Saturday, 16 February 2008

Unhealthy numbers

Yesterday, the SMH published a letter from Dr Glenn Pereira, Forbes, which started with:

I had the pleasure of being involved in the birth of an Aboriginal baby in our small rural hospital just after 9am on Wednesday, as the Prime Minister delivered his address to Parliament. My wife is also due to deliver in the next few days. It struck me that, statistically, the child born on Wednesday is about three times more likely to die in her first years of life than my child, and can expect to live 10 to 20 years less, even though they will have been born just days apart in the same hospital.

While the sentiments in favour of reducing the health disparities can only be a good thing, I suspect the doctor's mathematics/statistics is not so good. The much-quoted infant mortality and life expectancy figures that he seems to be using are national estimates. Applying them to children born in a particular hospital assumes that the situation there is like the national average, so they definitely can't be used to suggest that having been born in the same hospital doesn't make a difference!

I'm no expect on life expectancy and so on, and don't even know if anyone has figures for siutations this specific. Knowing how the disparity for kids in Forbes compares with the country as a whole might give the government a bit of an idea what health policies could actually help.

If only writing a thesis were as interesting as finding mathematical errors in the news...