According to the Alzheimer’s Society, ‘the number of those affected by dementia is soaring’.
But given that the Society is commenting an updated review into dementia prevalence, their statement is rather puzzling. The review should have taken into account the CFASII study published last year (the latest research on dementia), which suggests there is a much lower incidence of the disease than previously thought.
What is even more puzzling is that the report claims to take account of this research, and yet concludes that the prevalence of dementia is exactly the same as was previously thought.
Considering the political prominence of dementia, the huge emphasis on diagnosis rates and apparent ‘postcode lottery of diagnosis’ it is certainly important. I’m usually the first to admit that there could be other evidence out there that balances the findings of CFASII, and I’d like to take a look at the evidence myself, to test the scientific validity of these new numbers – only I can’t, because the methods behind the second Delphi consensus report are about as transparent as thick pea soup and unfit for a document that will be used to shape and justify health policy for years to come.
The whole rationale behind the Delphi method is more suitable to an old boys’ club than 21st century scientific method. A panel of Key Opinion Leaders is brought together. Each individually looks at a set of studies and considers the data.
They give their own estimates for the prevalence of dementia in each age group (based on whatever scientific method they like, we have no way of knowing – I suspect the rolling of dice).
These estimates are then submitted anonymously – so that not even the other panel members know if someone is behaving out of turn.
Finally, after giving the panel an opportunity to rethink, no consensus is agreed, but the anonymous results are averaged. Hey presto! We have the official, unquestionable and much-to-be-quoted figure.
Transparent? Nope. Accountable? I think not. Reading the report, it is not even clear which studies were included. There are references, for sure, but no clear list of the prevalence studies. We have no idea how widely the experts disagreed, or who said what.
If this were simply a study to help map resources to need it would not matter so much, but these figures will be used to give every GP in the country a diagnosis rate, and a new target rate of 75%. The CQC will start to get interested in these figures, and then the real impact of this report, and the harms of setting targets for diagnosis, will start to bite.
Then there’s the issue of conflicts of interest. If these Key Opinion Leaders are going to have such an impact on policy, do we not have a right to know if they have any competing interests? The Alzheimer’s Society doesn’t seem to think so. While I accept that this is only an overview, and the full report will be available later this year, it is simply not acceptable to publish a report of this importance without them.
More importantly, when people declare interests it behoves them to ask themselves if their position is too conflicted for them to take part.
For example it is my view that the expert panel should not have included the Dementia Tsar, Professor Alistair Burns. Prof Burns is far too close to the Prime Minister’s dementia challenge to be able to declare himself an independent expert. As national clinical lead for dementia, it has been his task to increase rates of diagnosis (as instructed by the dementia policy). But having made such a song and dance about how poor GPs were at diagnosing the condition, it would be embarrassing for the Government to find we were better at it than they thought.
When I challenged him on Twitter, Prof Burns declared that he was not conflicted. I’m now not sure he’s the right person to judge and I am left feeling that there is something not quite right about these figures. They really are just a little too convenient.
Dr Martin Brunet is a GP in Guildford and programme director of the Guildford GPVTS. You can tweet him @DocMartin68.