Page 1
REVES
27 May 2014
Pathways to Longevity: Health and Functioning Trajectories
of Exceptional Survivors
Jennifer Ailshire,1 Hiram Beltrán-Sánchez, 2 and Eileen Crimmins, 1 1School of Gerontology, University of Southern California, USA
2 Center for Demography and Ecology, University of Wisconsin-
Madison, USA
Page 2
“For everyone in his or her thirties
and younger, especially children,
life-spans of 95 or 100 years will
be common.”
~ Vaupel, 2000
“100 Plus: How the coming age of
longevity will change everything…”
~ Arrison, 2011
The “Longevity Revolution”
“90-Plus Is the New Old
Age, Are You Ready?” ~
Forbes, 2011
“Living to 100 and
Beyond” ~ Wall
Street Journal, 2011
Page 3
Increasing Age at Death
Bell and Miller. 2005. Life Tables for the US SSA 1900-2100.
Page 4
The Population of Exceptional Survivors
(90+) in the U.S. is Growing
Page 5
“Increased longevity without quality of life is an empty
prize” (World Health Organization 1997)
Is longevity accompanied by healthy/unhealthy years?
What are the health and functioning profiles of the
oldest-old?
What characteristics predict maintaining health and
functioning into oldest-old age?
The longevity revolution raises questions
about content/quality of added years
Page 6
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
65 67 69 71 73 75 77 79 81 83 85 87 89 91 93 95 97 99
1900 Birth Cohort - % Surviving at Ages 65+
What is “Exceptional Longevity”
Percent survival by age calculated from cohort-based life tables: Bell and Miller. 2005. “Life Tables for the
United States Social Security Area 1900-2100.” Actuarial Report No. 120.
0%
1%
2%
3%
4%
5%
90 91 92 93 94 95 96 97 98 99 100
Page 7
PATHS TO LONGEVITY
Pro
bab
ilit
y o
f D
isease
On
set/
Lo
ss o
f F
un
cti
on
Time (Age)
Hypothetical
Expansion
Compression/
delay
Page 8
Long-lived experience expansion of unhealthy years
High levels of frailty & dysfunction (e.g., loss of physical
and cognitive function), multimorbidity1-3
Longevity is predicated on compression
Increased resilience/decreased susceptibility to age-
related diseases4,5
“The longer you live the healthier you’ve been”6
Relatively less attention to within-population
heterogeneity
What does existing evidence say?
1Anderson-Ranberg et al. 1999; 2Anderson-Ranberg et al. 2001; 3Baltes and Smith 2003; 4Barzilai and
Shuldiner 2001; 5Cutler 1975; 6Perls 1999
Page 9
Study description
National study of U.S. adults 70+
Interviewed in 1993, 1995, 1998, every 2 years to
2010 (up to 18 years of follow-up)
Analytic sample
Born 1921 or earlier (potential to survive to age 90)
Exceptional Survivors - survived to at least age 90
(N=2,740) Baseline age: Mean= 81; Range =73-97 (99th percentile)
Death tracking through 2010
Using exit interviews, NDI, SS Master Death File
The Health and Retirement Study (HRS)
Page 10
Baseline Demographic Characteristics by
Survivorship
N
Age, years 77.1 80.9
Female 55.9 68.9
Race/Ethnicity
White 86.7 88.4
Black 8.7 7.0
Hispanic 3.4 3.6
Other 1.3 1.0
Education, years 10.9 10.9
Non-Survivors Survivors
3,543 2,740
Survivorship differences in covariate distributions are
significant at p < .05
Page 11
Count of doctor diagnosed diseases (0-6)
“Has a doctor ever told you
that you have…?”
1) High blood pressure or
hypertension
2) Heart problems
3) Diabetes
4) Cancer
5) Stroke
6) Lung disease
Count of ADL limitations (0-6)
“Because of a health or
memory problem do you
have any dif ficulty with…”
1) Walking across a room
2) Dressing
3) Bathing
4) Eating
5) Bed transfers
6) Toileting
Disease and Disability
Page 12
Novel approach for examining heterogeneity in the
aging experience
Identifies clusters of individuals who have similar
aging trajectories
Accounts for selective mortality
Group-Based Trajectory Model
𝑃 𝑌𝑖 = 𝜋𝑗 𝑝𝑗𝑡 (𝑦𝑖𝑡)
𝑇
𝑗
𝑦𝑖𝑡∗ = 𝛽0
𝑗+ 𝛽1
𝑗𝑎𝑔𝑒𝑖𝑡 + 𝛽2
𝑗𝑎𝑔𝑒𝑖𝑡
2 + . . + 𝜀𝑖𝑡
Each trajectory consists of an initial value β0, linear age
term β1, quadratic age2 term β2
Page 13
Delayers
Escapers
Survivors
(high)
Survivors
(low)
Predicted Disease Trajectories 0
12
34
5
Num
ber
of C
om
orb
iditie
s
72 82 92 1028074 76 78 84 88 90 94 96 98 10086
Age
Escapers, 18.6%
Delayers, 25.2%
Survivors (low), 40.3%
Survivors (high), 15.9%
01
23
45
Num
ber
of C
om
orb
iditie
s
72 82 92 1028074 76 78 84 88 90 94 96 98 10086
Age
Escapers, 18.6%
Delayers, 25.2%
Survivors (low), 40.3%
Survivors (high), 15.9%
01
23
45
Num
ber
of C
om
orb
iditie
s
72 82 92 1028074 76 78 84 88 90 94 96 98 10086
Age
Escapers, 18.6%
Delayers, 25.2%
Survivors (low), 40.3%
Survivors (high), 15.9%
01
23
45
Num
ber
of C
om
orb
iditie
s
72 82 92 1028074 76 78 84 88 90 94 96 98 10086
Age
Escapers, 18.6%
Delayers, 25.2%
Survivors (low), 40.3%
Survivors (high), 15.9%
Page 14
Women - survivors, delayers vs. Men - escapers
Hispanics less likely to live
with disease. oldest -old age.
No education or black-white
dif ferences.
Former smokers more likely
to live with disease compared
to non-smokers.
Overweight/obese have
greater disease burden throughout oldest -old age compared to
normal weight at baseline.
Demographic and behavioral correlates
of survival trajectories
01
23
45
Num
ber
of C
om
orb
iditie
s
72 82 92 1028074 76 78 84 88 90 94 96 98 10086
Age
Escapers, 18.6%
Delayers, 25.2%
Survivors (low), 40.3%
Survivors (high), 15.9%
Page 15
No dif ferences by region.
Older adults living in large
metro areas at baseline more
l ikely to escape disease.
Geographic differences in survival
trajectories
01
23
45
Num
ber
of C
om
orb
iditie
s
72 82 92 1028074 76 78 84 88 90 94 96 98 10086
Age
Escapers, 18.6%
Delayers, 25.2%
Survivors (low), 40.3%
Survivors (high), 15.9%
Page 16
Delayers
(short)
Delayers
(long)
Escapers/
(Recoverers)
Predicted Disability Trajectories 0
24
6
Num
ber
of A
DL
Lim
itations
72 82 92 1028074 76 78 84 86 88 90 94 96 98 100
Age
Delayers (long), 18.6%
Delayers (short), 25.2%
Survivors, 40.3%
Escapers, 15.9%
02
46
Num
ber
of A
DL
Lim
itations
72 82 92 1028074 76 78 84 86 88 90 94 96 98 100
Age
Delayers (long), 18.6%
Delayers (short), 25.2%
Survivors, 40.3%
Escapers, 15.9%
02
46
Num
ber
of A
DL
Lim
itations
72 82 92 1028074 76 78 84 86 88 90 94 96 98 100
Age
Delayers (long), 18.6%
Delayers (short), 25.2%
Survivors, 40.3%
Escapers, 15.9%
02
46
Num
ber
of A
DL
Lim
itations
72 82 92 1028074 76 78 84 86 88 90 94 96 98 100
Age
Delayers (long), 18.6%
Delayers (short), 25.2%
Survivors, 40.3%
Escapers, 15.9%
Survivors
Page 17
Women - l ive through very old age with increasing burden of
disability vs. Men - delay disability into their 90’s.
Hispanics and less educated e
more likely to live with
increasing disability through
oldest-old age.
No black-white dif ferences.
No dif ferences by smoking
status.
Overweight/obese have
much greater odds of living
through oldest-old age with increasing disability burden.
Demographic and behavioral correlates
of survival trajectories
02
46
Num
ber
of A
DL
Lim
itations
72 82 92 1028074 76 78 84 86 88 90 94 96 98 100
Age
Delayers (long), 18.6%
Delayers (short), 25.2%
Survivors, 40.3%
Escapers, 15.9%
Page 18
More delayers in South. Less
probability of survivors
living in Western U.S.
(mountains).
More short delayers in non-
metro urban areas. More
survivors in small
urban/rural areas.
Geographic differences in survival
trajectories
02
46
Num
ber
of A
DL
Lim
itations
72 82 92 1028074 76 78 84 86 88 90 94 96 98 100
Age
Delayers (long), 18.6%
Delayers (short), 25.2%
Survivors, 40.3%
Escapers, 15.9%
Page 19
As we march towards longer life
spans, what are we marching
towards?
Longevity is a prize… for some
Potential modifiable risk factors
associated with poor health and loss
of function in very late life
Excess body weight
Behaviors of exceptional survivors not necessarily
what we would expect (e.g., smokers)
Not much social patterning in health and functioning
Conclusions
Page 21
Age at death of 90+ survivors
05
10
15
Pe
rcent
90 95 100 105 110Age at Death
Page 22
Time in Study 90+ survivors
05
10
15
20
Pe
rcent
0 2 4 6 8 10 12 14 16 18Years in Study Interval
Page 23
Life Expectancy at Birth 1900-2010
47.3
68.2
78.7
40
45
50
55
60
65
70
75
80
85
90
1900 1950 1960 1970 1975 1980 1985 1990 1995 2000 2005 2010
Page 24
Example: Morbidity profiles of oldest -old members
in New England Centenarian Study (ages 97-119):
Paths to longevity (N = 424)
(Evert et al. 2003)
Categorization is ad hoc; unclear if patterns
represent larger U.S. population
Focused only on age of disease onset
Heterogeneity in Health Patterns of
Survivorship to Extreme Old Age
Age of Disease Onset Men (%) Women (%)
Escapers 100+ OR never 24 43
Delayers 80-100 44 42
Survivors <80 32 25
Page 25
Determining Number of Classes
16
000
17
000
18
000
19
000
20
000
21
000
BIC
.85
.86
.87
.88
.89
.9
En
tro
py
2 3 4 5 6Solution: Number of Classes
Entropy BIC
Page 26
Baseline Disease Prevalence by
Morbidity Pathway
0%
10%
20%
30%
40%
50%
60%
70%
80%
Hypertension HeartDisease
Stroke Lung Disease Diabetes Cancer
Survivor (low) Survivor (high)
Page 27
Baseline ADL Prevalence by
Disability Pathway
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
50%
Walking Dressing Bathing Bedding Toileting Eating
Recoverer Survivor
Page 28
Lifetime Disease Prevalence by
Morbidity Pathway
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Hypertension HeartDisease
Stroke Lung Disease Diabetes Cancer
Escaper Delayer Survivor (low) Survivor (high)
Page 29
Race/Ethnic Composition
01
23
45
Num
ber
of C
om
orb
iditie
s
72 82 92 1028074 76 78 8488 88 90 94 96 98 100
Age
Escapers, 18.6%
Delayers, 25.2%
Survivors (low), 40.3%
Survivors (high), 15.9%
Escapers
Delayers
Survivors
(low)
Survivors
(high)
Page 30
01
23
45
Num
ber
of C
om
orb
iditie
s
72 82 92 1028074 76 78 8488 88 90 94 96 98 100
Age
Escapers, 18.6%
Delayers, 25.2%
Survivors (low), 40.3%
Survivors (high), 15.9%
Baseline Smoking Status
• Current smokers more likely to escape or delay
Escapers
Delayers
Survivors
(low)
Survivors
(high)
Page 31
01
23
45
Num
ber
of C
om
orb
iditie
s
72 82 92 1028074 76 78 8488 88 90 94 96 98 100
Age
Escapers, 18.6%
Delayers, 25.2%
Survivors (low), 40.3%
Survivors (high), 15.9%
Baseline Weight Status
Escapers
Delayers
Survivors
(low)
Survivors
(high)
Page 32
Lifetime ADL Prevalence by
Disability Pathway
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
Walking Dressing Bathing Bedding Toileting Eating
Recoverer Delayer(long) Delayer(short) Survivor
Page 33
02
46
Num
ber
of A
DL
Lim
itations
72 82 92 1028074 76 78 84 86 88 90 94 96 98 100
Age
Delayers (long), 18.6%
Delayers (short), 25.2%
Survivors, 40.3%
Escapers, 15.9%
Gender Differences in Trajectories
71% Female
61% Female
74% Female
79% Female
Page 34
02
46
Num
ber
of A
DL
Lim
itations
72 82 92 1028074 76 78 84 86 88 90 94 96 98 100
Age
Delayers (long), 18.6%
Delayers (short), 25.2%
Survivors, 40.3%
Escapers, 15.9%
Race/Ethnic Composition
Escapers
Delayers
(long)
Delayers
(short)
Survivors
Page 35
02
46
Num
ber
of A
DL
Lim
itations
72 82 92 1028074 76 78 84 86 88 90 94 96 98 100
Age
Delayers (long), 18.6%
Delayers (short), 25.2%
Survivors, 40.3%
Escapers, 15.9%
Baseline Smoking Status
Escapers
Delayers
(long)
Delayers
(short)
Survivors
Page 36
02
46
Num
ber
of A
DL
Lim
itations
72 82 92 1028074 76 78 84 86 88 90 94 96 98 100
Age
Delayers (long), 18.6%
Delayers (short), 25.2%
Survivors, 40.3%
Escapers, 15.9%
Baseline Weight Status
Escapers
Delayers
(long)
Delayers
(short)
Survivors