Enough is enough! An appeal to reason
Civitas, 1 February 2013
By Mervyn Stone
Just before Christmas 2012, Sir David Nicholson initiated an ‘urgent, fundamental review‘ of Clinical Commissioning Group (CCG) funding formulas.
Just before Christmas, I was preparing to tell a meeting of The Community Voice about some particularly striking nonsense I had come across in a paper in the British Medical Journal.
It came from a group of 11 ‘health economists’ who were, so the paper suggested, hoping that Sir David would implement their formula―which had the acronym PBRA for Person-Based Resource Allocation. You can read my earlier post Doctors in the Waiting Room for more about PBRA.
On New Year’s Eve, I asked the group’s leader about something she had written that suggested that the BMJ paper was not their last word―I wanted Community Voice to get the latest version―but was told that PBRA was the last thing they had published.
So I told Community Voice members something about PBRA for the case of a 70 years old lady who had had hospital treatment for Parkinsonian dementia. According to PBRA, her GP’s formula-based account with the CCG for the following year would be debited £436, quite apart from any actual debit for the treatment itself. Yes, that is what the PBRA formula would do!
A week later, the Department of Health published the PBRA3 formula―the latest of a string of offerings―which had been over a year in the making and had plenty of debits to match PBRA. For example, a following-year debit of £653 for any individual who had had at least six different hospital-confirmed diagnoses over the last two years―a deduction from the GP’s allocation with no apparent connection to any payments for the past six treatments.
There was so much questionable logic in the DH publication that I wrote a 7-page critique of PBRA3’s logic entitled Plain Explanation or Special Pleading?―and sent it to Professor Sheena Asthana for authoritative comment, from someone whose knowledge of the academic health care industry is probably unmatched. She has urged me to explain things more plainly―which is what I have tried to do in the revised critique―not an easy task because questions of logic are not everyone’s cup of tea.
I have heard it suggested that the problems and efforts of fitting the models PBRA and PBRA3 (and those between) to a huge data base of 350 variables on each of 5 million individuals may been so fascinating and so exhausting, respectively, that little time was left in the contracted-research period for checking the logic of what has been achieved.
On 22nd January, I switched on Radio 4’s Today just in time to hear Evan Davis introducing an item on Big Data. Here are bits I have scribbled down from Listen Again:
a truly massive amount from multiple sources / a step change in the sophistication of processing programs that can crunch all this data and tease out the subtle patterns / more than the sum of its parts / the Prime Minister has been very keen to exploit the resource the NHS presents / combine genetic information with NHS patient lists / the semantic web / key to being able to handle massive data sets / updating predictions with trending.
The ever-sceptical Evan Davis asked his interviewee, Dr Shirley Ann Jackson, who is President Obama’s adviser on the exploitation of the ‘semantic web’, whether she was convinced that people know what they’re doing―that Big Data is manageable data?
On the evidence that I have been sweeping up for a decade into the Statistics Corner of Civitas and into the Civitas booklet Failing to Figure: Whitehall’s Costly Neglect of Statistical Reasoning (appreciably strengthened by sight of what the Department of Health may be doing with PBRA) I suggest that over here we should withhold assent to Dr Jackson’s response to Evan Davis’s question―a modest but optimistic Well, we are making progress!