The Today programme have said that there is ” fresh evidence that people who are admitted to hospitals in England at weekends face a significantly increased risk of death.”
4 days ago, the Telegraph reported that Andrew Lansley had decided to fund hospitals to work 24/7 with no weekend change to staffing, and the report cites Dr Foster’s data, which I have previously criticised for methodological failings here.
The source of the latest story is this paper in the Journal of the Royal Society of Medicine, here.
It seems that Bruce Keogh has used this data to suggest that hospitals should work to a 24/7 rota rather than a weekend one – I will link to his interview on the BBC when it is available.
There are a number of flaws with this leap. First is the concept that weekend staffing is 1)related and 2)responsible.
Currently, if a consultant is on -call at the weekend, she or he is dealing primarily with the patients on the ward. There are no management meetings, outpatient clinics or educational seminars. It’s unclear whether consultant presence at weekends is any less, on the wards, than it is during the week. It’s an illusion to think that, weekdays, consultants are doing nothing but being on the ward. We have no decent comparative data. (And Dr Foster still haven’t responded to that BMJ piece criticising their methods.)
The JRSM paper is in many ways a very good paper. They examined the 14,217,640 admissions into English hospitals in 2009/10. The researchers then looked at death rates in the following 30 days. This is what they found:
Hospital stays which included weekend days were less likely to result in death than hospital stays on midweek days.
Weekend days were associated with the lowest risk of subsequent hospital death for those admitted at the weekend.
Hospital admissions at the weekend resulted in a higher death rate at up to 30 days later. They say “For every 100 deaths amongst patients admitted on a Wednesday, we would expect 116 among otherwise similar patients admitted on a Sunday. However, for every 100 deaths among patients in hospital on Wednesday, we would expect to see 92 among similar patients already in hospital on a Sunday “.
Where the paper falls down is in making extrapolations to how the NHS is run at weekends. They have found some interesting things. However, we cannot explain this data. The authors also say, very interestingly:
“Our analysis was replicated on data from 254 leading US-managed care hospitals, finding increased risk of death for admission to hospital at the weekends and a reduced risk associated with hospital stays on weekend days compared to weekdays. This was despite differences in the organisation and delivery of health care between England and managed care in the US”
In otherwords, the presence of a 24/7 based hospital rota did not get rid of the finding that weekend admissions were associated with a higher death rate.
The authors of the paper then go on to say that
“Although our data does not allow a scientific analysis of the mechanisms underlying this phenomena, one could speculate that ….”
And then there are lots of speculations.
The bottom line is that this study gives us data, but is not able to explain it. There are many possible reasons for their findings, which include the issue that patients admitted at the weekend have different characteristics which places them at higher risk of death. Bruce Keogh has taken the view that staffing levels are the issue, and that we need 24/7 hospital rotas – but it would be just as possible to declare that we should stop admitting elective patients at the weekend, as this may be safer. It would also be possible to say that there is evidence that patients are better off being discharged after a weekend, rather than before it.
It’s a pity that good data has made unjustified leaps to changes in policy. It would be a disaster if vast sums were spent making non evidence based changes to working practices – and which they made no impact on death rates.