SARS Death Rate Lower In Countries Responding Aggressively To Initial Outbreak

An analysis accepted for August publication in the journal Emerging Infectious Diseases indicates that countries that quickly initiated control measures against SARS (severe acute respiratory syndrome) saw a slower spread and a lower fatality rate.

"The quickest spreading epidemics have also had the highest case fatality rates," said Nicholas P. Jewell, professor of biostatistics and statistics in the School of Public Health at UC Berkeley. "It's probably a reflection of better control measures, rather than a difference in infectivity or virulence of the disease."

Control measures that have proven effective include rapid hospitalization and intense care after infection, coupled with isolation of cases, infection-control measures in hospitals, and vigilant surveillance at the community and population levels.

The data the researchers analyzed from the World Health Organization (WHO) provide no evidence that the virulence of the disease changes over time or from country to country, despite fatality rates ranging from 9 percent in Vietnam to a current 19 percent in Taiwan.

"One of the main questions addressed by our statistical analysis was whether there's a different version of the virus that is more infectious and more virulent in some countries than in others," Jewell said. "I doubt that."

In line with a report last month in the journal Lancet, Jewell and colleague Alison P. Galvani, a Miller Post-doctoral Fellow in the Department of Integrative Biology, report an average case fatality rate of 14.7 percent for SARS.

The case fatality rate is the proportion of people who die, and is best estimated in emerging epidemics by the number of deaths divided by the sum of deaths plus confirmed recoveries. Early estimates of the death rate were calculated by dividing the number of deaths by the total number of infected individuals. Jewell said it's a valid measure for advanced epidemics, but one that underestimates the death rate in early-stage epidemics because it ignores the fact that some infected eventually will die.

Jewell and Galvani caution that the data used in their analysis, updated daily by WHO, are early surveillance data and are subject to country-by-country differences in reporting practices. This uncertainly underscores the need for case-by-case data from affected areas, the researchers say. The May 7 Lancet report used records from 1,425 cases in Hong Kong, but this raw data has not been made available to other researchers.

"Individual level data will give more precise estimates of the reproductive ratio, a number that gives you the ability to predict both the initial rise and time to subsequent decline of an epidemic," Jewell said. "Such epidemic models may provide a framework for evaluating alternative control measures."

The reproductive ratio, or R0, is the average number of secondary cases generated by one initial infection. A good estimate of this number requires detailed "chain data," Jewell said - who you caught it from, how long it took you to catch it and who you may have transmitted it to.

"As the virus propagates, this information allows you to create a chain that links each person infected," he said.

The researchers' conclusions come from comparing fatality rate to the epidemic's doubling rate - the time it takes for the number of cases to double - in the countries most affected by SARS: China, Hong Kong, Singapore, Vietnam, Taiwan and Canada. The country with the lowest fatality rate, Vietnam at 9 percent, responded to its initial outbreak in a hospital by basically locking down the facility, Jewell said, instantly reducing the spread. The doubling time was the longest of any country - 43 days.

Canada was one of the countries hit worst, with a case fatality rate of nearly 20 percent, followed closely by Hong Kong. Both countries were seeded by cases from China early in the epidemic, before the countries had a chance to institute control measures, which set the stage for community spread. Both ended up with a rapid spread -the doubling time was13 days - and a high fatality rate, though both outbreaks appear now to be under control.

"Containment of an outbreak at an early stage affords a greater chance of success than does a later response, and the former clearly puts less strain on the health care system," the epidemiologists wrote.

The paper's third author is Xiudong Lei, a graduate student in biostatistics at UC Berkeley.