What’ve 17,000 pregnant men and hospital mortality stats in common?


Pregnant men – the key to understanding NHS coding?

Most of the comments I’ve received on this blog about my articles on Hospital Standardised Mortality Ratios (HSMRs) has been very positive. Health workers, members of the public and the occasional journalist have contacted me to offer support or thanks and to affirm how my conclusions have matched their daily experience, or to ask questions to get a better grasp of the facts.

Of course, not all the comment has been positive, but in some cases an originally confrontational stance has modified as facts have been discussed.

HSMRs – the statistics that fuelled the completely misleading headlines about ‘hundreds of avoidable deaths‘ at Stafford hospital and which are now being used by Jeremy Hunt and the Parliamentary Tory party (aided and abetted by lazy and/or malicious media) to attack more than 10% of England’s hospitals – are a complex matter. Explaining why they’ve been wrongly calculated and are being wrongly used is an involved business and it’s easy to understand why people find the issues difficult to grasp.

But every now and then, you come across something which makes the complex very simple indeed.

HSMR calculations are based on statistics supplied by hospitals to NHS authorities for funding purposes – ‘payment by results’ (PBR). But a key weakness is the principle of ‘rubbish in, rubbish out’ – if the input is wrong, the resulting statistics are going to be wrong too. And, as I’ve shown in various articles, the input has been very wrong – and there’s lots of room for them to be even more wrong than I can currently demonstrate.

Some people have tried to contest that. However, I read something this morning, that is such a perfect illustration of the problem that it’s almost a question of..

Case closed

Last year, in the NHS, the HES – the hospital episode statistics which form the basis for PBR and (via Doctor Fosters Intelligence) – included 785,263 ‘episodes’ of ‘obstetric care’. Obstetrics is care relating to pregnancy.

Of those 785,263 episodes, 16,997 were coded as being given to male patients.

If hospital coding departments can be so overstretched, understaffed, underfunded, undersupported or underqualified (take your pick!) that not just one or two – or even one or two hundred – episodes of obstetric care were recorded as being given to men, then any objective assessment must surely conclude that a system so riddled with flaws is effectively useless. And it certainly isn’t a proper foundation for massive headlines, swingeing government policies, hospital closures and demanded resignations.

A hospital has a high HSMR score? By all means check that there are not big problems with the care it provides.

But when you read the screaming, strident headlines in large, bold letters telling you ‘hundreds died needlessly’ at this or that hospital because the NHS is so sloppy, callous, stupid, uncaring and in need of reform and we shouldn’t love the NHS as much as we do, do 3 things for me. Please.

  1. Remember the 17,000 pregnant men
  2. Find some salt and take a very large pinch
  3. Instead of thinking ‘how awful, something must be done!’, think instead ‘What’s motivating someone to make these statements, write these headlines, close hospitals?

Because to get things that wrong rarely happens by accident.

12 responses to “What’ve 17,000 pregnant men and hospital mortality stats in common?

  1. Thank you for your very thoughtful and measured analyses. Please hang in there!

  2. Just one thing to add Steve, its not just the coders. Most ICD10 coding errors/omissions stem from drs diagnoses being unclear, or lacking sufficient detail to match to the correct code. Two ways to rectify this – 1: electronic records where every dagnosis is supported by the ICD10 database, so you just find and select, 2. Med students need to be taught the whole process from making a diagnosis to that diagnosis being correctly coded.
    (I think the pregnant men thing is probably going to come down to a data handling error rather than to much human error, although its a nice story)

  3. Pingback: What've 17,000 pregnant men and hospital mortality stats in common? | Welfare, Disability, Politics and People's Right's | Scoop.it·

  4. Very well put as usual. I can’t help thinking that if everything was coded to sufficient depth and accuracy that all HSMRs (or equivalent) would oscillate between 95 and 105 with only very rare outliers.

    With regard to ward nurses in the NHS, the quote from John Bell of the Iona community is relevant:
    ” If you want people to retain their dignity then you have to ensure that the people who care for them have the time to be kind”.

  5. Pingback: 9/14 hospitals investigated for high deaths DON’T have high mortality? | The SKWAWKBOX Blog·

  6. Pingback: Then and now: the two faces of the BBC on Stafford NHS deaths | The SKWAWKBOX Blog·

  7. the coder may have done their job perfectly but if a female patient is listed on the hospital system as male, a coder may never be aware of the error, may never see it, “overstretched, understaffed, underfunded, undersupported or underqualified” or not

    • I understand. But for the system to have a hope of working properly, coders have to be willing and able to see discrepancies and go out and talk to medical staff to clarify them. As it stands, this is rarely possible – but please understand, this is a criticism of the system, not of coders. I think coders get a rough ride very unfairly.

  8. Pingback: Stafford nurses speak out | The SKWAWKBOX Blog·

  9. Pingback: Now govt uses Stafford myth to target 36% of NHS hospitals | The SKWAWKBOX Blog·

Leave a 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 )

Twitter picture

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

Facebook photo

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

Google+ photo

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

Connecting to %s