Association of social support with outcomes in incident dialysis patients.

Abstract

The association of social support with outcomes in ESRD, overall and by peritoneal dialysis (PD) versus hemodialysis (HD), remains understudied.
In an incident cohort of 949 dialysis patients from 77 US clinics, we examined functional social support scores (scaled 0 to 100 and categorized by tertile) both overall and in emotional, tangible, affectionate, and social interaction subdomains. Outcomes included 1-year patient satisfaction and quality of life (QOL), dialysis modality switching, and hospitalizations and mortality (through December 2004). Associations were examined using overall and modality-stratified multivariable logistic, Poisson, and Cox proportional hazards models.
We found that mean social support scores in this population were higher in PD versus HD patients (overall 80.5 versus 76.1; P < 0.01). After adjustment, highest versus lowest overall support predicted greater 1-year satisfaction and QOL in all patients (odds ratio 2.47 [95% confidence interval (CI) 1.18 to 5.15] and 2.06 [95% CI 1.31 to 3.22] for recommendation of center and higher mental component summary score, respectively). In addition, patients were less likely to be hospitalized (incidence rate ratio 0.86; 95% CI 0.77 to 0.98). Results were similar with subdomain scores. Modality switching and mortality did not differ by social support in these patients, and associations of social support with outcomes did not generally differ by dialysis modality.
Social support is important for both HD and PD patients in terms of greater satisfaction and QOL and fewer hospitalizations. Intervention studies to possibly improve these outcomes are warranted.

Full text

Page 1

Association of Social Support with Outcomes in Incident
Dialysis Patients
Laura C. Plantinga,* Nancy E. Fink,†‡Rachel Harrington-Levey,§Frederic O. Finkelstein,?
Nasser Hebah,¶Neil R. Powe,* and Bernard G. Jaar†‡**
*Department of Medicine, San Francisco General Hospital and University of California, San Francisco, California;
†Department of Epidemiology, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland;‡Department of
Medicine, Johns Hopkins University School of Medicine, Baltimore, Maryland;§Department of Psychology, Edith Nourse
Rogers Memorial Veteran’s Hospital, Bedford, Massachusetts;?Department of Medicine, Yale University, New Haven,
Connecticut;¶Dialysis Clinic, Inc., Nashville, Tennessee; and **Nephrology Center of Maryland, Baltimore, Maryland
Background and objectives: The association of social support with outcomes in ESRD, overall and by peritoneal dialysis
(PD) versus hemodialysis (HD), remains understudied.
Design, setting, participants, & measurements: In an incident cohort of 949 dialysis patients from 77 US clinics, we examined
functional social support scores (scaled 0 to 100 and categorized by tertile) both overall and in emotional, tangible, affectionate,
and social interaction subdomains. Outcomes included 1-year patient satisfaction and quality of life (QOL), dialysis modality
switching, and hospitalizations and mortality (through December 2004). Associations were examined using overall and
modality-stratified multivariable logistic, Poisson, and Cox proportional hazards models.
Results: We found that mean social support scores in this population were higher in PD versus HD patients (overall 80.5
versus 76.1; P < 0.01). After adjustment, highest versus lowest overall support predicted greater 1-year satisfaction and QOL
in all patients (odds ratio 2.47 [95% confidence interval (CI) 1.18 to 5.15] and 2.06 [95% CI 1.31 to 3.22] for recommendation of
center and higher mental component summary score, respectively). In addition, patients were less likely to be hospitalized
(incidence rate ratio 0.86; 95% CI 0.77 to 0.98). Results were similar with subdomain scores. Modality switching and mortality
did not differ by social support in these patients, and associations of social support with outcomes did not generally differ by
dialysis modality.
Conclusions: Social support is important for both HD and PD patients in terms of greater satisfaction and QOL and fewer
hospitalizations. Intervention studies to possibly improve these outcomes are warranted.
Clin J Am Soc Nephrol 5: 1480–1488, 2010. doi: 10.2215/CJN.01240210
S
health outcomes in a variety of chronic health conditions (1),
including heart disease (2,3) and diabetes (4); however, the
association of social support with health outcomes remains an
understudied problem in chronic kidney disease and ESRD,
with little previous evidence of an association of social support
with improved outcomes (5–8). Most previous studies of social
support have focused on mortality alone, and results have been
contradictory (7,8).
Better social support could increase dialysis patients’ satis-
ocial support is widely thought to have the potential to
improve patients’ health outcomes. Perceived social
support has been shown to be associated with improved
faction with their care and their overall perceived health-re-
lated quality of life (HRQOL). In addition, social support could
provide the means for better treatment, medication adherence,
and nutrition, leading to better clinical outcomes. Among those
receiving dialysis, levels of social support and associations of
social support with outcomes may differ between the dialysis
modalities, because hemodialysis (HD) and peritoneal dialysis
(PD) differ dramatically in the amount of self-care required.
In a national cohort of incident HD and PD patients, we
assessed the level of social support overall and by dialysis
modality. We also examined the associations of social support
with patient satisfaction, HRQOL, and clinical outcomes, in-
cluding modality switching, hospital admissions, and all-cause
mortality. Finally, we assessed whether these associations dif-
fered by dialysis modality.
Materials and Methods
Study Design
Our cohort for this study consisted of 949 incident dialysis patients
(enrolled October 1995 through June 1998) who were treated at 77
not-for-profit, free-standing outpatient dialysis clinics in 19 states
throughout the United States. The study cohort was assembled from
Received February 9, 2010. Accepted April 6, 2010.
Published online ahead of print. Publication date available at www.cjasn.org.
The content is solely the responsibility of the authors and does not necessarily
represent the official views of the National Institute Of Diabetes And Digestive
And Kidney Diseases or the National Institutes of Health.
Correspondence: Ms. Laura Plantinga, San Francisco General Hospital and Uni-
versity of California, San Francisco, 1001 Potrero Avenue, Building 10, Floor 3,
San Francisco, CA 94110. Phone: 415-206-6007; Fax: 415-206-5586; E-mail:
plantingal@medsfgh.ucsf.edu
Copyright © 2010 by the American Society of NephrologyISSN: 1555-9041/508–1480

Page 2

the previously described Choices for Healthy Outcomes in Caring for
End-Stage Renal Disease (CHOICE) study (9) cohort.
Data Collection
Social Support.
dated cross-sectional, self-reported questionnaire, the Medical Out-
comes Study (MOS) Social Support Survey (10), at study enrollment.
The survey consists of 19 items that measure functional social support
in four dimensions: Emotional/informational, tangible, affectionate,
and positive social interaction (Table 1). Overall and domain-specific
scores were calculated by averaging across all and domain-specific
items, respectively, and transforming the average scores to a 100-point
scale: 100 * [(average score ? minimum possible score)/(maximum
possible score ? minimum possible score)]. At least two item responses
within each subdomain were required to calculate subdomain scores,
and scores in all four dimensions were required to calculate an overall
functional support score.
Patient Satisfaction.
Patient satisfaction was measured as part of
the CHOICE Health Experience Questionnaire (CHEQ) (11), a patient
self-report questionnaire that was administered at study enrollment
and at 1 year. Patients were asked, “How would you rate the quality of
care you have received as a dialysis patient, overall?” and, “Would you
recommend your dialysis center to a friend or relative who needs
dialysis?” Patients responded on a five-point scale with categories of
excellent/definitely yes, very good/probably yes, good/not sure, fair/
probably not, or poor/definitely not.
HRQOL.
HRQOL scores at baseline and 1 year were also ob-
tained from the CHEQ (11), which included the SF-36 health survey
items (12). Physical component summary (PCS) and mental compo-
nent summary (MCS) scores were calculated from the scores on the
SF-36 domains (12).
Patients’ social support was assessed by a vali-
Modality Switching.
the modality (PD or HD) in use at 4 weeks after enrollment in the study
and was obtained from clinic records. Patients were considered to have
switched modality when they remained on the new modality for at
least 30 days (13). Patients were censored for time to switch at death,
transplantation, or last date of follow-up (December 31, 2004).
Hospitalization.
Number of hospital admissions was determined
from US Renal Data System inpatient files. Follow-up time was calcu-
lated from study enrollment or start of Medicare coverage to death,
transplantation, or last date of follow-up (December 31, 2004); time
during hospital admissions was excluded.
All-Cause Mortality.
All-cause mortality information was ascer-
tained from National Death Index records, clinic reports, medical
records, and Centers for Medicare & Medicaid Services (death notifi-
cation forms and Social Security records). Follow-up for mortality
continued until death, transplantation, or the last follow-up date (De-
cember 31, 2004).
Other Variables.
Data on patients’ demographics (age, gender,
and race) and socioeconomic status (education, employment, and mar-
ital status) were collected from a baseline self-report questionnaire.
Presence and severity of comorbid conditions were assessed at baseline
using the Index of Coexistent Disease (14,15). Laboratory values (albu-
min, creatinine, and hemoglobin) and height and weight (used to
calculate body mass index) were obtained from patients’ records at
baseline and from Centers for Medicare & Medicaid Services Form
2728, respectively. Referral to a nephrologist was obtained as described
previously (16).
Baseline dialysis modality was defined as
Statistical Analysis
The overall distributions of social support domain scores were ob-
tained and compared by dialysis modality using Wilcoxon rank-sum
Table 1. MOS social support survey instrument items
Support Domain Items
Emotional/informational Someone you can count on to listen to you when you need to talk
Someone to give you information to help you understand a situation
Someone to give you good advice about a crisis
Someone to confide in or talk to about yourself and your problems
Someone whose advice you really want
Someone to share your most private worries and fears with
Someone to turn to for suggestions about how to deal with a personal problem
Someone who understands your problems
Someone to help you if you were confined to bed
Someone to take you to the doctor if you needed it
Someone to prepare your meals if you were unable to do it yourself
Someone to help with daily chores if you were sick
Someone who shows you love and affection
Someone to love and make you feel wanted
Someone who hugs you
Someone to have a good time with
Someone to get together with for relaxation
Someone to do something enjoyable with
Someone to do things with to help you get your mind off things
Tangible
Affectionate
Positive social interaction
Additional itema
The overall question was phrased, “How often is each of the following kinds of support available to you if you need it?”
Possible responses were “none of the time,” “a little of the time,” “some of the time,” “most of the time,” and “all of the
time.” Adapted from reference (10), with permission.
aIncluded in overall social support score calculation.
Clin J Am Soc Nephrol 5: 1480–1488, 2010Social Support and Outcomes1481

Page 3

tests. Patient characteristics were examined by tertile of baseline social
support score, using Pearson ?2tests for categorical variables and
ANOVA for continuous variables. For associations of social support
with outcomes, we compared highest and lowest tertiles of social
support. Satisfaction scores at 1 year were dichotomized as highest
versus other ratings, and HRQOL scores at 1 year were dichotomized by
the median score; odds ratios from logistic regression models were
used to assess the relationships between social support and these
outcomes. Cox proportional hazards models were used to obtain rela-
tive hazards of modality switching and all-cause mortality for social
support scores; Poisson models were used to obtain incidence rate
ratios (IRRs). Variables were chosen as covariates for the adjusted
models when they were potential confounders. Because some variables
could be considered in the causal pathway (e.g., marital status, dialysis
modality, baseline satisfaction/HRQOL scores), variables were added
stepwise. Sensitivity analyses, including interactions by modality; ex-
amination of hospitalizations, mortality, and modality switching within
the first 6 months and first year; and examination of change from
baseline to 1 year in HRQOL and satisfaction, were also performed. We
accounted for possible dependence of observations within clinics by
performing modeling on the basis of robust variance–covariance ma-
trices. All analyses were performed using Stata 10.0 (Stata Corp., Col-
lege Station, TX).
Results
Participant Characteristics and Social Support Scores
A total of 949 (91.2%) of 1041 patients who were treated at
77 clinics in 19 states provided sufficient information to
compute both an overall social support score and all four
social support subdomain scores. Figure 1 shows the distri-
butions of social support scores in our cohort. Scores of both
overall functional social support (Figure 1A) and all four
subdomains of social support (Figure 1, B through E) were
highly right-skewed, with mean (95% confidence interval
[CI]) social support scores as follows: Overall 77.2 (75.7 to
78.6), emotional 76.1 (74.5 to 77.6), tangible 79.9 (78.3 to 81.4),
affectionate 81.0 (79.3 to 82.7), and social interaction 73.0
(71.3 to 74.8). Scores were higher for PD versus HD patients
for all social support domains; for affectionate support, the
difference was marginally significant (P ? 0.05).
Baseline patient characteristics by tertile of overall social
support score are shown in Table 2. Having social support
scores in the highest tertile was statistically significantly asso-
ciated with female gender, white race, married status, greater
chance of being treated with PD, greater satisfaction with care,
and higher MCS score; conversely, these participants also had
greater comorbidity than those with lower scores. These asso-
ciations were similar for the subdomains of social support (data
not shown).
Association of Social Support with Dialysis Patient
Satisfaction and HRQOL
Table 3 shows the association of highest versus lowest social
support scores with patient satisfaction and HRQOL at 1 year.
Those with overall functional social support scores in the high-
est tertile were approximately 1.5 times more likely to rate their
quality of care at 1 year as “excellent”; among the four subdo-
mains of social support, only higher positive social interaction
was statistically significantly associated with excellent ratings
of care. For recommendation of center, those with scores in the
highest tertile were 2.5 times more likely “definitely” to recom-
mend their center at 1 year, and this association held for overall
functional social support and all four subdomains (Table 3). For
HRQOL, the highest level of social support was associated with
greater MCS scores (approximately twice as likely to have
higher MCS at 1 year) for overall support and all four subdo-
mains. For PCS, results were not statistically significant, except
for highest level of tangible support being associated with
1.7-fold greater chance of having a high PCS score at 1 year
(Table 3).
Overall, adjustment for demographics, marital status, dialy-
sis modality, and comorbidity did not change the associations
seen; adjustment for the baseline satisfaction or HRQOL did
often render the results nonstatistically significant, but the mag-
nitude of the associations remained (Table 3). Changes in score
from baseline to 1 year for satisfaction and HRQOL were mod-
est and not associated with baseline social support (data not
shown). Further adjustment for late evaluation (referred to a
nephrologist ?4 months before dialysis) did not substantially
change any of the associations of social support with patient
satisfaction or HRQOL (data not shown).
Association of Social Support with Clinical Outcomes in
Dialysis Patients
Of the clinical patient outcomes examined, only hospital-
ization was significantly associated with social support, with
a nearly 15% decreased risk for hospital admission among
those with highest versus lowest social support scores (Table 4).
This association was statistically significant for overall, emo-
tional, and affectionate support. Adjustment for demograph-
ics, marital status, dialysis modality, comorbidity, and albu-
min did not change the associations seen; again, further
adjustment for late evaluation did not substantially change
the associations of social support with decreased risk for
hospitalization (data not shown). The association was con-
sistent when hospitalizations were examined in the first 6
months (IRR 0.77; 95% CI 0.56 to 1.06) and the first year (IRR
0.83; 95% CI 0.67 to 1.02) only. Modality switching and
all-cause mortality were not associated with social support
either before or after adjustment for confounders (Table 4). A
nonstatistically significant association of higher social sup-
port with lower first-year mortality only was seen (relative
hazard 0.67; 95% CI 0.40 to 1.13), whereas the association of
switching with social support was not different at 6 months
or 1 year from that seen over the entire follow-up. Dialysis
modality–stratified analyses of the association of patient out-
comes with overall functional social support provided little
evidence of effect modification by modality for any outcome,
with the exception of a marginally significant (Pinteraction?
0.07) effect for recommendation of center at 1 year, with HD
patients being more than three times more likely to recom-
mend their center with higher social support but no associ-
ation among PD patients.
1482 Clinical Journal of the American Society of Nephrology Clin J Am Soc Nephrol 5: 1480–1488, 2010

Page 4

Discussion
We found that social support scores were high in our cohort
and that scores were higher in PD than in HD patients. Higher
social support, regardless of domain, was associated with pa-
tient-centered outcomes, including greater patient satisfaction
and HRQOL and reduced hospitalization; however, mortality
and modality switching were not associated with social sup-
port.
The mean scores in our cohort were generally higher (by 7 to
10 points) than the scores reported in the cohort of 19- to
95-year-olds who had one of four chronic conditions (coronary
heart disease, diabetes, hypertension, and depression) and with
whom the MOS survey was developed (10). The higher scores
in our cohort may be due to more thoughtful consideration of
all aspects of social support by those who were undergoing an
intensive treatment such as dialysis, or it could be that higher
social support may have provided a survival advantage to
those with earlier stage disease. It also possible that the higher
scores seen here reflect a “healthy cohort” effect. As expected,
level of social support at baseline was higher for those who
were treated with PD versus HD; this is likely due to the strong
influence of perceived social support on patient selection of
dialysis modality (17,18).
Previous studies have shown that social support is an impor-
tant determinant of perceived HRQOL in patients with ESRD
(19,20). Similarly, we found that, even after adjustment, having
A
B
CD
E
Figure 1. Distributions of functional social support scores, overall and by dialysis modality: Overall functional social support (A),
emotional support (B), tangible support (C), affectionate support (D), and positive social interaction (E). All scores are scaled 0 to
100, with higher scores indicating higher level of social support in domain indicated. P for HD versus PD, by Wilcoxon rank-sum
test.
Clin J Am Soc Nephrol 5: 1480–1488, 2010 Social Support and Outcomes 1483

Page 5

the highest versus lowest level of social support was associated
with greater 1-year satisfaction and HRQOL. These associations
generally persisted after adjustment for baseline satisfaction
and HRQOL, showing that the association of social support
with these measures continues over time. In light of these
results and because patients believe lack of psychosocial sup-
port to be a top priority for both clinical care and research in
chronic kidney disease (21,22), interventions to increase sup-
port could potentially provide dramatic benefits for satisfaction
and HRQOL among those with ESRD.
Among the clinical outcomes examined, lower hospitaliza-
tion rates were associated with highest versus lowest levels of
social support (15% reduction associated with greater overall
social support). Many hospitalizations of dialysis patients are
due to infections, fluid overload, and other issues that could be
treated on an outpatient basis if caught early, which may be
more likely for patients whose social networks are likely to
recognize health changes and encourage treatment-seeking be-
havior. Also, patients with greater levels of social support may
be more likely to adhere to treatment, medications, and diet, as
has been shown for postacute coronary syndrome (23), which
could prevent certain complications that lead to hospitalization.
Lower hospitalization rates could result in substantial cost sav-
ings for an already overburdened health care system, given that
inpatient admissions of patients with ESRD from 2003 to 2007
cost the US health care system approximately $33 billion, ap-
proximately 35% of the total Medicare expenditure for ESRD
during the same period (24).
Modality switching was not influenced by social support in
this cohort. Those who were on PD had higher social support
scores than did those who were on HD, likely reflecting the
influence of social support on initial modality selection (17,18);
however, it may be that the reasons for switching over time,
such as recurrent infections, may be more strongly biologically
mediated, particularly for PD patients; in fact, patient choice
was not cited as one of the reasons for switching among PD
patients in this cohort (13).
Mortality was not associated with level of social support in
this cohort. Again, it is likely that biological factors that lead to
the progression of cardiovascular disease (the leading cause of
death in dialysis patients) and to death may outweigh the
effects of social support over time. Previous studies of social
support in Chinese PD patients (25) and of perceived family
support in a small study (n ? 78) of in-center HD patients (26)
showed significant associations with lower mortality; however,
a larger US study showed that giving, not receiving, social
support was associated with lower mortality (27). Similarly, a
Dutch study showed that the discrepancy between expected
and received social support but not level of social interaction
was associated with lower mortality (8). Thus, our results do
Table 2. Baseline dialysis patient characteristics, overall and by tertile of overall functional social support score
Baseline Characteristic
n
All
Tertile of Functional Support Score
LowMiddleHigh
Pa
Range of overall functional support score
Demographic
age (years; mean ? SD)
male gender (%)
white race (%)
graduated high school (%)
employed (%)
married (%)
Clinical
ICED (% score of 3)
BMI (mean ? SD)
HD (%)
late evaluation (% ?4 months from first
nephrologist visit to dialysis)
Laboratory
albumin (g/dl; mean ? SD)
creatinine (mg/dl; mean ? SD)
hemoglobin (g/dl; mean ? SD)
Satisfaction and HRQOL
definite recommendation (%)
excellent overall care (%)
MCS score (mean ? SD)
PCS score (mean ? SD)
–0.0 to 100.0 0.0 to 72.473.6 to 92.192.2 to 100.0–
949
949
949
942
949
948
57.7 ? 14.8
54.1
66.9
70.7
13.7
56.2
58.2 ? 14.5
58.9
62.0
71.9
13.0
42.1
57.4 ? 15.0
48.2
67.4
70.1
11.8
56.9
57.5 ? 14.9
55.0
71.3
70.1
16.3
69.6
0.76
0.03
0.04
0.85
0.24
?0.001
948
884
949
790
29.525.626.536.4 0.01
0.14
0.02
0.10
27.2 ? 6.8
75.2
30.1
27.4 ? 6.8
80.1
31.8
26.6 ? 6.9
75.1
33.5
27.6 ? 6.6
70.6
25.4
927
927
920
3.63 ? 0.38
7.29 ? 2.59
10.8 ? 1.3
3.65 ? 0.37
7.40 ? 2.51
10.8 ? 1.3
3.62 ? 0.35
7.24 ? 2.69
10.7 ? 1.2
3.62 ? 0.40
7.21 ? 2.58
10.8 ? 1.4
0.37
0.63
0.81
884
884
843
843
81.8
65.7
74.1
56.7
85.2
68.4
86.1
72.0
?0.001
?0.001
?0.001
0.21
46.6 ? 11.6
32.7 ? 10.0
44.3 ? 11.7
32.5 ? 9.4
45.9 ? 11.2
32.1 ? 10.0
49.4 ? 11.3
33.5 ? 10.5
ICED, Index of Coexistent Disease.
aBy ANOVA (continuous variables) or ?2test (categorical variables) across tertile of functional social support score.
1484Clinical Journal of the American Society of NephrologyClin J Am Soc Nephrol 5: 1480–1488, 2010