Download Impact of the Buffering Model of Social Support

Community Participation Following Cerebrovascular Accident:
Impact of the Buffering Model of Social Support
Margaret Newsham Beckley
This study examined data of older adults who had sustained a stroke and their ability to participate in the community following rehabilitation. There were 95 participants in this study to determine the interaction effect of
functional limitation and social support on community participation. The findings indicate that for people who
had sustained a cerebrovascular accident, community participation was more related to their ability to do for
themselves, rather than the support that was available to them. The implications of these findings for future
practice, programs, and research are discussed.
Beckley, M. N. (2006). Community participation following cerebrovascular accident: Impact of the buffering model of social
support. American Journal of Occupational Therapy, 60, 129–135.
erebrovascular accident (CVA), or stroke, is a major cause of adult disability in
the United States (American Heart Association, 2005). With almost 730,000
people now experiencing a CVA each year (National Stroke Association, 1999),
more outcome studies are being undertaken to help in the development of rehabilitation services to support CVA survivors. Starting in the 1970s, health care
researchers became interested in studies of social support outcomes. Until now, one
of the main focuses of outcome studies on social support in CVA has been a person’s ability to carry out activities of daily living (ADL). Although this is often a
prerequisite for discharge from hospital, it does not guarantee that a person will be
able to participate successfully in the community. This study builds on previous
CVA outcome studies by examining the impact of social support in a larger context, specifically testing the buffering model of social support on a person’s ability
to participate in the community after CVA.
C
Literature Review
Margaret Newsham Beckley, PhD, OTR/L, BCN, BCG, is
Assistant Professor, The Ohio State University, College of
Medicine and Public Health, School of Allied Medical
Professions, 453 West Tenth Avenue, Columbus, Ohio
43210; [email protected]
Social support has been found to have a positive impact on people’s health and
well-being, as well as their ability to adjust to the trauma of illness or injury
(Bisconti & Bergeman, 1999; Caplan, 1974; Cassel, 1974; Cobb, 1976; Hegelson,
2003; Lau & McKenna, 2001; Newsham, 1998; Tsouna-Hadjis, Vemmos,
Zakopoulos, & Stamatelopoulos, 2000). A concept that emerged in the 1970s,
social support has been described as the availability or provision of a relationship,
information, or assistance that empowers a person to manage their day-to-day life
effectively in the presence or absence of crisis (Newsham). The buffering model and
the direct effect model are those models most frequently cited in the social support
literature (Vaux, 1988). The buffering model rests on the hypothesis that an identifiable form of social support has a beneficial effect in the presence of stress. The
direct effect model rests on the hypothesis that ongoing social support resources
have a positive influence on health, whether stress is present or not.
The buffering model of social support represents social support as a specific
response to a stressor, with the purpose of counteracting or reducing the possibility of a negative outcome. Cohen and Syme (1985) report that the buffering model
characterizes support as exerting beneficial effects in the presence of stress by
The American Journal of Occupational Therapy
129
Downloaded From: http://ajot.aota.org/ on 07/28/2017 Terms of Use: http://AOTA.org/terms
protecting people from the pathogenic effects of stress. In
measuring support as depicted by the buffering model, one
would assess the availability of resources that help a person
respond to stressful events. Vaux (1988) reports the buffering model describes social support as acting to protect individuals from the effect of stressful conditions. “Generally,
this has been interpreted to mean that the relationship
between stressful life experiences and psychological distress
on physical illness would be diminished under conditions
of greater social support” (p. 92).
Figure 1 (Newsham, 1998) is a representation of the
impact of the buffering model of social support, incorporating the theoretical work of Cohen & Syme (1985) and
Vaux (1988), with regard to stroke survivors. The independent variable, functional limitation, is the stressor. Health
and well-being are represented by the dependent variable,
community participation. The diagram shows that social
support, by itself, may or may not be significant in regard
to a stroke survivor’s ability to participate in his or her community. The second relationship on Figure 1 shows that
functional limitation, by itself, may or may not be significant in regard to a stroke survivor’s ability to participate in
his or her community. However, based on the buffering
model, when social support and functional limitation interact with each other, there will be a significant relationship
between this interaction and a stroke survivor’s ability to
participate in his or her community (Newsham).
The direct effect model and the buffering model are
not the only models found in the social support literature;
however, these models have guided most of the research on
social support for the past 20 years. In this study, the buffering model is the model to be tested. The buffering model
requires the identification of a stressor (functional limitation due to stroke) and the presence of social support to
reduce the impact of the stressor on a person’s well-being (in
this case, the stroke survivor’s ability to participate in his or
Figure 1. Buffering model of social support and community
participation related to stroke (Newsham, 1998).
Based on the work of Cohen & Syme, 1985 and Vaux, 1988.
her community). The premise of this study is that the influence of social support on functional limitation enhances the
ability of a stroke survivor to participate in his or her community. There is an interaction between social support and
functional limitation that reduces the impact of the limitation on the outcome. It is the interaction of the social support variable that distinguishes the buffering model of social
support from the direct effect model.
Social Support and Cerebrovascular Accident
Social support is a multifaceted concept. As such, social
support studies in stroke recovery have investigated different aspects of the social support concept, such as the types
of support, the timing of support, and the structure of support. Although CVA research studies have only recently
begun to include social support as a predictor variable, it
has already been found to enhance CVA recovery
(Colantonio, Kasl, Ostfeld, & Berkman, 1993; Friedland
& McColl, 1992; Glass & Maddox, 1992; Lau &
McKenna, 2001; Tsouna-Hadjis et al., 2000).
In a longitudinal study (N = 44), Glass and Maddox
(1992) found that different types of support influence rehabilitation outcomes in different ways. Multivariate analyses
of variance indicated that patients with a high level of emotional support showed dramatic improvement in functional outcome, though they were often the patients with the
greatest impairment at the start of rehabilitation. Instrumental support, provided in moderate amounts, was most
closely related to positive outcomes. And informational
support was found to be mediated by the severity of the
stroke.
The timing of the provision of social support has an
impact in CVA outcomes (Helgeson, 2003). Social support
intervention provided at an inappropriate time during the
stoke recovery process may not have an effect on functional status (Glass & Maddox, 1992; Friedland & McColl,
1992). However, in a randomized trial with 48 experimental participants and 40 control participants, Friedland and
McColl (1992) hypothesized that, to be most effective,
social support needs to be provided as soon as a person is
medically stable. Measures were taken at three points: entry
into the study, immediately after intervention, and 3
months after intervention. No significant treatment effect
was found and no significant differences between the
groups regarding social support and outcome were reported. The investigators report that this lack of significant findings could have been due to the timing of the intervention,
the high incidence of psychiatric symptoms found in the
participants, the nature of the support given to the participants during the crisis period, or the older age of the participants. Another stroke study from McColl and Friedland
130
Downloaded From: http://ajot.aota.org/ on 07/28/2017 Terms of Use: http://AOTA.org/terms
March/April 2006, Volume 60, Number 2
(1989) (N = 85) assessed the type of social support in rehabilitation and demonstrated that social support from personal sources accounted for 79% of total variance
explained. The measures considered satisfaction and quality
of support along with practical and descriptive aspects of
support. In a prospective longitudinal study of 2,812 elderly participants (65 years of age and older) living at home,
Colantonio and colleagues (1993) examined the structure
of social support. The researchers found the social network
index to be a significant predictor of institutionalization
after CVA: The more securely an elderly patient was embedded in a social network before the stroke, the less need he or
she had to live in an institution.
Because the social support concept is multifaceted, the
impact of social support in stroke recovery needs to be
examined in order to identify the optimal type of support,
the phase of recovery in which the support is provided, and
the source of support. This will allow a more confident
interpretation of the impact of social support on health outcomes and community participation.
Methodology
The purpose of this study was to examine the impact of the
buffering model of social support on stroke outcomes. The
specific aim was to test the impact of buffering model on
reported functional limitation with regard to community
participation 3 to 6 months posthospitalization for stroke.
Based on the work of Cohen and Syme (1985), Newsham
(1998), and Vaux (1988), the hypothesis for this study was
that social support moderates the effect of functional limitation on community participation for CVA survivors who
are 3 to 6 months post hospital discharge.
Participants and Sample Size
The study was undertaken via interviews in the homes of
persons having a CVA and was conducted 3 to 6 months
after discharge from the hospital by occupational therapists.
Inclusion criteria were: (1) the CVA resulted from an
ischemic or hemorrhagic event, and (2) cognitive status was
such that participants were able to give informed consent to
participate (capable of attending to task, comprehending
questions or formulating appropriate responses, and understanding the purpose of the study). An exclusionary criterion was the presence of receptive or expressive aphasia. The
sample size was 95 participants.
Measurement of Variables
Dependent variable. Community participation, the depen-
dent variable, was measured with the Reintegration to
Normal Living Index (RNL Index) (Wood-Dauphinee,
Table 1. Variable Descriptions
Participation
Social Support
Quality
Quantity
Instrumental
Subjective
Functional Limitation
Gender
Male
Female
Race
Black
White
Monthly Income
$0–$700
$701–$1,400
$1,401–$2,100
$2,101–$2,800
Over $2,800
Age
N
Mean
SD
95
41.22
8.11
95
95
95
95
95
123.68
35.39
23.85
10.07
114.63
20.06
11.88
3.28
2.15
18.75
68.46
12.16
47
48
52
43
27
18
15
6
18
95
Opzoomer, Williams, Marchand, & Spitzer, 1988). It is an
11-item scale that measures managing in the home environment and community, such as the ability to move
around living quarters and the community, being occupied
in meaningful activity, participating in social activities, and
dealing with life events. The response categories are on a 5point Likert scale, ranging from “strongly disagree” to
“strongly agree.”
Independent variable. Social support was measured with
two different scales. The Social Support Inventory for
People With Acquired Disabilities (SSIPAD) (Friedland &
McColl, 1989) was used to measure the quality and the
quantity of social support. The Duke Social Support Index
(Hughes, Blazer, & Hybels, 1990) was used to measure
instrumental and subjective support. The SSIPAD measures
subscales for each of five sources of support (primary, family and friends, community individuals, community groups,
and professionals) for overall quality of support. The quality measure consists of 25 items and the quantity measure
consists of 10 items. For both scales, the greater the sum of
scores, the higher the quality or quantity of support. The
Duke Social Support Index (Hughes et al.) also measures
multiple dimensions of support. This index is a compilation
of four social support subscales: social network, social interaction, subjective social support, and instrumental social support. The two subscales used in this study were the subjective social support scale with 10 items and the instrumental
social support subscale with 13 items.
Functional limitation was recorded by asking the participants about their ability to perform specified ADL. The
interviewer then rated their reported performance according to the level of assistance needed to complete the activities on a scale of 1 to 7. The nonstandardized inventory
The American Journal of Occupational Therapy
Downloaded From: http://ajot.aota.org/ on 07/28/2017 Terms of Use: http://AOTA.org/terms
131
included 18 items incorporating activities in the areas of
self-care, bowel and bladder management, mobility, locomotion, communication, and social cognition.
Control variables. Income level was included as a control
variable because it is often associated with health status and
social connectedness—the greater the income, the greater
the potential for access to health care and social interactions
(Altman & Taylor, 2001). Income was measured in dollars
per month. The measurement of age—in years—was
included because mobility, activity levels, and frequency of
socialization can decrease as a person ages (Gallo, Reichel,
& Anderson, 1995). Race was operationalized as white or
black, as this urban population is predominantly white and
black. Race was included because these two races have been
shown to have different types of support systems (Dressler,
1990). Gender was included as men and women have differences in support and community participation (Treas &
Longino, 1997).
Analysis and Findings
The hypothesis for this study was that social support would
moderate the effect of functional limitation on community
participation for stroke survivors 3 to 6 months posthospital discharge. The findings of this study partly supported
the hypothesis. The interaction term that included subjective social support was found to moderate the relationship
between reported functional limitation and community
participation. This interaction effect means that the relationship between reported functional limitation and community participation depended on the level of subjective
social support. The interaction terms that included instrumental social support, quality of social support, and quantity of social support were not found to moderate functional limitation on community participation. Another way to
interpret the findings is that a stroke survivor’s ability to do
for him- or herself was more indicative of participating in
his or her community than the amount of support available
to that stroke survivor.
Analysis. The analysis included measures of central tendency and dispersion to determine the distribution of community participation and independent variables. A bivariate
analysis was done to determine the relationship between
community participation and the independent variables.
Ordinary least squares analysis was used to determine the
amount of variance the independent variables accounted for
in the dependent variable, community participation, and
was accomplished by regressing the dependent variable on
the independent variables. The independent variables
regressed on community participation included the interaction term (functional limitation * social support). The
interaction term was entered to determine if a preexisting
relationship between the two variables could account for
additional variance in participation. It was entered last in
the hierarchical model, so any variance explained was due
to the interaction term only after the functional limitation
and social support variables had been included as main
effects. Because there are four different aspects of social support to evaluate, four separate equations were used, each
being identical except for the social support variable (quality, quantity, subjective, and instrumental) and the interaction term that included the social support variable [(functional limitation * subjective social support), (functional
limitation * instrumental social support), (functional limitation * quality of social support), and (functional limitation * quantity of social support)].
To counteract problems of multicollinearity with the
regression models incorporating the interaction terms, the
independent variables were centered before using them in
the interaction terms. The variables age, gender, income,
and race were included as control variables in each equation
with regard to the dependent variable community participation. The variables functional limitation and social support were entered after the control variables as the main
effects on community participation. The interaction terms
for each equation were entered last in the hierarchical
model. Because of this procedure, any variance explained
was due to the interaction terms.
Findings. The findings of the statistical analysis for this
study and the relationship among the variables are presented as follows: univariate, bivariate, and multivariate statistics. The univariate results are presented in Table 1 and
include the number of participants (N), the mean, standard
deviation, skewness, range of values, and the percentage of
participants included in the analysis for each variable.
Bivariate results are presented in Table 2 and include the
correlation coefficient, standard error, and probability level
for each independent variable and the dependent variable
Table 2. Bivariate Results
GLM Procedure: Participation
With Each Independent Variable
Independent Variable
Age
Gender
Income
Race
Functional Limitation
Subjective Social Support
Instrumental Social Support
Quality of Social Support
Quantity of Social Support
Correlation
Coefficient
(r)
Standard
Error
p value
.071
.099
.001
.044
.570
.024
.387
.114
.176
0.07
1.72
0.56
1.75
0.04
0.42
0.25
0.04
0.07
.51
.36
.99
.69
.001
.82
.0002
.29
.09
Note. GLM = General Linear Model procedure is used to regress the dependent variable on the independent variables and is used when there is only one
dependent variable.
132
Downloaded From: http://ajot.aota.org/ on 07/28/2017 Terms of Use: http://AOTA.org/terms
March/April 2006, Volume 60, Number 2
was a normal distribution of the error term. Plots of predicted values versus error terms supported assumptions of
homogeneity and linearity. All of these analyses indicate
that the variables and equations used in this study were
sound, allowing for confidence in findings of the study.
The significant interaction term (functional limitation
* subjective social support) (p = .008) means that subjective
support was found to moderate the effect of functional limitation on participation. The model had an R 2 of 0.33,
accounting for 33% of the variance in community participation. The interaction terms (functional limitation *
instrumental social support), (functional limitation * quality of social support), and (functional limitation * quantity of social support) were not significant with regard to
participation.
community participation. Correlations among the independent variables are presented in Table 3. The multivariate
results are presented in Table 4 include the percentage of
variance explained (R 2) and the significance of the model
(F value) with the corresponding probability level. These
tables also include measures of standard error, probability
level, and the increment to the R 2 for each variable and the
interaction term in the model.
The hypothesis for this study was that social support
moderates the effect of functional limitation on community participation for CVA survivors 3 to 6 months after hospital discharge. Regression diagnostics show that the
assumptions of regression were not violated. Multicollinearity among the variables did not exist, as all tolerance values of significant variables were above 0.76. There
Table 3. Correlation Matrix of Independent Variables
Pearson Correlation
& Probability
Age
Functional
Limitation
Age
Gender
Race
Income
1.000
0.0
–0.05116
0.6439
1.0000
0.0
0.12587
0.2539
–0.2370
0.03*
1.0000
0.0
0.21593
0.0485*
–0.37120
0.0005*
0.49345
0.0001*
1.0000
0.0
Gender
Race
Income
Functional Limitation
–0.14932
0.1752
–0.00876
0.9369
0.10398
0.3465
0.10108
0.3603
1.0000
0.0
Quantity of Social Support
Quantity
of Social
Support
Quality
of Social
Support
Subjective
Social
Support
0.00409
0.9705
–0.3228
0.7707
0.03472
0.7539
–0.08847
0.4235
–0.00894
0.9356
1.00000
0.0
0.5825
0.5986
–0.5667
0.6087
–0.07728
0.4847
–0.15098
0.1704
–0.03041
0.7836
0.89402
.0001*
1.00000
0.0
–0.7264
0.5114
0.11966
0.2783
–0.35229
0.0010*
–0.00551
0.9603
0.18271
0.0962
–0.02523
0.8198
0.06241
0.5728
1.00000
0.0
Quality of Social Support
Subjective Social Support
Instrumental Social Support
Instrumental
Social
Support
–0.09484
0.3908
0.00499
0.9640
–0.16748
0.1278
0.03450
0.7554
0.24350
0.0256*
–0.13623
0.2166
–0.02821
0.7989
0.48812
0.0001*
1.00000
0.0
* significant relationships (p ≤ .05)
Table 4. Results of Ordinary Least Squares Procedures
Ordinary Least Squares
Procedure With Interaction
Term Functional Limitation *
Quality of Social Support
on Participation
Ordinary Least Squares
Procedure With Interaction
Term Functional Limitation *
Quantity of Social Support
on Participation
0.32
5.02
.0001
0.35
5.72
.0001
R2
F value
Probability Value
Independent Variable
Age
Gender
Income
Race
Functional Limitation
Social Support
Functional Limitation *
Social Support
Ordinary Least Squares
Procedure With Interaction
Term Functional Limitation *
Subjective Social Support
on Participation
Ordinary Least Squares
Procedure With Interaction
Term Functional Limitation *
Instrumental Social Support
on Participation
0.33
5.30
.0001
0.31
4.82
.0002
Standard Probability Increment Standard Probability Increment Standard Probability Increment Standard Probability Increment
Error
Value
to the R 2
Error
Value
to the R 2
Error
Value
to the R 2
Error
Value
to the R 2
0.06
1.6
0.6
1.67
0.03
0.04
0.001
.77
.32
.81
.34
.0001
.03
0.47
.004
.0106
.0031
.0028
.2480
.0457
0.06
1.5
0.57
1.64
0.03
0.06
.0047
0.002
.81
.26
.83
.47
.0001
.006
0.65
.004
.0106
.0030
.0028
.2480
.0778
0.059
1.55
0.59
1.83
0.03
0.03
.0018
0.001
The American Journal of Occupational Therapy
Downloaded From: http://ajot.aota.org/ on 07/28/2017 Terms of Use: http://AOTA.org/terms
.96
.19
.30
.13
.0001
.22
0.008
.004
.0106
.0031
.0028
.2480
.0101
.06
1.56
0.59
1.7
0.03
0.03
.0522
0.001
.67
.23
.79
.68
.0001
.05
0.54
.004
.0106
.0031
.0028
.2480
.0384
.0035
133
Again, the findings partly support the hypothesis that
social support moderates the effect of functional limitation
on community participation for CVA survivors 3 to 6
months after hospital discharge. The interaction term that
included subjective social support was found to moderate
the relationship between reported functional limitation and
community participation. More specifically, with every unit
increase of subjective social support, the parameter estimate
of functional limitation increased by 0.003. The other
interactions terms that included different aspects of social
support did not support the hypothesis.
Limitations of the Study
One of the limitations of this study is that it cannot be generalized, neither to other diagnostic populations nor to the
stroke population overall. As a result, it may be that the
findings are unique enough to these particular participants
that they may not provide accurate implications for other
populations. The study population had a mean reported
functional limitation score of 114.6 out of a possible 126.
The study population therefore represented approximately
40% of the overall stroke population (Ryerson, 1995). The
restricted range of reported functional limitation most likely influenced the results of the study. The participants that
met this criterion not only reported being able to perform
ADL with a fairly high degree of independence, but they
also had the necessary support at home to compensate for
what they were unable to complete for themselves.
Therefore, this study should not be generalized to stroke
survivors with lower functional limitation scores and different levels of support in the home.
The measurement of social support may be another
limitation. The Duke Social Support Index (Hughes et al.,
1990) and the SSIPAD (Friedland & McColl, 1989) were
used to measure instrumental and subjective social support
and the quality and quantity of such support, respectively.
Although these scales have been shown to be appropriate for
measuring specific components of social support, these specific components constitute only four aspects of the multifaceted social-support concept; other aspects of social support were not captured. Also, as noted previously, McColl
(2005) has shown that subjective criteria are used in measuring social support. This is implicit in measuring a concept that has both tangible and nontangible components.
A third limitation of this study is the attempt to capture aspects of a complex concept, participation, with a single measure. In this study the outcome scores, as measured
by the RNL Index (Wood-Dauphinee et al., 1988), were
favorable (mean score 41.22 out of a possible 55). The
results indicate that the participants were found to be capa-
ble of community participation. Over the course of the
interview however, participants often reported that they
were avoiding these activities for reasons such as “not feeling normal,” “not looking right,” feeling that “people don’t
want to be bothered,” and “feeling blue.” In considering
this discrepancy (participants reporting to be capable of performing the community participation tasks, yet not performing them), the initial positive appearance of the study
results may not be a reality in the participants’ day-to-day
lives. There appear to be other components of the participation process that were not captured with the RNL Index
and that may as a result limit applicability of these findings.
Implications for Practice, Programs,
and Future Studies
The findings of this study partly support the hypothesis
that social support moderates the effect of functional limitation on community participation for CVA survivors who
are 3 to 6 months post hospital discharge. These findings
suggest the need for occupational therapy professionals to
expand their practices to include the promotion of social
support and, in turn, increase the stroke survivor’s ability
to participate in the community. Intervention practices in
rehabilitation programs will need to go beyond traditional
clinical training and extend the focus of treatment to practices that foster social support and community participation. In addition, future rehabilitation studies would benefit from a focus on social support needs in discharge
planning. Although the participants in this study reported
to function at a fairly high level and had support available
at home, they often reported that they didn’t want to participate in community activities. Examining factors other
than functional level and support in the home may yield
results that could eventually support people who want to
participate in their community and, in turn, help occupational therapy professionals develop more comprehensive
discharge plans.
Finally, with almost 730,000 people now experiencing
a CVA each year (National Stroke Association, 1999),
future stroke outcome studies could build on the findings
presented here to help in the development of occupational
therapy services. Future studies could expand these current
findings by including participants with varying levels of
functional limitations, by initiating longitudinal studies
over participants’ remaining life spans, by determining what
types of support are most beneficial at each stage of rehabilitation, and by examining younger stroke survivors, who
typically have different social support systems from the participants in this study. ▲
134
Downloaded From: http://ajot.aota.org/ on 07/28/2017 Terms of Use: http://AOTA.org/terms
March/April 2006, Volume 60, Number 2
References
Altman, B. M., & Taylor, A. K. (2001). Women in the health care
system: Health status, insurance, and access to care. MEPS
Research Findings No.17. AHRQ Pub. No. 02-0004. Rockville, MD: Agency for Healthcare Research and Quality.
American Heart Association. (2005). Heart disease and stroke
statistics—2005 update. Dallas, TX: American Heart
Association.
Bisconti, T. L., & Bergeman, C. S. (1999). Perceived social control as a mediator of the relationships among social support,
psychological well-being, and perceived health. Gerontologist,
39(1), 94–103.
Caplan, G. (1974). Support systems and community mental health:
Lectures on concept development. New York: Behavioral
Publications.
Cassel, J. (1974). Psychosocial process and “stress”: Theoretical
formulations. International Journal of Health Services, 4,
471–482.
Cobb, S. (1976). Social support as a moderator of life stress.
Psychosomatic Medicine, 38, 300–314.
Cohen, S., & Syme, S. L. (1985). Social support and health.
Orlando, FL: Academic Press.
Colantonio, A., Kasl, S., Ostfeld, A., & Berkman, L. (1993).
Psychosocial predictors of stroke outcomes in an elderly population. Journal of Gerontology: Social Sciences, 48, S261–S268.
Dressler, W. W. (1990). Studies of stress and disease. In W. W.
Dressler (Ed.), Stress and adaptation in the context of culture:
Depression in a southern black community. New York: State
University of New York Press.
Friedland, J., & McColl, M. (1989). Social support and psychosocial dysfunction after stroke: Buffering effects in a community sample. Archives of Physical Medicine and Rehabilitation, 68, 475–480.
Friedland, J., & McColl, M. (1992). Social support intervention
after stroke: Results of a randomized trial. Archives of Physical
Medicine and Rehabilitation, 73, 573–581.
Gallo, J. J., Reichel, W., & Andersen, L. M. (1995). Handbook of
geriatric assessment (2nd ed.). Gaithersburg, MD: Aspen.
Glass, T., & Maddox, G. (1992). The quality and quantity of
social support: Stroke recovery as psycho-social transition.
Social Science Medicine, 34, 1249–1261.
Hegelson, V. S. (2003). Social support and quality of life. Quality
of Life Research, 12, 25–31.
Hughes, D. C., Blazer, D., & Hybels, C. (1990). Duke social support index: A working paper. Unpublished manuscript.
Lau, A., & McKenna, K. (2001). Conceptualizing quality of life
for elderly people with stroke. Disability and Rehabilitation,
23, 227–238.
McColl, M. (2005). Social support and occupational therapy. In
C. Christiansen & C. Baum (Eds.), Occupational therapy:
Performance, participation, and well-being (3rd ed.). Thorofare, NJ: Slack.
McColl, M., & Friedland, J. (1989). Development of a multidimensional index for assessing social support in rehabilitation.
Occupational Therapy Journal of Research, 9, 219–234.
National Stroke Association. (1999). Stroke statistics. Englewood,
CO: Author.
Newsham, M. M. (1998). Social support and community reintegration after stroke. Dissertation Abstracts International, 60
(02), p. 551. (UMI No. AAT 9918660)
Ryerson, S. D. (1995). Hemiplegia. In D. A. Umphred (Ed.),
Neurological rehabilitation (3rd ed., pp. 681–721). St. Louis,
MO: Mosby.
Treas, J., & Longino, C. F. (1997). Demography of aging in the
United States. In K. F. Ferraro (Ed.), Gerontology Perspectives
and Issues (2nd ed., pp. 19–50). New York: Springer.
Tsouna-Hadjis, E., Vemmos, K., Zakopoulos, N., &
Stamatelopoulos, S. (2000). First-stroke recovery process:
The role of family social support. Archives of Physical Medicine and Rehabilitation, 81, 881–887.
Vaux, A. (1988). Social support: Theory, research, and intervention.
New York: Praeger.
Wood-Dauphinee, S. L., Opzoomer, A., Williams, J. I.,
Marchand, B., & Spitzer, W. O. (1988). Assessment of global function: The Reintegration to Normal Living Index.
Archives of Physical Medicine and Rehabilitation, 69,
583–590.
The American Journal of Occupational Therapy
Downloaded From: http://ajot.aota.org/ on 07/28/2017 Terms of Use: http://AOTA.org/terms
135