How much of the current NHS and social care pressure is due to ageing, other causes of population ill health, changing public demands, changing supply, or other issues.
This is a question I keep being asked. I don’t know the answer. It is probably impossible to quantify precisely.
My effort is a complex answer, it comes in 5 parts:
- Population issues
- Indicators of illness
- Capacity and supply of health care
- Some common myths – ageing, demographic time bomb, prevention is pointless
- Summary points
The following blog is my own reflection on solutions to some of these problems, often things that aren’t featuring heavily in what we are writing in our plans.
spoiler – its not “the ageing population”, it may be morbidity and numbers of people but its not ageing per se.
1 Population issues
The population is getting bigger
The population in Sheffield is growing by 1.13% per year. The population growth isn’t even across all groups.
Population growth is different in different age strata, as is indicated in the graph below for Sheffield.
As can be seen the expected rate of growth out to 2037 is 7% in 0-19s, 7% in 20-64, 34% in 64-79 year olds and 76% in those aged 80 and above. Across all age bands, the population of Sheffield is set to grow by 13% out to 2037.
Both the changing birth rate, death rate and inward migration drive population growth.
2 Health and illness
Life expectancy and Healthy Life Expectancy
a) Life expectancy matters. It is rising. It has risen by about 3.5 years in the last 20.
Inequalities persist, there’s an 15-20 year difference between the best and worst. This is unacceptable. It is not just a geographical issue – differences in life expectancy between people with serious mental illness or learning disability and the population average. This is also unacceptable. It is rising on account of social change, better treatments, reduced lifestyle risks.
It’s worth a note that the long term consequences of austerity, welfare reform and increases in some risks say obesity may yet put this trend in improving life expectancy at risk.
Healthy life expectancy is a more useful metric, and matters a lot more
Healthy Life Expectancy is a metric that incorporates the length of life, but also the number of years lived with poor health. For example, the graph below shows that for women in Sheffield average life expectancy is 82, but approximately 20 of those years are lived with poorer than optimal health.
Healthy Life Expectancy is not improving. Unlike life expectancy, healthy life expectancy does not appear to be increasing in Sheffield (or England). Equally, it is not getting worse, so the notion that we are getting less healthy as an explanatory for increased pressure is difficult to make stick.
As is often reported this (often avoidable) illness and early death is not equitably distributed in any population. There is a c20-25 year gap between most and least deprived people in Healthy Life Expectancy, as indicated below.
The data for female healthy life expectancy shows a similar pattern. Again as with life expectancy, there are also substantial differences in the life expectancy in other vulnerable groups including those with a learning disability or with a serious mental health problem, and other populations with multiple disadvantage, and the wider population.
b) Causes of death, causes of illness and avoid ability
Causes of death
CVD – 25%, at least one third of this is preventable or delay able
Cancer – 30%, at least one quarter is preventable or delay able
This is death at any age, EARLY death will present similar picture – arguably larger slice preventable.
So of all deaths a large proportion are preventable. It is important to not just think in terms of death, even early death, it is also important ot think in term of the lost quality of life (and lost economic productivity) before death on account of illness.
Don’t just think of death – think of the illness and misery that goes prior to the death.
Key references on the epidemiology of common causes of death are referenced
Causes of illness
The pattern of illness shows a different picture. Main causes of disability (think reasons why people use health and care services). Many of these are entirely or mostly preventable. Avoidable illnesses are quantified here with the metric of Years Lived with Disability
The pattern of illness is different to causes of death; there are many things that lead to lost quality of life that don’t actually kill us. This is illustrated below using the metric of Years Lived with Disability (YLD) to quantify, again alongside the immediate risk factors:
Accordingly, the immediate causes of those years lived with disability are different to causes of death.
When this data is split by age groups, some very interesting trends emerge. For example (data not shown here) for working age population alcohol is the leading cause of lost health (expressed in DALY terms) – this is of obvious importance from a productive economy perspective.
Again is this getting worse and accounting for increased demand
e) Social gradient in multi morbidity
There is a 10 – 15 year difference in the age of onset of what is called multi morbidity – or when people have multiple long term health conditions. People living in the most deprived neighborhood develop multiple morbidities 10-15 years before those in the least deprived. Put another way, at age 50-54 18.3% of the population have more than one condition morbidity in least deprived populations compared to 36.8% in most deprived (Barnet Lancet 2012). As many of the illnesses are preventable, this further brings into question the “ageing population” issue and suggests that it is avoidable illness that causes problems, rather than age per se. The slide I often show is this one.
In my view, this graph, combined with the CHE (and other) report (referenced) firmly dispels the notion that the “ageing population” is responsible for the demand side of escalating care costs.
we know little to nothing about how the epi of multimorbidiy is changing over time – this is crying out for research
IT IS NOT, I REPEAT NOT, AN “AGEING” THING BUT A MORBIDIY THING
Lastly if you needs any more persuasion check slide 11 of this analysis of a 500k person real world dataset in Somerset
3 Supply side considerations
the “demand management” side of supply has been significantly degraded in recent years. The data below references General Practice. The same story can be told, more startly, with regard to social care.
As Professor David Haslam is often quoted “GPs are like the heat sink in your computer; you don’t know its there until it is broken, then your system breaks”. Arguably the same is true of social care.
Recent HSCIC data (national) on general practice workforce between 2004 to 2014 highlights that the UK population has risen by c2m since 2010 whilst GP numbers are relatively flat – thus GPs per 100,000 population has fallen from 62.4 to 60.0. Subsequent commentary had suggested that statistics on GP numbers do not take account of the increase in the non-clinical work of GPs.
During this time period, though data is only available from sample surveys, it is known that the average GP consultation rate in the population has increased significantly.
Hobbs et al recently published the results of a five year analysis of GP workload in the UK using 100 million consultations in the CPRD across almost 400 practices. The crude annual consultation rate per person increased by 10·51%, from 4·67 in 2007–08, to 5·16 in 2013–14. Consultation rates were highest in infants (age 0–4 years) and elderly people (≥85 years), and were higher for female patients than for male patients of all ages. The greatest increases in age-standardised and sex-standardised rates were in GPs, with a rise of 12·36% per 10 000 person-years, compared with 0·9% for practice nurses. GP telephone consultation rates doubled, compared with a 5·20% rise in surgery consultations, which accounted for 90% of all consultations. The mean duration of GP surgery consultations increased by 6·7%, from 8·65 to 9·22 min, and overall workload increased by 16%. These findings were taken as a substantial increase in practice consultation rates, average consultation duration, and total patient-facing clinical workload in English general practice. These results suggest that English primary care as currently delivered could be reaching saturation point. This data only explored direct clinical workload and not indirect activities and professional duties, which have probably also increased.
Capacity in hospitals is undoubtedly becoming more stretched
Perceptions of pressure and actual pressure on the urgent care system is inevitably going to increase as the NHS continues with a long term 50% reduction in the number of beds, the reduction is most marked in mental health and LD beds. This is most neatly illustrated by the Kings Fund.
During 2014/15 there were on average 135,616 beds available in wards open 24 hours a day in NHS hospitals in England. That is around half the beds available in 1989/90 and 6,855 fewer than in 2010/11. The decline since 2010/11 has been particularly pronounced in learning disability specialities (-31%). This long term decline is a result of deliberate policy to shift care closer to home and advances in technology.
Hospital beds per capita in England is low comparatively – we might carefully consider before we cut further
Looking across Europe the number of hospital beds per million population and the length of stay in England is low.
Hospitals are becoming busier.
There were 18.2 million finished consultant episodes (FCEs) in 2013/14, 2.5% more than in 2012/13.
More recently that has been an increase in the intensity of bed use, as measured by occupancy – occupancy rates for acute beds have increased from 87.7 per cent in 2010/11 to 89.5 per cent in 2014/15. The NAO have suggested that occupancy rates above 85% will be associated with regular bed shortages.
Workforce is often cited as a core challenge across all the policy problems in the NHS.
The specialist workforce is growing faster than generalist.
Many have highlighted the major disconnects between strategic goals and workforce trends. Though a point that is applicable to the whole of the NHS workforce, the increase in consultant body has considerably outstripped the rise in GP and other aspects of the NHS workforce over the last 6 years.
NB this data is out to 2014, I don’t have to hand more recent data (and I’m writing a blog not a PhD thesis).
This growth pattern in the medical and nursing workforce seems at odds with a policy ambition of moving care closer to home and might call into question whether the balance between generalist and specialist is correct. The important caveat to this is the large data gaps in the availability of information on workforce.
The supply and demand side together is complex, as is illuistrated below:
4 Three common myths
The “it’s all about the ageing population” that’s the reason why the system is overheated. The ageing population will sink us
The OBR report is helpful re numbers and macro sense – demographic pressure is a small factor in health care cost growth – the key issue is number of cases (prevention) and expensive technology combined with over diagnosis and heightened patient expectation
when the question of what is driving cost growth – over supply (demand expands to fill available supply, number of cases is growing (in many egs cost / case relatively stable, see Thorpe reference), a complex picture emerges. Its often suggested that cost growth is a function of “the ageing population2. .
The evidence to suggest the notion that “the ageing population will bankrupt the system” is weak at best on two main grounds. Firstly, the evidence is pretty cleartt’s not age per se, but proximity to death / last 18 months of life (whatever age that death is 50 or 90, its the last 18m). See the CHE York report (referenced) – activity growth rather than unit cost is what’s driving spend (cost / case vs new case incidence – nb some caveats here). Health care activity rises substantially for individuals in the period before death so that expenditure is driven substantially by mortality in the population (figures 13 to 16 from p20 on, re lifetime inpatient costs and in final year of life)
The older the age of death then the lower is the cost in the final year of life and the cumulative lifetime costs. A hidden benefit of prevention – and I presume as age of death increases then cost should be reducing!??
Secondly, expenditure is concentrated in individuals with multiple diseases. We don’t know whether or how this is changing over time. The prevalence and identification of complex medical conditions are important drivers of expenditure (nb social gradient).
The demographic timebomb myth
Again, related to the above. The old age dependency ratio is another critical variable in future planning – the doomsday scenario may not be as bad as many (including me) have suggested. There is also much that can (and should) be done with regard to both formal and (arguably more importantly) informal social care, neighbourhood capital and wider social economy considerations.
Spikjer and MacInness and Appelby in the BMJ and the Kirkwood study – 1) The extent, speed, and effect of population ageing have been exaggerated because the standard indicator—the old age dependency ratio—does not take account of falling mortality, 2) When measured using remaining life expectancy, old age dependency turns out to have fallen substantially in the UK and elsewhere over recent decades and is likely to stabilise in the UK close to its current level – v interesting 3) the capacity of healthcare systems to cope with increasing longevity will depend on the changing relationship between morbidity and remaining life expectancy and, in particular, the effect of education 4) population projections are often wrong, 5) Many quite well. Less poorly than commentators have us believe etc
Lastly long term broader social changes further compound. Reduction of nuclear family, slow decline of community or social capital etc…(the implication being that “in the old days we could rely on neighbours to look out for our old folk, or close living family members”)…..
It is worth considering the broader social trend and background issues on the dependency ratio – approp housing, education and employment opportunities, social cohesion, formal and informal care, meaningful activity and many other concepts are all important in the concept of healthy ageing from a social model perspective. This may offset the issues associated with age per se. Look at Japan for example – one of oldest societies in world, still with functional system . So looking from a much broader societal perspective there’s a good deal still to play for
That said, long term changes in dependency ratios underscore the important of carers and keeping carers well.
Healthy ageing is something to start in middle age, or younger. Mitnitski – The rate of aging: the rate of deficit accumulation does not change over the adult life span. Deficits accumulate mercilessly, like compound interest. To be fit at 80, be fit at 50. Deficit accumulation will, on average, double twice between ages 50 and 80 highlights the importance of health in middle age on late life outcomes.
“prevention is ultimately futile in broad economic terms as folk will likely live longer and get dementia” – (see my recent blog, referenced)
It’s a pretty poor argument against investment in prevention. The truth is that investment in prevention MAY be a net save in some areas, may not in others. Worst case in a macro sense is that preventive approach slows rate of growth compared to counterfactual of simply reacting when stuff happens (ie current model of NHS). Wanless made this case 20 years ago. Was accepted by Treasury then. Precise estimates are largely impossible when talking about a programme of interventions. There are a lot of moving parts, all of which act with uncertainty.
(on dementia per se not all cases are preventable, some are – a third of dementia is vascular – standard CV risks and most likely CV risk reductions apply. Some recent evidence from states (NEJM 2014) – the incident rate of dementia is falling. The prevalent rate may thus fall slowly (dependant on what happens with death rate), but the prevalent number will continue to increase with pop growth. As a third of dementia is vascular – standard CV risks and most likely CV risk reductions apply what’s good for the heart is good for the brain.
see also this excellent reference
Health care costs in the English NHS: reference tables for average annual NHS spend by age, sex and deprivation group
and compare cost profiles (here for men) across deprivation and age
more AGAINST a hypothesis of the “ageing population timebomb”
5 In summary
- The basic age structure of the population isn’t changing terribly fast, the number of older people is going up.
- Our pop profile & projections is that the absolute number of people aged >65 is increasing, but the overall age structure is not changing that quickly.
- Thus there are more “person years” in the population
- But are healthy life expectancy is not improving but not getting worse and the percentage of people with multi morbidity doesn’t seem set to fall anytime soon (good quality epi is missing). This basically leads to more “unhealthy person years” in a fixed capacity system
- This is not evenly spread across the population. There are inequalities, as described above.
- the three key drivers of health care cost growth are (not necessarily in order, apart from the first) population growth, deprivation (proxy for lifestyles then morbidity), over diagnosis, new technology often of marginal value, and age profile change (which following above I think is a proxy for no of people with a closeness to death).
Whilst demand on the system is increasing on account of the above, the supply in the system is falling, in some sectors faster than others. Also there may be an argument there is over supply in some areas of the service and under supply in others. This leads to nuanced solution that involved prevention, addressing inequalities and growing some forms of supply whilst holding other forms of supply.
The recent OBR and Nuffield Trust analysis concluded that in the long run the NHS is not unaffordable, and that cost growth driven by cost per case (mostly technology) and number of cases (our failure to prevent) NOT, demographic pressure.
Similarly, the Glasgow work found that the current problems in NHS in Scotland are as a result of
1) falling % of NHS £ to “community” led to hike in non elective (and a range of other issues)
2) problems not due to too little funding overall
3) worried well – especially in most affluent population
4) systemic and cultural incentives leading to over diagnosis
5) degradation of primary and community care over 10 years, including social care.
The slides from this Glasgow work are excellent and well worth a read. I have summarized them.
Thanks to Steve Laitner for vey helpful comments on an earlier draft
 Clinical workload in UK primary care: a retrospective analysis of 100 million consultations in England, 2007–14. Lancet Volume 387, No. 10035, p2323–2330, 4 June 2016. http://www.thelancet.com/journals/lancet/article/PIIS0140-6736(16)00620-6/abstract
 NHS England (2015); The King’s Fund estimate for 2011/12 to 2013/14. The number of hospital beds. 25 March 2015
 OECD comparison
 HOC Library – SN/SG/2641
 Nuffield Trust – financial sustainability of the NHS – http://www.nuffieldtrust.org.uk/blog/nhs-financially-sustainable
Population projections out to 2037 – https://www.sheffield.gov.uk/your-city-council/sheffield-profile/population-and-health.html
Key points on the epidemiology of the common causes of death are referenced.
Smittenaar et al. Cancer incidence and mortality projections in the UK until 2035
- Cancer incidence rates are projected to decrease by 0.03% in males and increase by 0.11% in females yearly between 2015 and 2035;
- thyroid, liver, oral and kidney cancer are among the fastest accelerating cancers.
- 243 690 female and 270 261 male cancer cases are projected for 2035.
- Breast and prostate cancers are projected to be the most common cancers among females and males, respectively in 2035.
- Most cancers’ mortality rate is decreasing; there are notable increases for liver, oral and anal cancer.
focused on mortality, morbidity, treatment and cost. Good information on epidemiology specifically wrt to incidence and prevalence (ie number of cases) is more difficult to see in this report.
Bhatnagar et al. The epidemiology of cardiovascular disease in the UK 2014
Lung disease in the UK | British Lung Foundation – https://statistics.blf.org.uk/
Prevention as a means of cost control
Prevention as long term health care cost control 1/3 (Mean Cost per case vs number of cases)
Prevention as long term health care cost control 2/3 (Continued under prevn decreases overall efficiency)
Prevention as long term health care cost control 3/3 (Prevention one person at a time vs policy focused approaches)
Thorpe KE, Treated disease prevalence and spending per treated case drove most of the growth in health care spending in 1987 – 2009. Health Affairs 2013 5 pp851-858. http://www.ncbi.nlm.nih.gov/pubmed/25972341
Impact of individual level interventions – should we go for more individual level, or for policies that prevent the need for pills. Prevention
Pills vs services vs public policy
prevention is futile as we will all get dementia anyway. Discuss. https://gregfellpublichealth.wordpress.com/2016/08/24/prevention-lets-not-bother-as-itll-never-save-money-how-much-money-will-you-save-in-our-stp-by-next-week-fell-12/
STP in macro economic context – remember Wanless. Reflecting on Wanless
What if we managed a 10% smoking prevalence, a though experiment. Do you want to grow the local economy by 7.5% and save lives all at once.https://gregfellpublichealth.wordpress.com/2016/06/11/scaling-up-prevention-an-stp-thought-experiment/
Ageing population issues, population projections and healthy ageing
It’s not all about the ageing population, here’s my take on whyhttps://gregfellpublichealth.wordpress.com/2016/03/21/is-the-ageing-population-is-accounting-for-escalating-health-costs/ and https://gregfellpublichealth.wordpress.com/2016/03/25/the-its-all-about-the-ageing-population-conundrum-where-next/
Mitnitski – ncbi.nlm.nih.gov/pubmed/25972341
https://flipchartfairytales.wordpress.com/2015/08/28/the-healthy-ageing-challenge/ – this blog is astounding: a broader view on healthy ageing. Explores the distinctions between life expectancy and healthy life expectancy, and the absolutely fundamental importance of this in both health, health care and social policy.
CHE spending in last 18 months of life – Spending in last 18 months vs “the ageing population” – York Centre for Health Economics. Medical Spending and Hospital Inpatient Care in England: An Analysis over Time
this is a critical reference
this is simply the best debunking of the ageing population myth I have ever seen
Aging Population and Its Potential Impact. What’s the Issue?
Spikjer and MacInness – Population ageing: the timebomb that isn’t? http://www.bmj.com/content/347/bmj.f6598
The Kirkwood Newcastle study – http://www.theguardian.com/society/2012/may/29/tom-kirkwood-research-dispels-myths-ageing.
Appleby – Population projections: why they are often wrong. BMJ2014;349:g5184