Simon Stevens on the lost art of analysis

It is fairly rare for incisive analysis to emerge publicly from the Department of Health. In part that is the result of trends dating back to the creation of the NHS Executive in the early 1990s, when empirical analysis and rigorous modelling were subordinated to managerial expediency.

But having spent seven years of my life consuming this stuff, I can also attest that it is partly because some of the best analyses never see the light of day, destined as they are to remain confidential advice to ministers.

Fortunately, every once in a while one of these breaks free from Richmond House, and in case you missed it, this just happened.

Explaining NHS Deficits is the most interesting DoH document in at least three years. So on the off chance it has not registered on your radar screen, my column this month is given over to it.

What makes it interesting is that it carefully and empirically manages to detonate many of the most powerful urban myths surrounding the contemporary NHS. Let's take two of them.

First, the claim that 'the reason for deficits in the South is that the weighted capitation formula gave too much weight to deprived areas'.

It turns out this explanation does not comfortably fit the facts. While new allocation targets that favoured Easington over Esher were introduced, the deficits that arose in 2005 did not correlate with the change in individual PCT resources.

Where the geographical argument does have more merit is that health service performance in the South was generally worse, so the distance to travel to meet a given standard was higher.

But numerous academic studies have failed to show that it is fundamentally a resourcing problem.

Second, what about the claim that 'NHS deficits have been caused by large pay rises'? Well it turns out that unit labour cost growth in the financially well-lubricated four years after the NHS plan of 2000 was fairly similar to that of the cash-starved four years prior to it.

So of the incremental growth available of£13bn, only£1.8bn went on above-trend pay. This supports the benign conclusion of DoH economists that 'on available evidence, the NHS largely avoided demand-pull cost inflation - one of the major macro threats posed by [the large NHS budget increase] - and succeeded in translating the [budget increase] into a substantial increase in capacity'.

But there of course is the rub. The overshoot was on the number of new staff rather than on the level of pay rises for existing staff. Because from 2000-05 the NHS recruited an additional 128,000 full-time equivalents over and above the historic trend in staffing growth.

So of the incremental growth available of£13bn,£11bn went on wages, of which a whopping£9.2bn went on extra staff, far in excess of what had been specified in the NHS plan.

Now on the one hand that can be seen as a good thing, addressing longstanding 'capacity' shortfalls. But on the other it can be hugely problematic if a) staffing expansion exceeds the budgetary growth to pay for the new posts and b) those new posts do not generate commensurate improvements in quality or output. Both of these have happened.

Now depending on your ideological beliefs about how to organise the NHS, those problems result from either a major managerial disconnect between finance, operations, and human resource planning both nationally and at hospital level; or a failure to fully embed an incentive-driven system that would ensure new capacity produced equivalent improvement.

But I think there is a deeper lesson: the kind of excellent analysis that the DoH has just produced is too little too late - two years after deficits first hoved into view and went on to claim the scalp of the department's permanent secretary. One of the great frustrations when the NHS plan was being drawn up seven years ago was the paucity of available analysis, particularly around workforce and labour markets. Those chickens have come home to roost.

So one of the tasks of the DoH's new official leadership is to generate a culture of first-class, fact-based analysis, with a set of capabilities to match. And then to apply them to the major policy questions still outstanding. Let me conclude with two examples.

- Hospital mergers: just how good is the evidence that merging neighbouring district general hospitals improves performance? Evidence from the US shows hospital mergers in the 1990s created savings but allowed them to hold purchasers to ransom - raising inpatient prices by 40 per cent or more where merging hospitals are closely located.

-- Practice-based commissioning: are we about to see the triumph of hope over experience? Optimistically assume perhaps half of GPs will be seriously interested in commissioning activities. And optimistically assume that perhaps two thirds of those will be competent commissioners. The combined probability is that only a third of GPs will be willing and able to make payment by results deliver in the way that its advocates suggest.

Don't get me wrong. I am a big supporter of practice-based commissioning, and will return to this theme in my next column. But shouldn't policy-makers honestly confront the fact now
that by itself it probably won't deliver the kind of sophisticated commissioning most people now accept we need across the whole country?

Or will we have to trundle forward a few years without that kind of hard-nosed confronting of reality on these and other topics, until we get a further DoH post mortem?

Simon Stevens is chair of UnitedHealth Europe and visiting professor at the LSE.
Simon_L_Stevens@uhc.com