Never, never on a Sunday...

We can have some fun with numbers with this one:

People have a ‘significantly’ higher risk of dying if they are admitted to hospital at weekends, new evidence reveals. Patients are 16 per cent more likely to die if they are admitted on a Sunday than a Wednesday, and 11 per cent more likely to die if they are admitted on a Saturday...

The analysis covered all planned and emergency admissions – more than 14.2million – to hospitals in England during 2009/10. It looked at 180,000 patient deaths within 30days of being admitted.

It's not clear if they made the obvious adjustment to the figures: "Another factor could be that seriously ill patients are more likely to be admitted on weekends. If they were less ill, they would have had their admissions postponed until a weekday.", and we aren't told how much lower admissions are on a weekend, so we'll have to work with what we've got.

a) 180,000 deaths out of 14.2 million admissions = 1.27% chance of dying.

b) Assuming same number of admissions every day (for lack of anything better), that can be split up into:
Mon - Fri = 1.22% chance of dying.
Sat - Sun = 1.38% chance of dying.
Check: (5 x 1.22) + (2 x 1.39) = 8.88, 8.88 ÷ 7 = 1.27.
Check: 1.22% + (13.5% x 1.22%) = 1.38%.

c) So your additional chance of dying if you are admitted on a weekend is only 0.17%.

Then there's the question of what-is-to-be-done?

Reduced staffing, fewer senior doctors on duty and poor access to diagnostic tests on weekends could have an effect, said researchers... Health secretary Andrew Lansley has said more could be done to increase the number of senior doctors working weekends.

What's the point of that then? It is clearly the case that it is far more expensive getting people to work at the weekend, because they want to be paid overtime etc (call it time and a half for sake of argument). So for a fixed salary budget, a hospital could employ ten fewer people Mon - Fri and seventeen extra people Sat - Sun*. There would be a modest reduction in deaths at weekends, but this would always be lower than the additional deaths during the week, so there'd be a net increase in deaths, end of discussion. So the answer is, just keep muddling along as best they can.

* Check: 10 people x 5 weekdays = 50 days' normal pay. 50 days normal pay ÷ 1.5 = 33.3 days' overtime pay. 33.3 ÷ 2 = 16.7, round it up to 17.