Caregiver Resentment Over Time:
The Influence of Race and Gender
Daniel W. Durkin Ph.D. & Gail Williamson, Ph.D.
Background
• Older adults will be increasing in
numbers and % of population
• Policy goal to keep older adults in
community for as long as possible
• Also consumer goal
• For these reasons (and others) the
number of family caregivers
expected to increase
Background Cont.
In order for older adults to remain in
the community it is essential to:
(1) identify caregivers in need of
support; and
(2) provide support to caregivers that
is appropriate to the context in
which the care is provided.
Caregiver Resentment
• Many caregivers of older adults feel
resentful.
• Caregiver resentment has been
understudied in the caregiving
literature.
• Previous studies have shown that
resentment has been positively
linked to a greater need for
assistance with activities of daily
living, care recipient behavior
problems, feelings of caregiver
burden and the quality of care
provided to the care recipient.
Pearlin Stress Process
Model
Benefits of Pearlin SPM
1) Most commonly used model in elder
caregiving research
2) Applicable across a broad range of
caregiving populations
3) Offers a framework for assessment
that accounts for individual variation
4) Provides a measurable model that
explains how these constructs relate
to each other
5) Interventions can be developed that
target single or multiple domains of
the caregiving experience
Research Questions
1)
2)
3)
How does caregiver resentment of caregivers
of older adults change over a three-year
period?
Are there differences between African
American and White caregivers and male
and female caregivers in changes in
resentment over time?
How do caregiver context variables, primary
stressors, subjective stressors, and
resources predict the trajectories of caregiver
resentment over three years?
Methods
• Secondary data analysis
• Descriptive analyses, bivariate
analyses
• Hierarchical Linear Modeling
• Data from 425 African American
(43%) and White (57%) caregivers
• 130 men (31%) and 295 women
(69%)
• Family Relationships in Late Life
(FRILL) 2 study
FRILL 2
• FRILL 2 (AG15321, P.I.: Gail Williamson)
was a NIA-funded multi-site longitudinal
study of informal care provided by
community-dwelling caregivers of elderly
care recipients with whom they co-resided.
• Three waves of data were collected over
three years approximately 18 months
apart.
• Both the caregiver and care recipient were
assessed at all three waves;
• However, only data from the caregivers are
included in this study.
• It should be noted that FRILL 2 did not
include an intervention component; thus
changes over time were not influenced by
researcher intervention.
FRILL 2
• Recruitment in and around the
communities of Athens, GA, Pittsburgh,
PA, and Tuscaloosa, AL
• Probability-based random digit dialing
(RDD) and word-of-mouth advertising.
• Primary goal was to oversample African
American caregiving dyads in order to
obtain data sufficient to address
significant gaps in research literature.
• Face-to-face structured interviews,
lasting between 1.5 and 2 hours, were
conducted in respondents’ homes by
pairs of carefully trained interviewers.
Benefits of HLM
• HLM has the capacity to analyze
information about the rate and the
pattern of change in targeted
variables over multiple time
points.
• Takes into account inter- and
intra-individual variability in
change and cross-level
interactions of time with
predictors (Raudenbush & Bryk,
2002).
Benefits of HLM Continued
Specific advantages of HLM over
standard repeated measures analyses in
this study were:
(1) Allowed the examination of predictor
variables that change over time;
(2) Allowed the use of data from
participants that did not have data from
all three time points; and
(3) Took into account variance in the
spacing of assessments across
participants.
HLM: Level 1
• 5 separate HLM analyses for each
outcome: anger, anxiety, depressed affect,
resentment, and self-perceived health
• Time
• ADL/IADL
HLM: Level 2
Caregiving
Context
Primary
Stressors
Secondary
Stressors
Resources
CG Age
Dementia Diagnosis
Perceived Income Adequacy
Social Support
Activity Restriction
Formal Services
CR Age
CG Sex
Years Caregiving
Spouse/Non-spouse
Race
CG Education
CG Income
Pre-Illness Relationship
Quality
HLM Initial Model
Revised Model
This initial model was then revised using the
following guidelines:
• Level 1 predictors that did not have significant
fixed or random effects on the outcome were not
included in the revised model.
• Level 1 predictors that had significant fixed
effects, but not significant random effects, on the
model were included in the revised model as
fixed effects.
• Level 1 predictors that had significant random
effects (whether the fixed effects were or were
not significant) were included in the revised
model as fixed and random effects.
• Level 2 predictors were added to explain the
variability in any level 1 coefficients that had
significant random effects.
• CG Gender added as a predictor
Final Model
The final model was used as the
basis for drawing conclusions
and included:
 Variables essential to answer the
research questions (sex, race
and time)
 ADL/IADL at Level 1 if there were
significant fixed or random effects
on the outcome; and
 Level 2 predictors that were
significantly correlated with the
outcome variable and had
significant fixed effects on the
outcome.
Pre-Illness Relationship
Quality
• Caregiver report of the Mutual
Communal Behavior Scale
(Williamson & Schulz, 1995)
• 10-item measure of communal
feelings between the caregiver
and care recipient prior to onset
of illness
• Higher scores indicate better preillness relationship quality.
Activity Restriction
• Activity Restriction Scale (Williamson,
Shaffer, & Schultz, 1995)
• 10-item measure of restriction in
caregiver activity due to the caregiving
role.
• Caregivers were asked how much
caring for the care recipient has
restricted them in 10 areas (e.g. “caring
for yourself,” “visiting friends”).
• Higher scores indicate more activity
restriction.
• Can be viewed as a proxy for social
isolation.
Caregiver Resentment
• Sum of responses to a 17-item
instrument (Williamson, Shaffer, &
FRILL, 2000; Williamson et al., 2005)
• Caregivers indicated how often (1 =
never, 5 = almost
• always) they felt resentful about such
circumstances as not having enough
time for themselves, having to give up
plans for the future, or that their care
recipients were overly dependent or
made unreasonable demands.
• Higher scores indicate greater
resentment.
Demographics
Descriptives
HLM Caregiver Resentment
Results
 Caregivers with greater activity
restriction reported higher levels
of resentment.
 Caregivers who reported a better
pre-illness relationship and more
social support reported less
resentment.
 There was a main effect of time
such that resentment lessened in
the later sessions.
Results Continued
 Change over time also
demonstrated a significant
quadratic pattern (β = .93, SE =
.48, p<.05) suggesting a
significant drop in resentment
between Time 1 and Time 2.
 Random effects were significant
for time (χ2 [223] = 338.52, p
<.01) suggesting that there is still
some variance unexplained by
time in the model.
Results Continued
 There was no significant main effect
for race suggesting that there are
no significant differences between
African American and White
caregivers in their report of
resentment over time.
 Main effect for sex such that female
caregivers (β = -4.20, SE = 1.12,
p<.01) reported more resentment
over time.
 Random effects were significant (χ2
[108] = 747.67, p <.01) suggesting
that there is still some variance
unexplained by the variables in the
Discussion: Pre-Illness
Relationship
• Caregivers who reported a higher
quality pre-illness relationship
reported less resentment over
time.
• Suggest that positive feelings
toward the care recipient earlier
in life may mitigate the negative
effects of caregiving on emotional
well-being.
Social Support & Social
Isolation
• Caregivers who reported more activity
restriction and less social support also
reported more resentment over time.
• Support previous findings that an
increased sense of social isolation and
lack of support from family and friends
contributes to negative emotional wellbeing.
• Findings suggest that these deficits may
have a cumulative effect over time,
especially for women.
Implications: Relationship
Quality
• Target relationship quality in
interventions throughout the life
course
• May pay dividends in the long-run
• Strengthen interpersonal bonds
• May buffer the effect of caregiving
stressors
Implications: Isolation/Social
Support
•
•
•
•
Expand informal support caregivers
Increase participation in activities
Identify informal support systems
Help caregivers understand the
important role of social support
• Strengthen already existing networks
of family and friends
• Religiosity/spirituality
Thank You