The presentation is available on the NHS Confed Website http://conference.nhsconfed.org/Pages/default.aspx

the cat video is:here

Click here to download:

presentation_set.pdf (1.06 MB)

(download)

Click here to download:

presentation_set.pdf (1.06 MB)

About a month ago Michelle Drage challenged me when I said that I thought commissioning was a flawed concept.  I was remarking on some international experience and the observation that what the most interesting GP 'commissioners' were up to looked good but actually about provision.  Here are some early thoughts in response to that challenge.  Dr Nick Hicks I hope is going to help me provide the second part of this analysis.  I am grateful to ideas I have got from David Stout, Dr Martin McShane, Sophia Christie, Professor Richard Saltman, Professor Martin McKee, Dr Jennifer Dixon and a number of others.  

Continued attempts to make commissioning work

The NHS has been trying to develop effective commissioning since 1991.  While it is true that the process has been hampered by the lack of an agreed definition of what commissioning is, the directive way it in which it has been managed and frequent reorganisations interspersed with periods of neglect, the results remain somewhat disappointing.  Following the World Class commissioning process where a very large amount  was spent on consultancy, development and assessment there were signs that some PCTs were beginning to make an impact.  But even here the results were patchy.

The current reforms appear to be a further attempt to reinvent the idea of commissioning and so it is worth asking whether the model of commissioning that has been operating is the right one and whether a more fundamental redesign is required.

There are several reasons to question whether the model of commissioning that has been in use was ever going to deliver what was expected and therefore what needs to be done to redesign it.

The model is not is use elsewhere

The English version of commissioning is very ambitious and includes measuring need, planning ways to meet it, procuring health services, making contracts and monitoring their performance.  It is also unique.  Where it departs from the role of payers in other health care systems is in the very high level of detail which commissioners attempt to impose on providers and therefore the extent to which they take on risks and responsibilities, often including responsibility for the financial viability of providers and their performance against detailed national and local targets.   This is much more directive and intrusive than is found in other systems and although some USA payers do attempt to control more elements of provider behaviour the level of resource available to them is very much greater and even so there is limited evidence about its efficacy.  So although a number of countries do copy policy innovations from the NHS such as NICE.  No attempts have been made to copy the English NHS model of commissioning.   

This approach is also very different from procurement in other parts of the economy, e.g. construction, where there has been a move away from input process based specifications to more focus on specifying the outcomes that are required.  

Commissioning reflects a continuation of NHS Culture not a challenge to it

The unique NHS approach to commissioning may be because it incorporates many of the basic assumptions that underpin NHS culture.  These have rarely been questioned because by definition culture is not noticed by those that are immersed in it.  Commissioning developed in a system where most of the relationships were between bodies that werepart of the same organisation.  Power was hierarchical, those running things were subordinate to those at the centre and regions, tasks got done by direct instruction and the focus was generally short term.  All of this was directly translated into purchasing as it was called before the term commissioning was coined.  The commissioning cycle that has been developed and the approach of using contracts, detailed specifications, pathway design and other techniques carries with it the baggage of 40 years of planning and hierarchy based direct management.   This means that commissioning in its traditional form may be more likely to inhibit the development of personalisation, choice and innovation than to promote it. 

Problems with the model

There are a number of issues with the underpinnings of commissioning that challenge its appropriateness and reduce its probability of success.

Firstly, there is the problem that a great deal of clinical care takes place in a ‘black box’ whose workings are invisible to the commissioner or if they are visible, are incomprehensible.   The information asymmetry between commissioners and providers is significant and is compounded by the complexity of much of the healthcare supply chain so that, even without the asymmetry problem, commissioners are likely to be overwhelmed with complexity if they attempt to coordinate care.  In doing so they will also assume responsibility for risks which they are poorly positioned to manage.  Admissions to hospital, referrals, treatment decisions and other key drivers of costs are taken by individual clinicians and there are limits on how far commissioners can have any significant influence on this.  They can shape the options available and the context in which these decisions take place but intervention in detailed decision making can be illegitimate and possibly hazardous.    

Secondly, there are limits to the extent to which contracts and incentives can be used to change provider behaviour.  For obvious reasons contracts in healthcare will always be incomplete. In other words, they will leave important items unspecified because the transaction costs of filling in the gaps are just too high.  Contracts can only be varied substantially by negotiation and periodically. The cost of making these changes is often high.  

 Experience in other systems with DRG payment mechanisms demonstrate that there are limits to how far financial incentives can be used to drive provider behaviour.  This is partly because the design of mechanisms that unequivocally produce the desired behaviours in providers, that do not produce unintended consequences and which are not highly costly to operate is very difficult.  Also, like targets, if there are too many micro-incentives and penalties providers may decide that the transaction costs of optimising these is not worth the return or that gaming and other undesirable strategies are adopted instead.

Thirdly, there are aspects of the healthcare system which undermine the ability of commissioners to operate as purchasers.  In particular, the highly concentrated nature of hospital provision and the lack of flexibility on the supply side of the system limits their power.  In much of the NHS there are limited alternative suppliers and clinical interdependences and an unfavourable ratio of variable to fixed costs so that savings made in one area lead to price increases in others to cover overheads.  The lack of spare capacity and new entrants has also meant that the scope for even effective commissioners to drive change has been limited.

Forthly, commissioning has tended to perpetuate an unhelpful and historic division between primary and secondary care – even having an explicit goal of shifting work from one area to the other as though they were entirely separate areas.   But, attempts to change activity in secondary care often depend on changes in the behaviour of primary care e.g. through changing referrals or the management of some conditions.   Research by the Audit Commission and the King’s Fund has founded limited evidence of success in this area.  There has often been a problematic relationship between commissioners and primary care and commissioners have had limited ability to influence the behaviour of GPs or to deal with problems of poor performance.  

Models from elsewhere suggest a different approach

History and experience in other countries suggests that commissioning may be more successful if it is focussed on setting very clear expectations about outcomes, standards and if necessary a small number of process measures that are particularly important to patients or for population health.  It can have an important role in creating high level incentives to ensure that these expectations can be delivered. They may have to take action to stimulate the provider side of the system to ensure that there are dynamic providers able to meet the needs of patients.  In some cases payers will need to take quasi regulatory action to ensure that complex multi specialty services that require planning and coordination (e.g. trauma, cancer services, stroke) are provided in a cost effective way. But none of this implies that they should be involved in the micro redesign of services except possibly where there are serious failures.   In other words payers leave most of the heavy lifting of sorting out how to deliver outcomes, organise complex systems of care delivery and co-ordinate different services to the providers.  They should also try and transfer as much of the risk as they can to the providers as generally they are best placed to manage it.  

The more integrated the provider the more the payer can pass these tasks and risks to the provider and the easier the design of the payment system.   However, the trade off is that the more integrated the provider the smarter the payer has to be at measuring outcomes and being able to compare the performance of different providers. 

In these models commissioners can cover large populations and there economies of scale such as reduced transactions costs, the ability to manage insurance risk, etc.  This partly explains the concentration of insurance funds in Germany and the Netherlands over the last few years and Denmark’s shift from councils to regions.  The advantage of commissioners covering a defined geography is that one part of the system has responsibility for the entire population, there is a clear denominator for making comparisons and incentives to invest for the long term in health promotion and prevention.  The danger of commissioners competing on risk selection also avoided. The disadvantage is that they may not be subject to sufficient pressure to improve as their population is captive.

Time to say goodbye to commissioning

The concept of commissioning looks rather unhelpful and includes such a wide range of aspects of planning and delivering healthcare that it is difficult to pin down where its boundaries are.  Most of all, in drawing the line between commissioning and provision too many provider responsibilities have remained with commissioners.  Policy makers have expected commissioners to solve problems and change provider behaviour through indirect methods in areas where managers of those providers have often struggled to get results through more traditional and direct methods.  This attempt to reinvent direct management with a contractual wrapper may explain the explosion of costs in PCTs in the last part of the previous decade. 

Even if they had not been burdened with a large number of other duties unrelated to commissioning, frequently reorganised and subject to initiative sapping top down management the expectations of what the this model of commissioning could achieve were unrealistic. 

The analysis above suggests that commissioning can be usefully disaggregated into:

·         Payer functions:  population and prevention focussed, looking at large populations -  assessing need, measuring gaps,  setting outcome goals, procurement and monitoring.  This includes the design of incentive frameworks and in some cases market management.   They need to engage with the public and a wide range of stakeholders to define priorities. 

·         Provider functions: patients and pathway focussed - creating networks of provision, developing pathways, organising care and improving provider performance.   They need to have highly effective ways of enaging patients in the design of care.

·         In some systems some functions relating to standard setting are vested with regulators, government or professional associations rather than with the payers. 

We have not had a commissioner-provider split.  Instead there has been overlap, confusion and the continuation of top down provider management accompanied by a lot of activity uncharitably characterised by a Scottish health board chief executive as ‘playing shops’.  

GPs as commissioners

The question is whether the reform proposals set out by the Secretary of State are just another attempt to reinvent commissioning but this time done by GPs. However, much faith one has in the ability of GPs this would seem to be unwise.  The very much reduced resource available to support GP commissioning makes it impossible.  Something different is required.  However, the various documents explaining the reforms are not very clear on what the answer is. 

The insight that GPs and other clinicians bring important knowledge and insight to the process of planning and managing health care provision is absolutely right.  The principle that GPs, and indeed all other clinicians, should recognise that every clinical decision has resource allocation consequences for which they are responsible is very important and a principle of the reforms that enjoys widespread support.

The Command Paper says ‘Commissioning by GP consortia will mean that the redesign of patient pathways and local services is always clinically-led and based on more effective dialogue and partnership with hospital specialists. It will bring together responsibility for clinical decisions and for the financial consequences of these decisions.’   This seems to reflect the way that a number of GPs conceptualise commissioning – as the redesign of provision, the part of the commissioning process that the analysis above suggests should be best left to providers.  

The decision to vest payer and provider function with GPs and to put a number of other provider functions across a divide creates some tensions and unnecessary complexity:

To be really effective GP consortia will have to get a grip on some elements of primary care.  There is evidence that there are significant variations in quality and these translate into increased hospital utilisation and poorer outcomes.  The ability of consortia to do this will be constrained by the need to avoid conflict of interest involved.

If the incentives to change GP behaviour are too direct there is a danger that patients may perceive that GPs decision-making is influenced in by the requirement to manage budget or more damagingly because they have a financial interest in the provider to which they are being referred. 

A lot of tortuous policy has had to be designed to curtail these conflicts of interest and provide assurance.  There is a distinct danger that this could dampen the innovation and change that the policy is designed to create.

Conclusion:

We need a different approach that achieves the government’s aims

That builds on the strengths of general practice

That doesn’t put secondary care in a different box

That will drive improvement in primary care

And which will unleash the entrepreneurial abilities of GPs and specialists working together

Next step is to say how……..

@mellojonny asks if anyone can explain why there are large cuts going on in the NHS now.  This is a good question because until 31 March NHS was still receiving levels funding which were relatively generous by historic if not recent standards.   There are a number of possible reasons, I've listed a few here and maybe others will add to them since I've done this quickly and I should really be doing something else.

1)      Unresolved structural issues in a number of places which means that they are unable to become more efficient or are locked into high cost facilities or they have developed a model of practising medicine and running the organisation that is high coster than similar places. The last decade of growth has meant that in many places this issue has not been dealt with.  The Secretary of State's decision to put a number of these changes on hold is will have increased the deficits in a number of places and led to a search for other more direct but probably less effective methods .

2)      In some cases the resource allocation formula may not be properly recognising needs in some areas while overcompensating others. The beneficiaries of overgenerous resource allocation tend not to advertise this.  The big issue may be that the formula does not recognise the impact of age as well as it should.  It may also be weak for areas with very unusual factors – for example very high levels of non-english speaking patients. 

3)      A number of strategies commissioners used to save money have involved moving work out of hospitals. The reason for doing is that the out of hospital care appears to be cheaper.  In some cases this may be genuine but often it is more likely to be an artefact of accounting practice and overhead allocation.  The direct costs may well be very similar.  The relationship between fixed costs and variable cost in hospitals is very unfavourable and so any loss of income leads to a large amount of fixed and overhead cost being left uncovered. The easiest solution to this is to increase the workload in other areas to replace the contribution to overhead and fixed costs that has been lost. The result of this is often that any savings that were made by arbitraging the difference between hospital and community settings is completely removed by the additional cost of new work in the hospital.  This can be seen in the phenomenon reported by the Audit Commission who observed that most of the increases in income generated by providers was in non- tariff activity where it is easier to negotiate both volumes and prices.  The effect of this is that when this work is decommissioned or demand is managed the amount that needs to be saved is much greater than the variable costs saved by stopping the work.

4)      The operating framework required the PCT should hold back 2% of their spending to be used non recurrently to deal with change. There was also relatively limited ability to pull down other retain surpluses. Both commissioners and providers are starting to see non-recurrent money running out and therefore taking action now.  This means less income for providers.

5)      Hospital tariffs were set at 1% below the average cost and a range of adjustments were made to tariffs which had the effect of embedding a 2% efficiency savings into it.  The target for non-tariff work was -1.5%. In addition to this the penalties for readmission, the measure which limits payment for emergency cases beyond the level in 2009/10 to 30% of the tariff and some other technical adjustments all add up to an extremely challenging pressure on providers.

6)      Providers can see years of this ahead of them and are getting ahead of the curve if they can

7)      Social care budget reductions are starting to impact on the ability of trusts to discharge

8)      Demand still seems to be rising

9)      Hospitals with large PFIs have reduced scope to make savings as the payment for the building and its maintenance are protected and there may be constraints on other services bundled as part of the PFI.

10)   When people start to declare big savings programme there is a tendency for there to be a herd response as it is useful to get an announcement out under the cover of the angst caused by others.  This also encourages statements about the initial savings  to be over stated.

I was talking to the chief executive of a large university hospital trust on Friday who has an enormous cost improvement programme. His observation was that in many places people have got out of the habit of living with stringent cost improvement. He thought that many cost improvement programmes in the last decade have been delivered using additional income rather than through improved efficiency and this is my impression as well. This certainly seems to be suggested by the sluggish performance in productivity that the NHS has demonstrated - putting aside the enormous methodological difficulties of actually measuring this. His trust, by contrast, has had very low levels of growth and therefore has needed to deliver 6 to 7% of genuine efficiency improvement in the last few years, something most other trusts only just come to terms with. His view is that because this is new and this type of change is difficult some of the responses to this may not be very subtle.

One of the hidden treasures that Health Policy people ought to be more aware of is the Annual London School of Hygiene and Tropical Medicine Health Services Research Lecture.  This year it was given by Professor Arnold Epstein of Harvard, NEJM, etc. and was on pay for performance systems (P4P).

There has been a growth in public reporting and P4P in the USA.  The move to P4P seems to be a response to the fact that while public reporting moved some quality measures it seems to have had little impact on others – in particular screening.  Prof Epstein observed wryly that the standard USA response to any complex problem is to use markets and incentives and to do otherwise would be un-American.

There are now a large number of P4P plans in the USA.  In data he presented he suggested that there 53% of plans covering 81% of enrolees had some P4P element.  The impact on physician income is 5-10% (compare the QoF at 20-25%, he didn’t comment on this but Martin Roland has recently suggested this is way to high and Prof Epstein hinted at this).  Hospitals have 1-2% of their income covered by P4P which is about the same as their total margin – so while it is a small percentage it’s big enough to get their attention. CMS is planning a large extension of P4P over the next few years.

The structure of schemes generally inclues a mix of clinical, administrative, service experience/ satisfaction and efficiency (age adjusted cost) and gain sharing with hospitals.  The QoF is still the leader in terms of scale and scope – but its not clear that the effect of adding all these additional measures is very powerful and could be counter productive.

The schemes have been extensively evaluated.  The findings are mixed:

·         P4P does make a difference in a number of areas but it is hard to design it to achieve the right results and the effects may be modest.  For example, in the Pacific System the scheme rewarded people who are already doing well even though they had made little or no improvement but did nothing for those near the bottom of the pile who had improved a lot.  

·         Good news is that, like Martin Roland’s findings, there is no significant adverse socio-economic gradient

·         The effect seems to attenuate over time and controls catch up – it is however not clear why

·         There are problems with ceiling effects

·         There are surprisingly few adverse effects and not much evidence of a crowding out effect in which incentives in one area reduce effort elsewhere.  Unfortunately, there is no evidence of positive spill over effects either

Nevertheless, P4P will be expanding in the USA - as it will in England.  Hospitals will be the first area for expansion.  Widespread public reporting will expand in the USA.  There will also be much more focus on efficiency.  There will be fewer schemes that are incentive only – there will be more penalties for failing to meet goals rather than just positive incentives.  There will be more focus on new incentives with the aim of creating a wider range of levers – including capitation payments, bundled payments, ACOs, payments based on best / most appropriate care, gain sharing, readmission penalties,  complication penalties, etc.

Overall my impression is that 1)  P4P works quite well for a limited range of things at any one time, if its too complex providers may stop bothering; 2) it needs to be used with other mechanisms 3) it is hard to design 4) it needs constant updating 5) it needs to go with the grain of the evidence on effectiveness and 6) it is not a magic bullet and that a mix of different methods will work better than one big idea. 

On interesting question is what happens when measures are retired or incentives change.  There seems to be some evidence that if the change seemed to be a good idea then while there is some loss many gains can be retained if the physicians think its worth it. 

See also:

http://healthcare-economist.com/2011/03/10/did-masshealths-p4p-program-improve-outcomes/

Lindenauer http://www.nejm.org/doi/full/10.1056/NEJMsa064964

Rosenthal  http://www.nejm.org/doi/full/10.1056/NEJMp0804658

Tim Ensor on the UK http://www.nejm.org/doi/full/10.1056/NEJMsa055505#t=article and http://www.nejm.org/doi/full/10.1056/NEJMsa0800310

Epstein on the Pacific scheme http://jama.ama-assn.org/content/294/14/1788.abstract

There has been a lot of discussion of this issue and apparently the government has executed a significant U-turn in amending the bill to remove the power of Monitor to set a maximum price. But, I don't think that the debate about price competition in healthcare has been properly aired and I don’t think this is the end of the story.

The reason why there is so much anxiety about price competition is because of the classic problem of information asymmetry between the payer and the provider and more fundamentally because for a significant amount of healthcare it is difficult to measure the outcomes and quality of care.  This leads to the reasonably well founded anxiety that if providers are squeezed on price by competition, or over aggressive purchasers, they will skimp on quality if it is not very carefully measured.  This measurement is made more difficult by the fact that there is not necessarily an agreed definition of what the product is, over what timescale outcomes should be measured or whose assessment of quality should be taken into account in calculating value.

There is a strong view amongst a number of health economists that this is a genuine problem and there does appear to be a reasonable empirical basis for this.  This applies even in areas where measurement is possible but providers deliver multiple products including some where quality is less visible to purchasers.  This is because there is the opportunity to cross subsidise from the areas with opaque quality to support price reductions in those areas where it is more transparent. 

It's worth pointing out that the entire Lansley programme is premised on creating reliable measures of outcomes for which the individual parts of system can be held to account and which will be used as a basis for making choices by patients. I think he has massively underestimated the scale, complexity and difficulty of the task and the difficulty of attributing outcomes to individual providers that form part of a complex pathway.  If this project were to succeed the most significant objection to price competition would have been removed.   It is therefore slightly surprising to see that while the government has rejected competition on price it does not seem to have noticed what this implies for one of the key pillars of the reforms. 

Therefore the objection to price competition seems to be based on pragmatism rather than principle. Even so, there are still some hazards from poorly managed price competition:

There are risks associated with the loss of economies of scope and scale when components of highly interconnected services are lost as a result of services being individually procured on the basis of price.  

An objection that has not been heard very much, rather surprisingly, is that the underlying idea behind the reforms is that patients will choose which provider they want to use on the basis of quality.  This would mean that referrers would need to be blind to price differentials on offer by different providers. This is not possible because of the combined role of commissioner and provider that GPs occupy. GPs have rightly been concerned that patients may perceive that their clinical judgement or referral decisions could be skewed by personal economic considerations.

But abandoning price competition also has some issues:

·         Fixed prices can reduce the scope for innovation particular those which have a high cost for the initial treatment but a lower life cycle cost or services that need to have a relatively high initial price to enter the market.  Prices that are fixed at a low level may re-enforce current monopolies

·         It may be very allocatively inefficient, allow the generation of surplus profit and distort decision making

·         It can remove the potential for value for money deals to use spare capacity at marginal cost

A final point is that only 60% of acute hospital activity is covered by the tariff and very little else.  Until this is resolved prices will be set by negotiation, local tariffs, competitive tender and other processes that will allow price to be a major factor in the decision that is made.   Also  removing the power to set a maximum price still allows flexibility.

The prior questions that need to be asked are:

·         Do we have good definitions of the clinical ‘products’ that are being purchased?

·         Can we price these accurately?

·         Can we measure the value that they create, the outcomes or at very least some good intermediate markers for a quality outcome?

·         Do we understand any economies of scope and / or scale which requires this service to be part of a wider range of services?

·         Are we confident that we can spot any cross subsidy or the dilution of quality in other areas?

·         Do we have the skills and resources for a negotiation on price?

There is some way to go to make this work.

After observing reform in a number of different countries and various versions of reform in the UK at different times I am left feeling that there is severe shortage of humility amongst most health care reformers. 

The advocates for reform that I have dealt with over the last 15 years – whether from politically positoned think tanks, political parties, trades unions or some disturbing libertarian economists I met recently  – have all shared a burning certainty that they have the answer.  So when Richard Horton challenges those of us worried about the reforms to define an alternative vision he presents me with a problem.

1)      We don’t know what works.  We have some ideas about what might work in some circumstances but ideas that have worked in one setting are not guaranteed to work elsewhere.  We are better at identifying what won’t work.

2)      The evidence requires qualifications, caveats, etc.  This is boring and confusing compared with the simple certainty of the policy entrepreneur

3)      Describing a static system which is elegantly constructed is relatively easy and can appear attractive.  It’s what happens when it is switched on and has to interact with the outside world that matters.  This is where a range of unanticipated and often unpleasant effects take place.

4)      Our lack of conviction that we are right, and I hope some residual conscience means that we are not willing to subvert the evidence, selectively cite sources or just make false claims to support our case

We don’t acknowledge what we don’t know and there is a shortage of research.   Where there is research it is often incomplete, observational and tends to be hard to find – there are few journals reporting on non USA health policy.   While I treat these uncertainties as a reason to be cautious the advocates for particular policies tend to view it as space to play in. This is assisted by:

·         A formulation that says the previous system was poor, so the new proposals will be an improvement

·         Once a solution is identified it is put forward continuously but rarely developed, details are generally not filled in.  It seems that Andrew Lansley developed his ideas between 2006 and 2008 and very little development has taken place since

·         Treating contrary evidence and alternative views as opportunities to criticise rather than to learn. 

All this means that my formulation is much less exciting. 

1)      Given the uncertainty more experimentation linked with proper evaluation is required.  Interesting for me, boring for politicians & the media

2)      Some of these will fail and measures will be required to prevent harm to patients.  I think this types of failure is unavoidable, politicians and the public expect risk free innovation.

3)      There is enough evidence to set some parameters for these experiments and to define some basic characteristics of the system.  These include:

·         A clear definition of what success for the system looks like and a valid theory of change and quality improvement

·         The ability to create a variety of types of competition as well as integration depending on the type of services

·         A mix of different incentives to drive improvement and to focus attention on the development of high quality processes and systems and for providers to rethink many of their existing business models and delivery methods

·         A clear view of where there are limits to markets and appropriate mechanisms to correct for market failure: e.g. inequality issues, services that are natural monopolies

·         A high level of clinical involvement in the management of important pieces of the system

·         High quality leadership able to operate with a high level of autonomy

·         Transparency and the availability of information for patients, clinicians and the public

·         Mechanisms that ensure patients are involved in / take responsibility for their own health

·         The removal of barriers between different sectors

·         Enabling policies e.g. a more flexible  workforce  and more flexible training mechanisms

·         Policies and mechanisms that shape patient behaviour including increasing self care

·         Regulatory and payment systems that supports all  this

·         An intelligent payer function (I am not sure how important this is)

Lansley’s reforms could provide some of the environment for  this experimentation but the design of the system has some rigidities and rules that need some attention. Its still very based on the divisions between primary, secondary and social care.

 What is needed is a more fluid and flexible approach and much less certainty that the current answer is the right one. 

These are my personal musings on the underlying ideas, assumptions and beliefs that seem to be behind the reform programme or that need to be true for it to work effectively and how solid these are.  The thoughts here are about some of the assumptions about the use of market mechanisms. I previously looked at GP commissioning.

The unit of competition

There are two assumptions here that may be only partially true.

Firstly, that there is competition for the delivery of clearly defined products and that these products are relatively easy to define.  This may derive from the way that elective care has tended to be the model policy makers tend to use.  But even a straight forward hip replacement might actually be one of a very wide range of offerings with different mixes of implant, rehab and even different surgical techniques.  At present a number of services such as outpatients are not even really a product, they are a component in a wider process not least because in many cases, while the patient or the referrer knows there is a problem, there is significant uncertainty about what it is.  In a significant number of cases it may be that this uncertainty remains after the patient has been extensively investigated: for example, a significant number of endoscopies find no medically explainable cause.  

Second, there seems to be an assumption that existing providers are the correct unit on which competition should operate.  This reflects the standard model of the NHS made up of hospitals and other large providers, but there are reasons to question whether this is the most appropriate way of looking at a number of services.  Hospitals have a significant monopoly hold on many local markets, especially for long term conditions where patients make frequent use of the hospital.  Competition at the level of groups of GPs, specialists, nurses and other professionals offering to manage a fcpatient’s complete care would create a dynamism that is likely to be lacking from competition just at the hospital level.  

Conclusion  These are difficult problems they may not have an easy solution however, it does mean that allowing more primary-secondary care collaboration and creating mechanisms to allow hospitals to change their shape quickly will be important.

The impact of new entrants

The reforms assume that new entrants will have a positive impact on improving quality and efficiency and will be an important source of innovation and change.  There are good reasons to suppose that this is correct.  However, there is an underlying assumption that the new entrants will be genuinely new – many of them may be GP companies or other incumbents in JVs with the private sector and in some cases the motivation will be to exclude real business model or technical innovation.   Secondly, there is a problem about whether there is space for new entrants.  If prices can’t be pushed down and the market isn’t growing the only way that space can be created is if existing providers either exit or radically reduce their costs.  

Conclusion:  This assumption seems to reflect the way the reforms were developed during a period of growth.  The reduction in management cost creates a small amount of headroom and the efficiency target of 4% a bit more.  It’s not clear whether this, combined with a certain amount of chaos and dislocation makes the health market that attractive.   Standard contracts, detailed service specifications and inflexible pricing methods have also proved to be an effective way of preventing innovative models from emerging. 

Prices and tariffs

Health policy in England has put a great deal of store in using tariffs and pay for performance as a mechanism to drive different types of provider behaviour.  This assumption is carried forward into the new White Paper although in a more subtle and less self-contradictory way than some previous policy.  There are some features of how tariffs work and some important lessons that mean that the White Paper may still be over-optimistic about the uses of tariffs.  Firstly, there is a limit to the number of objectives it is possible to pursue.  Secondly, to be effective as incentives tariffs have to be continually adjusted and the DH is woefully under resourced to be able to do this.  Thirdly, tariffs have to send clear signals to providers about what they need to do and be sufficiently high powered to make it worth taking the trouble to respond, but, like targets, not so high powered that it leads to gaming or other distortions.  Fourthly, fee for item of service tariffs are very effective if the objective is to increase the quantity of what is produced – unfortunately there is quite a lot of activity where a reduction is required.  Bundled tariffs are the answer here but these are difficult to create.  Finally, it seems that designing and calibrating a pay for performance system is surprisingly difficult. 

Overall the principle that cleverly designed pricing systems are useful is correct, the devil is in the detail of the design and it is important to have a limited number of fairly clear objectives.

Market signals and planning

The White Paper envisages the NHS as a regulated market rather than a managed system which has market elements.  There are a number of features of this that are quite different from the current approach and which rely on certain conditions applying.  The most significant of these are about how local health services made up of different and potentially competing providers come together to create a coherent whole. 

There appears to be an assumption in the White Paper that system management and strategic planning are not very necessary except for the role of the economic regulator in relation to competition regulation and the designation of essential services.  The assumption seems to be that strategic change is more the result of the decisions of individual agents rather than of a deliberate planning.    There are a number of reasons to be cautious about the ability of market mechanisms to solve certain types of problem in healthcare.  With the correct regulatory framework, adequate information and intelligent purchaser markets can drive significant improvements in efficiency, responsiveness and quality.  However,  it is much less clear that markets are good at solving complex problems such as the configuration of services.  There are a number of reasons why this is the case. Many hospital services are made up of a number of different specialties:  for example, trauma, cancer, emergency medicine are composites of other specialties and clinical services rather than simple products.  In these cases it is probable that the market signals relating to the individual component services will determine who provides them and where they are located and these will swamp signals in the market for the composite products.  In other words, market mechanisms may lead to neurosurgery, vascular surgery and other specialist services being distributed across several hospitals while the relatively small market for trauma would be signalling that a single site is required.  Payers have to actively manage the market to get the desired result – it is unlikely to emerge. 

It is interesting to see in recent speeches that the SoS has started to talk about Health & Wellbeing Boards providing ‘strategic coherence’.  The introduction of the tests for reconfiguration which tasks with GPs with signing off the case for change also seems to suggest some acknowledgement that somewhere strategic decisions will need to be taken.  It is quite likely that increasingly providers may be driving this as they respond to workforce, specialisation and competitive pressure.

Agile providers

There are similar problems with wider questions of hospital configuration, complicated by the interdependence between many services.  The market might signal that a provider should exit an area of business, for example elective orthopaedics, but this may be costly or impossible if it is still required to provide emergency services.  Markets assume that suppliers can make quick adjustments in their prices, costs or what they supply.  This assumption does not translate into the NHS as it is currently constructed and the new world does not appear to have machinery that deals with this.  While income may fall very quickly if patients choose other hospitals or GP Consortia shift work, costs fall very much more slowly.  Businesses in the commercial sector can respond to this in several ways 1) they can win business from other customers, 2) they can increase the income from their existing customers 3) they can reduce their costs.  1) and 2) are much less available in a system with fixed prices and with a fixed pot of money in the market.  3) is difficult because a very large proportion of the price of hospital care represent overhead costs that cannot be shed easily.   

Integration and fragmentation

This is a whole topic on its own and it is interesting that at the UCL Partners / Monitor Conference the Secretary of State thought it worth dedicating a significant amount of time explaining how his reforms would drive integration and that AWP did not require fragmentation.  This reflects a problem with the White Paper which while it asserted a desire for integration it was very weak in explaining how this would come about.  The focus on choice and competition seemed to confirm a prejudicial view of integrated delivery which is sometimes seen as an excuse to exclude competition.  It still seems that the policy framework to promote competition for the market, for bundled care or years of care is poorly developed.

 Equity

The White Paper seems to make an interesting assumption about equity.  While the title of the White Paper is Equity and Excellence, very little reference is made to equity: the implication is that the NHS is already equitable and the problem is one of excellence.  This is a very questionable assumption, the distribution of GPs, services and outcomes are not equitable.  Markets are blind to questions of equity.

Final thought

 Markets drive innovation and change.  The most powerful mechanism for this is the ability of new entrants to offer new products, different business models and other ways of disrupting the system.  To do this they need headroom which can come from the market expanding or existing providers shrinking or exiting.  The environment in which the reforms are being introduced make this more difficult.  It seems that the use of market mechanisms to disrupt the NHS is a strategy for a two term government in a period of growth.  Both these assumptions looked reasonable when the reforms were designed. 

These are my personal musings on the underlying ideas, assumptions and beliefs that seem to be behind the reform and how solid these are.  The thoughts here are about some of the assumptions about GP consortia. I will post a further set about those related to markets and accountability.  They are not the NHS Confed’s view – it’s just me thinking.

GP commissioning will be different

Given the scale of the management cost reduction and some of the clues in the White Paper and its associated rhetoric it appears that GP commissioning will be different from that carried out by PCTs.  It's not precisely explained and presumably it will be for the Pathfinders to work this out.  In the HSJ Dame Barbara Hakin is reported as saying ‘she was concerned GPs were expecting the NHS to “marshal commissioning support in a way that faintly looks like PCTs currently”.  If this happened, she said commissioning would not be turned into “the professional and highly attuned system that it needs to be” to tackle rising acute demand.’  So although she confirms that this time commissioning will be different – she doesn’t explain either why or how. 

In a number of places the Secretary of State – for example in his  letter to GPs on 24 September 2010 - refers to commissioning as making decisions about individual patients but he also makes reference to shaping services to meet the needs of patients. It is not at all clear how far this shaping is intended to be as a result of planning and deliberate pathway redesign as opposed to providers responding to signals sent by changes in GP behaviour and patient choice.

Conclusion: The underlying assumption seems to be that commissioning is more about the collective and individual decisions of GPs and creating an framework in which patients chose and much less about planning.  The issue may be that there are quite a few difficult questions where a planning approach may be required.

GPs as an agent for the patient

Markets require customers and in the way this market has been designed this will usually be the GP acting as an agent of the patient. 

The foreword to the White Paper states its first principle as :  ‘First, patients will be at the heart of everything we do. ….. Patients will be in charge of making decisions about their care’.   But the third principle is  ‘We will empower health professionals. Doctors and nurses must to be able to use their professional judgement about what is right for patients. We will support this by giving front­line staff more control’.   Both these aspirations are laudable but they are not always going to be consistent.  Nowhere in the White Paper is there any acknowledgement of the principal-agent problem - the idea that the interests of an agent may not always be aligned with those of the person that they are serving.   GPs try to be advocates for their patients and there is a large area of shared interest but there are legitimate reasons why the interests of patients will diverge from those of their practitioner.  I am sure that Angela Coulter would point out that there is also a tendency to regard empowered patients as resource hungry, demanding and troublesome.  She would say that this is wrong but there is a long standing view in the NHS that  empowering patients is a zero sum game where professionals lose power.  Stephen Thornton reports finding these views even amongst the otherwise inspiring GP leaders he worked with.  He reports them reluctant to cede power and say ‘There was much talk of ‘persuading patients’, ‘explaining to patients’, ‘managing patients’ expectations’, but little about ‘listening to patients’ or ‘empowering patients’. Those few who did speak of sharing of decision making with their patients and of the importance of supported self management were clearly regarded as rather radical by their colleagues.’

The GP’s role as agent may become more difficult if there is a significant amount of price competition between providers as patient choices would have different cost implications which might influence referral behaviour.  Just the implication that it might could be damaging to patients’ trust in GPs.  

Conclusion: This assumption is only partially correct and will be a challenge for GPs.  It becomes more of a problem when the next 2 assumptions are examined.  

GPs view of population health

There seems to be an implicit assumption by the Secretary of State that because GP practices see a high proportion of their registered patients in a year that this provides a population health view point.  Little mention is made of public health input into consortia to inform commissioning (as opposed to health improvement).  However, the sum of all these experiences, while powerful and important, does not necessarily equate to a complete population view.  This is not just because of the relatively trivial problem that there will be some categories of patients that are under represented (young men) or missing (the homeless, prisoners, etc).  There are two more fundamental issues.  Firstly, there will be weak signals that are hard to see amongst all the individual consultations and which require aggregation to spot (as seems to have been the case in Mid Staffs).  Secondly, and rather more arcane, it is very hard, if not impossible, to derive the preferences of a population from an understanding of individual views about complex choices from multiple options (Arrow’s impossibility problem). 

Conclusion: A worry about the assumptions here but the good news is that many GPs seem to recognise the need for epidemiological and other public health skills to provide a population health perspective.    

Rationing is not a problem

In addition to the agency problem the taxpayer also brings a third set of interests that have to be balanced – ensuring that scarce resources are efficiently and fairly allocated.  In fact, one of the most significant  and positive changes that the White Paper is aiming to achieve is to get GPs to take responsibility for the financial consequences of their referral decisions.   But there seems to be assumption in the White Paper that GPs will not have to take difficult decisions about groups of patients, as opposed to the difficult decisions they already take about individuals: these include the management of treatment thresholds and difficult prioritisation decisions that will require a population perspective.

Conclusion:  This clearly is not a safe assumption and GP consortia will need clear decision approaches that are robust in the face of judicial review.   

Primary care

The White paper assumes that primary care is not an issue. The impact assessment gives a list of poor outcomes as a reason for reform.  Many of these are directly related to the quality of primary care or at least the lack of integration between specialists and primary care.  The opportunity to use the best primary care providers to improve the quality of the rest would seem to be key success and one of the potential advantages of the reforms and one that some GP leaders are keen to embrace.  Perhaps the government is being smart in not alerting GPs that they may find themselves accountable to colleagues who are less easily fooled than lay PCT managers and who might even be personally at risk if they don’t deal with poor performance brought to their attention.  Even more reason not to mention the other policy tool for improving primary care – any willing provider competition. 

Conclusion: It seems obvious that to make commissioning decisions stick and to have any effect on demand GPs leaders will have to deal with the clinical and organisational practice of their colleagues.  If this is a hidden agenda it is probably a more powerful and benign one than some of those attributed to the reforms.  However, while using GPs to improve primary care may be effective when the problem is one of clinical quality it may be harder where the issue is over provision or relatively expensive services. 

Some others?

Demands on GP’s time: not very great, only a minority will be leading.  Yes but leaders need people to follow and this needs to be more than grudging or passive acceptance.

Management & planning:  not much of this required. This seems more to do with an anti-management stance than reality.

Governance: Leadership by election.  This is fine if you don’t need to make too many tough decisions.  Its generally been abandoned within hospitals.

Allocations and fair shares.  Changing to personal based allocations and a new formula and the increased number of commissioning organisations will increase the overall distance of areas from their fair share capitation target.  This will cause some consternation.  I wonder if it will also start to focus attention on the differences in spend between practices, especially when it is time to distribute the GP consortium quality premium payment. 

Conclusion

There are a number of assumptions about the way this new world will operate that are at least questionable.  There are others where the answer is contingent on the systems, processes and governance that GP consortia select.  Given the uncertainty a systematic programme of properly designed action research is required – not just a programme to share learning.

The Consultation closes on 19 January.  As is the case with my previous post this represents my personal view and not those of my employers or members

Having discovered that the proposals for value-based pricing drugs left a lot to be desired I thought it was worth revisiting the consultation on the cancer drugs fund.  We have all been very busy dealing with the White Paper but this policy is worth attention. There have been a number of indicators that there may be a problem.

First alarm bell the BBC quotes the authoritative voice of Alan Maynard:

“You are undermining NICE and all the work it does to target resources where we get the biggest health gain, the bang for the buck. And this drives a horse and cart through logical and systematic rationing of health care resources," he said.  Prof Maynard said he was worried that the cancer fund will see cancer being given preference over other conditions because a combination of drugs companies, patients and campaigners has put the government under pressure.  "More money for cancer, means less money for motor neurone disease. More money for cancer means less to treat patients who have got Alzheimer's."  He said many of the cancer drugs were portrayed as wonder drugs when they only extend a patient's life by three to four months.

Second alarm: The Lancet whose headline was unambiguous: ‘New £50 million cancer fund already intellectually bankrupt’:

“A very modern triumph of political expediency over rationality…. what this fund represents is not the victory for patient groups that some believe. Rather, it is the product of political opportunism and intellectual incoherence…let us be clear: it not only undermines NICE, it undermines the entire concept of a rational and evidence-based approach to the allocation of finite health-care resources..”

Third alarm bell, James Raftery points out that the consultation’s impact assessment reveals a lack of support for some of the key assumptions underpinning the policy. 

Does the cancer drugs fund undermine cost effectiveness analysis? Yes,  insofar as it privileges cancer without good reason. No, if good reasons exist for treating cancer drugs differently. The impact statement reviewed if the public valued cancer treatments more highly. It concluded that :  “no evidence had been found for prioritising cancer above other severe conditions or to prioritising drug treatments above any other interventions for cancer.”

The biggest criticism of the cancer drugs fund is that it takes money away from patients with other diseases. The impact assessment estimates this opportunity cost at 4,000 QALYs. Although based on optimistic assumptions about the relation between cost and QALYs, it provides a first estimate. The impact statement acknowledges that this could only be justified if the public attached a higher value to these cancer QALYs compared to other diseases. As noted above it found no evidence for this

Some important questions

Why £200m?  The figure was a pre-election pledge and seems to be based on the costs avoided by the NHS of not having to pick up a 0.5% increase in employers national insurance.  The Rare Cancers Forum did estimate the shortfall and this is used to provide a somewhat tortuous ex-post facto description of how this figure relates to an estimate of what the NHS would need to spend to achieve average European usage levels in the consultation document. There is no particular reason to presume that the average is the right amount.  The BBC investigation suggested that the amount of money put aside was too small (by a factor of 3).

Is this a good investment?  The impact assessment says:

Diversion of £200m p.a. to fund will displace 8,000 QALYs, worth £480m from patients elsewhere in NHS.  Costs of fund administration committees, £4m p.a. (valued at £10m p.a.)…. Cancer patients in the NHS will receive 4,000 QALYs p.a., with an unweighted value of £240m. Shareholders in pharmaceutical industry receive £77m additional profits, of which £4m to UK.

This does not look like a good investment.  Obviously from an individual patient perspective the answer may be different, it certainly is if you are a shareholder.  But the question here is about the global allocation of resources and the case could be made if there is a good reason to give particular weight to care near the end of life - even then this would have to give particular value to cancer as opposed to other conditions, such as heart failure of COPD.  The impact assessment is helpful again:

While there may be support in principle for greater weighting of QALYs provided to patients with severe conditions, there is currently no robust evidence in the literature to support a particular magnitude of weighting. It should also be noted that no evidence has been found for prioritising cancer above other severe conditions, or for prioritising drug treatments above any other interventions for cancer….

Although the consultation document has a section called ‘Why Cancer?’ it does not actually provide an answer to this question at all.

Will it impact on the UK’s poor cancer survival record?  As the consultation document seems to imply(para 2.14) the answer unfortunately is almost certainly no as many of the drugs involved have limited effectiveness and certainly would produce less impressive results than investment in early detection, surgery, radiotherapy and perhaps I guess even primary prevention. This seems to be close to what Sir Mike Richards told the Independent.  In fact, rather surprisingly for those of us who have believed that that UK is universally a laggard in the adoption of new medicines Sir Mike Richards’ review suggests a more complex picture.  We are low in some areas, but it is not clear what impact this has on outcomes or what the right level should be.  Low drug use might be compensated for by higher investment in other parts of a patient pathway.

Are the other assumptions robust?

The impact assessment says that it is assumed that

·         The mechanism for disbursement of the fund will prevent over-spending.

·         Pharmaceutical companies will not increase the prices of their products 

·         After 3 years arrangements will be put in place to set the prices of medicines - including those affected by the cancer fund

These do not look very robust to me.  The assumption that value based pricing will take over this role seems to be contradicted by the consultation which suggests that it will only apply to new drugs brought to market from 2014.

Regional Panels

Basing a policy for the next three years on organisations known to only have two years life is possibly a first.   There are some interesting reasons put forward.  Firstly, cancer is organised at a supra PCT level.  Secondly, the Impact Assessment notes that “The funding decisions of regional funds would be much less closely scrutinised by international purchasers – so firms may be more willing to offer prices that maximise benefits to patients, while still generating profits. Lower drug prices mean that more cancer patients can benefit from the fund.”  This assumes that the regional level has good procurement and negotiation skills.  I am not sure whether this is the case.

There is a proposal for regional panels to take the lead in assessing the evidence for the effectiveness of particular classes of drug. It is not at all clear why this is necessary. The conclusion of a number of commentators that this policy does represent a significant attack on NICE does seem to be given some weight by this.

Conclusions

Although the points about undermining  NICE and the rational basis for resource allocation are a concern and are important there is still a problem.  NICE does not have a remit for off label applications of drugs and its process does mean that there is a delay between drugs coming to market and undertaking an appraisal.  There is a limit to how far this can be speeded up.  There may be some drugs that just can’t get into the NICE process because they serve such a small patient subgroup.  This does mean that there are a number of drugs that clinicians may want to use but which are not incorporated into the tariff payment.  But this is not just a problem in cancer, a number of other specialties have the same issue.  It might be rational to put aside some funds to support some spending in this area, but there is no rationale for just choosing cancer.   If we are going to do this there is no point pretending that it is anything other than a political choice and that it won’t skew resource distribution.  Such choices can be made by elected politicians, that’s part of their role and there is an argument that there is sufficient uncertainty, different perspectives and value judgements to be made that this may be an area where judgement has to be used and there is a limit to economics.  

I think that oncologists and other specialists should take responsibility for their use of very high cost drugs and be thinking about the trade-offs involved between the different treatment options available to them.  The system of exception panels allows clinicians to pass responsibility to a third party which can be blamed and, for those that want to, it allows a difficult conversation to be put off. And yet people with the expertise, the understanding of the science, the risks and benefits and knowledge of the patients are the specialists. Following the same principle applied to GP commissioning, oncologists should manage the budget for these medicines and take responsibility for overspends. 

The consultation is entirely about how to make the cancer drugs fund work effectively. A broader consultation on mechanisms for high cost specialist drugs of marginal effectiveness might be more appropriate.  This should include how to determine what should be included and what level of sub-optimal investment we are prepared to make to allow for rescue and other compassionate interventions. 

Please let me know what you think.

The consultation and impact assessment

http://www.dh.gov.uk/en/Consultations/Liveconsultations/DH_120834

Some additional comment

http://www.healthpolicyinsight.com/?q=node/526