Time to break the silence about unequal health spending

23 Aug 16

Fears that devolution of NHS spending will lead to greater inequality miss the point. Variation is already built into the system thanks to a nonsensical formula

A few days ago, I was listening to a lively argument on LBC about whether or not a particular factor, one of many, was the cause of a particular ill, when an economist phoned in. He said there was an easy way of deciding the question – just press the ‘regression’ key on a PC without even having to understand anything about the relevant data-base – and the black-box machinery would ‘take account of’ all the other factors. In his elegantly informative paper Devo-Health: What & Why for the Institute of Public Policy Research, economist Harry Quilter-Pinner shows he is well outside the class of the caller to LBC, but nonetheless manages to remain ignorant of data that bears heavily on the devolution question.

The combined authority of Greater Manchester has been promised greater control of its £6bn health and social care NHS budget, which Quilter-Pinner sees as ‘delegation’ rather than devolution. The £6bn covers the resource allocations of the dozen clinical commissioning groups in Greater Manchester. For 2016-17, the per capita allocations vary little – from £1,124 to £1,392. For the 200 or so CCGs of England, however, they go from £957 to £1,414 – a ratio of 1.5 to 1. With an average of £1,302, Greater Manchester is comfortably close to the top of that range. The huge variation in allocations is ignored by Quilter-Pinner when he downplays the risk that devolution would increase the “variation in the quality of care and the health outcomes” which, he says, “already exists across England under our more centralised system”.  

For Quilter-Pinner, ‘variation’ means variation in CCG outcomes such as the cancer survival rates elegantly displayed in one of the 63 maps of the NHS Atlas of Variation. The variation of per capita allocations about which Quilter-Pinner is silent is what NHS England imposes on CCGs on the basis of a target formula whose provenance and design is treated with contempt by leading statisticians and which has even been proved, rather than subjectively judged, to be utterly nonsensical. The formula derives from a widely practised abuse, by Department of Health finance officers and the health economists they engaged, of the statistical method of ‘regression’. But Quilter-Pinner makes no reference to the formula. Along with many others, he is complicit in the silence of those in NHS England who either believe or purport to believe in the trustworthiness or ‘fairness’ of the formula as a target measure of what a CCG needs to service the healthcare needs of its GP-registered patients.

In 2007, faith in the then-current version of the formula was enhanced when Simon Stevens (then of United Health) claimed that management incompetence rather than a defective formula was responsible for primary care trust deficits.  The accounts-based evidence of the financial performance of CCGs in 2013-14 (strengthened by the 2014-15 accounts) is that Stevens was right, and I owe him an apology for suggesting otherwise in Failing to Figure.

The National Audit Office did not question the formula in its 2014 paper for the Public Accounts Committee, which proudly displayed a weak correlation between financial performance and ‘distance from target’ (actual allocation minus formula). NAO research could not, however, find any correlation between the distance variable and a handful of key health outcomes. Nor did it discover the strong correlations that can now be seen between the Atlas of Variation performance measures and per capita allocation, either target or actual – and which, it should be noted, are irrelevant to the question of formula validity.

More to the point, the NAO failed to discover that the weak correlation they presented to the PAC was almost certainly the outcome of CCG finance officers’ response to large distances above or below target. Once the allocation variable has been properly taken into account in a two-variable regression, the target formula is almost irrelevant for most of the 26 Atlas performances for which there is strong correlation with allocation. The ‘distance from target’ variable values for individual CCGs result from a pull-or-push ‘political’ tussle over the years between target and allocation and the 200 or so values are effectively random numbers. All of which affirms the nonsensical roots and character of the formula itself.  

It is unlikely that politicians will ever relinquish control of NHS funding and possibly undesirable that they should. The IPPR paper is frankly political in urging the Labour candidate to make further health devolution a pledge of his campaign to be mayor of Manchester, and in giving a leading role to the former Labour health secretary Alan Milburn in an IPPR working programme with the same objective. The irony of such a role for Milburn is that he was the minister who introduced the 2002 revision of the formula for PCTs with the assurance that it could somehow legitimately “take account of unmet health needs” and would be “fair to all parts of the country”.

The IPPR paper envisages just two possible risks of devolution. Firstly, that it could increase the inequalities in health outcomes thereby making the variation so unacceptable that “we will lose the ‘N’ in the NHS”; secondly, that “there are very real concerns that ‘devo-health’ will ultimately lead to finger-pointing between central and local government as the next round of public sector cuts hit”. The paper claims that the first risk has been exaggerated, on the grounds that the inequalities are due to disobedient and therefore corrigible management. The second risk pictures, like the Sistine chapel ceiling, a Whitehall finger responding to a pointing finger that somehow represents all the local authorities that have very different per capita allocations.

The IPPR paper is unable to envisage the risk that the dam of silence about the formula will break one day, and that fingers will then be pointed in all directions by local authorities, particularly at Manchester with its currently happy £1,302 per head. The inundation would give new and well-informed life to the long-standing arguments about the fairness of the succession of formulas that has generated such inequalities.

Did you enjoy this article?

AddToAny

Top