1.  The 2003 Healthcare Conference Actuaries Adding Value 5-7 October 2003 Scarman House, The University of Warwick

2. Workshop Session C1 : Risk Factors in Insured Lives Mortality and Critical Illness Experience nNeil Robjohns nRajeev Shah

3. Risk Factors in Insured Lives Mortality and Critical Illness Experience A comparison of Critical Illness versus Mortality n Historic Trends n Pointers to Future Trends n Select Patterns n Causes of Claim n The impact of claim delays and IBNR

4. Risk Factors in Insured Lives Mortality and Critical Illness Experience A comparison of Critical Illness versus Mortality n Historic Trends n Pointers to Future Trends n Select Patterns n Causes of Claim n The impact of claim delays and IBNR

5. Historic Trends n Population trends n Impact of falling prevalence of smoking n Variations by socio-economic group

6. Summary of Trends in CI Incidence and Mortality Best Estimate Avg Change % pa, England & Wales, 1980-2000 Men, aged 40 - 60 Size of balls represents relative importance of CI measured by incidence rates

7. Summary of Trends in CI Incidence and Mortality Best Estimate Avg Change % pa, England & Wales, 1980-2000 Men, aged 40 - 60 Size of balls represents relative importance of CI measured by incidence rates

8. Summary of Trends in CI Incidence and Mortality Best Estimate Avg Change % pa, England & Wales, 1980-2000 Women, aged 40 - 60 Size of balls represents relative importance of CI measured by incidence rates

9. Summary of Trends in CI Incidence and Mortality Best Estimate Avg Change % pa, England & Wales, 1980-2000 Women, aged 40 - 60 Size of balls represents relative importance of CI measured by incidence rates

10. Overview of CI Trends research Impact of Changes in Smoking Prevalence Rough estimates of the risks faced by ‘typical’ smokers relative to those who have never smoked are : n Heart Attack risk - 2 to 3 times higher n Stroke risk - 2 to 4 times higher n Overall Cancer risk - around double n Lung Cancer risk - 10 to 15 times higher

11. Summary of Trends in CI Incidence and Mortality Rough Estimate Avg Change % pa, England & Wales, 1980-2000 Men, aged 40 - 60

12. Summary of Trends in CI Incidence and Mortality Rough Estimate Avg Change % pa, England & Wales, 1980-2000 Women, aged 40 - 60

13. Summary of Trends in CI Incidence and Mortality Estimates for 40 – 60 age group, England & Wales, 1980-2000 At aggregate population level : n Mortality rates have fallen 2½%pa for men, 2%pa for women. n CI incidence fell 1%pa for men, but has risen ½%pa for women. So, for smoker-segregated rates in the 1990’s the picture looks far worse : n CI incidence rose by ½ to 1%pa for men n CI incidence rose by 1 to 1½%pa for women. n Mortality rates fell by 2%pa for men, 1½%pa for women. But : n Changes in smoking prevalence accounted for falls of a little over 1%pa for men and ½%pa for women in the 1980’s. n This effect is waning and was much smaller in the 1990’s. n Even after allowing for this smoking prevalence effect, trends for the 1990’s were worse than for the 1980’s.

14. Relative CI Rates by Deprivation Category Scotland, 1989 – 93, Ages 40 – 59, Males

15. Relative CI Rates by Deprivation Category Scotland, 1989 – 93, Ages 40 – 59, Females

16. Summary of Trends in CI Incidence and Mortality Avg Change % pa, Scotland, Population & Insured (Dep Cat 1 - 3) Men, aged 40 – 60, 1981 - 2000 Size of balls represents relative importance of CI measured by incidence rates

17. Summary of Trends in CI Incidence and Mortality Avg Change % pa, Scotland, Population & Insured (Dep Cat 1 - 3) Men, aged 40 – 60, 1981 - 2000 Size of balls represents relative importance of CI measured by incidence rates

18. Summary of Trends in Cancer Incidence and Mortality Avg Change % pa, Scotland, Population & Insured (Dep Cat 1 - 3) Men, aged 40 – 60, 1981 - 2000 Size of balls represents relative importance of cancer site measured by incidence rates

19. Summary of Trends in Cancer Incidence and Mortality Avg Change % pa, Scotland, Population & Insured (Dep Cat 1 - 3) Men, aged 40 – 60, 1981 - 2000 Size of balls represents relative importance of cancer site measured by incidence rates

20. Summary of Trends in CI Incidence and Mortality Avg Change % pa, Scotland, Population & Insured (Dep Cat 1 - 3) Women, aged 40 – 60, 1981 - 2000 Size of balls represents relative importance of CI measured by incidence rates

21. Summary of Trends in CI Incidence and Mortality Avg Change % pa, Scotland, Population & Insured (Dep Cat 1 - 3) Women, aged 40 – 60, 1981 - 2000 Size of balls represents relative importance of CI measured by incidence rates

22. Summary of Trends in Cancer Incidence and Mortality Avg Change % pa, Scotland, Population & Insured (Dep Cat 1 - 3) Women, aged 40 – 60, 1981 - 2000 Size of balls represents relative importance of cancer site measured by incidence rates

23. Summary of Trends in Cancer Incidence and Mortality Avg Change % pa, Scotland, Population & Insured (Dep Cat 1 - 3) Women, aged 40 – 60, 1981 - 2000 Size of balls represents relative importance of cancer site measured by incidence rates

24. Summary of Trends in CI Incidence and Mortality Population & Insured (Dep Cat 1 - 3) Estimates for 40 – 60 age group, Scotland, 1980-2000 For the ‘insured group’ relative to the population : n Rates for heart attack and stroke are around 75% to 80%. n Rates for lung cancer are around 65% to 70%. n Rates for other cancers are around 95% to 110%. So, n Cancer is a larger part of total cost for insured lives than population. And for trends on each major CI, the ‘insured group’ : n Fared better in the 1980’s (bigger falls, lesser rises). n Fared worse in the 1990s’.

25. Risk Factors in Insured Lives Mortality and Critical Illness Experience A comparison of Critical Illness versus Mortality n Historic Trends n Pointers to Future Trends n Select Patterns n Causes of Claim n The impact of claim delays and IBNR

26. Simplistic Model of Disease Timeline

27. Simplistic Model of Disease Timeline Prevention - Possible Outlook for Trends n Generally positive for both Mortality and CI, n… but beware lifestyle changes and obesity. n Likely steady trend, n… but possible leaps forward (eg: Genetics, Polypill, ??). n Positive impact of falling prevalence of smoking will run out of steam. n Reduced impact on insured lives ? n Little impact on cancer ? n Strong positive impact of prevention measures on heart attack rates over last 20 years.

28. Simplistic Model of Disease Timeline Detection - Possible Outlook for Trends n Almost all negative for CI (earlier claims, less severity), n But positive for mortality, n Probable step change / shocks. n Even if not developed into national screening programmes, unleashes clear scope for anti-selective behaviour by insureds. n CI definitions without a severity underpin are clearly vulnerable.

29. Simplistic Model of Disease Timeline Treatment - Possible Outlook for Trends n Expect improvements in mortality / morbidity, but after CI claim event. n May reduce CI rates where CI definitions have a credible, effective severity underpin. n If morbidity post CI-event is reduced, benefits become nearer to “windfall” status.

30. Pointers to Future Trends n Consider how our interests are aligned with medical advances and social change n Outlook for mortality remains very positive n But impact of medical advance on CI is far less clear nPotential steady gains from prevention measures, nBut likely step-increases from disease / event detection, nPossible gains from treatment mainly accrue to insureds under current definitions.

31. Risk Factors in Insured Lives Mortality and Critical Illness Experience A comparison of Critical Illness versus Mortality n Historic Trends n Pointers to Future Trends n Select Patterns n Causes of Claim n The impact of claim delays and IBNR

32. Select Patterns n Population Comparisons n Male / Female Differentials n Initial Selection n Smoker Selection n Selection by Distribution Channel

33. Comparison against population rates for Males

34. Comparison against population rates for Females

35. Male-Female Differentials

36. Initial Selection

38. Smoker differentials

39. 100A/Es Split by Sales Channel - Males

40. 100A/Es Split by Sales Channel - Females

41. Risk Factors in Insured Lives Mortality and Critical Illness Experience A comparison of Critical Illness versus Mortality n Historic Trends n Pointers to Future Trends n Select Patterns n Causes of Claim n The impact of claim delays and IBNR

42. Causes of Claim n 100 A/E n Major Causes

43. Claims by Cause - Males Critical Illness Groupings Cancer Heart Attack Stroke Kidney Failure All* All neoplasms Heart Attack & CABG Stroke Kidney Failure The above & MS, MOT, TPD Cause of Death Groupings Cancer Heart Attack Stroke Kidney Failure All** All neoplasms Acute MI, Other IHD & HD & Other Circulatory Cerebrovascular & Hypertensive Cirrhosis of liver & Nephritis/Nephrosis The above & other causes analysed

44. Claims by Cause - Females Critical Illness Groupings Cancer Heart Attack Stroke Kidney Failure All* All neoplasms Heart Attack & CABG Stroke Kidney Failure The above & MS, MOT, TPD Cause of Death Groupings Cancer Heart Attack Stroke Kidney Failure All** All neoplasms Acute MI, Other IHD & HD & Other Circulatory Cerebrovascular & Hypertensive Cirrhosis of liver & Nephritis/Nephrosis The above & other causes analysed

45. 100A/Es - Critical Illness vs Life Assurance - Males 100A/E Cause of Claim / Death Life Assurance Critical Illness

46. 100A/Es - Critical Illness vs Life Assurance - Females 100A/E Cause of Claim / Death Life Assurance Critical Illness

47. Claims by Cause as percentage of All Claims Critical Illness vs Life Assurance - Males % Cause of Claim / Death Life Assurance Critical Illness

48. Claims by Cause as percentage of All Claims Critical Illness vs Life Assurance - Females % Cause of Claim / Death Life Assurance Critical Illness

49. Risk Factors in Insured Lives Mortality and Critical Illness Experience A comparison of Critical Illness versus Mortality n Historic Trends n Pointers to Future Trends n Select Patterns n Causes of Claim n The impact of claim delays and IBNR

50. IBNR n Diagnosis to Notification n Diagnosis to Settlement n Diagnosis to Settlement*

51. Diagnosis – Notification (By Benefit Type) Cumulative Claim (%) Delay Stand Alone Full Acceleration

52. Diagnosis – Notification (By Sex) Cumulative Claim (%) Delay Males Females

53. Diagnosis – Notification (By Cause) Cumulative Claim (%) Delay CABG Cancer Heart Attack Multiple Sclerosis Stroke TPD

54. Diagnosis – Settlement (By Benefit Type) Cumulative Claim (%) Delay Stand Alone Full Acceleration

55. Diagnosis – Settlement (By Sex) Cumulative Claim (%) Delay Males Females

56. Diagnosis – Settlement (By Cause) Cumulative Claim (%) Delay CABG Cancer Heart Attack Multiple Sclerosis Stroke TPD

57. Diagnosis - Settlement* (By Benefit Type) Cumulative Claim (%) Delay Stand Alone Full Acceleration

58. Diagnosis – Settlement* (By Sex) Cumulative Claim (%) Males Females Delay

59. Diagnosis - Settlement* (By Cause) Cumulative Claim (%) CABG Cancer Heart Attack Multiple Sclerosis Stroke TPD Delay

60. Diagnosis - Settlement* (By Cause – Amounts) Delay CABG Cancer Heart Attack Multiple Sclerosis Stroke TPD Cumulative Claim (%)

61. Workshop Session C1 : Risk Factors in Insured Lives Mortality and Critical Illness Experience nNeil Robjohns nRajeev Shah nQuestions and Discussion