Social vulnerability and survival
22. Inzitari M, Newman AB, Yaffe K et al. Gait speed predicts
decline in attention and psychomotor speed in older adults:
the health aging and body composition study.
Neuroepidemiology 2007; 29: 156–62.
23. Kang HG, Dingwell JB. Separating the effects of age and
walking speed on gait variability. Gait Posture 2008; 27:
572–7.
24. Lamoureux E, Sparrow WA, Murphy A, Newton RU. The
effects of improved strength on obstacle negotiation in
community-living older adults. Gait Posture 2003; 17:
273–83.
25. Morrison MW, Gregory RJ, Paul JJ. Reliability of the finger
tapping test and a note on sex differences. Percept Mot Skills
1979; 48: 139–42.
26. Gill DM, Reddon JR, Stefanyk WO, Hans HS. Finger tapping:
effects of trials and sessions. Percept Mot Skills 1986; 62: 675–8.
27. Hausdorff JM, Yogev G, Springer S, Simon ES, Giladi N.
Walking is more like catching than tapping: gait in the elderly
as a complex cognitive task. Exp Brain Res 2005; 164: 541–8.
Received 11 June 2010; accepted in revised form
4 July 2011
Age and Ageing 2012; 41: 161–165
© The Author 2012. Published by Oxford University Press on behalf of the British Geriatrics Society.
doi: 10.1093/ageing/afr176
All rights reserved. For Permissions, please email: [email protected]
Published electronically 26 January 2012
The impact of social vulnerability on the survival
of the fittest older adults
MELISSA K. ANDREW1, ARNOLD MITNITSKI2, SUSAN A. KIRKLAND3, KENNETH ROCKWOOD1
1
Geriatric Medicine, Dalhousie University, Halifax, Nova Scotia, Canada
Medicine, Mathematics and Statistics, Dalhousie University, Halifax, Nova Scotia, Canada
3
Community Health and Epidemiology, Dalhousie University, Halifax, Nova Scotia, Canada
2
Address correspondence to: M. K. Andrew. Tel: +1 (902) 473-4995; Fax: +1 (902) 473-1050. Email: [email protected]
Abstract
Background: even older adults who are fit experience adverse health outcomes; understanding their risks for adverse outcomes may offer insight into ambient population health. Here, we evaluated mortality risk in relation to social vulnerability
among the fittest older adults in a representative community-dwelling sample of older Canadians.
Methods: in this secondary analysis of the Canadian Study of Health and Aging, participants (n = 5,703) were aged 70+
years at baseline. A frailty index was used to grade relative levels of fitness/frailty, using 31 self-reported health deficits. The
analysis was limited to the fittest people (those reporting 0–1 health deficit). Social vulnerability was trichotomised from a
social vulnerability scale, which consisted of 40 self-reported social deficits.
Results: five hundred and eighty-four individuals had 0–1 health deficit. Among them, absolute mortality risk rose with increasing social vulnerability. In those with the lowest level of social vulnerability, 5-year mortality was 10.8%, compared with
32.5% for those with the highest social vulnerability (adjusted hazard ratio 2.5, 95% CI: 1.5–4.3, P = 0.001).
Conclusions: a 22% absolute mortality difference in the fittest older adults is of considerable clinical and public health importance. Routine assessment of social vulnerability by clinicians could have value in predicting the risk of adverse health
outcomes in older adults.
Keywords: frailty, survival, mortality, frail elderly, aged, social vulnerability, elderly
Introduction
At any age, on average, older people have more health deficits (broadly defined to include illnesses, symptoms and
functional limitations) than do younger people. Even so,
some people live into old age having accumulated very few
health deficits. This group of apparently very healthy
elderly people is of some interest. In particular, they might
offer some insight into the overall health of a population.
This is because in the absence of known problems with
161
M. K. Andrew et al.
their own health (intrinsic factors), the health and mortality
outcomes of these fittest people of a society may depend
rather on extrinsic factors such as the attributes of that
society and the environment in which it is located.
The age-related accumulation of health deficits can shed
some insights on frailty, i.e. on the problem of heterogeneity
in health outcomes of people of the same chronological age.
Health deficits are defined broadly, to include symptoms,
signs, diseases, disabilities or laboratory abnormalities.
A number of studies have found that the more deficits that
people have the more likely they are to experience adverse
health outcomes, such as death, institutionalisation or worsening in their health [1–7]. This relationship to adverse health
outcomes is true across the entire range of health deficits, so
that people in whom deficits are not demonstrable (whom we
refer to as being in the ‘zero state’ of frailty) have the lowest
rate of death, institutionalisation and worsening health.
Social conditions, including socio-economic status,
social support, social engagement and mastery, have powerful influences on health [8–13]. However, the influence of
social factors such as socio-economic status in older age is
debated [14–16]. Nevertheless, we have previously reported
results from both the Canadian Study of Health and Aging
(CSHA) and the National Population Health Survey
showing that a holistic social vulnerability index, akin to the
frailty index, which includes social factors from various
domains including socio-economic status, social support,
social engagement and mastery, predicts both mortality and
cognitive decline among older adults [17, 18]. Here, our objective was to investigate the impact of social vulnerability,
quantified using a social vulnerability index, on the survival
of only the fittest older adults, defined using a frailty index.
Methods
Ethics statement
Ethical approval was obtained from the ethics committees
of each of the 18 CSHA study centres. Written, informed
consent for participation in the CSHA was obtained from
each participant or proxy respondent.
The sample came from the second phase of the CSHA-2
(conducted in 1996–97). The CSHA is a representative study
of dementia and other health problems in older Canadians
aged 65 and older [19]. At baseline (CSHA-1, 1990–91), participants (n = 10,263, of whom 9,008 were community dwelling) were sampled in a population-representative manner
from English- and French-speaking older Canadians, though
those living in the Yukon or Northwest Territories, residents
of Aboriginal Reserves or military bases, and those with an
immediately life-threatening illness were excluded. The sample
was clustered within five Canadian regions and stratified by
age, with over-sampling of those aged 75 and older. The
baseline data collection occurred in 1991, with follow-up by
interview and/or clinical examination or vital status verification at 5 (CSHA-2) and 10 (CSHA-3) years [19]. Of 5,703
participants in the CSHA-2 screening interview, 729 (13%)
were missing frailty status. People with missing frailty data
162
were older (83.2 versus 78.5 years, P < 0.0001), often relying
on proxy respondents. The fittest individuals were defined as
the 584 who were in the ‘zero state’ of frailty, i.e. those in
whom either 0 or only 1 health deficit were reported. Of
these, 541 (93%) had social vulnerability status data allowing
for a social vulnerability index to be calculated. The cohort
was followed for 5 years, at which time vital status was known
for all but 3 of these 541 individuals; 95 (17.6%) had died.
Frailty and social vulnerability were based on self-report.
The social vulnerability index has been described in detail
elsewhere [17]. It includes 40 items addressing various social
domains including living situation, marital status, social engagement, social support, feelings of mastery and empowerment and socio-economic status. Our previous investigations
of its properties have supported its use as a holistic measure;
in particular, no single item or group of items has been found
to drive associations with health outcomes [17, 18]. Responses
to 40 social variables were assigned a value of ‘1’ if representing a deficit and ‘0’ otherwise. The sum of this deficit count,
divided by 40, is the social vulnerability index, so that the theoretical range is from 0 (none of the 40 social deficits) to 1
(all 40/40 social deficits); higher scores indicate greater vulnerability [17]. Social vulnerability was divided into tertiles of
low, intermediate and high index values. Individuals missing
data for one or two of the social deficits were assigned to the
tertile of social vulnerability based on their existing data. The
43 individuals who were missing more than one to two variables were coded as missing social vulnerability status and
therefore excluded from further analysis.
The frailty index was operationalised using 31 health deficits (illnesses, symptoms and functional problems). Both
the frailty and social vulnerability indices have been validated
and lists of their constituent variables have been published
[6, 17]. The association between level of social vulnerability
(independent variable) and mortality (dependent variable)
was analysed using Cox regression. The absolute risk of
mortality was calculated for the three strata of social vulnerability, and Kaplan–Meier survival curves were generated.
Exercise, smoking and alcohol histories were obtained
from a self-reported questionnaire conducted at the first
phase of CSHA, 5 years prior to the baseline data for this
analysis. These variables were not updated in later phases of
the CSHA, so previous reported history is all that was available. Exercise was categorised as high (regular exercise, three
or more times per week, more intense than walking), intermediate (regular exercise but either less frequent or of walking
intensity or less) and low (no regular exercise reported). This
definition of exercise has been validated in the CSHA [20,
21]. Smoking was defined as a lifetime history of ever having
‘smoked cigarettes, pipe or cigars regularly (nearly every day)’.
Alcohol intake was similarly defined as ever having been a
regular drinker of beer, wine or spirits.
Results
Of 4,974 people with known frailty status, 584 (12%)
reported only 0 or 1 deficit. Compared with those with two
Social vulnerability and survival
or more health deficits, those in the zero state sample were
younger (76.3 versus 78.8 years old), less likely to be currently married (43.2 versus 52.6%) and they had slightly
higher MMSE scores (27.1 versus 26.7) and more education (11.3 versus 10.5 years; all P < 0.001). Fewer women
(46.8 versus. 61.6%; P < 0.001) were in the zero state
group. The fittest individuals were more likely to report
both exercising (P < 0.001) and alcohol consumption (P =
0.009). There was no difference in smoking history.
Demographic and lifestyle characteristics of the
zero-state group by level of social vulnerability are presented in Table 1. Those with high social vulnerability were
older, less likely to be married, and had lower educational
attainment and MMSE scores. There were no differences in
the lifestyle factors. Adjusting for age and sex, high social
vulnerability was associated with an increased risk of death
(HR: 2.5, 95% CI: 1.5–4.3, P = 0.001) (Figure 1). Among
those in the zero state of frailty, the absolute risk of mortality rose with increasing social vulnerability. Among those
with low social vulnerability, 32 of 296 died (10.8%), when
compared with 37 of 165 (22.4%) with intermediate social
vulnerability. In the high social vulnerability group, 26 of 80
(32.5%) had died within 5 years.
Previously reported exercise, previous history of smoking
and previous alcohol intake were not statistically significantly
associated with mortality among those in the zero state of
frailty. In the survival models adjusted for these lifestyle covariates as well as age and sex, the association between high
social vulnerability and mortality was strengthened (HR: 3.22,
95% CI: 1.80–5.78, P < 0.001).
twice as likely to die as the third with the lowest social vulnerability. This increase in risk represents a 22% absolute
mortality difference between those with low and high social
vulnerability.
While these findings are limited by being based on a
single, relatively small sample of older adults, the effect
size is striking and warrants further investigation. In particular, replication in samples from countries in different
stages of development may be instructive. In addition,
possible gender differences and the respective impact of
mid-life versus older age social conditions especially
warrant further inquiry. The use of self-report data is a
further limitation of this study, although we have shown,
in relation to the frailty index, that self-report, observer
assessed and test data give comparable estimates in the
average rate of deficit accumulation and in the maximum
observed values [5]. Moreover, for social vulnerability, it
Discussion
Figure 1. Kaplan–Meier curves showing survival those in the
zero state of frailty (having 0–1 health deficit) by level of
social vulnerability (low, intermediate and high). SV, social
vulnerability.
We found that, among the fittest Canadian older adults, the
third with the highest social vulnerability were more than
Table 1. Baseline demographic and lifestyle factors according to level of social vulnerability in the ‘zero-state’ sample (those
with 0–1 health deficit)
Low social
vulnerability
(n = 296)
Intermediate social vulnerability
(n = 265)
High social vulnerability
(n = 80)
Statistical significance of
differences
....................................................................................
Age: mean (SD)
Sex: % female (%)
Education: mean (SD) years
Marital status: % married (%)
MMSE score: mean (SD)
History of smoking (asked 5 years
previous) (%)
Alcohol consumption (asked 5 years
previous) (%)
Exercise (asked 5 years previous) (%)
Regularly, >3/week, more intense than
walking
Regularly but less frequent and/or less
intense
No regular exercise
75.2 (4.9)
45.6
12.0 (3.7)
73.3
27.7 (2.3)
52.6
76.7 (5.4)
43.6
10.9 (3.9)
46.1
27.3 (2.2)
51.8
77.9 (5.5)
52.5
9.8 (3.6)
26.3
26.7 (2.7)
60.0
P < 0.001
P = 0.42
P < 0.001
P < 0.001
P = 0.001
P = 0.53
49.8
46.8
34.4
P = 0.09
24.8
22.9
21.3
P = 0.7
57.1
56.2
52.5
18.1
20.8
26.2
163
M. K. Andrew et al.
may be that self-perception of one’s social circumstances
is particularly important.
Frailty and social vulnerability status were missing for
some of the CSHA cohort. Those who were missing these
indices were older and had more illnesses than those with
complete data. They more often relied on proxy reports,
making collection of many self-report items impossible. This
is unlikely to have affected the findings of the current study,
because the cohort was limited to those in the ‘zero state’ of
frailty, who had 0 or 1 health deficit, because those with
missing full frailty index data generally had many health problems. However, it is possible that the exclusion of the 43
individuals (7.4% of the zero state sample) who were known
to be in the zero state of frailty but were missing social vulnerability status may have affected the analysis of the association between social vulnerability status and health.
The lifestyle data were collected 5 years prior to the
current baseline data, and we have no means of verifying
whether the previous self-reports correlate with current behaviour, although they do correlate with outcomes [20, 21].
Nevertheless, it is interesting to note that the association
between social vulnerability and mortality that we have
identified seems to be robust to an initial consideration of
confounding by lifestyle factors, though the influence of
health-related behaviours on the association between social
circumstances and health in older people remains an area
for further investigation.
The influence of social factors, usually limited to
consideration of socio-economic status, on health in older
age has been debated. A recent US study found that
socio-economic differences in mortality became insignificant
at older ages (ages 70 years and older); this was felt to relate
to important differences in health and mortality experienced
at earlier ages, possibly leading to a ‘survival of the fittest’
phenomenon [15]. In the Whitehall studies in the UK, the
magnitude of association between occupational status and
health was found to diminish after retirement, while the importance of other socio-economic indicators was maintained
[16]. Others have argued that while the importance of more
limited indicators of SES such as income and education may
diminish at older ages, consideration of other factors such as
financial assets remains important [14]. These sorts of findings have contributed to discussion of how to best measure
SES in older age, where occupational status, income and education may have more limited import due to retirement,
pension schemes and cohort educational norms [22]. Here,
we have broadened our consideration of social factors to
create a more holistic representation of aggregate social circumstances, not limited to consideration of a few socioeconomic factors, which may explain the differences between
our findings and some previous literature reports.
The outcomes of people in the zero state of frailty are of
considerable interest. That is because changes in the health
of individuals are the sum of ambient changes, plus changes
associated with their given state of health [23]. Although
mortality outcomes will always be higher as people grow
older, the environment will undoubtedly be important—the
164
mortality of the fittest older adults in Somalia, for example,
will be higher than that of their counter-parts in
Saskatchewan. The outcomes of people in the zero state of
frailty provide a quantitative estimate of these background
effects, an application of considerable potential. For example,
the chance of dying for an individual with three health
deficits is a function of the chance of the healthiest people
dying—those in the zero state—plus the chance of dying as
the number of deficits increases to three [23].
Older people who have aged without accumulating
health deficits can be considered examples of healthy
ageing. Study of this group of individuals presents an important opportunity to understand predictors and facilitators of healthy ageing, which could in turn lead to new
ideas about how to improve the health and quality of life of
our ageing populations.
Investigating how social factors affect the healthiest
gives a quantitative estimate of those influences on the
ambient health of the population. Limiting the analysis to
the fittest individuals may allow for improved isolation of
what factors in a society make it healthy or unhealthy. We
suggest that survival in the zero state of frailty is a candidate marker for the health of populations.
Key points
• Changes in the health of individuals result from both
intrinsic and extrinsic factors.
• Studying the fittest older people allows for investigation
of extrinsic vulnerability related to social circumstances.
• Social vulnerability captures the degree to which a person’s
social situation leaves them susceptible to further insults.
• Among the fittest and healthiest older adults, social vulnerability meaningfully influences survival.
• The survival of the fittest older adults is of potential interest as a candidate marker for the health of populations.
Conflicts of interest
K.R. has applied for funding to commercialise a version of
the frailty index based on electronic capture in a comprehensive geriatric assessment.
Funding
The analyses were supported by a Canadian Institutes for
Health Research (CIHR) operating grant (MOP- 64169)
held by A.M. and K.R. All analyses were conducted by us.
The CSHA data are held in house. M.K.A. was supported
by a post-doctoral research fellowship from the CIHR.
K.R. receives career support from the Dalhousie Medical
Research Foundation as Kathryn Allen Weldon Professor
of Alzheimer Research.
Self-rated and physician-rated health
References
1. Goggins WB, Woo J, Sham A, Ho SC. Frailty index as a
measure of biological age in a Chinese population. J Gerontol
Med Sci 2005; 60: 1046–51.
2. Kulminski A, Yashin A, Ukraintseva S et al. Accumulation of
health disorders as a systemic measure of aging: findings from
the NLTCS data. Mech Ageing Dev 2006; 127: 840–8.
3. Kulminski AM, Ukraintseva SV, Kulminskaya IV et al.
Cumulative deficits better characterize susceptibility to death in
elderly people than phenotypic frailty: lessons from the Cardiovascular Health Study. J Am Geriatr Soc 2008; 56: 898–903.
4. Rockwood K, Andrew M, Mitnitski A. A comparison of two
approaches to measuring frailty in elderly people. J Gerontol
Med Sci 2007; 62: 738–43.
5. Mitnitski A, Song X, Skoog I et al. Relative fitness and frailty
of elderly men and women in developed countries, in relation
to mortality. J Am Geriatr Soc 2005; 53: 2184–9.
6. Mitnitski AB, Song X, Rockwood K. The estimation of
relative fitness and frailty in community-dwelling older adults
using self-report data. J Gerontol Med Sci 2004; 59:
M627–32.
7. Yang Y, Lee LC. Dynamics and heterogeneity in the process
of human frailty and aging: evidence from the U.S. older adult
population. J Gerontol Soc Sci 2010; 65: 246B–55.
8. Putnam RD. Bowling Alone: The Collapse and Revival of
American Community. New York: Simon & Schuster, 2000.
9. Kawachi I, Berkman LF. Social cohesion, social capital, and
health. In: Berkman LF, Kawachi I, eds. Social Epidemiology.
Oxford: Oxford University Press, 2000; 174–90.
10. Berkman LF. Social support, social networks, social cohesion
and health. Soc Work Health Care 2000; 31: 3–14.
11. Andrew MK. Social vulnerability in old age. In: Rockwood
K, Fillit H, Woodhouse K, eds. Brocklehurst’s Textbook of
Geriatrics and Gerontology. Philadelphia: Saunders Elsevier,
2010; 198–204.
12. Veenstra G. Social capital, SES and health: an individual-level
analysis. Soc Sci Med 2000; 50: 619–29.
13. Woo J, Goggins W, Sham A, Ho SC. Social determinants of
frailty. Gerontology 2005; 51: 402–8.
14. Robert S, House JS. SES differentials in health by age and
alternative indicators of SES. J Aging Health 1996; 8: 359–88.
15. Crimmins EM, Kim JK, Seeman TE. Poverty and biological
risk: the earlier ‘aging’ of the poor. J Gerontol A Biol Sci
Med Sci 2009; 64: 286–92.
16. Marmot MG, Shipley MJ. Do socioeconomic differences in
mortality persist after retirement? 25 year follow up of civil
servants from the first Whitehall study. BMJ 1996; 313:
1177–80.
17. Andrew MK, Mitnitski A, Rockwood K. Social vulnerability,
frailty, and mortality in elderly people. PLoS One 2008; 3:
e2232.
18. Andrew MK, Rockwood K. Social vulnerability predicts cognitive decline in a prospective cohort of older Canadians.
Alzheimers Dement 2010; 6: 319–25.
19. Rockwood K, McDowell I, Wolfson C. Canadian study
of health and aging. Int Psychogeriatr 2001; 13 (Suppl. 1):
1–237.
20. Middleton LE, Mitnitski A, Fallah N, Kirkland SA,
Rockwood K. Changes in cognition and mortality in relation
to exercise in late life: a population based study. PLoS One
2008; 3: e3124.
21. Hubbard RE, Fallah N, Searle SD, Mitnitski A, Rockwood
K. Impact of exercise in community-dwelling older adults.
PLoS One 2009; 4: e6174.
22. Grundy E, Holt G. The socioeconomic status of older
adults: how should we measure it in studies of health inequalities? J Epidemiol Community Health 2001; 55:
895–904.
23. Mitnitski A, Bao L, Rockwood K. Going from bad to
worse: a stochastic model of transitions in deficit accumulation, in relation to mortality. Mech Ageing Dev 2006; 127:
490–3.
Received 2 November 2010; accepted in revised form
7 September 2011
Age and Ageing 2012; 41: 165–171
© The Author 2011. Published by Oxford University Press on behalf of the British Geriatrics Society.
doi: 10.1093/ageing/afr161
All rights reserved. For Permissions, please email: [email protected]
Published electronically 16 December 2011
Self-rated health and physician-rated health
as independent predictors of mortality
in elderly men
ERIK J. GILTAY1, ALBERT M. VOLLAARD2, DAAN KROMHOUT3
1
Department of Psychiatry, Leiden University Medical Center, P.O. Box 9600, 2300 RC, Leiden, The Netherlands
Leiden University Medical Center, Department of Infectious Diseases, Leiden, The Netherlands
3
Division of Human Nutrition, Wageningen University, Wageningen, The Netherlands
2
Address correspondence to: E. J. Giltay. Tel: +31 71 5263448; Fax: +31 71 5248156. Email: [email protected]
165