Making people work for their health care

The Economist this week carries an article and an editorial piece about what they are calling “The Arkansas experiment“.  In January 2018 President Trump announced that there would be federal waivers to allow states to introduce a test of ‘community engagement’ for entitlement to Medicaid.  Medicaid is the means-tested system offering support in the US for health care for people of working age; ‘community engagement’ means, more or less, a work test, requiring people to be working, ‘volunteering’, studying or responsible.  Arkansas is so far the only state to implement this, but the Economist notes that 14 other states have applied for similar waivers.

The Economist expresses some doubt about the policy: it is complicated, engagement is difficult to prove in a world of precarious work, and incentivisation is perverse.  The main thing that sick people need before they can work is to be healthy.  But they start with a rather questionable statement of principle:  “The theory behind tying cash benefits to work requirements is sound. Asking people to do something in exchange offer a payment can build political support for welfare programmes”.    Conditionality may well be the price that politicians have to offer to get a programme accepted; that’s not the same as saying that conditionality leads to greater support.  If anything, the polities where people are most determined to impose conditions on the poor are also usually the ones where support is most tenuous.

The “theory” behind work requirements, if it deserves to be called a theory,  is highly questionable.  ‘Activation’ policies, which are supposed to prod unemployed people into work, are based on a series of false premises – that benefits used to  promote unemployment ‘passively’, that the answer to unemployment is more vigorous job-seeking, and that people will not move into work without a spur.  Empirically, activation doesn’t improve job matching; there is some evidence that it can make lead to mismatches, or even slow down the rate at which people move in to employment.  ‘Activation’ for people who are sick – a policy we’re now seeing in the UK, reflected in the treatment of sick people on ESA and Universal Credit – is worse still.  People on these benefits have to ready themselves for work nevertheless – sickness is no excuse.  It’s only a small step from there to the extension of the same principle to health care.  Depression?  Ulcerative colitis?  Congestive heart failure?  Pull yourself together!



The NHS long-term plan for England is not really a ‘plan’

I’ve not been close to work with the NHS for some years, and I’ve been away from England rather longer; but I think I know what a plan looks like, and the NHS Long Term Plan isn’t it.  The supposed plan is a long, rambling shopping list, lacking in structure, priorities or the means of achieving any aspirations people might have.  It’s all very well to say that the service will “dissolve the divide between primary and community health services”, or that people will have “more personalised health care when they need it”, but we need to know why, how, and how we will know if it’s happened.

One of the implications of the lack of structure is that there will be no way to tell whether or not this document has made any difference.  Being told, in bold lettering,  that “The health  service will continue to support implementation and delivery of the government’s new five-year action plan on Antimicrobial Resistance” or that “every trust in England with a maternity and neonatal service will be part of the National Maternal and Neonatal Health Safety Collaborative” tells us nothing at all – and if those objectives aren’t achieved, the gaps will be covered up by the rubble from hundreds of other equivalent objectives which might be achieved on the whole, or in part, or not at all.    The problem with long lists of policies,  Aaron Wildavsky wrote in Speaking truth to power,  is that they become a way of burying the things that don’t get achieved – “mechanisms for avoiding rather than making choices”.  The NHS Long Term Plan doesn’t identify problems, aims, methods, process, outcomes, or how to get value for money.  That doesn’t inspire confidence for the future.

Is the NHS the best health care service?

An international ranking of health services in 11 countries rates the UK NHS at the top.  The Commonwealth Fund, an American think-tank, ranks health systems on five main criteria:  Access, Equity, the Care Process, Administrative Efficiency and Health Outcomes.  Each of those criteria is based in turn  on a range of subordinate indicators:  the “Care Process”, for example, takes into account prevention, safe care, coordination, and patient engagement.  It’s backed up further by more detailed assessment;  for example, the US does badly on infant mortality and premature death, but relatively well in relation to doctor-patient relationships and the management of stroke.  But speaking as a carer, I find it hard to believe that the state of our mental health services really represents the best that anyone can do.

The main purpose of the report is to give a critical perspective on health care in the US, which is outstandingly expensive as well as being the least effective of the systems; but there are questions to raise about other countries, too.  For the UK, we might wonder how it is that the health care system is ranked top of the league while the UK’s health outcomes are the second worst in the table.   The neo-liberal Institute of Economic Affairs commented, acerbically: “the NHS’s provision of care is equally poor for everybody, irrespective of income.”


The crisis in the NHS

Leaving aside the question of whether the NHS has a ‘humanitarian crisis‘, which sounds apocalyptic, there’s much about the current state of the NHS which is based in long-standing problems.  The first problem is the lack of spare capacity in the hospitals. The effect of insisting that beds have to be fully occupied is to create inflexibility and bottlenecks.

However, the problems which show themselves in the hospitals are not necessarily problems which can be addressed through the hospitals.  The second key issue has been the retrenchment of social care.  Social Services Departments, or Adult Care Departments, have radically reduced the scope of their involvement with the public: figures are difficult to find, but between 2008 and 2013 the numbers of people being served fell by a quarter.

The third problem has been an apparent failure of GP coverage.  This is puzzling because the figures seem to imply the opposite.  Currently there are 5.8 GPs per 1000 patients per practice.  That   averages out to just over 1700 people per GP, in the worst cases rising above 2300.  When I started in this game the ratio of GPs to patients used to be 2200, in some areas going up to 3500, and that was in the days when GPs also had to come out at night.  However, there appear to be more people on GP lists overall than there are in the population – suggesting that general practices and CCGs are not very good at keeping their records up to date, possibly because it’s not in their interests to do so.  I’ve also not been able to find respectable figures for how many people are not registered with any local GP, which may be marginal (the same people are less likely to get access to any health care) but is potentially important in the demand for direct access to A and E.

It’s understandable that the government is focusing on GPs, because it’s the most immediate response that could affect the numbers of people coming into A and E without directly requiring new capital investment to do it.   Demanding that GPs change their office hours, however, is not likely to make much difference; this redistributes the times when people get seen, but it’s no guarantee that more will be seen and where, for example, a GP is taken off a Monday rota to go on a Sunday rota, it may mean (depending on the practice setting) that fewer are.  There may be other implications.  GPs do much more than talk to patients; they also coordinate continuing care and the multidisciplinary team.  (I understand this may be different in England, where GPs have been complaining that they’re more remote from community nursing.)  If at least one GP has always to be seeing patients, when can the practice ever have team meetings to discuss care management?

The fourth problem concerns how we respond to the population in need.  We should dismiss one of the common explanations: that the ageing population itself implies a greater burden.  There are theories about the ‘expansion of morbidity’, suggesting that people are ill for longer; there’s a contradictory view, the ‘compression of morbidity’, which says that people are healthier for longer – but frankly the evidence isn’t convincing for either of them. (The issues are discussed in a WHO report, Global health and ageing.)  However, it is true that local population movement does increase local demands in some places – the South East of England is overcrowded while some areas of Scotland are depopulated.  That’s  a different kind of issue.  We need to give more thought to the kind of services that are available for a mobile and often transient population.

Rationing isn’t best done by charging for services

A discussion on the Radical Statistics list sent me off looking through some old material.  The cash crisis in the National Health Service has fuelled calls for charges to be introduced, as a way of quelling demand. Charges, however, are not a good way of doing this.

This table was part of a paper I never published – I cannibalised sections for a book, and put the arguments differently.

Normative criteria applied to different forms of rationing


Distributive equity Access for a target population: avoiding errors of exclusion Reducing aggregate demand Selectivity:
avoiding errors of inclusion
Denial of service – closing the doors ?
Rationing by price ?
Filtering/ referral processes ? ?
Eligibility criteria
Priority weighting ? ? ?
Dilution – less service for everyone ? ? ? ?
Deterrence ? ?
Delay – waiting times ? ?


Some of the ratings might be questioned – I think, for example, that there is a case to show that making people wait has bad distributive consequences, not just questionable ones.  The central point to take from this is that there is no reason to suppose that price rationing has any intrinsic superiority over other methods. If anything, it is rather inadequate as an instrument of public policy: it has nothing to prevent inappropriate inclusion or exclusion, and while some people think it’s a fair procedure (which is debatable), it is unlikely to be fair in distribution. Waiting lists and queues are not much better. The two methods of rationing which relate best to most of the criteria are eligibility rules and filtering, such as triage or using referral processes.

Dementia: we shouldn’t expect miracles from drug cures

The media do love to run reports about wonder drugs.  The latest is aducanumab,  which is being widely reported as a breakthrough in the treatment of Alzheimer’s disease.  The Telegraph   hails a ‘new drug that halts mental decline’.  The Mail headline asks:  Could this be the end of Alzheimer’s?  Revolutionary drug ‘may stop the disease from ever developing.’  For those with access, the actual report of the trial is in Nature.

There are several reasons why this is over-hyped, and it’s not just that this is based on a small sample from first stage trials.

  1.  Dementia is not a disease, but a syndrome; the symptoms of dementia are known to be the result of a variety of  different pathologies.
  2. It’s far from clear that Alzheimer’s is a single disease with a common cause.  It’s widely used as a residual category for dementias that have not been explained in other ways.
  3. It’s not clear how far the amyloid-β proteins that the drug is removing are actually responsible for dementia.  It’s a plausible hypothesis; but, as point 2 implies we should expect, there are some apparent cases of Alzheimer’s without amyloid-β plaques, and because the proteins (Aβ and tau) can build up for 15-20 years without symptoms emerging, the presence of the proteins does not mean that the person with a build up will suffer  dementia.
  4. It’s  not clear that whether, if the hypothesis is correct, the plaques impair cognition which might be relieved,  or cause permanent  damage.  This study claims, with few direct details,  that cognitive decline is slowed – but the comparison they make is not between drug receipt and placebo, but between those cases who had both received the drug and recovered better from plaques from those who received the drug without that effect.

Lastly, it ought to be noted that many people dropped out of the trial, at least 20 because of side effects.  The way that drug trials work – or should work – is that the tests should be able to refine the identification of those for whom the drug is appropriate, and those for whom it isn’t.  That takes time and good science, and hype and the search for miracle cures tends to get in the way of that.

Additional note, 5th November 2019:  The i newspaper carries the page 1 headline:

The woman who defied Alzheimer’s: Patient with rare genetic mutation raises fresh hope that a cure for dementia can be found.

I started to cover this on the blog, but this earlier post already says what needed to be said.  All together, now:  there is no cure for dementia.  The much-maligned NHS website says it plainly:

There is currently no “cure” for dementia. In fact, because dementia is caused by different diseases it is unlikely that there will be a single cure for dementia.

Further additional note, 6th November 2021: 

New Scientist has finally caught up with the argument.

 “Alzheimer’s, as currently understood, is really not a single disease. … This new approach comes not a moment too soon … Tunnel vision has held us back for too long.  With this new approach, a single blockbuster drug might well be out of the picture, but instead, there is a real possibility that we could dismantle Alzheimer’s by a thousand cuts.”

Snakebite medicine is running out

The news that we are running out of medicine to deal with snake venom in sub-Saharan Africa is not what economists call ‘market failure’, but it is a failure of the market. According to MSF, five million people are bitten by snakes each year, and 100,000 die.

Part of the problem is, of course, that victims are poor.  MSF comment that “antivenom treatment can cost up to US$250-500 per victim, representing the equivalent of four years of salary in the countries concerned.”    In this case, Sanofi  have decided to focus their production on something else that is important (rabies treatments), but will yield them better profits. It’s a commercial decision, which is no ground for criticism; they are doing what private producers do. Market theorists make a great deal of the market as an engine of choice, but they tend to focus on the choices of consumers; they often forget that producers get to make choices, too. The nature of market provision is that there will always be some degree of selection, and so that there will always be some kind of gap. That is why nothing that really matters can ever be left to markets altogether.

I must defend the Conservatives: they're right to set spending against GDP

The Conservative Party has been criticised for claiming that they have halved the deficit, when in monetary terms they have not.  I have to leap to their defence.  The test they have used, treating the deficit as a proportion of GDP,  is entirely legitimate and appropriate.   It sets spending against income, which is how it needs to be seen.   I have used the same indicator myself in this blog.

Let me, however, take the case a little further. The figures in the first two lines come from the OECD and the World Bank.  It’s not surprising that total social expenditure has been cut; but many people will not have realised that, by the test the Conservatives are now applying, spending on health care, which was supposed to be protected, has also suffered.

2009 2010 2011 2012 2013
Social expenditure 23.9% 22.8% 22.7% 23% 22.5%
Health care 9.9% 9.6% 9.4% 9.2%
Benefits and tax credits
– working age and children
6.1% 6.0% 6.0% 6.0% 5.7%
Tax reliefs 18.3% 18.8% 20.5% 21.3%


The third line comes from the DWP.  It’s true that benefits have increased over the years, but that is mainly down to two factors:  pensions, and Tax Credits for the low paid.  If we look only at the DWP benefits for people of working age, which include the ‘out of work benefits’,  another story emerges.  The figures are

1979  2.2%
1984  3.6%
1989  2.7%
1994  4.2%
1999  3.3%
2004 2.8%
2009  3.3%
2014  2.9%

There have been some fluctuations with economic conditions, as we might expect; but the high point was twenty years ago, during the Major government, and the narrative we often hear, of increasing dependency and a relentless increase in costs, is hogwash.

The last line in the table, drawn from a National Audit Office report earlier this year, shows the cost of tax reliefs as a percentage of GDP over the same period.  It helps to make it clear where the government’s priorities lie.

The most direct way to improve health: more money

A new study by the Scottish Public Health Observatory has reviewed the potential effects on health of a series of interventions.  The effects on health are considered mainly in terms of mortality and the need for hospital care.  The measures in relation to incomes, cutting alcohol, tobacco or obesity, and encouraging cycling and walking to work.

Health interventions

The biggest benefits, by far, would be gained by the introduction of a living wage; the next biggest, though it is even more important for the poor, would be an increase in minimum income benefits.  Some of the measures they are being compared to seem restrained – a weight-management service rather than measures to reduce sugar, short-term interventions on alcohol rather than taxation by unit – but the size of the difference is so big that it’s unlikely to  tilt the table.

A pensioner goes walkabout

A report in the Dundee Courier this morning concerns a manhunt (or womanhunt, if you prefer) for an 86-year old  from Dunbartonshire.  When her care workers arrived at 7 am yesterday, she wasn’t there.  A search has been mounted with police dogs, a helicopter and the co-operation of the Ministry of Defence.  It’s not unusual for the police to be involved in searches for missing pensioners – in a case last year in Birmingham, they”searched the local area, made enquiries with relatives and checks with taxi firms, public transport and health care providers” – but the scale of this search raises questions about how we ought to respond.

There is nothing unusual about people with dementia going out.  People with dementia tend to do things that they’ve done throughout their adult life.  It’s often the context that makes the actions seem disturbed  – people going for a walk because they think they’d like to take a walk, cooking in the middle of the night, going to the shops when everything is shut, lighting a fire when they’re cold, deciding to visit their family while not remembering that their family don’t live where they used to.   In the bad old days, which are not that long gone, older people would routinely be incarcerated in psycho-geriatric wards, typically with locked doors.  When we committed ourselves to keeping people in their own homes, we took on a degree of risk.

What kind of response would be proportionate?  I don’t know; I don’t even begin to have an answer.   But there has to be a better way than helicopters and dogs.