Sociodemographic Effects on Taskspecific ADL Functioning at the Oldestold Ages

1
Sociodemographic Effects on Task-specific ADL
Functioning at the Oldest-old Ages

• Danan Gu1 and Qin Xu2

1. Public Policy, Duke University, 2. 3. China
Research Center on Aging
2
Outline

• Introduction objectives
• Data and methodology
• Results
• Discussion

3
I. Introduction objectives- Introduction

• ADL has a hierarchy Previous research has
  suggested that Individual loses functioning of
  each six task probably has ordering in that each
  task involves different required capability to
  perform (e.g., Dunlop et al. 1997Jagger et al.
  2001 Katz et al. 1963)
• Studies systematically examining the dynamic of
  task-specific ADL and its associates are very
  rare.
• None of previous studies, which focus on ADL
  task-specific, explore the correlations with
  socio-demographic factors.
• There is no study to examine the dynamics of each
  ADL task-specific focusing on the oldest-old in
  developing countries.
• One of major limitations is the underestimation
  of disability in the study of ADL functioning at
  the old ages (Gill et al. 2002 Zeng, Gu, and
  Land 2004)
• The conventional method which does not include
  the ADL change before dying overestimate effect
  of age, gender, ethnicity, and living alone on
  disability transitions while underestimates the
  effects of SES (Gu Zeng 2004).

4
I. Introduction objective- Objective

• By including the ADL information before dying,
  this study tries to capture more robust effects
  of sociodemographic factors on task-specific ADL
  functioning over a two-year survey interval.

5
II. Data source methodology - Data source

• The Chinese Longitudinal Healthy Longevity Survey
  (CLHLS) in 1998, 2000, and 2002. All three waves
  of data are pooled together.
• Persons who reported age 106 or higher at Time 1
  are excluded from this study.
• Total 19,778
• Male 8,142
• Female 11,636

6
II. Data source methodology- methodology

• Time 1 the start-point of each two-year
  interval.
• Time 2 the endpoint of each two-year interval.

7
Measuring onset and recovery of disability in
each ADL task
Bathing
Toileting
Dressing
ADL
Indoor transferring
Eating
Continence
8
II. Data source methodology-continue
not disabled(or no need help)
each ADL task
disabled (or need help)
9
onset of task disability
the information of ADL changes before dying
task recovery
10

• An individual is considered having an onset of
  disability in one task
• if he/she moved from not disabled (no need in
  help) in that task at Time 1 to disabled (need in
  help) in the same task at Time 2 or before dying
• A person is classified as having a recovery
  from disability in one task
• if he/she moved from need in help in that task at
  Time 1 to no need in help in the same task at
  Time 2 or before dying

11
six items into two categories

• requires both upper and low extremities
  (high-level tasks/items )
• include bathing, toileting, dressing, indoor
  transferring
• requires upper extremities only or continence
  (low-level tasks/items )
• include feeding, continence
• Factor analysis confirms this classification

12
Sociodemographic factors

• Age 80-89, 90-99,100-106
• Sex Male, Female
• Residence urban, rural
• Ethnic group Han, other minorities
• Education 0 year (illiteracy), 1 years of
  schooling
• Economic independence
• economic independence mainly relying on
  pension or own financial source
• economic dependence namely relying on other
  resources such as spouse, children or other
  family members, or government subsidy and others

13
Sociodemographic factors -continue

• Primary lifetime occupation
• for males agriculture, non-agriculture
• for females agriculture, housewife,
  non-agriculture
• Marital status
• currently not married (never married,
  divorced, and widowed)
• currently married
• Living arrangement living alone , not living
  alone

14
Control variables

• The confounders include
• (1) Family/social support and connection
  Religious activities, proximity with children,
  playing card, and who take care of the elder in
  the daily life are considered as the proxy of
  family and social supports/connections in this
  study.
• (2) Health practice alcoholic consumption,
  smoking, diet, and doing exercise
• (3) Health conditions cognitive function,
  self-reported health, chronic disease conditions,
  hearing and visual function, and functional
  limitations.
• (4) The length of surviving between Time 1 and
  Time 2, and whether the respondent died or not
  before Time 2 were also included in the model to
  adjust their influence.

15
Statistical models

• models for males and females separately
• those who lost to follow-up respondents at Time 2
  are excluded
• Random effect logit model after correcting
  intra-person correlation was applied

16
III. Results
17
Table 1 Percentage distributions of
socio-demographic variables by gender ()
18
Table 2 Dynamics changes of each ADL task by
gender
19

• The functioning of the high level tasks declines
  faster than the low level tasks.
• The incidence rate for bathing is almost two
  times more than that of continence.
• The recovery rates present an opposite pattern,
  and the recovery rate for bathing is 50 less
  than that of continence.

20
(No Transcript)
21
(No Transcript)
22
Table 3 Odds ratios of gender and age on
the onset of disability and recovery from
disability of each ADL task
Notes (1) Male is the reference group. Age is
measured at the beginning of each two-year
interval. (2) N is the number of observations.
(3) The results are based on controlling other
confounders mentioned in the text. (4) Odds
ratios and their significant level were corrected
by intra-subject correlations due to some
subjects contributing two or threemore than one
observations to the pooled data set at a specific
time (Time 1or Time 2). (5) , plt0.1 , plt0.05
, plt0.01 , plt0.001.
23

• Gender differential effect is significant among
  nonagenarians and centenarians, and it reaches
  the biggest among nonagenarians. The gender
  differential patterns are similar across six
  tasks.
• The effect of increasing each additional age on
  onset of and recovery from task-specific
  disability is different both across tasks and
  across age groups.
• The effect of increasing each additional age on
  onset of task disability is gradually diminishing
  with the advancement of age.

24
Table 4 Odds ratios of the onset of
disability of each ADL task
25
Table 4 Odds ratios of the onset of
disability of each ADL task
Notes (1) Reference group of each covariate is
listed in the parentheses. (2) All variables are
measured at the beginning of each two-year
interval. (3) The results are based on
controlling other confounders mentioned in the
text. (4) Odds ratios and their significant
level were corrected by intra-subject
correlations due to some subjects contributing
two or threemore than one observations to the
pooled data set at a specific time (Time 1or Time
2). (5) , plt0.1 , plt0.05 , plt0.01 ,
plt0.001.
26
Table 5 Odds ratios of the recovery of each
ADL task
27
Table 5 Odds ratios of the recovery of each
ADL task
Notes (1) Reference group of each covariate is
listed in the parentheses. (2) All variables are
measured at the beginning of each two-year
interval. (3) The results are based on
controlling other confounders mentioned in the
text. (4) Odds ratios and their significant
level were corrected by intra-subject
correlations due to some subjects contributing
two or threemore than one observations to the
pooled data set at a specific time (Time 1or Time
2). (5) , plt0.1 , plt0.05 , plt0.01 ,
plt0.001.
28

• Among males, compared to octogenarians,
  nonagenarians and centenarians have a higher risk
  to develop task-specific disability in high level
  of hierarchy, while they share a similar risk to
  develop task disability in the low level tasks in
  the hierarchy.
• On the other hand, female nonagenarians and
  centenarians have a significant higher risk to
  develop task-specific disability in all tasks
  compared to female octogenarians. The
  disadvantage of older ages in effect on task
  recovery is less apparent than its effect on
  onset of task-specific disability (see Table 5).
  Such results indicate that disability
  deceleration is more obvious for males than for
  females at oldest-old ages.

29

• Result presented in Table 4 reveals that
  urban/rural residential difference in onset of
  task-specific disability is homogeneous rather
  heterogeneous although urban oldest olds have a
  higher chance to have onsets of task-specific
  disability and lower chance to get recovery on
  general.
• The difference of residential effect on
  task-specific disability between females was
  found. Whereas urban males have same risks to
  have onset of task-specific disability as rural
  males, urban females experience 10-20 higher
  risks to develop task-specific disability than
  rural females.
• Furthermore, Table 5 indicates that urban
  oldest-olds have a lower chance to get recovery
  from high level task disability for both males
  and females.

30

• Table 4 illustrates that Minority ethnicities
  have better ADL functioning over Han, the
  majority. Minorities have 20-45 lower risk to
  develop task-specific disability except eating
  for both males and females.
• Table 5 further indicates that although
  minorities have higher chances to recover once
  they have onset of disability across all tasks
  for both males and females, such higher chances
  are not statistically significant possibly due to
  highly individual variation.

31

• SES has limited effect on both onset of and
  recovery from task-specific disability. SES
  demonstrates the similar pattern across tasks and
  across genders. Living alone has significant
  lower risk to develop disability across tasks for
  males, but such effective is not significant
  among females (Table 4).
• On the other hand, the positive effects of living
  alone on task-specific recovery are not all
  significant across tasks despite all the odds
  ratios are much higher than 1 except eating for
  females (Table 5). The protective effect of
  marriage on task-specific disability and recovery
  was not detected.

32

• Additionally, for males, chronic condition
  and functional limitation will raise the risk of
  developing disability across tasks.
• But they dont have effect on getting recovery
  from disability across tasks.
• Generally speaking, persons who are in absence of
  limitation in high level tasks in the hierarchy
  will postpone his/her onset of disability of the
  low tasks in the hierarchy.
• Such pattern is similar for both males and
  females (Table 4).
• The results listed in Table 5 show that it will
  not help male oldest-olds to get recovery from a
  task disability if they have no limits in any
  other task.
• But for female oldest-olds, they will have a
  higher chance to get recovery from a task
  disability if they dont have functioning limits
  in other task.

33
IV. Discussion

• Slightly higher recovery from each task
  disability was observed. The main possible reason
  is that China has a higher mortality, which makes
  the oldest-old in China more robust compared to
  their counterparts in western nations.
• Our results show that each sociodemographic
  factor has somewhat different role in dynamics of
  ADL functioning across tasks.
• Consistency and inconsistency with previous
  studies.
• Although the results of this study are far from
  conclusive and deserve further investigation,
  they provide new knowledge about the general
  patterns of dynamics of task-specific disability
  of the oldest-old, which contributes to a better
  understanding of the social and demographic
  factors associated with task-specific onset of
  and recovery from disability.

34

• One strength point of this study lies its
  systematically examination of socio-demographic
  effect on dynamics of six task-specific ADL
  functions at the oldest-old ages, which is rarely
  touched by previous studies, but is more
  informative.
• One limitation of the present study is that
  information on ADL status before dying for the
  dead persons before the second interview is
  obtained from next-to-kin, which may involve some
  bias.
• We only collected one episode of dynamic change
  for each ADL task both among survivors and among
  the deceased persons, which will not fully
  capture the whole story of dynamics of each task
  given.