A Grauntian perspective on contemporary British mortality

1. A Grauntian perspective on contemporary British mortality Mike Murphy, London School of Economics Prepared for symposium: Mortality Past and Present: Celebrating the 350th anniversary of the publication of John Graunt's Bills of Mortality Gresham College, Barnard's Inn Hall, London 29 November 2012

2. Graunt’s legacy Mortality analysis & Life tables Sex differentials Cause of death analysis Trend analysis Cause of death classification Use of administrative data

3. Mortality analysis & Life tables: Period life expectancy at birth by sex

4. Years of life expectancy gained by age in subsequent decade, England and Wales, based on ONS dataMalesFemales

5. Further expectation of life by age and year reached age, based on ONS dataMalesFemales

6. Sex differentialin period life expectancy at birth

7. Cause of death analysis:Data quality " ... all dying thereof die so emaciated and lean (their Ulcers disappearing upon Death) that the Old-women Searchers after the mist of a cup of Ale, and the bribe of a two-groat fee, instead of one, given them, cannot tell whether this emaciation, or leanness were from a Phthisis, or from an Hectic Fever, Atrophy, &c or from an Infection of the Spermatick parts ... ". Graunt J. Natural and political observations mentioned in a following index, and made upon the Bills of Mortality. Facsimile edition. New York: Arno Press, 1975.

8. Leading causes of death by sex in 2011 & 2001, England and Wales

9. Early Epidemiological Transition? 14. The second observation is, that of the said 229,250 dying of all diseases, there died of acute diseases (the Plague excepted) but about 50,000 or 2/9 parts. 15. The third observation is, that of the said 229 thousand about 70 died of chronic diseases, which shews (as I conceive) the state and disposition of the country (including as well its food, as air) in reference to health, or rather to longevity ... the chronic diseases shew the ordinary temper of the place, so that upon the proportion of chronic diseases seems to hang the judgment of the fitness of the country for long life. Graunt J. Natural and political observations mentioned in a following index, and made upon the Bills of Mortality

10. Leading causes of death by sex and age group, England and Wales, 2009 Source: Office for National Statistics

11. Data quality 10. In brief, when any dead body is found in England, no algebraist, or uncipherer of letters, can use more subtle suppositions and variety of conjectures to find out the demonstration, or cipher, than every common unconcerned person doth to find out the murderers, and that for ever, until it be done. Graunt J. Natural and political observations mentioned in a following index, and made upon the Bills of Mortality

12. “There is only one cause of death at old ages ...” 7. … in case a man of 75 years old died of a cough (of which had he been free, he might have possibly lived to ninety) I esteem it little error (as to many of our purposes) if this person be in the table of casualties, reckoned among the Aged and not placed under the title of Coughs. Graunt J. Natural and political observations mentioned in a following index, and made upon the Bills of Mortality

13. Population Ageing 18. “In the foregoing observations we ventured to make a standard of the healthfulness of the air from the proportion of acute and epidemic diseases, and of the wholesomeness of the food from that of the chronic. Yet forasmuch as neither of them alone do shew the longevity of the inhabitants, we shall in the next place come to the more absolute standard, and correction of both, which is the proportion of the aged, viz. 15,757 to the total 229,250. That is of about 1 to 15 or 7 per cent.” “… I conceive, that in countries subject to great epidemic sweeps men may live very long, but where the proportion of the chronic distempers is great, it is not likely to be so; because men being long sick, and always sickly, cannot live to any great age …” Graunt J. Natural and political observations mentioned in a following index, and made upon the Bills of Mortality

14. Period life expectancy for men at 65: by NS-SEC, 1982 to 2006 Source: Longitudinal Study, Office for National Statistics

15. An optimistic view of the future? “In recent years, UK life expectancy at birth has been growing at a rate of around a year every five years. For males in the UK, it is the fastest growth in longevity anywhere in the world. If it carries on, British males could become the world’s longest-lived men a few years from now” Mosaic society: Complex changes in UK society bring with them policy challenges Professor Ian Diamond, Chief Executive, Economic and Social Research Council, 2007.

16. Life expectancy at birth, EU 2009MalesFemales Source: WHO HFA Database

17. Life expectancy at age 65, EU 2009MalesFemales Source: WHO HFA Database

18. A Potential Decline in Life Expectancy in the United States in the 21st Century “as a result of the substantial rise in the prevalence of obesity and its life-shortening complications such as diabetes, life expectancy at birth and at older ages could level off or even decline within the first half of this century.” (p. 1142) S. J. Olshansky, D J. Passaro, R C. Hershow, J Layden, B A. Carnes, J Brody, L Hayflick, R N. Butler, D B. Allison, and D S. Ludwig. New England Journal of Medicine 352;11 March 17, 2005 Ageing populations: the challenges ahead “If the pace of increase in life expectancy in developed countries over the past two centuries continues through the 21st century,most babies born since 2000 … [in] countries with long life expectancies will celebrate their 100th birthdays..” (p. 1139) Kaare Christensen, Gabriele Doblhammer, Roland Rau, James W Vaupel Lancet 2009; 374: 1196–1208. Expectations about life expectancy

19. Estimated & projected rate of improvement of SDR, England & Wales 1901-2050 Source: Based on ONS data

20. Estimated annual rate of improvement (%) of ASMR, England & Wales 1920-2010 MalesFemales Source: Based on HMD data

21. Hypothesised explanations for the ‘Golden cohorts’ include: • Differences in smoking patterns between generations • Better diet and environmental conditions during and after the Second World War • Those born in periods of low birth rate facing less competition for resources as they age • Benefits from the introduction in the late 1940s of the Welfare State • Benefits from medical advances which have increasingly affected older people. Karen Dunnell, 2008, Ageing and Mortality in the UK – National Statistician's Annual Article on the Population. Population Trends 134: 6-23, p. 19

22. Cohort life expectancy at birth by sex

23. In 2000, Smoking responsible for: Percent of deaths in causes Distribution of smoking deaths by cause & sex, all ages (%) Source: Peto, Lopez, Boreham, Thun (2006)

24. Does smoking provide the best explanation for observed patterns? There is strong cohort effect for lung cancer (& smoking) Cohort approaches provide a better basis for: elucidating role of smoking on deaths in England and Wales investigating cohort effects in adult mortality Analysis using 20th and 21st Century Mortality Database England and Wales of deaths by sex, age (35+ yrs) and cause of death in the period 1950-2007.

25. APC model To estimate the role of changing patterns of smoking overall on adult cohort mortality we use an APC formulated as: For any cause of death For details, see Carstensen, B. (2006). Age-period-cohort models for the Lexis diagram. Statistics in Medicine, 26:3018-3045. Murphy, M. & Di Cesare, M. (2012) Use of an age-period-cohort model to reveal the impact of cigarette smoking on trends in twentieth-century adult cohort mortality in England and Wales. Population Studies, ISSN 0032-4728

26. Model characteristics Demographically-informed model with constraints: • Age-specific rates for reference cohort • Relative risk for cohort (RRc), incorporates linear “drift” component, & constrained to value 1 for reference cohort • Relative risk for time period (RRp), constrained to average value of 1 over fitting range (& has no linear trend) • Age, period and cohort functions modelled as cubic splines

27. Age Period Cohort models for mortality: cohorts born 1885-1955, England and Wales

28. The approach • Most variation can be accounted for by age and cohort functions, i.e. the period function is approximately constant • Users are often concerned with the interpretation of current trends, i.e. rates of change

29. APC – Lung Cancer Cohort with the highest levels (females) Inflection point Cohort with the highest levels (males)

30. For these causes, the model simplifies to

31. Annual percentage rate of change of cohort relative risk for those born 1885-1955, England and Wales

32. Annual change in cohort standardised mortality rate for those born 1885-1955, England and WalesMalesFemales

33. Regression of annual change of total cohort mortality on change of lung cancer cohort mortality, actual and fitted values, for those born 1885-1955, England and Wales:

34. Conclusions: Improving mortality in Britain in Twentieth century Rate of improvement increasing to present level of just over 2% p.a. on average Large infant mortality improvements over whole period Initially little old age mortality improvement; recently increases at these ages Smoking-related mortality has an major contribution to explaining mortality change, both sex differentials and to cohort patterns The future?

35. Thank you

36. The final model We argue that the first term is cyclical (see earlier) and the second term is long-term monotonic (ETT), which we fit as a linear term leading to the final model for k: With estimates