Flu Death Estimates Are Only a Guess, Experts Say

Government projections that as many as 1.9 million Americans could die in a global flu epidemic amount to a guess that could prove to be highly inaccurate, several public health experts say.

“The problem with all the numbers is that nobody knows,” said Dr. Arnold S. Monto, a University of Michigan epidemiologist who specializes in infectious diseases. “I try to avoid coming up with these numbers as much as possible. I know they are based on imperfect information, extrapolating from the past.”

The figures in a draft of the government’s pandemic preparedness plan have heightened public alarm about a highly virulent strain of bird flu from Asia that could undergo a genetic mutation and become easily transmissible between humans.

While agreeing that preparation for a pandemic must be an urgent national priority, Monto and other experts say the death toll cannot be reliably estimated at this point, because no one knows how lethal a mutation of the virus might be.

“Usually, as a virus adapts to human-to-human transmission, it becomes less virulent,” said Ira M. Longini Jr., who analyzes flu and other types of epidemics as a biostatistics professor at Emory University. “A virus that kills the host cannot transmit itself as well. From the virus’ point of view, it wants the host to live.”

The worrisome strain of bird flu, known as H5N1, has spread rapidly among domestic and wild birds in Asia. More than 100 people have been infected with it, and about half have died. However, in virtually all the cases, the human victims are thought to have caught the virus directly from birds.

Perhaps the clearest indication of uncertainty in the mortality estimates comes from the government itself.

Just last year, the Department of Health and Human Services estimated that a flu pandemic could cause 89,000 to 207,000 deaths. Those numbers were contained in an earlier draft of the national pandemic preparedness plan. The worst-case figure in that report was about 10 times lower than the comparable number in the 2005 draft.

The new draft, first reported by the New York Times, estimates 209,000 to 1.9 million deaths.

A spokeswoman for the department declined to discuss the difference in the estimates, and suggested that the latest numbers were under review. The government’s final plan for dealing with a flu pandemic is expected soon.

Trust for America’s Health, a nonprofit organization that focuses on public health issues, has also produced estimates of a pandemic’s toll. In a report released in June, the group predicted that 180,000 to 1.1 million Americans could die.

The lower number represents a relatively mild pandemic, such as occurred in 1968. The higher number forecasts what could happen in a severe pandemic, such as the 1918 flu, which killed an estimated 20 million to 50 million people worldwide.

The government and private estimates were all calculated using a computer model developed by the federal Centers for Disease Control and Prevention. What varies are the assumptions.

“They range from very conservative to worst-case Armageddon,” said Kim Elliott, deputy director of Trust for America’s Health. “There is no definitive way of knowing how severe a strain is going to be.”

A death toll of 1.9 million would be “a pretty doomsday scenario,” she said.

The group’s report illustrates a mid-level pandemic, about three times more severe than the 1968 virus. Under those conditions, it estimates, 541,000 Americans could die out of a total of 67 million infected.

In California, the most populous state, the death toll could approach 61,000 in a mid-level pandemic, the group said.

Longini, the biostatistics professor, said he considered estimates of 1 million or more U.S. deaths “a bit off the mark.”

Other alarming bird flu predictions have been issued recently, only to be withdrawn. In September, the United Nations’ pandemic readiness coordinator, Dr. David Nabarro of the World Health Organization, suggested that as many as 150 million people could die worldwide. The following day, the agency backtracked, saying that 2 million to 7.4 million was a reasonable working estimate.

Dire scenarios hark back to the 1918 pandemic and a much different world.

“Things are very different now in terms of medical care and what we know about the influenza virus,” Longini said. For example, many flu deaths are caused by bacterial infections that take hold in weakened patients. These can be treated with antibiotics, which did not exist in 1918. Still, health experts say they see no real harm in issuing dramatic predictions if doing so nudges people to take influenza seriously.

The U.S. lags some Western countries in preparing for a flu pandemic. Stocks of antiviral medications are low, even though the drugs could help keep an outbreak in check while scientists develop a vaccine. Many state and local governments, which would bear the brunt of a public health emergency, have inadequate policies.

Even during what is considered the “normal” annual season, 36,000 to 40,000 Americans die of the flu -- about the same number killed annually in traffic accidents. But whereas most people now buckle their seat belts, many do not get the recommended annual flu shot.