Social Support 63in predicting HIV disease progression. Among participants
with more advanced symptoms at baseline, longevity was
positively associated with network size and informational
support. Among participants with asymptomatic disease sta-
tus at baseline, a large network size predicted more immediate
onset of symptoms. The authors suggest that the negative in-
”uence of network size may be related to the stress of disclo-
sure of HIV status to others or to poor health habits. Miller,
Kemeny, Taylor, Cole, and Visscher (1997) conducted a three-
year longitudinal study measuring the association between so-
cial integration (de“ned as the number of close friends), the
number of family members, and the number of groups or or-
ganizations to which the participant belonged, and HIV pro-
gression (immune parameters, AIDS diagnosis, death from
AIDS). Contrary to other studies, they did not “nd an associ-
ation between social support and HIV progression.
Leserman and colleagues (1999) followed a cohort of
82 HIV-infected men without symptoms of AIDS every
six months for 5.5 years. Satisfaction with social support was
evaluated, as well as the number of support persons. AIDS
progression was de“ned as the point at which the person met
Center for Disease Control (CDC) AIDS surveillance case
de“nition. Some confounding factors (age, education, race,
baseline helper cells, tobacco use, and number of retroviral
medications) were controlled in the analysis. For each point
decrease in cumulative support satisfaction, the risk of AIDS
progression increased by 2.7 times. Number of social support
persons was not related to AIDS progression. Thornton and
colleagues (2000) studied long-term HIV-1 infected gay men.
Perceived support was measured using the Interpersonal Sup-
port Evaluation List (ISEL), and participants were followed
for up to 30 months. Survival analyses indicated that social
support was not related to a transition to AIDS-related com-
plex (ARC) or AIDS.
In summary, studies linking social support to HIV pro-
gression to AIDS have shown mixed results. Social support
may have a protective effect among individuals with more ad-
vanced symptoms, although “ndings have been inconsistent.
Mechanisms for social support, including health behaviors
and medical adherence, also need further study. A potential
mechanism may be adherence to medical appointments. For
example, Catz and colleagues (1999) found greater outpatient
appointment adherence among patients with more perceived
social support.
Coronary Disease
Social isolation, de“ned as having inadequate social support
or social contact, has been implicated in decreased sur-
vival time following a myocardial infarction (MI). Studies
have suggested that a lack of support places patients at
increased risk for cardiac mortality after an MI (Berkman,
Leo-Summers, & Horwitz, 1992; Case, 1992; Ruberman,
Weinblatt, Goldberg, & Chaudhary, 1984; Welin, Lappas, &
Wilhelmsen, 2000). Further evidence sustaining the link be-
tween support and cardiac mortality has been provided by
interventions that provide emotional support and stress re-
duction. These studies have been shown to result in reduced
incidence of MI recurrence over a seven-year follow-up pe-
riod (Frasure-Smith & Prince, 1985).
However, a secondary analysis of data from Frasure-
Smith•s Canadian Signal-Averaged ECG Trial indicated that
neither living alone, having close friends, nor perceived
social support were signi“cantly related to cardiac events,
acute coronary syndrome recurrences, or arrhythmic events
(Frasure-Smith, Lesperace, & Talajec, 1995). The authors
explain the lack of a negative “nding by proposing that their
inclusion of a measure of negative emotions (e.g., depressive
and anxiety symptoms) had a stronger relation with cardiac
events, and may have accounted for much of the association
between social support and cardiac events. A later study by
the same team evaluated a potential moderating effect for
social support on the consistent association between depres-
sive symptoms and cardiac mortality (Frasure-Smith et al.,
2000). In this study, social support was not associated with
cardiac mortality. However, the interaction between depres-
sion and perceived support indicated that among patients
with very low and moderate levels of perceived support, the
impact of depression on a one-year prognosis was signi“cant.
For patients in the highest quartile of perceived social sup-
port, there was no depression-related increase in cardiac mor-
tality. Further analyses evaluated whether the buffering effect
of perceived support was produced by reducing depressive
symptoms over time. Results supported this hypothesis:
Among one-year survivors who had been depressed at base-
line, higher baseline social support predicted improvements
in depressive symptoms over the one-year post-MI follow-up
period. Future studies should more carefully control for po-
tential covariates as well as elucidate potential mechanisms
for support•s impact on prognosis after MI. Orth-Gomer and
Unden (1990) have found a second potential factor that, com-
bined with social isolation, predicts mortality among men
post-MI. In their study, the combined effects of lack of social
ties and the coronary-prone behavior pattern were a better
predictor of mortality than social isolation alone, explaining
almost 70% of the mortality.
One study has linked social support with the incidence of
and death from coronary artery disease among general popu-
lations of individuals who were not previously diagnosed
with coronary artery disease. Orth-Gomer, Rosengren, and
Wilhelmsen (1993) measured emotional support from close
relationships (labeled attachment) and social support by an