Monday’s elevation of the terror alert level was followed by Tuesday’s scary headlines: "Anthrax attack could cost 100,000+ lives" and "Modeling of bio-attack on large city predicts mass casualties with prior distribution of antibiotics."

The headlines were spawned by the publication (Mar. 17) of a new study in the Proceedings of the National Academy of Sciences predicting that a large-scale anthrax attack on a major U.S. city could cause 123,000 deaths, given our current level of preparedness.

This estimate is based on a mathematical model of a scenario where 2.2 pounds of anthrax spores are released from a height of 328 feet over an urban area of 11.5 million people and no one is treated until 48 hours after the attack.

The model predicts that 13.1 percent of the population (1.49 million) will become infected and that 8.3 percent of those infected will die (123,000) in an area 120 miles long and 11 miles wide.

The researchers made a number of public health recommendations, including expanding the size of the health care network and home stockpiling of antibiotics such as Cipro. They claim 60,000 lives could be saved if antibiotics are immediately available.

Unfortunately, the mathematical model employed has significant, if not show-stopping shortcomings. Such models are only as good as the thinking that goes into them. Otherwise, they are about as useful as crystal balls.

The model assumed that all million billion anthrax spores in the hypothetical 2.2 pounds released are capable of causing infection. This probably would not be the case.

Anthrax spores must be sufficiently small to infiltrate the alveolar spaces of the lung. This would require finely milled, and specially treated spores — known as weaponized anthrax which is not easy to manufacture, especially in large quantities.

Once released into the air, many spores would combine with dirt and dust to form clumps too large to penetrate the lungs.

The model seems to assume the entire target population just happens to be outside at the time of the attack, making them especially vulnerable to falling spores. But most people tend to be indoors at any given moment. Any spores that did manage to enter buildings would tend to settle harmlessly.

The mathematical model pretty much ignores the likelihood of unfavorable meteorological conditions, instead relying on a constant breeze blowing at 10 miles per hour.

But a stronger or shifting wind would probably disperse spores into harmless concentrations. Too little wind, and the spores will fall to the ground, clump with dust and dirt and not rise again in harmful concentrations.

After all, simple exposure to a lone anthrax spore doesn’t lead to infection. Thousands to tens of thousands of spores are needed to start a life-threatening infection through inhalation, the most worrisome form of anthrax exposure.

The researchers acknowledged in their study but apparently not to the media that their model "may be too simplistic to monitor and predict the spatiotemporal anthrax concentrations after an actual attack."

Moreover, the model’s predictions conflict with the only large-scale population exposure to aerosolized anthrax. In 1979, anthrax spores escaped from a Soviet bioweapons facility, infecting 79 people and killing 68 in an urban area of about 1.2 million people.

It’s not a trivial toll, but it’s not thousands of casualties either.

Mass bioterrorism with anthrax is simply not a high percentage play. The terrorist group Aum Shinrikyo dispersed anthrax aerosols in Tokyo on several occasions, failing to produce any illness.

It’s one thing to manufacture and distribute a small amount of aerosolized anthrax — as happened in the October 2001 mail attacks — but its quite another to effectively deliver massive amounts from the sky.

Perhaps the most disturbing deficiency of the model and the one that raises questions of a potentially sensationalistic motivation on the part of its designers is that it wasn’t designed to provide a range of results depending on varying conditions. As is, the model essentially depicts a single scenario based on an unlikely accumulation of worst-case conditions.

Better that is, less scary and less newsworthy modeling would provide potential results for a variety of scenarios (e.g., varying weather conditions, spore size, infection rates and population distributions) and assign probabilities for their occurrences.

Such results would provide government and public health officials with more and better information on which to make preparedness decisions. If a catastrophic mass anthrax attack is highly unlikely, we shouldn’t spend much time worrying about it.

This model seems designed to shock and scare — not too inform. Sounds like another kind of terrorism?

Steven Milloy is the publisher of, an adjunct scholar at the Cato Institute and the author of Junk Science Judo: Self-defense Against Health Scares and Scams (Cato Institute, 2001).

Respond to the Writer