The latest life expectancy data from the Office for National Statistics doesn’t provide much to celebrate. The UK’s life expectancy continues to stall and its poor standing in international life expectancy league tables calls for urgent action, especially as mortality is the tip of a much bigger disease iceberg. Action needs to be informed by an accurate assessment of mortality changes and how the root causes vary in their impact on different population groups and areas, so that government policy and services can respond with measures that are evidence-led.
Recent trends in life expectancy have attracted much debate about both the direction and causes of these changes. Reports even differ about whether life expectancy is stalling, falling, or improving but more slowly than before, depending on how official data are interpreted. Here we use the latest ONS data to clarify how life expectancy is changing.
Interpretations of mortality trends vary depending on the measure used, and this has contributed to inconsistencies in reporting. For example, the number of deaths is an easily understood figure, but it predominantly measures the population’s size and age structure. Even if death rates are unchanged or fall, the number of deaths can rise when the population grows or ages – the UK population is doing both currently. For example: 57,000 more deaths were registered in England and Wales in 2018 compared with 2011 (a 12 percent rise), but the age-standardised death rate fell by 13.2 per 100,000 (a one percent decrease).
Measures such as age-standardised death rates and period life expectancy at birth are more reliable for interpreting trends because they adjust for these demographic changes. Both are summary measures of the same underlying data, providing alternative, but mutually consistent and widely used ways of comparing mortality reliably over time and between different populations. Life expectancy can be measured on a period (point in time) or cohort (for a specific group) basis.
Different measures can present different views of mortality. For example, there have been recent reports that life expectancy is falling. These were references to work by the Institute and Faculty of Actuaries, who recently revised down their Continuous Mortality Investigations (CMI) model projections of long-term improvement in cohort life expectancy at age 65 (a quite different measure to current, period life expectancy at birth). However, even on this basis, the Institute did not say life expectancy overall is falling, nor is it projecting a fall in life expectancy for any group – it is instead projecting a slower rate of improvement among people aged 65+ than previously assumed: “almost all users of the CMI Model expect that mortality will continue to improve, even if this is at a slower rate than in the first decade of this century.” Similar reports of falling life expectancy followed when ONS’ 2016 projections of life expectancy improvements were revised downward from previous projections.
Even the national and foreign media are now reporting that life expectancy in Britain is falling. ONS data released this week shows that life expectancy at birth in UK overall (and in England) isn’t falling, the average for 2016-18 being slightly higher than previous years.** However, after many decades of steady improvement, the current decade has seen a dramatic reduction in the pace of improvement. This slowdown started around 2011: between 2011 and 2018 life expectancy in England increased by just 0.5 years in males and 0.2 years in females, compared with 2.2 and 1.7 years respectively in the preceding seven-year period 2004 to 2011.
And there are some other adverse trends. Although male life expectancy has shown a steady if small rise since 2011, the virtual flat-lining of female life expectancy since then is particularly worrying given that UK’s life expectancy compares poorly with similar European countries, especially for females. Moreover, although improvements in life expectancy have slowed across much of Europe, the slowdown has been greatest in UK – hence UK’s life expectancy disadvantage relative to European counterparts has widened, especially for females.
The bad news doesn’t stop there. Although the slowdown in mortality improvements is affecting all types of neighbourhoods, the impact is greatest in more deprived areas, causing inequalities to widen. In the poorest ten percent of areas in England, female life expectancy actually fell by 100 days of life between 2012-14 and 2015-17, compared with an increase of 84 days in the least deprived areas.
Commenting on life expectancy trends in England, Public Health England noted “the overall slowdown in improvement is due to factors operating across a wide range of age groups, geographies and causes of death”, with decelerating mortality improvements from cardiovascular disease being one of the significant contributors. Public Health England’s analysis and other evidence to date suggests several factors are driving the slowdown, making longer-term trends less predictable.
In summary, life expectancy at birth among males continues to improve, but much more slowly since 2011; female life expectancy has been fluctuating annually and, overall, has changed little since 2011. But there are no signs yet of a trend of falling life expectancy, or that there will not be a return to improvement in the longer term.
That said, there is no room for complacency. Urgent action is needed to reinvigorate mortality improvements and reduce inequalities. The analyses to date provide clear pointers for action, but further investigations of the underlying drivers are needed. It is imperative that these are scientifically robust and that the facts about such salient trends in the health of the population are accurately reported.
Veena S Raleigh, senior fellow, The King’s Fund
Peter Goldblatt, senior advisor, UCL Institute of Health Equity
Competing interests: None declared
** ONS report small (second decimal place) declines in life expectancy for males and females in 2018 compared with 2017, but this is likely to be within year-on-year variability.