1. WP 2:Future disease patterns and their implications for disability in later life C. Jagger, R. Matthews, J. Lindesay Leicester Nuffield Research Unit

2. Dynamic simulation model Based on MRC Cognitive Function and Ageing Study (MRC CFAS) Has two stages: –Transition builds on earlier work modelling the impact of diseases on the onset of disability and death (Spiers et al 2005) –Projection applies transition rates to ‘age’ the population

3. Proposed work for WP2 Four strands Gender-specific projections –Refitting models separately by gender Projections of DFLE –Add DFLE as output to projections Range of measures of disability –Hierarchy of FL/IADL/ADL Further scenarios - diseases and ethnic minorities –Literature review diabetes, ethnic minorities

4. Proposed work for WP2 Four strands Gender-specific projections –Refitting models separately by gender Projections of DFLE –Add DFLE as output to projections Range of measures of disability –Hierarchy of FL/IADL/ADL Further scenarios - diseases and ethnic minorities –Literature review diabetes, ethnic minorities

5. Changes made to the simulation program Disease prevalences: Modelling of disease prevalence by five year age group Extrapolation of prevalence model to estimate disease prevalence by two year age group GAD mortality improvements Now included more systematically

6. Allowing gender-specific projections Currently being tested Allows different disease prevalence at baseline for males and females Assumes disease prevalence trends by age same for males and females Baseline transition probabilities are assumed to be the same for males and females (small numbers prevents modelling separately for each gender) Allows different assumptions about change in prevalence and transition probabilities for males and females.

7. Proposed work for WP2 Four strands Gender-specific projections –Refitting models separately by gender Projections of DFLE –Add DFLE as output to projections Range of measures of disability –Hierarchy of FL/IADL/ADL Further scenarios - diseases and ethnic minorities –Literature review diabetes, ethnic minorities

8. Scenarios - Ageing alone Age-specific prevalence of diseases is constant prevention strategies and effective treatments simply offset the negative influences of obesity and other cohort trends Incidence of and recovery rates to dependency remain the same with no further effect of treatments Mortality rates continue as GAD principal projections

9. Ageing alone – total population 44% increase from 2006 to 2026 80% increase from 2006 to 2026

10. Ageing of the population – disabled population 86% increase from 2006 to 2026 127% increase from 2006 to 2026

11. Ageing alone – LE and DFLE %DFLE/LE 90% 86% 85% 80% 73% 66%

12. Scenarios Improving population health –decline in risk factors, particularly smoking and obesity –new treatments or technologies emerge to reduce the disabling effects of arthritis, dementia, stroke and CHD and make further gains in survival Poorer population health –obesity trends of 2% increase annually continue increasing prevalence of arthritis, stroke and CHD –Treatments continue to focus on reducing the mortality from diseases rather than reducing the disabling effects Disease specific –Reduction in prevalence of stroke, CHD, arthritis and cognitive impairment of 1% every 2 years

13. LE and DFLE at age 65 in 2006 and 2026 %DFLE/LE 90% 86% 90% 86% 90% 87% 90% 86%

14. LE and DFLE at age 85 in 2006 and 2026 %DFLE/LE 73% 66% 72% 64% 73% 69% 73% 67%

15. Proposed work for WP2 Four strands Gender-specific projections –Refitting models separately by gender Projections of DFLE –Add DFLE as output to projections Range of measures of disability –Hierarchy of FL/IADL/ADL Further scenarios - diseases and ethnic minorities –Literature review diabetes, ethnic minorities

16. Harmonisation of disability All WPs use different surveys for disability –CFAS, BHPS, GHS, ELSA, FRS Calculated age and standardised prevalence at 65 and 75 –With cutpoints difficulty and help –Investigated proxy respondents

17. Harmonisation of disability - future PCA of items in Individual surveys (BHPS, ELSA, GHS) to determine dimensionality Mokken analysis of scale items to determine hierarchy –by gender, to check for differences (discard items that do not work the same for males and females), then for all –Repeat for those living alone only, to check ordering of items is consistent Same analysis for CFAS, but also include: –With and without those in institutions –Wave 0 v. wave 10 –Longitudinal analysis

18. Harmonisation of disability - future Keep items that fit hierarchy and are same by gender Compare across studies with age standardised prevalence to determine suitable cut point Compare with measures others are presently using (eg. measure used in WP5)