Evolution, not revolution

Despite hysterical claims to the contrary in many commentaries, there is no doubt that the proposed National Health Service (NHS) reforms were evolutionary, not revolutionary (a phrase to which Le Grand can claim authorship, at least in this context, appearing in a letter he published in the Financial Times on 29 October 2010). As Timmins argues in the opening ‘Act’ (he terms the chapters of his book Acts, subdivided them into Scenes), the changes have their origins in what were a truly revolutionary set of quasi-market reforms: those introduced by Kenneth Clarke in 1991, following the seminal White Paper, Working for Patients (Department of Health, 1989). These included the setting up of an internal market, the splitting of purchaser from provider, the formation of independent hospital trusts, the replacement of hospital global budgets with fee-for-service or payment by activity and GP commissioning (under the name of GP fundholding).

In the event, these reforms yielded some improvements in NHS quality and efficiency, but a review conducted by Le Grand, Nicholas Mays and Joanna Mulligan concluded that “the incentives were too weak and the constraints too strong” for them to have a more significant impact (Le Grand et al., Reference Le Grand, Mays and Mulligan1998, p.130). In Never Again?, Timmins argues that the New Labour Government that took power in 1997 reverted almost immediately to ‘command and control’. In fact, under the new Secretary of State, Frank Dobson, although the Conservatives' internal market was not dismantled in its entirety (in particular, the purchaser/provider split was retained, although GP fundholding was abolished), there was movement towards a greater reliance on what one of us has termed the ‘trust’ model of public service delivery: one where front-line providers are trusted to deliver the goods without any form of external pressure or institutional incentives, and where the language of co-operation and networks replaces that of competition and markets (Le Grand, Reference Le Grand2007). It was not really until the perceived failure of that model around 2000 that both the model and the Secretary of State were replaced: the Secretary of State by Alan Milburn, and the trust model by its converse – a command and control or ‘mistrust’ model – whereby external pressure to improve was applied directly by government through the setting of targets and the micro-management of performance.

Although targeting and performance management (or ‘targets and terror’ as it became known) were broadly successful at reducing waiting times (Propper et al., Reference Propper, Burgess and Gossage2008), the Government, to which Le Grand was then an adviser, felt that this model had too many adverse side-effects (gaming, the distortion of priorities, professional demoralisation) to be a long-term solution to the underlying problems with the NHS. However, they did believe that some form of external pressure was necessary and that a development of some form of competition in a quasi-market might be a source of that pressure instead of, or as a complement to, that from targets. Therefore the government introduced a batch of market-oriented reforms quite similar to those of Kenneth Clarke, including patient choice, foundation trusts, payment by results (actually payment by activity) and a version of GP fundholding under the name of practice-based commissioning. However, there were two significant differences from the earlier reforms: private providers were encouraged to enter the quasi-market, and prices were fixed. Moreover, there were two important differences in the reform process from the Coalition's reforms that were to follow: they were introduced incrementally and they did not require legislation.

At the time, as with Working for Patients, these reforms were roundly criticised, with apocalyptic predictions of disaster and the imminent collapse of the NHS. However, the King's Fund, in a comprehensive review led by Nicholas Mays, found that “the evidence … shows broadly that the market-related changes introduced from 2002 by New Labour tended to have the effects predicted by the proponents and that most of the feared undesirable impacts had not materialised to any extent” (Dixon et al., Reference Dixon, Mays and Jones2011, p.191) – though it added that the improvements were not as great as those induced by the target and performance management regime.

As Timmins shows, Andrew Lansley, the new Secretary of State for Health in the Coalition Government that came to power in 2010, took as a basis most of the New Labour reforms, developed them and filled in some of the gaps. Commissioning under New Labour by Primary Care Trusts (PCTs) had been much criticised. Hence, PCTs were replaced by GP-run clinical commissioning groups, many of them based on the practice-based commissioning consortia that had grown up under the previous regime. The idea that private or non-profit providers could win NHS contracts was extended to ‘any willing provider’, and an economic regulator was introduced to encourage competition and to regulate the quasi-market (a notable absence from the New Labour reforms). An administrative layer was replaced (Strategic Health Authorities) by a national commissioning board and some powers were to be given to the local government with respect to public health. And that was (largely) it.

So why all the fuss? Timmins puts this in large part down to Mr Lansley: his refusal to acknowledge the New Labour foundations on which he was building, his own framing of the changes as ‘a challenging and far-reaching set of reforms’ (Department of Health, 2010, p. 6) and his desire to enshrine these in a legislative ‘big bang’. The last of these was particularly significant, for it provided both a strong incentive for the latent opposition to the previous governments’ market-oriented reforms (which had never really gone away) to mobilise and also provided a forum (several fora, in fact) to express that opposition.

As Timmins shows, the opposition and the way it was dealt with created massive political damage for the Coalition. Worse, it probably damaged the outcome of the reforms. The commissioning groups have been hobbled with a much greater role for local government and, more significantly, for local provider interests; the role of the regulator, Monitor, has been changed, from promoting competition to simply trying to prevent anti-competitive behaviour; and the language of competition has been replaced by that of ‘integration’. The principal danger of all this is not that the outcome will be a disaster; rather that it will result in a positive but very weak performance of the same kind that, as we noted earlier, characterised the Conservative internal market. And this may happen for the same reason: the incentives are too weak and the constraints are too strong.

Evidence-based policy or policy-based evidence?

Before the pause in the legislation, in a front page article in the Financial Times, Timmins (Reference Timmins2011) cited Cooper who suggested that, although fixed-price competition had been demonstrated to improve clinical performance, price competition could actually harm patient outcomes. The evidence on price competition was based on research on the impact of price competition in New Jersey and Carol Propper's work on the Conservatives’ own internal market (Volpp et al., Reference Volpp, Williams, Waldfogel, Silber, Schwartz and Pauly2003; Propper et al., Reference Propper, Sutton, Whitnall and Windmeijer2008). This research suggested that price competition could indeed drive down costs, but also, on occasion, could drive down quality.

In the book (pp. 80–81), Timmins draws attention to Cooper's comments, as well as the views of Carol Propper and Le Grand, and notes that this contributed to the government's eventual ruling out of price competition. However, further research evidence produced by Cooper, Propper and colleagues, and based on the New Labour quasi-market, had a rather different impact. Cooper et al. (Reference Cooper, Gibbons, Jones and McGuire2011) found that, in line with the predictions of economic theory, hospitals in England that were located in more competitive areas reduced their mortality more quickly during the period when patients could choose their provider. This evidence was confirmed by analyses performed by Gaynor et al. (Reference Gaynor, Propper and Moreno-Serra2010). The mechanisms by which this occurred were elucidated in a piece by Bloom et al. (Reference Bloom, Propper, Seiler and Van Reenen2011), who found that competition was associated with better hospital management and lower death rates.

Moreover, the English experience was not an aberration. Research on US experience with fixed-price competition also illustrated that it improved clinical performance. In fact, the magnitude of the effect that Cooper and colleagues and Gaynor and colleagues found in England was nearly identical to the treatment effect measured in the United States by Kessler and McClellan (Reference Kessler and Mcclellan2000).

Unsurprisingly, this research – illustrating, as it did, that choice and competition improved performance in the NHS – was seized on by proponents of the government's reforms. The work of Cooper and colleagues was cited by the Prime Minister in several speeches, it was referenced in debates in the House of Commons and the House of Lords, and it was referenced by the NHS Futures Forum (Cameron, Reference Cameron2011; Bubb, Reference Bubb2011).

Equally unsurprisingly, given its profile, the research attracted some vehement criticism. Allyson Pollock and colleagues launched an attack (Pollock et al., Reference Pollock, Macfarlane, Kirkwood, Majeed, Greener, Morelli, Boyle, Mellet, Godden, Price and Brhlikova2011) in a comment piece in the Lancet. This was followed by various opinion pieces in the Guardian newspaper and named and unnamed blog posts and much Internet discussion. Although there was some scholarly discussion of the work, for example, the commentary by Bevan and Skellern (Reference Bevan and Skellern2011), the vast majority of the debate was ill considered and sensationalist, often revealing a basic misunderstanding of the data and quantitative methods used by the researchers (Bloom et al., Reference Bloom, Cooper, Gaynor, Gibbons, Jones, McGuire, Moreno-Serra, Propper, Van Reenen and Seiler2011).

Of particular interest was the tone of the debate. This was often abusive, involving accusations of unprofessionalism and implications of conspiracy. Some of this was obviously fuelled by ideological objections to the market orientation of the reforms, and by a political antipathy to the Government. However, there was also occasionally an underlying suggestion that academics should not be doing this kind of thing; that the academic's role should be primarily that of an independent social critic; and that academics should not provide ammunition that supported the direction of government policy, because that in some way compromised their independence.

This raises a number of issues. One concerns research methods: how to discuss, debate and dissect research evidence that involves complex methodology and econometric estimation. Increasingly, as big data become available and econometric techniques evolve, data driven decision making and statistically driven causal inference will become commonplace. However, the rise of econometrics and big data create a ‘black box’ where smaller groups are unable to determine the methodological veracity of research. This is a double-edged sword. Gaps in knowledge allow those without familiarity with the research techniques to level critiques of the research, some of which may sound plausible to the public, but which sound ludicrous to informed researchers. However, the researchers themselves become increasingly immune to challenge; at the extreme, they develop a monopoly in information, and, as with all monopolies, there is a danger that this can be abused.

It is not clear what can be done about this. Peer review is clearly an important guard against methodological mistakes or misjudgements. However, journal reviewers often do not have the time and resources fully to check all aspects of a heavily quantitative piece of research and the timing of the peer review process often means that topical research is not published in time to affect policy debates. One answer might be, where possible, to require replicability, or indeed actual replication, of policy-relevant research, where independent researchers use the same data and methods to check the results. Something like this happened with the hospital competition work, where two quite independent research teams used very similar methods and data – and did indeed come up with similar results. However, the duplication here was by accident rather than design. Already much policy-relevant research arrives too late to be of any use in the formulation of policy; a replicability requirement would delay things further. Perhaps the only long-term strategy is for more training for journalists, policymakers and academics in statistical inference so they can make well-informed assessments of their own.

A second issue is more general and concerns the role of evidence-based policy. As the argument for this goes, better evidence that is policy relevant will increase the likelihood that policies are driven by substance, not ideology. As the body of research supporting a particular policy grows, the case for policy action should become more convincing. This was precisely the case in England. Evidence showed that price competition in both the United States and United Kingdom could lower quality. However, three articles found that fixed-price competition in the United Kingdom improved outcomes. These results were consistent with the magnitude and direction of the US evidence. Moreover, both empirical results were consistent with economic theory, which suggests that price competition may lower quality, whereas fixed-price competition should improve it.

However, the experience of those looking at the impact of competition on the NHS has illustrated that, far from depoliticising policy debates, the academic evidence itself got politicised. Advocates of evidence-based policy may believe that their work might help decide between policy alternatives, but what this experience suggests is that, however extensive the evidence may be, it is unlikely to sway those whose fundamental beliefs about the way the world works are being challenged. In fact, it seems the best that can be hoped for is that evidence can at least be used to improve the design of the alternatives – as, for instance, the evidence on price competition contributed to that form of competition being dropped.

Finally, what about the role of the academic? This is relatively simple. Academics should not be lackeys of either the government or its opposition. Rather they should aim to produce and publish objective research on the key economic and social relationships within the policy area concerned – as best they can. If this can be produced in time to inform government decisions over the relevant policy, so much the better. However, academics should not overestimate the likely impact of the research on the fundamental issues that underlie many policy decisions.

Framing reform

In a sense, both of the principal issues to which we have drawn attention in this piece are problems of framing. As Timmins argues, if the reforms had been framed in terms of evolution and not revolution, they would not been implemented without controversy, but probably would have been put in place in their entirety, much more quickly, and with far less political damage. Similarly, if research is framed as ‘the answer’ to a political debate, it is doomed to be politicised and thereby potentially dismissed. Conversely, if it is framed as an instrument for improving policy design, it is more likely to be listened to, at least by those in a position substantively to shape policy. Behavioural economics experiments in the laboratory has shown us that framing is important in understanding why people behave as they do; what the process of NHS reform has illustrated is that it is of critical importance in the messy real world of policy formation.