• Another postcode lottery 12 July 2016 | View comments

  • 12th July 2016 | By David Jolley

    There are papers and papers – so many papers published now that are relevant to dementia, people with dementia, or older people with difficulties of mental health in later life: some are really special.

    I think Carole Brayne’s papers are wonderful and amongst these ‘A two decade dementia incidence comparison from the Cognitive Function and Ageing Studies 1 and 2’ www.thelancet.com/journals/lancet/article/PIIS0140-6736(13)61570-6/abstract is especially important.

    The review of relevant literature reminds us of the Domesday forecast for the future of a world drowning in dementia. These forecasts are informed by extrapolation of prevalence rates and incidence rates identified in studies 20 years and more ago. There are a number of studies from the USA and Europe which hint of changes – lower incidence and lower prevalence maybe. These studies have shown trends but most have lacked power to obtain statistically significant, robust findings.

    The CFAS studies drew on large samples drawn from Primary Care Registers 1989-94 and then 2008-11. An earlier paper – Lancet 2013 Oct 26; 382(9902) 1405-1412 had found that the prevalence of dementia in Cambridgeshire, Nottingham and Newcastle (summed) had fallen from 8.5% to 6.5% between these two time slots. Extrapolating the 8.5% estimate to the population of England and Wales gives a national prevalence of 884,000. Using the actual prevalence discovered 2008-2011 the figure becomes 670,000. This is a reduction of 214,000 or 32%.

    This most recent paper looks at incidence – the number of new cases of dementia developing during a two year period from the baseline assessments. The findings are so interesting and so important:

    The easy take-home message is that incidence fell by 20%.

    The detail says that the fall was small in women, but evident in all but one of the age-bands included. For men the fall was dramatic – evident in every age-band and most emphatically amongst those who are 80-84 years and 85+.

    Estimates based on incidence findings from the early 1990s had forecasted in the order of 250,000 new cases per annum nationally by 2015. Using the updated findings, the calculation yields only 209,600 – a reduction of 40,000 or approaching 20%.

    So the picture is changing – it seems likely that education and health promotion, based mostly on cardiovascular risk factors, is having an impact.

    Most interesting to me, though, is the additional detail: falls in incidence are seen in Cambridgeshire and Nottingham, but not in Newcastle, where incidence rose from 16.7 cases/1000 person years to 20.6.

    Looking at the population by Townsend’s Deprivation Index and postcodes – incidence of dementia fell amongst the wealthiest and middle band of postcodes – but rose amongst the most deprived. The figures from the 2008+ cohort show a fairly steep gradation: 14.0 new cases /1000 person years amongst the wealthy, 18.7/1000 person years in the middle band and 20.6/1000 person years in the most deprived.

    The incidence of dementia is shifting – it can be shifted further if individuals take note and take action.

    On a bigger scale there is encouragement to provide better lives for the most disadvantaged so that they are less likely to develop dementia, with all that means to individuals and families.

    But dare we say it – the economics have to favour this egalitarian move, for reducing the number of people with dementia by a third will be a massive cost-saver. Any medication that had such robust evidence of benefits would be supported by NICE without hesitation.

    Do the people who can make these sort of changes know about this?

    Are these changed findings of incidence and prevalence, and their relationship to deprivation, reflected in the dementia map/calculator of NHS England?

    Should there be a revision of target numbers for GPs taking these new findings into account?

    « Back to archive
  • Leave your comment

  • Name:
    Enter the code shown above: