Lies, damned lies and Jeremy Hunt

It’s bad enough when the press misinterpret data to produce shocking headlines, when a secretary of state does it it’s time to call foul.

 

So, it seems you’re much more likely to die in hospital at the weekend than during the week, right?

Wrong. In fact more patients die on a Wednesday mainly because mid-week is when hospitals are at their peak of occupancy. A higher number of patients would logically imply a higher number of recorded deaths.

But the actual daily mortality rate is not even considered in the research upon which both the Health Secretary and the Prime Minister are relying for their ‘facts’.

The research is a two part study, first published in 2012 in the Journal of the Royal Society of Medicine and then updated in the BMJ in 2015.

The study looked at the records of 14,217,640 admissions into hospital during 2009/2010 and specifically counted deaths occurring within 30 days of admission classified by the day of admission – not the day on which the death occurred.

This is a completely different statistic from the ‘weekend death rate’ headlines that have been thrown out to the public by the press and politicians as an urgent matter of concern.

What the study found was that 187,337 patients died within 30 days of admission to hospital, that’s only 1.32% of the more than 14 million who were admitted. That’s good news, meaning that only just over 1 in every 100 patients died within that time frame.

This would suggest, if spread evenly across a seven day week by simple process of maths, that 26,762 patients (14.29%) who subsequently died within 30 days would have been admitted on each of the days.

Of course, the first question one might ask is why we might assume that mortality is spread evenly at all. There may be some practical reasons or simply the results of custom that make this unlikely. Human behaviour is rarely broken down into simple patterns of symmetry, why should the health service be any different?

When classified by day of admission, however, the study found little significant difference between the ‘death rates’ on particular days of the week except for a measurable increase in those admitted on a weekend when it increased by 11% on a Saturday and 16% on a Sunday.

On the face of it those figures do seem horrifying, 11% and 16% are significant increases, but they need to be considered in context as do all the other figures here.

There are currently 168 NHS Trusts in the UK, many of them run more than one hospital so it is difficult to actually give a figure for the exact number of hospitals, so for simplicity we can use the Trusts as a measure of the number of healthcare providers involved in the process.

Spread evenly across the week and the year (both of which are huge presumptions) the overall mortality rate would equate to just over 514 deaths within 30 days for each admission day of the year, or slightly more than 3 deaths per day for each NHS Trust (actually 3.06).

Because we are dealing with a defined number of mortalities (187,333) the fact that the researchers identified a higher level at weekends means that those on weekdays are actually below that average. The adjusted weekday levels and the subsequent percentage increases (of 11% and 16% respectively) actually pan out as follows:

 

Weekday mortalities within 30 days of admission

(Each Weekday)

Per year 25,768; Per day 496; Per Trust 2.95

Saturdays

Per Year 28,834; Per day 550; Per Trust 3.27

Sundays

Per Year 29,891; Per day 575; Per Trust 3.42

 

So the actual averaged increases on any given day equate to 0.32 and 0.47 deaths per trust. Not even one a week.

That hardly seems to indicate, to my eyes at least, a huge increase in negligence or inefficiency. Wouldn’t we expect a truly compromised service to cause much more damage?

Nonetheless the 16% higher rate justified the “16% more likely to die”, “significantly increased risk” headlines. And it does, admittedly, add up to nearly 7,000 more patients a year nationally. But that is 7,000 more than would be expected on an even spread of mortality across the week and, by its very nature, is matched by a 7,000 reduction across the rest of the week below that average figure.

Nonetheless it led to the classification, by Jeremy Hunt and his supporters alone, of this statistical difference as ‘excess deaths’, implying that these patients would not have died had they been admitted on any other day of the week.

That is, in itself, a breath-taking conclusion to be borne out of no supportive data whatsoever. The 2012 paper made no attempt to link the apparent ‘weekend effect’ to any measurable cause and even suggested that the complexity of the factors involved meant that it would be premature to consider the effect ‘addressable’ at all without further and deeper investigation.

But the myth of a dysfunctional weekend hospital had been born; feeding directly, as it does, into the ‘seven day a week health service’ mantra that Jeremy Hunt was already looking to impose on GP’s and a public that seems frankly lukewarm to the whole idea.

Fast forward now to the updated report of September 2015 co-authored in part, as was the initial research, by Professor Sir Bruce Keogh, the Medical Director of NHS England and a long-time proponent of measuring clinical outcomes to improve performance.

He is also known to be a supporter of seven day working at all levels within the NHS.

The new report looked at data from 14.8 million hospital admissions in 2013/2014.

They found that the ‘increased’ mortality figures for the weekend were roughly similar to their initial report at 10% and 15% on Saturday and Sunday respectively (compared to the initial findings of 11% and 16%) but added that they found an increased level of 30 day mortality on Fridays and Mondays particularly for oncological (cancer) patients.

The revised report therefore refers to a more generalised “weekend” effect from Friday to Monday which, it suggests, equates to 11,000 more deaths within 30 days of admission over this period than across the rest of the week.

This seems to ignore, however, the fact that they have now isolated the rest of the week to a three day period between Tuesday and Thursday. It does make you wonder at which point a ‘weekend’ stops being worth the name.

This is the statistic that has allowed David Cameron and Jeremy Hunt to claim that the initial claim of 6,000 was an underestimate. What they fail to say is that the new figure is spread across four days, two of which are not officially the weekend and would be unaffected by any new contract arrangements.

They also fail to mention any of the many caveats that the authors have included in the report.

They note, for example, that 50% of admissions on a Saturday are emergency cases a figure that rises to 65% for Sundays. That Saturday and Sunday represent only 6% and 8% respectively of all the cases reviewed.

Perhaps most importantly they warn “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”.

Yet the government and Jeremy Hunt in particular seem to want to make this the central plank of their platform for imposing a new contract on Junior Doctors and casting the dissenting medical professionals as somehow not caring about patients.

The Health Secretary said in Parliament in response to a question about doctors contracts; “Every year, there are 11,000 excess deaths as a result of inadequate cover at weekends, and we do not want that to continue”.

The editor of the BMJ has written to Jeremy Hunt to seek his assurance that he has understood the nature of the conclusions drawn by the researchers and over 2,000 complaints of misrepresentation have been made to the cabinet office by Doctors incensed at the implications of the health secretary’s rhetoric.

Whether one agrees with the principle aim of a seven day a week health service or not the justification lies not in the working practices of Junior Doctors who already work seven days a week (as, indeed, do most consultants as required). If anything, the research shows how robust the service is at dealing successfully with emergencies and non-elective procedures at weekends.

And it certainly cannot lie in the blatant misuse of statistics and misrepresentation of facts that we have been witnessing.

It’s an insult to the medical profession, to the Health Service and, most of all, to you and me.

 

 

Subscribe to MHM Newsletter


If you enjoy finding out more about health matters then why not get the very latest news delivered fresh to your inbox every week? It's a completely FREE service so join up now, simply enter your email address below:

Title
First Name
Last Name
Email Address
Confirm Email

Please rest assured, we will never pass your email address to any other company. These emails are free and without obligation. You can unsubscribe any time you choose.

Copper Heelers
Featured News
Copyrights ©2015 - Meta Health Monitor