Wednesday, July 15, 2009

Scientists call for more accurate measure of swine flu death rate

July 15, 2009

More accurate ways of measuring mortality from swine flu are needed to help health officials to make the right decisions about managing the pandemic, scientists said yesterday.

Standard methods of calculating death and hospital admission rates for the influenza A (H1N1) virus are unreliable and can both underestimate and overestimate the true figures, according to research at Imperial College, London.

One of the key measures of the virulence of any disease is its case fatality rate (CFR), which is generally calculated by dividing the total number of known cases by the total number of deaths. A good estimate of the CFR for swine flu, and of its hospital admission rate, is needed to help healthcare planners to decide on appropriate measures for containing infection, such as school closures, and on vaccination strategies.

Any changes in the mortality rate must also be monitored closely to determine whether the virus is mutating to become more virulent, which may require policy changes.

At present the CFR is crudely estimated at 0.14 per cent in Britain and Europe, 0.57 per cent in the US and 1.03 per cent in Mexico. These estimates, however, are based on a simple calculation that is flawed for three reasons, a study from Imperial’s MRC Centre for Outbreak Analysis and Modelling has shown.

The total number of deaths can easily be underestimated because deaths from heart attacks and strokes to which swine flu has contributed will not always be counted as swine flu deaths. This can produce a CFR that is too low. The CFR can also be underestimated because of the time lag between infection and death.

On the other hand, the total number of infections is much higher than the total number reported because few asymptomatic or mild cases are reported. This means that mortality may be lower than present estimates suggest.

The study, published in the British Medical Journal, advises that a more sophisticated approach to investigating mortality is needed to equip ministers and health officials with more reliable information.

All figures must be adjusted to take into account the time that passes between people falling ill and starting to die, which raises the CFR in Britain to about 0.24 per cent.

Calculation methods should be changed more accurately to reflect the real numbers of mild swine flu cases, and all deaths in which the virus is a factor.

Scientists need to study groups of a few thousand people in detail to estimate the overall attack rate and hospital admission rate and extrapolate the results to the general population. Death rates can then be obtained by studying in detail patients admitted to hospital and combining the two sets of data.

Tini Garske, who led the study, said that research proposals had been submitted to funders to start collecting data in this way.

“We should be doing studies trying to estimate the extent of mild infection, and keep a close watch on severe and hospitalised cases,” she said. “We must also collect data on how long it takes for people to die after they become infected. These data should put us in a position to get more reliable estimates.

“Reliable estimates of the case fatality ratio are important for healthcare planners to put the right policies in place. If the CFR is high, and many people are going to die, you may want to put into place more aggressive policies.”

She said that she thought it more likely that the present crude CFR overestimated mortality, rather than underestimated it.

“I would guess that there are probably a lot more cases undetected, making it an overestimate. However, there will be people who have died with flu as a contributing factor who have not been picked up. That will have happened, and it will happen. We need to be aware of this uncertainty.”

Accurate assessment of the CFR would be particularly important for tracking whether the virus was mutating to become more virulent, Dr Garske said.

No comments: