When a responsible UN official says that a global pandemic of avian flu might kill between 5 and 150 million people in the world, it is time to put away those rose-coloured glasses and pay attention. Especially since this man was Dr David Nabarro, the scientist recently selected by the UN to coordinate the world's response to such a crisis. True, the World Health Organisation immediately said that their official estimate of the global death toll remained at between 2 and 7 million. But the higher figure had the feeling of a cat coming out of a bag. Could it be right?
Avian flu invites yawns. Who cares if a few chickens start sneezing in Thailand? Here is why you might care. The current strain of Asian bird flu (H5N1) is perhaps the most virulent ever. So far, 55% of the people who have caught the bug direct from poultry have died from it. But the bug cannot be readily passed between humans, so we are all OK for now, if we promise not to sleep with chickens.
The real fear is that the bug will be caught by a human being (or a pig, if you must know) who already has influenza. Being devilishly clever, H5N1 might indulge in some gene-swapping with the boring old flu bug, and consequently learn to hop between human beings.
Everything then depends on whether the new form of H5N1 maintains its lethal nature. That will determine whether we get a relatively minor flu pandemic such as those of 1957 and 1963, or a major disaster such as the Spanish flu of 1918, which killed an estimated 40-60 million people. (That is three times the number killed in the first world war, by the way, and twice the total number who have so far died from Aids.)
The death toll from a flu pandemic depends first on the attack rate, or the proportion of people who become infected with the illness. This varies between 5-10% for a normal pandemic, rising to 25-35% for a major event. Also critical is the case mortality rate, or death rate among those who contract the illness. This is harder to pin down, but seems to hover at maybe 0.1-0.2% for the minor variety to perhaps 2.5-5% for the real McCoy.
To calculate likely deaths in a pandemic, take the population, multiply it by the attack rate, and multiply the result by the case fatality rate. (You might think this is all a bit basic but, take my word for it, scientists are not very far ahead of you on this.) For the world as a whole, this method produces a central estimate of about 7 million deaths in the mild pandemic, and 74 million in the bad version. The higher figure is not far from the centre of the Nabarro range. The equivalent numbers of deaths in the UK alone would be 60,000 and 650,000 respectively.
Scientists agree that one of these two outcomes is probably looming. Some virologists say the minor episode is the more likely, but most simply say "sorry, don't know, squire". Fortunately, the British government, well ahead of the US and most other developed countries, has placed an order for enough Tamiflu (an anti-viral drug which might reduce attack severity) to treat a quarter of the population if needed. Which leaves only three-quarters of us to worry about those sneezing chickens.
