Economics, Politics, and Psychology: The Case of Avian Flu

There is concern about the possibility of a flu pandemic that would be as or more lethal than the 1918-1919 Spanish flu pandemic, which may have killed as many as 50 million people worldwide; 500,000 died in the United States. A strain of avian flu first detected in 1997 has infected some 150 million birds, including chickens, ducks, and geese, mainly but not only in eastern Asia. More than 100 human beings have been infected, of whom about half have died. The victims were infected by contact with diseased birds rather than by contact with infected humans. As long as the only transmission is from birds to humans rather than from humans to humans, there will be no human pandemic. But the flu virus is notoriously mutable; if the current strain of avian flu mutated into a form that made it transmissible from one infected person to another, it might spread rapidly through the human population. Stocks of vaccine for immunizing people from the avian-flu strain, and of drugs (mainly Tamiflu) for treating already infected people, appear are inadequate. The Swiss pharmaceutical manufacturer Roche, the only producer of Tamiflu, has been reluctant to license its production to other manufacturers.

The probability of a pandemic is unknown, but probably significant because of the vast number of infected birds and the increasing number of infected human beings, in whom the virus might mutate into a form in which it was transmissible to other human beings. Flu pandemics have been frequent. There were two in the twentieth century besides the Spanish flu pandemic. They occurred in 1957-1958 and 1968-1969, and each killed more than a million people worldwide. All three twentieth-century pandemics involved strains of avian flu. There was also the swine-flu pandemic scare in 1976; the failure of a pandemic to materialize has engendered some skepticism concerning the likelihood of an avian flu pandemic. One of the most foolish forms of commentary on issues of public safety is to note the number of false alarms and infer from that number--entirely illegitimately--that there is nothing to fear.

The world in general and the United States in particular are unprepared for a flu pandemic. Although the current strain of avian flu was discovered eight years ago, vaccine development and production are just beginning, along with stockpiling of Tamiflu. Apparently there is at present only enough vaccine for 1 percent of the U.S. population. Roche has only a limited capacity for producing Tamiflu and, as mentioned, is reluctant to license other pharmaceutical firms to produce the vaccine. The President recently announced a $7.1 billion program for improving the nation's defenses against flu pandemics, but it will take years for the program to yield substantial protection.

So we are seeing basically a repetition of the planning failures that resulted in the Hurricane Katrina debacle. The history of flu pandemics should have indicated the necessity for measures to assure an adequate response to any new pandemic, but until an unprecedented number of birds had been infected and human beings were dying from the disease, very little was done.

The causes are the familiar ones. People, including policymakers, have grave difficulty taking measures to respond to risks of small or unknown probability. This is partly because there are so many such risks that it is difficult to assess them all, and the lack of solid probability estimates makes prioritizing the risks inescapably arbitrary, and it is partly because politicians have truncated horizons that lead them to focus on immediate threats to the neglect of more remote ones that may be more serious. ("Remote" in the sense that, if the annual probability of some untoward event is low, the event, though it could occur at any time, would be unlikely to occur before most current senior officials leave office.) But by the time a threat becomes immediate, it may be too late to take effective response measures.

There is also a psychological or cognitive impediment--an "imagination cost"--to thinking seriously about risks with which there is little recent experience. Wishful thinking plays a role too. There is the inverse Chicken Little problem: the illogical reaction that because the swine-flu pandemic never materialized, no flu pandemic will ever materialize. Another example of wishful thinking is the argument that most people afflicted by the Spanish flu in the 1918-1919 pandemic died not of flu, but of bacterial diseases such as pneumonia that the flu made them more vulnerable to. But, first, is is far from clear that "most" died of such diseases, and, second, the current strain of avian flu appears to be more lethal than the Spanish flu. Only about 1 percent of Spanish flu victims died, whereas 50 percent of known human victims of the current avian flu have died. That percentage is probably an overestimate because many of the milder cases may not have been reported or may have been misdiagnosed; but it is unlikely that the true fatality rate is only one-fiftieth of the current reported rate. It is estimated that even a "medium-level" flu pandemic could cause up to 200,000 U.S. deaths and a purely economic impact (that is, ignoring the nonpecuniary cost of death and illness) of more than $150 billion.

A specific problem with respect to preventing flu pandemics is the difficult economics of flu vaccines. Because of the frequent mutations of the virus, a vaccine may be effective for only one season, in which event the manufacturer must recover his entire investment in the vaccine in just a few months. The expected cost of the vaccine to the manufacturer is increased by his legal liability (a form of products liability) for injuries due to the side effects of the vaccine. If a large population is vaccinated, a percentage of the population, amounting to a very large number of people, will in the normal course experience illness in the months following the vaccination. Many of them will be tempted to sue, and uncertainty about the causation of an illness may enable a number of persons to recover damages who would have become ill anyway. This problem can be solved in a variety of ways: by requiring proof of negligence rather than imposing strict liability for side effects of vaccination; by increasing the burden of proving causation in vaccination suits; or by the governmen's undertaking to indemnify the producers for damages attributed to the vaccine. Even if such steps were taken, there would be a strong case for the government's financing vaccine development and procuring large quantities of vaccines for distribution as needed.

Measures along these lines are now being taken; and the government's agreeing to indemnify manufacturers for damages resulting from vaccine side effects would be a natural evolution from the National Vaccine Injury Compensation Program, created in 1986, which provides relatively modest "no fault" compensation for injuries caused by vaccination but does not preclude lawsuits against the manufacturers of the vaccine. However, measures not begun until the threat of a pandemic is imminent may be too little, too late.

A difficult question is compulsory licensing of patented or other proprietary flu vaccines. On the one hand, compulsory licensing would speed the production of vaccine; on the other hand, it would reduce the incentive of firms to develop new vaccines in the first place. The answer may be to combine compulsory licensing with generous research subsidies.

Hurricane Katrina and now the danger of an avian flu pandemic--one an actual, the other a potential, catastrophe for which the nation failed or is failing to prepare adequately--underscore the need for institutional reforms that will overcome policy myopia based on inability to plan seriously for responding to catastrophes of slight or unknown probability but huge potential harm.