This site is intended for health professionals only

At the heart of general practice since 1960

pul jul aug2020 cover 80x101px
Read the latest issue online

Independents' Day

Named or shamed? The absurdity of dementia diagnosis rates

  • Print
  • 10
  • Save

As part of his on-going, military-styles engagement with the enemy that is dementia, and with great fanfare, Jeremy Hunt has launched a new Government website to name and shame regions of the country over their diagnosis rates for the condition. But I have a problem - and I would like our Health Secretary to help me here; to lend me a portion of his clinical judgement; aid me in the application of scientific scrutiny - because I am not sure if I should be ‘named’ or ‘shamed’.

You see, for the government - with powerful vested interests in the big business that is dementia care whispering in their ear - more is always better and high diagnosis rates can only be a mark of good practice. Keen to do well, it was with the air of a schoolchild being handed back their test results that I checked the statistics for my practice - to find that we have a diagnosis rate of 126.7%! Can more always be better? Is it possible to have too much of a good thing? How can we have diagnosed more people than are out there?

My first instinct was that the figures must be wrong, but Mr Hunt doesn’t get things wrong, surely? Had our diagnosis rate been a woeful 40% I am sure he would have been quick to dismiss my pitiful bleating about misleading figures and declared me to be one of the laggards that he so eloquently derides in his video on the Department of Health website. For the sake of consistency, therefore, we must conclude that the figures are 100% accurate. Binscombe Medical Centre, according to the calculator, should have 86 patients with dementia. The precision of the Government’s calculations is remarkable; not a patient more, not a patient less, 86 it shall be - but we have diagnosed 109. What should I make of these 23 patients that should not exist? Should we conduct an urgent internal review of all our patients with dementia to reassess the validity of their diagnosis? Have we, in our zeal, rank ignorance and need for extra training (GPs always need extra training) been misdiagnosing at a greater rate than anyone else in the country? Should we report ourselves to the GMC? The CQC? The new Chief Inspector of General Practice? Please advise me, Mr Hunt!

Or maybe, just maybe, these diagnosis rates, used to create headlines about ‘Shocking Variations in Diagnosis’, beat up doctors for their lamentably poor performance and focus commissioners away from patients and on to yet another target, are about as reliable as the economic forecasts were in the years before the crash. Maybe we should remember that they are not proven scientific facts, but estimates that are inherently prone to statistical error. Perhaps I could point out that the figures have been derived from studies that are all over 20 years old, and that this data - placed so emphatically on the internet and yet derived from an era when the cyber-world was still the uninhabited domain of computer geeks - has been extrapolated to the present day to give imaginary figures for 2013/14? Might it be worthwhile pointing out that the latest study, bringing the figures up to date and published last year, suggests that there could be over 200,000 fewer people living with dementia in the UK than previously thought, as the incidence has been falling dramatically in the last two decades?

Could I refer to the danger of taking estimates designed for a population of over 60 million, and  applying them to a practice population like ours, of under 11,000? How statistical error always increases as sample size decreases, and so to reduce the figures by over a thousand-fold in this way might be expected to bring up the odd anomaly? How the presence or absence of a 40-bedded dementia care home in your practice area might just have a huge impact on your local figures when you are only meant to have 86 patients in the first place?

Of course, Mr Hunt will concede none of these things. After all, to admit for even one second that the figures for one practice could be out by as much as 30 percentage points would fatally undermine his ability to lambast the laggards. So, as he parades these naked figures for all to see, presented with a scientific validity that borders on the absurd, surrounded by the courtiers of Big Pharma and industrial vested interests; as he stands proudly with his Dementia Tsar on one side and the Alzheimer’s Society on the other, it falls to me, and those like me, to shout out from the crowd. As we look around at our patients, and worry about the effect of targets, misdiagnosis and resources directed away from the care that they really need, towards diagnosing more and more, earlier and earlier, all we can do is to try and draw attention to the Emperor’s peculiar attire. No doubt guards will come and bundle us away before we disrupt the proceedings.

Dr Martin Brunet is a GP in Guildford and programme director of the Guildford GPVTS. You can tweet him @DocMartin68

Readers' comments (10)

  • The concerns that Dr Brunet raises are very real.

    In my personal view the language used in talking about dementia reached a low point with the G8 Summit. Nobodys language was worse than the "prepared" words of David Cameron & Jeremy Hunt.

    We must first note that the figures that today "map" dementia are twenty years old! The DELPHI figures were general and crude and never part of an Alzheimer Society "battle". A "battle" that is most certainly Jeremy Hughes style.

    The 'map' that these "shocking variations" are based upon is absurd in that it is printed free of what Mary Midgley terms 'scientific pluralism'

    Have vested interests and politicisation overwehlemed the real voices and the real truth of those living with dementia?

    Need Iask?

    Unsuitable or offensive? Report this comment

  • Emer MacSweeney

    Congrats on the 126.7% Martin!
    Is your practice unbelievably good at diagnosing dementia? Or, is it an outlier in terms of prevalence versus national average? Or is the figure an artefact of clinical coding, or movements in list size? It would be interesting to hear your view on why the Delphi-based estimate of 86 is so much lower than the actual figure of 109, as in your practice's case the algorithm has produced an under-estimation.

    Unsuitable or offensive? Report this comment

  • Tom Caldwell

    Brilliant piece

    Unsuitable or offensive? Report this comment

  • Brilliant or forgetful?

    Unsuitable or offensive? Report this comment

  • Emer MacSweeney asks valid questions. It would be helpful to have Dr MacSweeney's views on whether outside interests should be made completely transparent on this drive to increase diagnosis.

    Unsuitable or offensive? Report this comment

  • To give fuller explanation of my personal view on this, please read my review of 2013:

    Part I: The "epidemic" of metaphors, considers language used in talking about dementia:

    Part II: Who is in the driving seat?, is a look back at policy developments regarding dementia in the UK.

    Unsuitable or offensive? Report this comment

  • In reply to Emer MacSweeney.

    Thank you for your comment. However, I'm not sure I should be congratulated. My point with the Delphi figures is that they were never intended to be extrapolated to practice level, or used to generate diagnosis rates or targets. To do this is a complete misuse of the research and makes the figures meaningless (for all practices, not just my own) - I should no more be congratulated at rolling a double 6 in dice than 'achieving' a high diagnosis rate.

    What worries me greatly is for those practices (and their patients) who have been given low figures. Whether or not I am congratulated is irrelevant, but a practice given a low diagnosis figure will be put under huge pressure to 'improve', despite the fact that their figure will be more a quirk of local demographics than a marker of poor care. This pressure will not be good for their patients and is likely to lead to overdiagnosis just in order to get numbers up.

    Unsuitable or offensive? Report this comment

  • As Dr Emer MacSweeney confirms, she is CEO for Re:cognition. It also seems she has widely promoted CANTABmobile. Details of this here:

    There is of course a lot to be gained for these companies in promoting diagnosis.

    The doctor who put together this pinterest board appears to have significant concerns about the misleading promotion of CANTABmobile and the lack of transparency regarding the vested interests of some of those promoting said.

    Unsuitable or offensive? Report this comment

  • Ivan Benett

    There seems to be a lot of excitement generated by this piece. What a shame these attitudes are promoted by a VTS organiser in charge of young minds.
    There are several explanations for variation from the mean. First of all the demography of this practice - it has a care home in its catchment area so of course many of the people with dementia will be registered at that practice and fewer with surrounding practices. In Central Manchester the practice with a similar profile also looks after many of the care home residents too. Secondly, all populations follow a Normal Distribution with a standard deviation and built in range - it's normal to have prevalences higher and lower than the mean. Finally, yes there will be variation in how well people are diagnosed and the put on the practice register - the 'rule of halves' is alive and well in all long term conditions.
    So there could have been a useful discussion on each of these three points, and hopefully Dr Brunet will have used this example to have a more constructive and intelligent educational session on his VTS programme. Unfortunately this piece treats a serious subject with cheap rhetoric and simplistic headlines. Whether Dr Brunet and others like it or not people with cognitive impairment and dementia get ill quicker, stay ill longer, and are less likely to return to premorbid function. So we need to know about them in order to intervene quickly to prevent them deteriorating with even the simplest of illness. We need to optimise management of risk factors for deterioration, and begin discussions with them and the family about the future.
    Need any more ideas for your VTS? There's a whole term's worth of stuff. But please, cut the cynicism.
    PS I have no vested interests in this subject. My own parents are cognitively sharp, but many of yours wont be and we need to create a Health Service that will look after them well

    Unsuitable or offensive? Report this comment

  • In reply to Ivan Benett.

    Of course there are explanations for the discrepancy in our practice figures - and care homes is a huge factor here. My point, however, is that this only shows how inaccurate the calculator is at practice level. For those practices which are inner city and have no nursing homes the figures will be at the other end of the spectrum - but those in charge of the figures will not look for a sensible explanation like demographics, but be quick to call them laggards and demand an 'improvement in performance.' This will lead to a pressure to diagnose in order to reach a target, which will lead to overdiagnosis and misdiagnosis - a very serious issue indeed.

    Diagnosing dementia is important, but it should only, ever, be for the best interest for the patient and not to hit a target.

    As for the young minds in my care on the VTS - they are a pretty savvy lot and I'm sure they will be able to tell for themselves whether or not I am spouting nonsense.

    Unsuitable or offensive? Report this comment

Have your say

  • Print
  • 10
  • Save