What the research on hospital death rates really says

There’s a been a big media splash today about research which shows an increase in the chance of death if you’re admitted to hospital over the weekend, compared to during the week.

Most of it has been reasonably accurate and clear, like the Guardian piece. But some has been misleading, like this Telegraph piece which states that you’re ‘twice as likely’ to die if admitted at the weekend, which seems to be nothing short of fictional.

The actual research paper can be found here [subscription required]. So what does it actually say, in simple terms?

The researchers looked at just under 15 million UK hospital admissions for 2013-14, of which around 280,000 led to death.

They worked out that the chances of a patient dying within 30 days of being admitted were 10% higher if they were admitted on a Saturday and 15% higher if they were admitted on a Sunday, compared to if they were admitted during the week. This sounds pretty damning, and frankly scary.

But it’s more complicated than that.

The first thing to note is that these numbers are what we call relative risks, i.e. the difference in risk compared to the same risk for another group of people. Relative risks can sound impressive, and they make for good headlines.

But the absolute risk of dying if admitted to hospital, i.e. the risk all by itself, is very small. Overall it was 1.8%, rising to roughly 1.98% for Saturday admissions and 2.07% for Sunday admissions. Not such a punchy headline anymore.

The researchers were also well aware that patients who are admitted over the weekend are sicker to start with. In fact, 50% of Saturday and 65% of Sunday admissions were emergencies, compared to 29% during the week.

The researchers did try to correct for this problem. They reanalysed their data after excluding all patients who died within 3 days of being admitted, to try to limit the impact that emergency admission had on the results. The relative risk of dying if admitted at the weekend dropped – to a 7% increase for Saturday admissions and a 10% increase for Sunday admissions.

As the study itself points out, we should be seriously cautious about interpreting these numbers, especially when it comes to blaming them on staff who prefer a ‘Monday to Friday’ working culture, as Jeremy Hunt is so keen to do.

Simply excluding everyone who dies within 3 days of admission isn’t a great way of excluding all emergencies. As any doctor (or friend, or relative) will tell you, many severely unwell patients don’t die as soon as they reach hospital. Modern medicine can support them for days if not weeks, so they could still be contributing to that increased weekend admission death rate.

Furthermore, during my days as a general medical doctor, it was routine to see patients admitted at the weekend not as critical emergencies, but as the result of deteriorating slowly over the previous few days, not being able to see a GP on a Friday, and eventually ending up in the option of last resort, an A+E bed on a Sunday, in worse shape than if their issue had started on a Monday. There’s a real case for poor access to other services being the real cause of increased weekend admission deaths, not a ‘Monday to Friday’ culture in hospitals.

This tallies with the fact that Saturday admissions do better than Sunday admissions – because they’ve only had to survive one day’s deterioration without their GP before coming into hospital, not two. If poor care in hospitals was really the problem, we’d expect Saturday admissions – who’d have to spend two whole days in hospital over the weekend – to do worse. But they don’t.

It’s important to note that no more deaths actually occurred on weekend days compared to weekdays. Weekend staff seem to be doing an amazing job of keeping people alive once they make it to hospital.

Another thing that any doctor will tell you is that care in the first few days of an admission isn’t likely to be much different if you’re admitted on a weekend compared to a weekday. A+Es and medical assessment units – the first two stops during most admissions – routinely have as many staff on at the weekends as during the week, and urgent tests and procedures still happen quickly. It’s only when a patient has been moved on to a general ward, maybe 2 or 3 days later, that they might feel the effects of the weekend slow-down, as routine investigations have to wait. This doesn’t tally with staffing factors being the cause of increased weekend admission deaths.

Don’t misunderstand me; doctors couldn’t be keener to ensure that patients get good care on every day of the week. Contrary to popular belief, only 1% of consultants opt out of weekend working. And I’d never deny that junior doctors can feel overworked and undersupported at the weekends, meaning that there is room for improvement in senior doctor hours.

But we have to understand what statistics really mean before using them as the basis for policy and contract changes which will profoundly affect our health service for decades to come. To assume that the increased death rate for weekend admissions is preventable, in the words on the researchers themselves, would be ‘rash and misleading’. Take note Mr Hunt.

We already have a 7-day NHS. Don’t believe the hype.


About Alex Langford
I am a psychiatrist (now an SpR) based in Oxford after 3 years working in South East London. Before I went into psychiatry, I used to be a general medical doctor, and I also have a BSc in psychology. I'm particularly interested in improving the public face of psychiatry, evidence based medicine, teaching and patient rights. Don't mention cricket unless you've got the next fortnight free to discuss it.

One Response to What the research on hospital death rates really says

  1. Reblogged this on BethCanReflect's Blog and commented:
    Really informative (and reassuring!)

Leave a Reply to bethcanreflect Cancel reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: