- The British health secretary wants to impose new working conditions on doctors.
- He says we’re more likely to die if we’re admitted to hospital over the weekend than during the week.
- The ‘weekend effect’ has been shown, but weekday and weekend admissions are very different patients.
- While staffing levels are low at weekends, the secretary is wrong to interpret the evidence as supporting a change of shift patterns.
It gave me a view I never wanted of the first doctors’ strike in Britain for 40 years. The dispute about the new contract the Secretary of State for Health, Jeremy Hunt, has recently imposed on junior doctors rumbles on, and will lead to more acrimony and more strikes in the coming months. Hunt says his reforms are necessary to address the ‘weekend effect’. He claims that patients admitted over the weekend are more likely to die than those admitted during the week because there are less doctors on duty over the weekend.
Who is Jeremy Hunt?
A quick look at Hunt’s Wikipedia page reveals a fairly typical career for a Tory minister. He was privately schooled before reading politics, philosophy and economics at Oxford – a degree he shares with several of his cabinet colleagues – and was elected to Parliament after a few unspectacular business ventures.He was the Secretary of State for Culture, Olympics, Media and Sport during the build-up for London Olympics in 2012. He carried out his responsibility with such diligence that he didn’t notice that the firm he’d outsourced security to hadn’t employed the staff they would need until a few weeks before. The army had to be drafted in at the last minute to do the job and they’re still arguing about who needs to pay who what.
If most of us had to tell our boss that we’d messed up so badly that he needed to call in the army to pick up the pieces, we wouldn’t expect to have a job by the time we’d finished the sentence. Prime Minister David Cameron is evidently a forgiving soul. He reshuffled Hunt to the more prestigious role of Secretary of State for Health, putting him in charge of the National Health Service.
It is of course entirely coincidental that Hunt and Cameron took the same degree at the same university at the same time.
It would be unfair to assume that Hunt’s understanding of the health sciences is deficient due to his having no background whatsoever in them, and it would be ungenerous to assume that the Olympic debacle is evidence of incorrigible incompetence on his part. We all make mistakes and we try to learn from them, so if the several million pounds oftaxpayer money he lost didn’t buy the security for the Olympics, it may have bought him some wisdom.
Let’s give him the benefit of the doubt for now, and turn our attention to the numbers behind the weekend effect.
What is the weekend effect?
The evidence comes from a paper published in the British Medical Journal, which is unfortunately behind a paywall. Fortunately, such things can be begged, borrowed or blagged though if anyone knows somewhere it can be legitimately linked to, I’d appreciate it if you could put it in the comments.
First of all, let’s look at the data the authors analysed. It was gleaned from the NHS records of all hospital admissions in England in 2013-2014. It was restricted to England because the Scottish, Welsh and Northern Irish NHS are run by the regional assemblies and parliaments.
The authors then compared patient outcomes by the day of the week that they were admitted, using Wednesday as a baseline. They found that people admitted on a Saturday were 10% more likely to die within 30 days than people admitted on Wednesday. For people admitted on Sunday, the figure went up to 15%.
So does that prove that junior doctor contracts need to be radically changed to increase staffing levels over the weekend? Well, perhaps if you studied politics, philosophy andeconomics at Oxford. Anyone else would ask for a bit more information.
First of all, what is that 10% or 15% of? To give the raw figures, the weekend effect is accounted for by around 8,000 deaths among the 292,277 deaths the authors considered out of 15,859,508 admissions. In other words, only 1.8% of people admitted to hospital died within 30 days, or one in fifty-six. When they say risk increases by 15% if people are admitted on Sunday than Wednesday, they’re talking about 15% of not very many, which is still not very many.
For all that, 8,000 deaths still require explanation. The lead author, epidemiologist and biostatistician Nick Freemantle, told journalist John Ware on BBC Radio’s Report program, 8,000 deaths are ‘not a little accounting error’.
Who is admitted when?
The next question is whether the patients admitted over the weekend are comparable to patients admitted during the week. Asking whether the people admitted on Wednesday are likely to do better or worse than the people admitted on a Saturday is not the same as asking whether you are likely to be treated better if you come in on a Wednesday or a Saturday, which is what matters.
The paper shows a big difference in admission patterns during the week and at the weekend. More than twice as many people are admitted on any weekday than on Saturday or Sunday. We would expect that, as scheduled surgeries are only carried out during theweek. Someone spending a night in hospital to recover from having their tonsils out is much more likely to survive the next 30 days than someone ambulanced in after a car crash or a heart attack. The figures in the paper support that: only 29% of admissions on a weekday are emergencies but among the much smaller number of admissions over the weekend, 50% on Saturdays and 65% on Sundays are emergencies.
What’s more, people admitted over the weekend are likely to spend longer in hospital than people admitted during the week, which implies whatever puts people in hospital over the weekend is more serious than whatever put them there during the week. Most people admitted on a weekday go home the next day, while most people admitted at the weekend stay for at least three days.
Length of stay is a crude measure of how ill the people are, but the pattern held up when the authors only considered the most serious cases.
Why is there a weekend effect?
While we may make some inferences about the different patient populations, the authors couldn’t go beyond the information in front of them. They identified a pattern that raises important questions about what is going on, but did not leap to conclusions about why it’s going on:
It is not possible to ascertain the extent to which these excess deaths may be preventable; to assume that they are avoidable would be rash and misleading.Freemantle put it more succinctly to Ware:
There is a substantial challenge at the weekend. We’re not sure what the cause of it is.
Jeremy Hunt is not a man to be deterred by a fear of being rash and misleading. He said:
We have around 200 avoidable deaths every week in our hospitals.
Ware said they had asked Jeremy Hunt’s office to comment on how he got that from the paper to that conclusion, but did not receive an answer.
Hunt was criticised and challenged by both the British Medical Association and the Parliamentary opposition, led by Labour Party leader Jeremy Corbyn, for misrepresenting the paper. The answer to the criticisms was provided not by Hunt but by Cameron himself at the weekly shouting match of Prime Minister’s Questions. He said he had contacted National Medical Director Bruce Keogh, who was a co-author on the paper that all the furore is about:
Now we’ve had time to go into these figures in more detail, I can tell the House this: that the health secretary was indeed guilty. He was guilty of an understatement. The true figures for excess deaths at the weekend are eleven thousand, not six thousand.He gave no further explanation, presumably reasoning that having got the figure wrong in the first place, his authority would be unassailable when he quoted a different figure. Later in the session, he told Corbyn that Cameron’s mum would tell him to, ‘Put on a proper suit, do up your tie and sing the national anthem’.
The Oxford lecturers who taught him his debating skills must be so proud.
Where did those 11,000 deaths come from?
When Cameron talked about going into the figures in more detail, he conjured the image of himself and Hunt poring over spreadsheets and databases, applying the analytical skills they learned from their politics, philosophy and economics studies to uncover profound truths that were opaque to mere epidemiologists and biostatisticians.
In fact, the figure of 11,000 comes from the same paper that produced the figure of 8,000. And yes, it was 8,000, not 6,000 as Cameron said. He couldn’t even get that right. What Cameron presumably meant was that he’d badgered Keogh to explain it to him in words of one syllable.
The authors’ analysis wasn’t confined to the weekend. They compared every day of the week to Wednesday. It turned out that while 30-day mortality among Tuesday and Thursday admissions was similar to Wednesday admissions, it was 2% higher among Friday admissions and 5% higher for Monday admissions.When Cameron added the figures from Friday and Monday to the figures from Saturday and Sunday, he was assuming a four-day weekend. Unfortunately, the cameras did not capture the expression of Iain Duncan-Smith, Minister for Work and Pensions and the Tory Party’s answer to the Dickensian factory owner Mr Gradgrind, when he woke up to what Cameron was actually saying. It would certainly be a surprise to anyone with a proper job.
Cameron did not explain how low staffing levels at weekends lead to excess deaths among people admitted on weekdays. He simply added the excess mortality for Monday and Friday to the excess for Saturday and Sunday. To my mind, the Monday and Friday excess is evidence against weekend staffing levels being to blame rather than evidence for it, but it is not conclusive either way.
What goes on in hospitals at weekends?
Contrary to the impression given by the government’s calls for a ‘seven-day NHS’, hospital Intensive Care and Accident & Emergency departments are fully operational 24 hours a day, 7 days a week.
In the Report, Ware asked Julian Bion, a professor of intensive care medicine, about staffing levels over weekends. He acknowledged that there are only about as quarter as many consultants (senior doctors) on duty over the weekends, although they spend substantially more of their time on patient care than they do during the week.
They’re under a lot of pressure. They’re working very, very hard but they’re doing it with fewer staff very often. Somewhere or other, there’s either too many people around on a Wednesday andSunday’s the model we should go for or we’re cutting corners, and my guess is we’re cutting corners.
Bion is currently leading a study that will explore the weekend effect in more detail than Freemantle was able to. By focusing on 20 hospitals, he and his colleagues will be able to get a better idea of who the patients are and why they die. Hopefully it will answer the question of whether staffing levels are involved or not, although it won’t be finished in time to influence the current dispute.
Does the evidence support Hunt or not?
Based on the evidence available now, should we believe that changing junior doctor contracts will make the weekend effect go away? Ware referred to an internal Department of Health report, apparently leaked to the Guardian, saying even Hunt’s own department are dubious:
[The Department of Health] cannot evidence the mechanism by which increased consultant presence and diagnostic tests at weekends will translate into lower mortality and reduced length of stay.
In fairness, the quoted sections don’t refer to junior doctors but Bion made the point that if the problem really is that hospitals are understaffed at weekends, that staff is made upby many more people than the junior doctors who are the subject of the current dispute. Pharmacy staff, clerical staff, the laboratory staff and radiographers who carry out diagnostic tests and many other professionals would need to be considered.
Ahead of the evidence
The report also said that recruiting enough doctors to fully staff hospitals over the weekend would be ‘challenging’, implying that it could only be done by either rearranging shifts to spread the problem throughout the week rather than concentrating it over the weekend, or by demanding longer hours from junior doctors. Exhausted doctors are unlikely to improve matters.
Health economist Rachel Meacock summed it up to Ware:
[There is] no evidence to suggest yet that introducing the seven-day services would reduce weekend deaths, or that that could be achieved without increasing weekday deaths, and there is no evidence that such a service extension would be cost effective.
Meacock was clear, professional and non-judgemental, and avoids the direct criticism of Hunt that Ware must have wanted to get on air. He pressed her, and she went a little further:
The policy seems to have gotten ahead of the evidence.
No doubt Ware would have liked her to say ‘Hunt is lying his arse off’, but Meacock was not taught to debate at Oxford so kept it civil.
What do you think? Have Hunt and Cameron genuinely come up with the solution to the weekend effect? Or are their pants on fire?