Brains working towards machine consciousness

A naive model of the swine flu infections (figures available at BBC news) in the UK shows the confirmed cases growing at an exponential rate (which is what you would expect). 

The good news is that as the population has increasingly been exposed to the virus, the infection rate can be expected to fall - or at least, one might think that is the good news.  The virus is mild at the moment, and with this model of the current rate of infections, we would expect the whole population to be infected by day 156 (well. actually before then, probably by 6 days, but this is a 'safety margin') which would be the 24th September 2009.  That would be likely to equate to approximately 30,000 deaths. at the current mortality rate in the UK from the disease.

But if the suggestions that the virus could become more lethal if it infects people during the normal flu season are correct, this would at least mean that the entire population had been exposed to the virus before that season starts - thus reducing the risk of further fatalities. 

There are complicating factors, of course.  It can be reasonably hoped that there will be a vaccine available around September.  But there is also at least one case of Tamiflu resistant H1N1 (the swine flu virus) which has been reported.

In my opinion (and I am not a virologist, of course) the use of Tamiflu to try to reduce the spread in the early days of the pandemic was a mistake.  In some cases it makes patients asymptomatic, but, apparently, still infectious (albeit at a lower rate of infection).  The problem is, just like over-use of anti-biotics, this sort of use can also lead to resistant strains building up.  And if we get a more lethal version of H1N1 develop, the last thing we want is for it to be more resistant to the anti-virals we have available.

Personally, I would rather take my chances with contracting swine flu now while the mortality rate is relatvively low, and going into complete isolation while it runs its course - but I would not like to offer that as advice to anyone else.

As I said, the model above is naive - as more of the population has been exposed to the virus, the infection rate should slow, as people are less likely to come into contact with others who are still infectious (I think).  But if the disease becomes more lethal, that may be a bad thing for us.

I would welcome comments on the reasoning - I cannot, off hand, see anything wrong with it, but perhaps someone with a better level of experience in epidemiology can point out any mistakes.

Update

I have revised my model slightly - the one shown above was (as I rather expected, to be honest) over estimating somewhat.  The new one is a nice simple 10e^0.1t

he solid black line is the old prediction, the reddish line is my new one.  This model would predict everyone would be infected by 25^th September if the infection rate continued as it is.  Of course, we are having a change in treatment and in attempts to isolate, so we may well see the infection rate increase initially, but they should also then start to drop off again (as mentioned above - exponential growth is not sustainable in a limited population)