interesting article on what should be the denominator.....
WHO's Methodology May Understate
SARS Death Rate, Some Officials Say
By KAREN RICHARDSON and BETSY MCKAY
Staff Reporters of THE WALL STREET JOURNAL
Is the real death rate for SARS higher than the public has been led to believe?
The World Health Organization maintains that the mortality rate for severe acute respiratory syndrome is currently about 5.6%. That figure, often cited by public health officials and the media, represents the number of known SARS-related deaths divided by the number of probable cases world-wide.
But some medical officials believe the real mortality rate may be 10.4%, or nearly twice as high. That's because in its calculation, the WHO includes not just known cases of recovery from the disease but also patients who remain hospitalized -- in other words, people who may yet die.
A more accurate method of calculating the mortality rate for a disease, some disease experts say, is to use only known outcomes: divide the number of deaths by the number of recoveries plus deaths. That excludes sick SARS patients whose fate is still unknown, as well as those who are listed as "suspected" cases but may not have SARS at all. "It's a much better measurement," Joel Cohen, a professor of populations at Rockefeller and Columbia universities, says of using known outcomes. "If patients are still in the hospital, you don't know what number will die."
Henry Niman, a research fellow at the Shriners Burn Center in Boston who operates an e-mail discussion group on the disease, called the methodology used to reach the 5.6% fatality rate "very optimistic."
The differing views raise an important question in tracking an emerging disease: how soon can its impact be properly measured? In the early days of AIDS, all cases were eventually deadly so determining a mortality rate wasn't relevant. But in the case of SARS, a reliable number can be difficult to determine because so many patients are still fighting the disease and most eventually recover. Death rates also vary from country to country.
Clouding the picture further are the varying criteria countries use to define a case of SARS. Until last week, for example, the U.S. reported only "suspected" cases of SARS to the WHO. Now it has tightened its definition, reporting probable cases instead. It makes a big difference: About 228 people in the U.S. are considered possibly to have SARS, but 38 of those have signs of pneumonia or acute respiratory distress, and are therefore considered probable cases.
Calculating the mortality rate using only deaths and recoveries, the death rate for Hong Kong is currently 17.7%, according to data posted Monday on the WHO Web site. Canada's is about 18.2%, and Singapore's is around 13.3%. No deaths have been reported in the U.S. Some outlying areas of China, Dr. Niman says, appear to have mortality rates of nearly 50%.
A WHO spokeswoman in Geneva defends the organization's methodology. "We have no plans to change the method of calculating the death rate at the moment, because that would make all previous records and [records] for different countries uncomparable," says Maria Cheng, the spokeswoman.
At the Centers for Disease Control and Prevention in Atlanta, officials also say the current methodology is appropriate. Calculating a rate using "deaths divided by total number of cases [regardless of whether or not people have recovered] is pretty standard epidemiologically," says Steve Ostroff, acting deputy director for the CDC's National Center for Infectious Diseases. He says the final death rate shouldn't differ much from current figures because the proportion of ill patients who will die isn't likely to change.
Hong Kong Director of Health Margaret Chan said Monday that now is "not the right time" to calculate the mortality rate. Acknowledging that using known outcomes was the more established and traditional method to calculate fatality rates, she urged the public to be patient until the outbreak of SARS is "resolved" and a "very accurate death rate" can be determined.
In Canada, some doctors support calculating the mortality rate by deaths and recovering patients, and excluding those in early phases of the illness. Using those criteria, Canadian doctors estimate that the mortality rate is about 1% for patients under 50, and about 25% for patients over 65.
"This is a three-week illness," with many patients hitting the most critical period from days 11 to 16, says Allison McGeer, director of infection-control at Toronto's Mount Sinai Hospital. Some patients who become critically ill or eventually die of the illness don't seem all that sick at first. "We don't have any good markers at onset to tell us how serious a case is going to get," says Dr. McGeer, 50, who is now at home recovering after spending almost three weeks being treated in the hospital for the illness.
Politically, the death rate hits a tender spot in Hong Kong, as government officials struggle to boost public confidence among people already deeply concerned about the economic impact of SARS on tourism, retail sales and foreign investment.
"Calculation of the death rate is quite a sensitive question at the moment," says Guan Yi, a microbiology professor at the University of Hong Kong, who agrees that the formula that uses actual outcomes should be used when the outbreak is over in Hong Kong. Before that, he says, "it'll make people scared, and they'll run away" from treatment."
In Singapore, officials have brushed aside the issue altogether. "We don't see this as a very meaningful exercise at this stage," said Singapore Health Minister Lim Hng Kiang at a news conference last week.
-- Elena Cherney and Richard Borsuk contributed to this article