5 viruses based on the behaviour of H3 viruses a dicey enterprise.
And the tests aren't standardized.
One of the tests for human flu strains is done in red blood cells drawn from chickens; the same test for H5N1 is done in horse red blood cells. Vaccine expert Dr. John Treanor admits it is questionable as to whether identical levels of antibodies against H3 virus (based on the chicken blood test) and against H5 virus (tested in horse blood) will offer equal protection to people.
"There's almost no way to prove that, and it's an apples and oranges comparison," says Treanor, who heads a vaccine evaluation unit at the University of Rochester, N.Y.
H5N1 viruses are considered poorly immunogenic, meaning they don't trigger the strong immune reaction needed to guard against infection.
But Webster and Treanor have conducted studies that provide tantalizing clues that H5N1 vaccines may be more protective than they are given credit for with current measurement techniques.
Treanor and some colleagues recently rounded up 37 people who had been vaccinated with an H5N1 vaccine in 1998 and who were willing to get a booster shot against the virus.
Blood tests showed no evidence of H5N1 antibodies before the volunteers were re-immunized. But the booster shot generated a strong antibody response -- much stronger than in volunteers who also got one shot but who hadn't been previously immunized. That suggests the immune system still harboured defences against the virus, defences not registered by the standard tests.
And working in ferrets, Webster and a team from St. Jude's showed that animals that had previously been vaccinated against an H5N3 virus survived exposure to what should have been a lethal dose of H5N1, even though testing showed the ferrets didn't have the antibody levels thought to equate to protection. (Ferrets are considered the best animal model for studying human flu.)
"We can have protection in the animal models with no hemagglutinin antibody at all," says Webster. "So maybe it's the test that's wrong."
Goodman agrees these findings do raise questions about whether H5N1 vaccines are conferring protection that isn't being measured using the currently devised tests.
But if science doesn't know how to measure whether something exists, will public health authorities or regulatory agencies take what amounts to a leap of faith and assume these vaccines are more protective than they seem?
"How confident are you that protection in ferrets is going to really mean that you're protected?" Treanor wonders.
"You're asking yourself whether you're confident making major public health decisions based on animal data that may or may not be true."
