ANALYSIS Ж News: An older, sicker US will need more specialists, study says (BMJ 2013;347:f6711) Ж Analysis: Japan’s answer to the economic demands of an ageing population (BMJ 2012;345:e6632) Ж News: Ageing population is not a drain on economy or NHS, says report (BMJ 2012;344:e2934) Ж Editorial: How to meet the challenge of ageing populations (BMJ 2011;342:d3815) POPULATION AGEING The timebomb that isn’t? Jeroen Spijker and John MacInnes argue that current measures of population ageing are misleading and that the numbers of dependent older people in the UK and other countries have actually been falling in recent years opulation ageing is a concern in all tive to those who pay for them. Although the Contribution of age group developed countries. For the first phased raising of the state pension age (from ≥ years - years - years time, there are now more people over 65 for men and 60 for women) to 68 for both . the age of 65 in the United Kingdom sexes, which will keep 3.5 million people in . than there are children under 15 the working age, will initially cause the ratio . Pyears.1 Over the past century, the proportion to decline from today’s 31% in England and . of over 65s has grown from about one in 20 to Wales, by 2035 it will have risen to 37%. by decade in years . around one in six. Although declining birth However, the old age dependency ratio is rates and infant mortality formed the basis for a poor measure of the burden of an ageing . Increase in life expectancy this growth from the end of the 19th century population. It counts neither the number of s s s s s s until the second world war, since the 1970s dependent older people nor the number who Decade increasing life expectancy has been an addi‑ sustain them. It merely takes a cut‑off point tional driving force (fig 1). This population (the state pension age) and assigns adults to Fig 1 | Change in life expectancy (in years) in ageing has worried policy makers because for the two sides of the ratio accordingly. This England and Wales within each decade broken every worker paying tax and national insur‑ might be a useful rule of thumb if the relative down by age group, 1950-2010 (data from www. ance there are more older citizens, who make size of these two age groups tracked the vol‑ mortality.org). Life expectancy increased by 2.6 years between 2000 and 2010, 0.1 years of which greater demands on social insurance, health, ume of old age dependency, but it does not. came from improved infant and child survival and welfare systems and have increasing mor‑ We propose an alternative measure that gives and, respectively, 0.5 and 2.0 years from lower 2‑4 bidity and disability. a more accurate and very different picture mortality among 5-64 and ≥65 year olds The standard indicator of population ageing and consider the implications of our results is the old age dependency ratio. It takes the for health policy. ulation did not climb above 25 until the start number of people who have reached the state of the last century, since when it has risen to pension age and divides it by the number of Counting dependent older people 40. Life expectancy has also increased across working age (16‑64 years) adults in order to Since the main factor behind the ageing pop‑ this period, so that although the median age estimate the proportion of older people rela‑ ulation is increasing life expectancy, age is of 24 in 1900 carried a life expectancy of 39 a poor measure of its burden. In 1900 mean more years, those at the median age of 40 in KEY MESSAGES life expectancy for a 65 year old woman in 2009 could expect to live a further 42 years. The extent, speed, and effect of population England and Wales was 11 years. Today she In aggregate terms, the population of 2009, ageing have been exaggerated because the could expect to live another 21 years (10 and despite being much older as measured by standard indicator—the old age dependency 18 years respectively for men). We can best cap‑ years lived, was nevertheless younger than ratio—does not take account of falling mortality ture this changing importance of age by realis‑ that of 1900 in terms of years left. This is cru‑ When measured using remaining life expectancy, old age dependency turns out ing that the age of a population comprises two cial, because many behaviours and attitudes to have fallen substantially in the UK and components: the years lived of its members (including those related to health) are more elsewhere over recent decades and is likely to (their ages) and their years left (remaining life strongly linked to remaining life expectancy stabilise in the UK close to its current level expectancies). than to age.6‑8 The capacity of healthcare systems to cope In a period of lengthening lifespans, not The old age dependency ratio defines all with increasing longevity will depend on the only does the average age of the population people above the statutory pension age as changing relationship between morbidity and increase, so too does the remaining life expec‑ dependent, regardless of their economic, remaining life expectancy and, in particular, tancy associated with each age.5 This has social, or medical circumstances. This over‑ the effect of education substantial effects. The median age of the pop‑ looks the fact that rising life expectancy makes 20 BMJ | 16 NOVEMBER 2013 | VOLUME 347 ANALYSIS % of population in age groups with Real elderly Old age % of population aged ≥ remaining life expectancy ≤ years dependency ratio dependency ratio % of population in age groups with UK Old age dependency ratio (%) Italy remaining life expectancy ≤ years Old age dependency ratio adjusted for Germany Old age dependency ratio (%) pension age (%) France Real elderly dependency ratio (%) USA Japan Percentage Ratio (%) Percentage Year Year Year Fig 2 | Proportion of the population aged ≥65, Fig 3 | Old age dependency ratios, proportion of the Fig 4 | Real elderly dependency ratio and old age proportion of the population at ages with population at ages with remaining life expectancy dependency ratios in various countries, remaining life expectancy of ≤15 years, and the old of ≤15 years, and the real elderly dependency ratio, 1950-2010 (data from www.mortality.org and age dependency ratio, England and Wales 1900- England and Wales 1950-2050 http://stats.oecd.org) (data from www.mortality.org and England and Wales censuses (www. 2011 (see appendix on bmj.com for calculations) ons.gov.uk)). See appendix on bmj.com for details of calculations these older people “younger,” healthier, and employed, irrespective of age, for the denomi‑ third over the past four decades (fig 3). Look‑ fitter than their peers in earlier cohorts. Many nator in our calculation. ing into the future, the old age dependency have accumulated substantial assets. Currently ratio will continue to rise, even if we adjust for over one million are still working, mostly part Different picture of ageing the planned changes in the state pension age. time, many with valuable experience or spe‑ According to the standard old age dependency However, the real elderly dependency ratio is cialist knowledge. The spending power of the ratio, for every person aged ≥65 there were 7‑8 set to fall further, stabilise for several years “grey pound” has risen inexorably. Many do adults of working age (15‑64) until about 1910 then gradually increase, although at no point volunteer work vital to the third sector or look (fig 2). The ratio subsequently increased almost through to 2050 will it regain the levels expe‑ after grandchildren. We know that most acute linearly until about 1980, when there were just rienced for most of the last century. Moreover, medical care costs occur in the final months of under four adults of working age for every older our projection is a conservative one, since we life, with the age at which these months occur person. These data underpin the population have held employment rates constant over having little effect.9 10 At least some forms of ageing debate, which typically concludes that time, when they are likely to increase because disability are being postponed to later ages. the increasing burden of elder dependency will of the rise in the state pension age, disincen‑ Good data on population health by age are strain health and social care systems, limit the tives to early retirement, and further progress available only for the last decade, but remain‑ aspirations of universal healthcare, and require on sex equality. It is thus probable that our ratio ing life expectancy data are a robust substitute raising the state pension age to limit the costs will stabilise near its current level. because they provide a more accurate picture to the welfare state. of the extent of ageing. Therefore, following However, if we define the dependent older International comparisons Sanderson and Scherbov and others5 10‑ 14 we population as people with a remaining life Figure 4 shows how our new measure and use a remaining life expectancy of ≤ 15 years expectancy of ≤15 years, the trend is differ‑ the conventional old age dependency ratio as the threshold of dependency in our calcula‑ ent. Figure 2 shows that from the late 1970s compare over the past half century for several tions, rather than a fixed age boundary, and do improvements in old age mortality have countries. While the old age dependency ratio so for each sex separately.