Flu forecast is based on suspect data

http://www.philly.com/inquirer/opinion/20090902_Flu_forecast_is_based_on_suspect_data.html

COMMENTARY

Flu forecast is based on suspect data

There is no solid evidence indicating that H1N1 will meet dire projections.

By Michael Fumento
The President's Council of Advisors on Science and Technology issued an alarming report on swine flu last week. A typical front-page article about it began, "Swine flu could infect half the U.S. population this fall and winter, hospitalizing up to 1.8 million people and causing as many as 90,000 deaths."

But the council's "plausible scenario" involving those alarming figures is based on three main assumptions, and all three are highly suspect.

First, the report says that "based on past pandemics," a sharp increase in infections "is likely ... starting in September and peaking in October." That's important, because the first shipments of vaccine aren't expected until mid-October.

The problem is that swine flu isn't comparable to those past pandemics. As I wrote in June, it's only called a pandemic because the World Health Organization changed its definition of the term so that "enormous numbers of deaths and illness" are no longer required.

Furthermore, true pandemics have been severe because the viruses involved were alien to our immune systems. But H1N1 virus, of which swine flu is a variety, "has been circulating every year since 1977," noted Peter Palese of Mount Sinai School of Medicine in New York, a member of the president's council. It's consistently part of the seasonal-flu vaccine. The swine flu variety is different in some ways, but this is "something our immune systems have seen before," Palese said.

In fact, older people seem especially protected, apparently from exposure to a strain that circulated before 1957. That helps explain why swine flu disproportionately afflicts younger people.

Palese, who is none too happy with the council's report, calls an October peak "highly hypothetical." James Chin, a professor at the University of California, Berkeley, and former WHO epidemiologist, is even more skeptical. "Just six weeks from now, about a quarter of the U.S. population will have to have been infected, or close to an average of two million infections per day," he noted. It took almost five months for the flu to reach an estimated two million infections.

Faulty figures

A second suspect assumption in the report is that, in the absence of a vaccine, swine flu will infect vastly more Americans - 30 to 50 percent of the population - than seasonal flu does. Why? "Because most of the population lacks significant immunity to a new pandemic strain," says the report. Once again, this overstates the newness of the virus.

The report's third dubious assumption is that the rate of death from swine flu "appears to be similar to" that of seasonal flu. There's no source for this claim in the report, so I contacted Dr. Harold Varmus, president of Memorial Sloan-Kettering Cancer Center in New York and a co-chairman of the president's council. In an e-mail, he cited figures from the British government and two journals, BMJ and Eurosurveillance.

The U.K. figures were described as a worst-case scenario. The rates reported by BMJ came from the United States, Mexico, and three other areas; the high end was from Mexico, the U.S. rate was far lower, and those from the other areas were lower still. Portraying them as a range for the United States, as the council report did, is misleading. Finally, the figures attributed to Eurosurveillance come from four different sets of data. Three of these have maximum death rates far below the council's minimum, and the fourth barely exceeds the council's minimum estimate.

Worried well

On the other hand, comparing current U.S. swine flu deaths (about 550) to the number of estimated infections (two million) suggests that the virus is far less lethal than seasonal flu. A new New York City estimate suggests seasonal flu is 10 to 40 times more deadly.

Chin noted that Australia is having its flu season now. But with swine flu cases having peaked in July, and with no swine-flu vaccine, the government is reporting a flu epidemic not discernibly worse than in recent years. Indeed, swine flu "appears to be replacing the current seasonal H1N1 virus," meaning there could be fewer deaths. Chin told me, "My bet is that the coming [U.S.] season will not be too severe - at or below that of a usual flu season."

The only realistic aspect of the scenario put forward by the council is that emergency facilities could be swamped. But rather than with truly sick people, it will be with the mildly ill and the worried well - as happened during the media fright fest last spring. This time, much of the panic will be due to the council's self-fulfilling prophecy.

Michael Fumento is director of the Independent Journalism Project, where he specializes in health and science issues. He can be contacted at fumento@pobox.com.