Skip to Main Content

You’ve likely seen a version of the image above. Some people observe two faces in profile. Others see a vase.

The same phenomenon can occur when scientists look at data, particularly when they try to weigh the benefits and risks of individual vaccines. Vaccines protect huge numbers of people, generally children, from serious diseases, but in rare cases, certain vaccines can tragically cause harm. How do those scientists figure out which to value more?


This dilemma was at the center of last week’s decision by an expert committee advising the World Health Organization to sharply scale back use of a controversial vaccine called Dengvaxia, the first to protect against dengue infection.

Two years ago, many of the same experts concluded the vaccine was safe to use in children 9 and older in places where dengue infection is almost unavoidable — even though there were strong theoretical concerns the vaccine might put some vaccinated kids at higher risk of developing a severe form of dengue. Severe dengue can lead to internal bleeding, shock, and even death.

Late last year, theory was shown to be reality. After reviewing the data, the WHO’s Strategic Advisory Group of Experts on Immunizations — knows as the SAGE — shifted its stance, recommending last week that the vaccine be given only to children who test positive for a previous dengue infection. A point-of-care blood test doesn’t currently exist, leaving the vaccine’s future in limbo for now.


Benefits trumped theoretical risks. But real risks trumped real benefits — even though the latter outnumbered the former.

A variety of factors influence these decisions, according to interviews with public health experts and ethicists who have made or studied them. The severity of disease being prevented and the treatability of the side effects being caused are crucial, as are ethics, public perceptions, and politics. And critically, these days, so is the likelihood that any negative attention generated by one vaccine might stain the reputations of others.

Dr. Art Reingold has been involved in countless debates of this type, having served for more than a dozen years on first the SAGE, and later on the corresponding body that advises the Centers for Disease Control and Prevention, the Advisory Committee on Immunization Practices, or ACIP.

“I would say that for many individuals having to vote or being on these committees and weigh in, many of them spend an awful lot of time looking at data and reviewing it and struggling with it when it’s not a slam dunk,” said Reingold, who teaches infectious diseases epidemiology at the University of California, Berkeley’s school of public health.

Reingold, whose term on ACIP concluded at the end of last year, noted that while many of that panel’s votes are close to unanimous, some recommendations — or decisions not to recommend a vaccine — are decided by a narrow vote. “And obviously in that instance pretty smart people have come down on different sides of the same question by weighing effectively all the same evidence,’’ he said.

As diseases fade, the vaccine calculus changes

In some cases, these decisions aren’t that difficult. A flu shot can trigger Guillain-Barre syndrome, though this adverse event is very rare. But influenza infection can also provoke this disorder, a form of progressive paralysis from which most people recover.

In other circumstances, the position of the scales — which side is higher, which is lower — tilts over time.

Oral polio vaccine has saved untold millions of children from paralysis in the more than half-century it has been in use. But in rare instances the vaccine also paralyzes, a fact that became clear in the first year of its use, back in the early 1960s.

Sometimes the paralysis occurs in the child who got the dose of vaccine, or in a close contact of that child, as was reported in 1962. Other times the weakened viruses in the vaccine circulate among unvaccinated children, who ingest them in water or food contaminated with traces of feces. As the viruses travel from gut to gut, they can go rogue — regain the power to paralyze. That phenomenon was first observed in 2000. Yet oral polio vaccine is still used today.

As polio eradication efforts have driven down the polio case count to very low levels, the toll of vaccine-related paralysis can surpass the damage caused by the viruses themselves. Last year there were 22 children in the world paralyzed by polioviruses; vaccine viruses crippled 96.

The shifting of the risk-benefit ratio for oral polio vaccine led the United States to switch exclusively to injectable polio vaccine in 2000. In 1996, when the decision to phase out oral vaccine was made, eight or nine children a year were being paralyzed by the vaccine, though polio itself hadn’t paralyzed a child in the United States for over a decade.

The risk posed by the oral vaccine became intolerable, given there was a safer, albeit more expensive alternative — the injectable polio vaccine does not paralyze. “Obviously thinking on these things can change. The risks and benefits relatively speaking can change,” Reingold said.

In 2016, the formulation of the oral vaccines was altered to drop the component that protected against type 2 polioviruses. That part of the vaccine was the most likely to regain the power to paralyze. Type 2 viruses had disappeared in 1999; there was too little benefit and too much risk associated with keeping them in the vaccine.

In the case of Dengvaxia, the calculus is not as clear cut. Evidence suggests that in places where about 70 percent of people have been infected at least once with dengue, the vaccine would prevent seven children from getting sick enough to need hospital care for every additional hospitalized case it provoked. In places where 85 percent of people have been infected, there would be 18 hospitalized cases prevented for every one the vaccine created.

Some dengue experts have argued those benefits should not be ignored. Others argue those risks cannot ethically be incurred.

The SAGE deliberated over whether it was permissible to use rates of local dengue infection as a substitute for individual testing — in other words, could the vaccine be given, as it had previously recommended, in places where studies show most people have been infected at least once?

They concluded both options pose real-world challenges, given the current lack of a rapid, accurate test. But they also noted there is no evidence to date that children who have never been infected with dengue — the ones the vaccine could harm — would ever experience a benefit from Dengvaxia. They worried that wide-scale dengue vaccination programs might be “ethically problematic and have adverse implications for trust and the long-term success of public health programs.”

A climate of mounting distrust

Why have oral polio vaccine risks been tolerated, but Dengvaxia’s deemed serious enough to effectively shelve the vaccine? Here the factors named above plus timing surely play a role.

Experts making these types of decisions these days are doing so in a climate of litigiousness and mounting vaccine refusal and hesitancy. Headlines questioning the safety of one vaccine threaten to fuel rejection of others. The government of the Philippines, where Dengvaxia has been given to more than 800,000 children, has threatened legal action against its manufacturer, Sanofi Pasteur.

“The public has a whole new understanding of science, data, facts, and fake news,” said Michael Osterholm, director of the University of Minnesota’s Center for Infectious Diseases Research and Policy. “We’ve surely had an anti-science movement well before the current situation, but it’s never been as acute.”

Then there’s the issue of who gets vaccinated. Most of these products are designed to protect children, who hold a special position in discussions of the ethics of medical treatments.

“When you have that child population put knowingly at risk, it gets really hard from the ethics point of view to ignore that.”

Art Caplan, New York University bioethicist

The ethical bar must be placed higher when it comes to kids, because they cannot make an informed decision for themselves, said Art Caplan, a professor of bioethics at New York University’s school of medicine. “So I think the issue is not just: Could we accept huge benefits for small risks? Because I think the answer to that is yes. But I think it’s: Can we accept huge benefits for small risks to very vulnerable children?

“When you have that child population put knowingly at risk, it gets really hard from the ethics point of view to ignore that,” he said.

And society’s tolerance of risk has changed, Caplan argued, pointing to the so-called Cutter incident to make his case.

In 1955, it was discovered that children had been mistakenly injected with polio vaccine that contained live viruses. The process by which the viruses in the vaccine were supposed to be inactivated — killed — hadn’t worked. Fifty-one children in the U.S. were paralyzed and five died.

But in the 1950s, polio was an enormous threat. Parents lived in fear their children would end up in an iron lung. Polio vaccination resumed. The company that made the vaccine, Cutter Laboratories, didn’t even go out of business. If a similar incident were to happen today, Caplan said, “it would have shut everything down forever.”

Compare that to the case of RotaShield, the first vaccine licensed to protect against rotavirus infection. These common viruses cause devastating bouts of diarrhea in young children, who can end up in the hospital as a result. Every year some children died of these infections in the U.S., but rotaviruses did more damage in the developing world, where stricken kids didn’t have easy access to hospital care.

RotaShield was approved in the U.S. in 1998. A year later, Wyeth Laboratories withdrew it from the market after studies showed babies who got it were at greater risk of developing intussusception, a type of bowel blockage that can kill if it isn’t corrected in time. The CDC estimated that for every 10,000 children vaccinated with RotaShield there would be one or two additional cases of intussusception over what is normally seen.

Some experts argued that the vaccine should still be marketed in the developing world, where the number of lives saved would far outstrip the cases of intussusception. The WHO estimated that in 2004, more than half a million children died from rotavirus infections, the lion’s share in South Asia and sub-Saharan Africa.

“When you think about the risk-benefit equation in a poor country, almost certainly it would have been far better in terms of illness — preventable deaths averted, cost of care reduced — to continue to use that initial rotavirus vaccine or rather to introduce it and use it in poor countries,” Reingold said.

But the vaccine’s fate was sealed.

“The fact is that the politics around the thing — this vaccine isn’t good enough for rich white children in the United States but it’s OK for poor black children in poor countries — were a non-starter,” he said. “I mean, it just doesn’t sell. Even if that’s still the wisest thing to do.”

It was nearly six years before another, safer rotavirus vaccine made it to market, six years during which more than half a million children a year died from rotavirus infections.