The fatality rate for severe acute respiratory syndrome, or SARS, has been widely reported as 4 percent. But many experts take issue with the way the CDC is calculating the death rate -- and say that the infection may be much more, or less, deadly.
An accurate fatality rate is impossible to determine at this point in the epidemic, experts say. The actual death rate will depend on how many people who are now sick die, and how many people suspected of having the disease turn out to have just a bad case of the flu.
Some researchers believe the 4 percent figure is too low because it is not being accurately measured. The CDC is calculating the death rate by dividing the number of deaths by the number of probable cases.
But because the number of probable cases includes some people whose outcome is not yet determined, some say the denominator in the CDC's equation is too large.
"Mortality rates are usually calculated on patients whose outcome is known," said Dr. Henry Niman, a surgery instructor at the Harvard/Massachusetts General Hospital. "(The CDC's number) would be accurate only if all patients hospitalized survived, which has not been the history of the disease in the more heavily affected areas."
Instead, some say the rate should be calculated by dividing the number of deaths by the sum of the number of deaths plus the number of people who recovered, which would exclude those people who are still sick, resulting in a higher death rate.
According to Niman's math, the death rate in Hong Kong would be 25 percent, in Canada 21 percent, in Singapore 15 percent and in Vietnam 10 percent. In China, the death rate would be only 5 percent.
Researchers have been unable to determine why SARS has caused more deaths in places like Canada and Hong Kong than in the United States, which has had no fatalities from the disease. Because of this variability, they say it's too much of a generalization to state a worldwide death rate. Worldwide, SARS has infected 3,461 and killed 170, according to the World Health Organization.
Niman's rationale is that including the entire number of probable cases in the death rate equation is diluting the number, because some of those people could still die.
But the other side of the coin is that because no available diagnostic test is 100 percent accurate, many of the people included in the count of suspected cases could be removed from the list.
With an accurate diagnostic test, researchers might also discover people who are infected with SARS but have no symptoms -- and so were not included in the current count of probable cases. This would decrease the death rate.
"We do not have an infection-specific death rate," said Dr. John Zaia, chair and professor of virology at the City of Hope Beckman Research Institute in Duarte, California. "That would be very valuable, because it could be that hundreds are getting infected but only small percentage are getting sick. That was the case for polio."
Another factor that could decrease the death rate is the possibility that recovered victims have not yet been reported as such, so the number of recoveries may rise.
"Only a small proportion at present have a clear enough outcome that we can declare them recovered," said Dr. David Freedman, a professor of medicine and epidemiology at the University of Alabama at Birmingham.
A foolproof diagnostic test is closer now that researchers in Canada, the United States, Singapore and Hong Kong have separately decoded the genome of the SARS virus.
But it will be weeks before a test is available for research purposes, and even longer before researchers can verify that the test is accurate enough to use in a doctor's office.
"In order to use a test to make a decision for a patient, it is subject to a bit more scrutiny and more evaluation than just something that we put out as an epidemiologic tool," said Julie Gerberding, director of the CDC.
Until a diagnostic test is proven accurate, the SARS fatality rate will remain an open question.
