Perverse Incentives:
An NHS notebook

Maureen Mackintosh and Pam Smith
The NHS reforms have brought some expensive forms of competition into the cash-limited health service, and have created some curious incentives, not least for increasing costs, and for 'treatment not
care'. Maureen Mackintosh and Pam Smith have
been keeping a notebook.

From Soundings issue 4 Autumn 1996

Incentives and the NHS

Quotations and events
'Until very recently the public services ran on perverse incentives and without any bottom line.... there was no reward for good performance and no sanctions for poor performance. The result was that there was simply no motivation for people to do better. Markets offer real incentives; they reward success and they penalise failure ...In the NHS,

Eric Caines, now Professor of Health Services Management at Nottingham University, was personnel director of the NHS 1990-April 1993, through the early process of reform, so he should know what they thought they were doing. His summary of the central idea of the NHS reforms is quite clear: to create a market in order to give people the incentives to 'do better'. the managed market was created in an attempt once again to bring pressure to bear for better use of resources.'
Eric Caines, The Health Summary 1995, p8

We have watched over the last few years as the language of incentives, sanctions and rewards has become embedded in political thinking on public policy. The discourse of incentives has become associated with 'reality', with an absence of naivety: watch how the above quotation moves quickly to associate motivation with incentives, and 'real' incentives with markets and material rewards. The previous or non-market NHS had incentives, but they were not 'real' and were also 'perverse', a word redolent of obstinacy, improper behaviour, facing in the wrong direction.
This is a notebook of short tales about 'perverse incentives'. It is a powerful phrase, that. It has a precise technical meaning: it means a financial incentive to act in a way which runs against the objectives of the institution where you work - or indeed against what you know is right. And it has - as just suggested -other layers of meaning, which denigrate by association worlds where the incentive structures are not properly 'aligned', that is, where it is not clear who has the (financial) interest in doing what. In particular, it is used, as above, to tar the 'old' NHS with inefficiency and ineptitude, and to imply, as Caines does, that markets are sorting the problems out.
But the situation is not as simple as that. It is not that the 'old' NHS had no inefficiencies: who could think that? (Yet somehow, these days, one cannot talk about the new NHS of'managed markets' unless one first protests that one is not idealising the old system.) Instead, the notebook illustrates the way that the new, reformed NHS, where governmental 'purchasers' and GPs buy health care for all of us, is itself full of'perverse incentives'. Though it has - as we will also show - desirable new incentives too. But, perhaps most of all, we want to suggest that financial incentives have come to matter more than before in what people do. In terms of Eric Caines' meaning of ‘incentives' we are now more 'incentive-driven'. That may make incentives more real. It is not clear it makes health care more 'efficient': indeed the meaning of efficiency is one of the problems we consider here.

Raising or lowering costs?

Quotations and events
Ms Primarolo: 'To ask the Secretary of State for Health what estimates her Department has made of the annual cost of negotiating and managing NHS contracts.'
Dr Mawhinney: 'None'.
Written parliamentary answer 15.7.95 quoted in Health Care UK 1994/5 p6

It is one of the most striking facts of the NHS reforms that a government which, in most fields, is only too anxious to cost activity, and to evaluate it in financial terms, has steadfastly refused to collect information on the cost of 'marketising the health service.

There are two different - but in principle compatible - things that we might mean by Eric Caines' 'better use of resources': doing things more cheaply; and doing different and mare desirable things. It is fair to say that the NHS reforms were supposed to do both. The reforms separated 'purchasers' of health care (local health authorities and GPs) who ‘held the NHS budgets, from 'providers' (hospitals, ambulance services, district nursing services, GPs, all organised as autonomous contractors to the service) who bid to them for contracts to provide us all with health care. (Yes, we know GPs are in both those lists: we come back to that below!)
This 'market' framework was supposed to drive down over-all costs, because providers - it was hoped - would compete with each other for contracts by reducing their prices. At the same time, it was supposed to ensure services were more responsive to patients' needs, because it was thought that the purchasers would concentrate on identifying and buying care for those needs - undistracted by the problems of running services. In the jargon, services would be 'purchaser-led' not 'provider-led': run for us, not for the convenience of hospitals.
So let us start with the central theme of the reform: costs. Have the reforms reduced them? Do they provide incentives to reduce them? To tackle those seemingly straightforward questions, we need to define what we are talking about. The key argument is about unit costs, that is, how much it costs to treat individual patients with particular ailments. If you treat more people for the same money (increased 'throughput' as it is unsuitably called) unit costs have come down. Unfortunately, the statistics are a minefield.

Quotations and events
'New figures from the government's statistical service show that in the past year alone, NHS hospitals treated an extra 455,000 patients. That is a 4.7 per cent increase.'
Virginia Bottomley, then Secretary of State for Health, Hansard, 25.11.94

These are not people! The numbers refer to 'finished consultant episodes' and completed day cases. Each time you see a consultant, that is an 'episode'. NHS trust hospitals are paid by the 'episode'. Measured episodes have certainly risen, but we do not know how much of that is more patients seen, and how much is counting one person several times.

Hospitals do respond to purchasers' pressures to raise 'throughput' by counting 'episodes' more carefully. That increases their money in one year. It doesn't raise the total sum available to the health service, however, it just redistributes it at higher administrative cost. And, as Julia Drown went on to point out, it gets harder over time to find ways to show an upward trend.

Quotations and events
'Hospitals are getting better at ident¬ifying activity. It's really a specious justification for demanding more money.'
Health authority manager, 1996 (interview with authors)

A trust hospital finance director realised she could boost out-patient numbers by ten per cent by adding in attendances at hearing aid clinics. 'Our performance appeared to have improved dramatically and the purchasers were delighted with us. Nothing had changed of course.1
Julia Drown, Guardian, 24.4.96

Purchasers look for falling prices each year, when trusts renegotiate their contracts. This is part of so-called 'efficiency savings'. If trusts can find a trick to increase apparent throughput, they have reduced their price per unit without sacking staff. We have therefore given our health 'providers', such as hospitals, a strong incentive to use their resources to count their activity minutely, and to push people through the system as fast as possible.
'Throughput' has certainly risen since the reforms - and was rising before them. The average length of stay in hospital fell from 11 days in 1989 to 8 in 19941, but most of that drop came from increases in day cases, where people come in the morning and leave in the evening.2 We do not know if this apparent rise in 'efficiency' is all gain for patients. If people leave too soon, they may be more likely to be readmitted. But we do not know if readmissions are rising: the 'government statistical service', as just noted, doesn't track people.
Against this pressure to (appear to) reduce costs per case, we have to set the costs of managing the reformed system.

Quotations and events
'The amount of paperwork has increased dramatically over the past year, which takes time away from client contact'
A health visitor

'The amount of paperwork has become ridiculous, some managers think it should take precedence over patient care.
A district nurse3

These (representative) quotations, from a study done over 1991-1994, reflect the increase in paperwork experienced by nursing staff in both the hospital and community sectors. The increase seems to be particularly felt by those working in community care.

Ask people in the NHS what is the worst aspect of the reforms, and many will say either 'paperwork' or 'bureaucracy'. What we have done in effect is introduce into the health service a billing system which it previously did not need. Have you been asked your postcode recently by a hospital? Have you been told apologetically that you need a GP letter now for each visit for recurrent treatment? How much does all this additional paperwork cost? And where is the money coming from?
As Dawn Primarolo MP confirmed, the government does not (want to) know how much it costs. Numbers of NHS general and senior managers went up from 4639 in 1989 to 17,900 in 1992, and continued to rise. The Audit Commission - a government body which audits the NHS - has added up higher level management costs, and come up with a figure of about £900m or 4 per cent of total NHS expenditure.4 Adding in junior finance and clerical staff brought it up to £2.3bn. or 10.5 per cent of total current spending - and this doesn't include the increase in nurses' paperwork. Everyone agrees administration and management costs have risen fast with the reforms; no one knows how fast. A Labour opposition health spokesman offered a guesstimate of £ the administrative costs of the reforms, which is 3.8 per cent of NHS 1994 expenditure.5
The government fluctuates between claiming the cost increases are justified by previous under-management, and declaring wars on bureaucracy (including 'reclassifying' people designated managers back into professional categories).6 Trust hospital managers agree the reforms are expensive: one estimated the additional costs of finance staff and auditing alone for one hospital at £150,000 a year, plus another £40,000 for the trust board costs.7 But the single most expensive reform - in administrative costs per patient treated - is probably GP fundholding.

GPs: whose side are they on?
Ours, we hope! We all rely on general practitioners. But 'GP fundholding' has created some odd incentives for GPs. What 'fundholding' means is that GPs hold a budget for buying care for their patients from hospitals, district nursing services, even in some cases ('total purchasing') from ambulance services and Accident and Emergency services. As pointed out above, this means that GPs are both 'purchasers' and 'providers'. Fundholders now care for around 60 per cent of us.
GPs have always been independent self-employed contractors to the NHS, so had to do a lot of paperwork. Fundholding, however, adds an expensive fund manager. It also means that hospitals have to negotiate individual contracts for each type of service with each GP, and keep track of and charge for each GP's patients. This is far more expensive to administer than a 'block' contract with a health authority for a particular treatment for a large number of patients in one area. One estimate of these 'transactions costs' of fundholding is about £81,600 per practice, but as the authors say gloomily, there are no estimates available to allow comparison with health authority transactions costs.8 The Audit Commission calculates that £232m has been provided so far for staff, computers and equipment for managing fundholding, as compared to £206m in 'efficiency savings' which can be retained by the GPs, and may be used for the benefit of patients.9
There is no satisfactory evidence on benefits to patients from fundholders' 'savings'. The Audit Commission found 60% of the savings spent so far had gone on premises (which GPs usually own) and equipment. Some have spent them on more patient care, including better levels of district nursing care, or extra clinics. There are some good incentives for GPs in these reforms.

Quotations and events
'At least the reforms mean that you haven't got the constant haemorrhage of funds into hospitals. And we now have more influence over acute services. Previously, hospitals were fairly impervious to local needs.' 'The key benefit of fundholding is that for the first time we are being asked for our views on services.'
A GP (not a fundholders) in a deprived area, author interview
Fundholding GP, Audit Commission 1996, pl9 (see note 8)

Both fundholding and nonfundholding GPs report that they can now influence the services offered by local hospitals in ways which are beneficial to their patients. Fundholders contract with hospitals, and can threaten to go elsewhere. Many non-fundholding GPs in inner city areas collaborate with their health authority in local commissioning groups, and have a seen real change in the responsiveness of consultants to their patients' needs

The greater influence of GPs over hospital consultants appears to be the best new incentive to emerge from the reforms. However, a combination of the new GP contract and fundholding have also produced, or increased, perverse incentives for GPs. Both systems involve capitation: payments per patient. The weight of capitation in GP incomes has risen, and fundholding budgets are set on a capitation basis.

Quotations and events
'Capitation fees are supposed to vary according to social circumstances, based on Jarman indicators of deprivation. But this is not enough to compensate for the much greater use of services in poor areas. My consultation time per patient is twice that of a GP in a richer area, and I have less time to do other things which bring in income. The census also greatly underestimates deprivation in some wards.'10
GP in a poor inner city area, author interview

If you have a fixed budget per patient, and commitments to pay for, or to provide, a given range of types of care, then the more you can exclude the expensive and demanding from your lists - the elderly, the severely mentally ill - then the easier your professional life, the better the service to the rest of your patients, and the greater your potential financial savings. Exclusion of this sort is called 'cream-skimming or 'biased selection.

There is no agreement about the extent of cream-skimming which is actually going on. Anecdotes abound about GPs refusing to accept patients onto lists, or rejecting unco-operative patients (those for example who refuse vaccinations, and hence reduce the income of GPs). There is also evidence of ethical rejection by GPs of such behaviour. But the incentives for cream skimming do appear to have risen. 'Cream-skimming in the specific context of fundholding is both technically feasible and financially attractive ... Practices with a disproportionate share of very ill patients will find themselves with resources well below the level allocated to other practices, whose population has a more favourable health profile.'11 This is one route for greater inequality of treatment to be brought into the health service. And it reinforces the effects of the established bias in fundholding towards the better organised practices, which are disproportionately in the suburbs and shires.
It is also true that, despite GPs' close links to patients, not all their decisions on rationing care may be for the best. Doubts include the dangers of ‘empire building': fundholders can benefit patients by bringing services closer to home; however they may prefer to be paid to provide services themselves, even if they would be provided more cheaply and of better quality in hospitals. This is the main incentive problem arising from the GPs' position as both providers and purchasers. Another doubt is the balance between very serious and less
serious care.

Quotations and events
'Fundholders determine the Com¬munity Psychiatric Nurse caseload... In many practices therefore there is a move from on-going care for people with severe and persistent mental illness to counselling for minor mental health) problems.'
M. Muijen and T. Hadley, 'The incentives war' Health Services Journal, 9.3.95

GPs, individually, come across certain kinds of severe and expensive illness only occasionally. Where they hold the budgets for it, therefore, they find it expensive when they do meet it. Meanwhile, they may tend to divert resources to more common - though still important - demands.

The point of having large health authorities was that they could spread risks and ensure that rarer acute conditions were properly catered for. They didn't necessarily achieve it, but there were no evident incentives against. This remains true of health authority purchasing. However, when purchasing passes to much smaller units, such as GP fundholders, the risk-spreading ceases to work so effectively and GPs may find it hard, or be unwilling, to pay for very expensive cases.

Ceasing to care
Once you divide up budgets into small amounts, used by different agencies for specific things, you risk boundary disputes. The worst one going on in the health service at the moment is over responsibility for the continuing care of people needing nursing and other forms of care because they can no longer care for themselves, especially the elderly infirm.

Quotations and events
'A disabled patient who requires help with bathing is having to endure the indignity of having his top half washed by staff from Camden Council's social services department and his lower half by district nurses.'
Hampstead and Highgate Gazette, 9.2.96

'One of the most worrying things about the NHS changes is what they have done to the attitudes of normal, decent professional people'
Christine Hancock, General Secretary, Royal College of Nursing (commenting on the above story), Hampstead and Highgate Gazette, 9.2.96

This humiliation could happen because since 1994 the NHS provides 'medical' care, and the local authority social services provide 'social' care. Neither is legally allowed to cross the boundary, both are strapped for cash, and no one can agree on which side of the boundary falls many a (much needed) bath. These boundary disputes cause pain to staff as well as users.

The Act which brought in the NHS reforms also provided for a huge change in the provision of residential and 'social' care for the elderly. It gave local authorities a fixed budget to pay for this care for needy local residents, whereas previously residential care had been paid for by the social security budget on a means tested basis, and there had been relatively little - but cheap or free - care offered at home by 'home helps'.
The background to the disputes over boundaries and responsibilities long predates the reforms: it lies in the rise in the amount of care needed by elderly people, as people live longer and can find less unpaid support. But the NHS has been withdrawing from 'continuing care' of the severely infirm elderly, and the social services have accused the NHS of ‘cost shunting' by leaving the problems with them. The incentives here are certainly perverse: to waste time trying to push the costs into someone else's budget. But Christine Hancock's comment nails an even more serious and widespread worry: that in a context where people's primary motivation is about caring for patients, this kind of dispute induces cynicism and inappropriate behaviour. The net effect is deeply demoralising.

Quotations and events
'Leeds Health Authority, in common with many other authorities, no longer has any long stay medical beds or contracted arrangements for such beds in private nursing homes.'
Health Service Journal, 10.3.94 (our emphasis)
Perverse incentives

This fact became news because the health service commissioner William Reid had just ruled that Leeds Health Authority had acted wrongly in ceasing to treat a stroke victim who could neither eat, move or talk, and had ordered the authority to pay the man's L33O a week nursing home costs.
Since the reforms NHS hospital trusts have been closing 'long stay' beds, and many now have none left. As a result of the case just quoted, the Department of Health issued 'Guidance' which, in the words of one commentator, formalises trusts' practice into policy.12 The closure of long stay beds is not new, nor is the rise of private nursing homes. The number of beds in private and voluntary sector nursing homes, clinics and hospitals rose from 18,200 in 1983 to 145,500 in 1994,13 while total NHS beds in England fell from 343,000 in 1983 to 220,000 in 1993-4 with half the decline happening before 1990.14 It is hard to find figures on the drop in numbers of continuing care beds for the elderly, but Stephen Dorrell, the Health Sectary, stated that they had fallen from 55,000 in 1979 to 37,000 in 1996, at a time or rising demand.15 Some observers think the decline has been sharper.
The reforms have not created the incentive to reduce continuing care of the elderly, but the financial pressures on trusts have reinforced it, the reforms have given the trusts powers to do it, and the community care aspect of the reforms have provided an opportunity. A current management catch phrase is that the NHS is about 'treatment not care'. 'Care' is increasingly means-tested, and paid for by families and patients.
If we consider therefore these two trends, the retreat from long term care of the infirm elderly and the rising management costs of the system, then in a clearly definable sense one has helped to finance the other. One cannot balance the costs and 'savings' directly one against the other: what has happened is that total real NHS spending has risen (slowly), so (less slowly) has demand for treatment and care, and so (sharply) have management costs. The retreat from 'care' has made some of the financial and political room for more administrative costs.

Perverse efficiencies

Quotations and events
'OxDONS syndrome ... this depart¬ment, which in 1989-90 could perhaps have been considered the ideal department - lean, efficient, good, even excellent, with no waiting list and superb staff morale ... had become by late 1994 unsafe, hyper-efficient, with exhausted and demoralised staff.'
C.B.T Adams discussing the crisis in Oxford's department of neurosurgery, quoted in British Medical Journal, 9.12.95

The previous head of neurosurgery, who resigned over the crisis, attributes 'OxDONS syndrome' to pressure to base prices on costs, which drove the trust to use private patient income to subsidise price per NHS patient rather than to expand facilities. A reply by the NHS executive director of finance attributes the problem entirely to bad management of private profits by the trust.

At issue here is not the rights and wrongs of a particular dispute, but the widespread perception that hospitals too find some perverse incentives in NHS accounting rules. The problem exercising trust finance directors is that - like GPs - if a hospital starts out efficient, it is still expected to make the same rate of efficiency savings each year. There is therefore an incentive to 'pad' costs. Otherwise low costs can lead to 'financial meltdown' in the words of one surgeon.

Quotations and events
'The surgeons claimed there was a perverse incentive to slip into financial crisis in order to secure extra resources.'
Health Services Journal, 23.5.96, reporting a British Medical Association (BMA) press conference.

'St James hospital trust, Leeds has agreed to operate a L5m deficit beyond the end of this financial year, despite NHS financial rules that trusts should break even. United Leeds Teaching Hospitals trust is also expecting a multi-million pound shortfall.'
Health Services Journal, 9.5.96

There is a double bind implicit in trying to 'manage markets' as a way of running a health service. The NHS rules involve setting prices based on costs to try to prevent trusts with local or regional monopolies from exercising power to push up prices. Associated with: a requirement to break even after paying charges on the capital passed to trusts when they were established; the relentless demand for specified unit cost/price reductions in each round; and (slowly) rising real wages in a labour-intensive service, this is driving hospitals into the red.
In principle this problem impacts first on the most efficient - with less scope for cuts, they start to get 'hyperefficient' - trying to cope with dangerously high 'throughput' - and quality drops. If they then see other hospitals being allowed to work with financial deficits, the incentive to pursue further efficiencies weakens still more. The incentive structures over the medium term reduce the search for efficiency they were supposed to promote.

Incentives and us
The problem with emphasising incentives is that we may become driven by them, and behave in ways we would not choose. The more people and institutions are paid according to 'performance indicators' the more they focus on getting those right. For example, the patients' charter focuses on waiting times, the government has emphasised waiting list targets, so the numbers waiting over two years for elective operations have been almost eliminated. Result: doctors charge that urgent cases are delayed to meet 'targets', and trusts are repeatedly accused of delaying putting patients onto waiting lists. So now data are being collected on waiting times for first consultation - we are chasing yet more data.
Information is useful - and it is expensive. Incentives are useful - and they are distorting. All health systems embody complex incentives - just like the rest of life. But the proponents of the NHS reforms cannot have it both ways. If incentives matter, then we have to get them close to right. We can't decide they don't matter after all just when they become problematic. The 'cream skimming' argument illustrates the point.
'The scheme's [fundholding] reg¬ulations offer little incentive to do it, even should a practice GPs and fund manager conspire to act against a fundamental moral principle of general practice.'
The Audit Commission on 'cream-skimming' by GPs in What the Doctor Ordered, op. cit

'Readers inclined to reject our analysis on the grounds that the relevant lessons from the United States are not applicable to Britain .
should be reminded that the ethics of medical care provision are partly shaped fry the financial incentives in operation.'
Matsagannis and Glennerster (op. cit.) anticipating the tone of this objection.

Medical ethics are a powerful force, but one cannot appeal to them too freely while simultaneously urging attention to financial incentives. One is shaped by the other, and the greatest danger of perverse incentives is that they promote cynicism. The use of the word 'conspire' in the above quotation has an over-trie-top feel: it suggests doubts in the writer. We don't need to conspire actively: a process of giving in, in perhaps small but mutually reinforcing ways, is enough to embed perverse incentives in behaviour.
There are no perfect cheap health systems that are wholly free of perverse incentives. The reforms have hugely increased the administrative costs of the NHS within a budget with a tight lid on it: the money has had to come from somewhere. By squeezing care in favour of administration , we appear to be rapidly worsening the financial crisis of a cash-strapped system. What we have done is institute an expensive form of competition for stringently restricted resources. It seems unlikely that the financial benefits accruing to patients in terms of better services have outweighed the costs, though we simply don't have the information to tell - and neither does the government.
Single-payer health care systems like the NHS - with finance channelled through general taxation or a single health fund - are the cheapest and among the most equitable in the world. Not perfect - but better than others. We are however putting impossible levels of strain on ours, both financially and ethically.
The notebook draws on published sources, and also the author's primary research in health and community care. Our thanks to our interviewees, and to John Clarke for comments on an earlier draft.

1. OHE (Office of Health Economics), Compendium of Health Service Statistics, 1995. The huge long term drop in length of stay in hospital reflects changing clinical practice.
2. Radical Statistics Health Group, 'NHS "indicators of success"; what do they tell us?' British Medical Journal, 22.4.95, pplO45-5O.
3. Both quotations from M. Traynor and B. Wade, The morale of nurses working in the community, NHS Trust Series Report III, Daphne Heald Research Unit, Royal College of Nursing 1994.
4. OHE (see note 1), Audit Commission, A price on their heads: measuring management costs in the NHS, HMSO 1995.
5. Financial Times, 31.5.96.
6. Guardian, 7.12.94, and 22.5.96.
7. Guardian, 24.4.96.
8.J. Dixon and H. Glennerster, 'What do we know about fundholding in general practice?', British Medical Journal, 16.9.90. 9. Audit Commission, What the Doctor Ordered, HMSO, 1996.
9. Audit Commissio, What the Doctor Oredered, HMSO 1996
10. Because of the influence of the Poll Tax on census completion in 1991.
11. M. Matsagannis and H. Glennetster, The]oumal of Health Economics, 1994, p52
12. D. Price, reporting on a study of the response to the HSG (95)8 circular on continuing care for older people, Health Services journal, 28.3.96.
13. D. Price, see note 10.
14. CSO (Central Statistical Office), Health and Social Services Statistics 1995, HMSO 1996.
15. Health Services Journal, 16.5.96.