Matthew Edwards, Head of mortality and longevity at Willis Towers Watson, discusses the most interesting mortality trends from 2016 in his latest article for The Actuary.
Recent mortality experience in England and Wales has been surprising, with little improvement in mortality rates over 2012 and 2013 and a significant worsening over 2015.
In Working Paper 93, the CMI presented a ‘near year end’ estimate of 2016 experience. By combining provisional deaths data to 14 October 2016 with an estimate for the remainder of the year, in conjunction with exposure based on ONS 2015 mid-year population estimates, the CMI concluded that total 2016 deaths in England and Wales were expected to be around 1% lower than in 2015, and the exposure around 1% higher, implying a 2% improvement in mortality rates.
Now that the provisional weekly death data for the whole of 2016 is available, we can attempt to update the CMI’s estimate. Analysing weekly deaths data since 15 October, and comparing with the equivalent period in 2015, shows that 2016 deaths in that remaining period for ages 65 and above are up 6.5%. Adjusting for the above exposure increase, the mortality rate for that period and age band has increased by around 5%.
This crude comparison shows that, if anything, the CMI’s estimate based on a mid-October cut is likely to have underestimated deaths for the year 2016, and therefore overestimated the variation from 2015. All other things being equal, the recalibration of the CMI_2016 improvement model to full 2016 data is likely to lead to lower life expectancies than were estimated in WP93 – and these were already materially lower than the CMI_2015 model calculations on 2015 data.
Looking at younger ages (age band 15-64), and considering the year as a whole, a similar comparison shows male mortality rates to have increased by of the order of 1%, with an approximate 2% increase in the female mortality rate. Although the numbers themselves are too crudely calculated to be reliable, the direction of the apparent variation is remarkable.
Dementia and Alzheimer’s now the leading killer in the UK
Recent data from the Office for National Statistics provides some insights into the cause of recent unexpected mortality experience, with an unforgettable headline. Their latest analysis covering deaths in England and Wales, released in November 2016, showed that deaths due to dementia and Alzheimer’s disease have now replaced ischaemic heart disease as the leading cause of death in England and Wales, accounting for 11.6% of deaths in 2015 (see Table 1).
Table 1. Top 5 causes of death in 2015
||Cause of death
||% of all deaths
||Ischaemic heart disease
||Cerebrovascular disease (incl. stroke)
||Chronic lower respiratory disease (e.g. bronchitis)
At a broad level, cancer remains the leading disease group by death numbers, followed by circulatory disease. However, this masks trends in the underlying causes of death.
At a more granular level, ischaemic heart disease has historically been the main cause of death in England and Wales. Over the last 15 years the rates of death from heart disease have halved. More recently we have seen a dramatic increase in deaths due to dementia and Alzheimer’s, due in part to a better understanding of the disease (affecting diagnosis and death reporting), but also owing to people living longer and therefore being more susceptible as they reach older ages.
Underlying much of these statistics is diabetes, which has increased alarmingly in recent years, and is directly linked to heart disease. There is increasing acceptance also of a linkage between diabetes (or rather, the underlying condition of ‘metabolic syndrome’) and Alzheimer’s, to the extent that Alzheimer’s is sometimes referred to as ‘type III’ diabetes.
Another piece of interesting UK news emerging in 2016 was the recognition of severe financial difficulties in National Health Service funding. The ‘health of the health service’ is a major driver of longevity for the vast majority of the population, as all the main aspects of medical care – primary care, screening, pharmaceutical treatment, surgery – depend on suitable financing.
In September 2016 the Office for Budget Responsibility projected that health spending would decrease in per-capita real terms by 0.9% (cumulative) over the four years 2016-2020. This contrasts with the previous 36-year period which saw real per capita spending increase by a factor of approximately 3.3. Note that this estimate will understate the per-capita impact for over-65s, owing to demographic shifts.
Financial pressure in the health service is being exacerbated by a number of causes, in addition to what actuaries might reasonably expect (for instance, medical treatment costs materially outstripping RPI, or the ever increasing strain from diabetes, which accounts for approximately one-tenth of health expenditure). These include:
- Reductions in local authority social care budgets (£4.6bn down since 2011), with knock-on effects on A&E and hospital services;
- The operation of the ‘Better Care Fund’ such as the £3.8bn transferred from NHS funding in 2015, funding additional community care in place of hospital care;
- Private Finance Initiative repayments on NHS / Foundation Trust debts, running at around £2bn annually;
- Operational strain of administering the internal ‘free market’ system, which has been estimated as costing of the order of billions of pounds annually (for example the Centre for Health and the Public Interest estimated the cost as around £5bn annually);
- Cuts ensuing from changes to the ‘payment by results’ system, relating to the payment of hospitals by treatment. These aim to achieve of the order of one-half of the £20bn ‘efficiency savings’ the NHS was trying to achieve under the last Government’s ‘Nicholson Challenge’;
- Cuts to the NHS repairs budget of £1.1bn in the 2016 budget.
NHS trusts and foundation trusts in England ended the 2015-16 financial year with nearly £2.5bn of deficits. According to much press coverage in late 2016, the NHS budget was set for a £22bn shortfall (cumulative to 2021) with further reductions necessary to cope with this position. The outlook for medical care for those without recourse to private health looks bleak.
Evidence points to a major failure of the health system exacerbated by failing is social care
The point has been brought to light in research published very recently in the Journal of the Royal Society of Medicine regarding the causes for high 2015 mortality, which concluded that ‘the evidence points to a major failure of the health system, possibly exacerbated by failings in social care’ rather than particular problems with, for instance, flu vaccine.
Emerging mortality trends in the USA
It is interesting to compare recent UK experience with that of the USA. A report released in late 2016 by the Centers for Disease Control and Prevention (‘CDC’) announced that overall US mortality had increased in 2015 from 2014, the first such increase for around twenty years. The age-adjusted death rate increased by 1.2%, driven largely to increases in unintentional injuries (see poisonings below), Alzheimer’s disease and deaths due to liver disease.
Also, the improvement in age adjusted deaths due to heart disease has slowed down (and in fact deaths due to heart disease have increased in aggregate in recent years due to shifting age mix), and no longer provides enough cushion to offset deterioration from other diseases.
Deaths directly attributable to diabetes have also increased although, of the main causes of death, that showing the largest relative increase was Alzheimer’s with deaths up 16% from the previous year.
There has been particular interest in the mortality of middle-aged (ages 45-54) white non-Hispanic men and women. This group has seen increasing mortality since around the year 2000, although this has come to prominence only recently owing to the overlay of differing ethnic effects (with mortality improvements from the other racial categories leading to overall improvements in that age band).
- All-cause mortality from this group increased by 9% over the period 1999-2013, with marked increases in poisoning, suicide and chronic liver disease swamping gains from decreasing cancer mortality. Poisoning relates to drug overdose / abuse, with particular concern in recent years around prescription drug overdoses and opioid abuse. Synthetic forms of heroin such as fentanyl are of the order of a hundred times ‘stronger’ than heroin, making lethal overdoses much more likely, while access seems to be becoming increasingly easy.
- A major part of the contribution from liver disease relates to hepatitis C. In 2016 the CDC announced that Hepatitis C-related deaths had reached an all-time high in 2014, surpassing the total combined deaths from 60 other relatively common infectious diseases (such as HIV, pneumococcal disease and tuberculosis). The sharp recent increase (a doubling in new cases of hepatitis C infection from 2010 to 2014) is partly attributable to more injectable drug use. The lack of obvious symptoms has led to worries that the real underlying problem may be more severe than is yet appreciated.
- There is a further contribution to liver disease from increasing alcohol usage in women. In 2013, more than a million women in the USA wound up in emergency rooms as a result of heavy drinking, with women in middle age apparently the most likely to suffer severe intoxication. According to The Washington Post, the rate of alcohol-related deaths for white women ages 35 to 54 has more than doubled since 1999, accounting for 8% of deaths in that age group in 2015.
2016 has also seen growing realisation in the US of the existence of a material gap in mortality improvements between higher and lower ends of the socio-economic spectrum, including studies that show differing trends in life expectancy from 1980-2010 for different income quintiles. Our analysis of these studies indicates an approximately 2% difference in improvement between top and bottom quintiles for age 50.
The Zika crisis
Finally, let’s look at global threats from disease, and in particular the disease epidemic that made the most headlines in 2016 - Zika.
The Zika virus has actually been known since 1947, when it was first isolated in the Zika Forest of Uganda (hence its name). Zika virus is related to dengue, yellow fever, Japanese encephalitis, and West Nile viruses. It is transmitted by daytime-active mosquitoes, although the transmission mechanism has now been seen to include sexual transmission.
After consideration of a number of microcephaly cases (and other neurological disorders) linked to the Zika virus, primarily in French Polynesia and Brazil, the World Health Organisation (‘WHO’) declared a Public Health Emergency in February 2016. Although the virus is not associated with adult deaths, the prospect of natal deformity in pregnant women caused widespread concern, particularly in the US because of proximity to South America and for those involved in the Rio Olympics. Cases in Europe have been rare, with Spain being the only country with a material number of cases.
In positive news, the WHO removed the Public Health Emergency status in November, as it was no longer felt appropriate given the understanding gained since February’s ‘panic’. This is reassuring given that there are at present no vaccines or cures are available.
Thinking about the implications for actuaries, it is expected the mortality / longevity impact of the Zika virus will be negligible, but there may be some increase to Critical Illness claims through Guilllain-Barré syndrome, associated with Zika.
This article was originally published by The Actuary.