Time for concern, not panic
Update 0813, 17 July 2009: This post was updated overnight - it now includes a table and a different explanation of the technical language.
Journalists like numbers. But given a range, the tendency is to go for the extremes. It's just the way we are, I'm afraid.
So given several potential death tolls for the UK from H1N1 swine flu, you can be sure that the headline writers will go for the biggest.
The figure of 65,000 deaths was given in a briefing to reporters by the Chief Medical Officer for England, Sir Liam Donaldson. It's based on what are called "mortality planning assumptions" for the NHS. It's very important to remember that these are not predictions.
What does it mean? Well, the case fatality rate is the percentage of people who may die after getting swine flu (this is a different method of predicting fatalities than that used by Imperial College London - see my previous post, What are the chances of dying from swine flu?).
The clinical attack rate is the percentage of people likely to get swine flu in the population in the first wave - some time between now and the end of winter.
So let's assume, for planning purposes, that 0.1% (one in 1,000) of those who get swine flu, die from it. If 5% of the population get it (the clinical attack rate), that means there would be around 3,100 deaths based on a population of around 60 million. And the other columns give other potential attack rates.
Now, if we take a higher case fatality rate of 0.35% (3.5 people in every 1,000), we get a different set of figures, ending with a potential death toll of 65,000.
You can be sure that this is the one that the headline writers will pick and probably the figure that most people will remember.
The plain fact is that we simply don't know yet how many people will die - but the figure of 65,000 seems extraordinarily high given what's happening around the world and taking into account that number-crunching from Imperial College.
Extreme and unlikely it may be - but it's not impossible, and it was given out by the chief medical officer, so that makes it okay for headlines? Discuss.
While we are dealing with numbers, remember that seasonal flu kills - it's a virus which is much underestimated. The last epidemic, in the winter of 1999-2000, probably caused around 21,000 excess deaths, and in an average winter, flu may kill 6,000 people.
The real difference we note when comparing pandemic to seasonal flu is the groups targeted. The biggest group who die from seasonal winter flu is made up of the frail elderly. They are least likely to catch H1N1 swine flu - but if they do, they have a greater than average risk of complications.
For some, it's when they read about young adults dying with swine flu and children in hospital that they begin to get really worried. It is a time for concern, especially for those with underlying health conditions (for a list of these, see posts from 12 June and 26 June). It's also a time in which we should all take personal hygiene seriously: "catch it, bin it, kill it".
There will be some deaths among people who were previously entirely healthy. But we can get a different perspective if we compare that to other risks we run each day, like road accidents, which kill around 3,000 people each year, or around 4,000 deaths from accidents in the home [53Kb PDF].
The vast majority who get H1N1 swine flu will be fine - and many won't realise that they've had it.
A lot will feel lousy for a few days. A minority will have complications and some will die. That is the reality of infectious disease. We are better prepared than ever to fight this pandemic. Antivirals like Oseltamivir (currently best known as the brand name Tamiflu) are effective, and a vaccine is coming - though it will be touch-and-go whether most at-risk groups are immunised before the peak of the first wave of infection.