Trolleyology and the Dengue Vaccine Dilemma

“Trolleyology” refers to a series of moral dilemmas that reveal the tensions between utilitarianism — the idea that a behavior is moral if its consequences maximize public good — and our individual intuitions about right and wrong.¹ In a classic scenario, an observer is standing by a railway track where he spots a runaway trolley barreling toward five people who will soon be killed. He’s standing by a switch that, if pulled, would divert the trolley away from the five but cause it to kill one person in its new path. In a related scenario, the observer is on a footbridge when he spots the trolley, and he’s standing by a very large man. If he pushes the man off the bridge, the body will stop the trolley; the five people would be saved, but the man would die. What should the observer do? Most people believe he should pull the switch in the first scenario but shouldn’t push the man in the second. Philosophers, psychologists, and even neuroscientists, have spent the past several decades debating why. One common explanation, familiar to most people who practice medicine, distinguishes between the immorality of intentional harms and the acceptability of merely “foreseen” ones. We wouldn’t, for instance, prematurely end the life of a man who’s had a disabling stroke, even though his organs could be used to save five other patients’ lives.

. .

Medicine also highlights, however, why utilitarianism’s limits cannot be fully explained by intent, since no matter the magnitude of potential benefit, not all foreseeable harms feel morally equal. Let’s say the stroke victim had atrial fibrillation, was started on oral anticoagulation, but a few months later had an intracranial hemorrhage and died. His physicians would probably feel regret but wouldn’t conclude that appropriately targeted oral anticoagulation, which confers a net public health benefit, is immoral for the population at large. But consider a vaccine for a common and sometimes deadly virus. Assume that the vaccine’s public health benefit is clear: if given to 1 million children over 9 years of age, it can probably prevent some 11,000 hospitalizations and 2500 severe cases of disease. But what if, in achieving this population benefit, the vaccine also caused 1000 hospitalizations and 500 severe cases of disease in children who would not otherwise have fallen ill? Would you initiate a widespread vaccination program in a disease-endemic region? If you hesitate, why do these scenarios — both resulting in public health benefits — seem morally different?

These are some of the questions facing governments in regions of South Asia and Latin America where dengue virus is endemic and where, as described by Sridhar et al. (pages 327–40), a dengue vaccine with roughly these characteristics exists. Dengue is a mosquito-borne flavivirus that affects up to 400 million people worldwide and, according to WHO, results in about 3.2 million severe cases annually and about 9000 deaths.² Sanofi Pasteur’s Dengvaxia, which entered the market in 2015, is the only available vaccine. In 2016, the World Health Organization (WHO) recommended that countries in which dengue is highly endemic consider introducing Dengvaxia for children over 9 years of age.² Clinical trial data suggested a signal of vaccine-related harm among younger children. Since younger children are less likely to have been exposed to dengue virus, it was hypothesized that older age in these studies may have been a surrogate for prior dengue infection. It was thus thought that the vaccine could act like a primary infection in children not previously exposed to dengue, potentially contributing to a more severe illness when subsequent exposure to natural infection occurred.

The report by Sridhar et al., who reanalyzed the serum from trial participants to assess their baseline serostatus, supports this hypothesis: the vaccine is protective among those who were previously exposed but increases the risks of hospitalizations and severe illness among the unexposed. The analysis also reminds us, however, that even in endemic regions, seronegativity can persist among children older than 9. The challenge is that many initial infections are asymptomatic, and no point-of-care test is currently available to determine serostatus. Nevertheless, the utilitarian calculus remains clear: in endemic regions, vaccinating children over age 9 reduces the rate of severe infection and hospitalization by 80%. It’s the moral calculus that still feels untenable.

Indeed, the vaccination of some 830,000 children in the dengue-endemic Philippines, which has provoked a massive public backlash, will probably further chill the utilitarian cause. The Philippines Department of Health launched the Dengvaxia vaccination program in April 2016, only to suspend it at the end of 2017 when Sanofi disclosed the findings of harm among the unexposed. In an already vaccine-weary public, the disclosure prompted fear and outrage. The Philippines’ former public health undersecretary wrote in a Facebook post that has since been removed, “This is the biggest government funded clinical-trial-masked-as-a-public-health program scam of an experimental drug in the history of the DOH . . . . This was reckless and irresponsible from the start and the public was deceived into thinking this vaccine would protect children from dengue.”

Of course, the vaccine does protect children from dengue; thanks to the vaccination campaign, there are far fewer severe cases than there would otherwise have been. But the story of illness averted doesn’t stand a chance against the one involving innocent children, a pharmaceutical giant, and a disease that takes a brutal, bloody toll on its victims. At the core of the outrage lies the challenge plaguing all prevention efforts: the faces of the children who were vaccinated and in whom a severe case of dengue subsequently developed will be highly visible in media coverage, while the many more vaccinated children who were spared severe illness will remain unseen.

Governments in dengue-endemic regions, seeking to protect as many children as possible, are now in a bind. The most ethical population approach probably involves informing families about Dengvaxia’s benefits and risks and letting them decide whether to vaccinate. But we should recognize that such deference to autonomy may come at the cost of lives. In a seminal study on perceptions of vaccines, the psychologists Ritov and Baron showed that people are reluctant to vaccinate a hypothetical child when doing so poses a risk of death, even if the risk of death is higher without vaccination.³ Furthermore, the reluctance to vaccinate intensified if people were told that a test existed that could determine who was at risk but that it was unavailable. That we fear acts of commission more than those of omission, even if inaction portends greater harm, doesn’t bode well for Dengvaxia uptake. But should we therefore assume that the more utilitarian approach, involving massive vaccination campaigns, is the best way to save the most lives?

Probably not: in real life, unlike in trolleyology, behaviors’ consequences extend beyond the numbers of lives immediately saved or lost. Consider people’s preferences regarding autonomous vehicles. Such vehicles promise to save lives by removing error-prone humans, but human-designed algorithms will determine the vehicles’ responses to situations in which harm is unavoidable.⁴ What should a vehicle do, for example, when faced with the trolleyesque dilemma of running over 10 pedestrians or swerving but sacrificing the passenger? A recent study found that people want others to drive cars programmed to maximize the greater good. But for their own cars, people tend to prefer programming to protect passengers “at all costs.” Participants were generally opposed to regulation requiring utilitarian standards, which they said would discourage them from purchasing an autonomous vehicle. The researchers conclude, “Regulating for utilitarian algorithms may paradoxically increase casualties by postponing the adoption of a safer technology.”⁴ Similarly, Dengvaxia vaccination campaigns may cause public backlash leading not only to decreased dengue-vaccination rates, but to heightened skepticism about all vaccines.

Although it’s impossible to know how many additional lives are risked by potential anti-vax spillover, further quantification will not reveal the morally tenable solution. Trolleyology is thus prophetic — not because it provides an answer, but because philosophers, psychologists, and scientists have failed to find one. Though a “rational” decision entails maximizing benefit and minimizing harm, any such calculus must consider the way our choices make us feel long after decisions are made. To that end, the WHO’s recent recommendation to implement a “pre-vaccination screening strategy,” vaccinating only people who test seropositive, seems, for now, the most socially acceptable solution.⁵ Because this strategy requires a readily available and accurate point-of-care test, ideally the recommendation will motivate industry and governments to invest in the infrastructure for such testing. Until then, the challenge remains as practical as it is philosophical. To maximize the gains of medical advances, we must begin to address the barriers that reflect less the limits of science than the quirks of our intuitions.