222 AIDS/HIV
identi“es •downstream,Ž • midstream,Ž and •upstreamŽ be-
havioral approaches to prevention. Downstreamapproaches
are those interventions that are targeted toward persons who
already exhibit high-risk behavior or who have already
contracted HIV or another sexually transmitted disease. Mid-
streamapproaches refer to interventions targeted toward
de“ned populations for the purpose of changing and/or
preventing risk-behavior; midstream interventions tend to
involve structured organizations (e.g., school, community-
based organizations) as well as entire communities. Up-
stream approaches are larger, macrolevel public policy
interventions designed to in”uence social norms and support
health promoting behaviors. They tend to be more •univer-
sal,Ž targeting entire populations rather than just groups
engaged in high-risk activities. Most downstream and mid-
stream approaches have been face-to-face interventions.
Some midstream and most upstream interventions target
communities or larger social units.
Downstream Approaches
Downstream interventions target populations engaging in
•high riskŽ behavior. Thus, prevention programs delivered in
settings that provide sexual health or drug abuse services
provide interventionists with access to individuals who are
likely to be at the highest risk for acquiring HIV. Such sites
afford •teaching momentsŽ for individuals who could bene“t
greatly from HIV risk reduction programs.
Sexual Health Settings
Settings that provide HIV counseling and testing (C&T),
family planning, or sexually transmitted disease (STD) treat-
ment all serve clients who are likely to have engaged in
behaviors that confer high risk for HIV infection. Such sex-
ual health settings, where it is normative to discuss sexual
behavior and encourage risk reduction, is an ideal place for
sexual behavior change interventions.
HIV C&T is the most heavily funded prevention activity in
the United States, and research to determine whether it reduces
risky sexual behavior has been abundant. Our group com-
pleted a meta-analysis of the studies completed through 1997
(Weinhardt, Carey, Johnson, & Bickham, 1999), and learned
that C&T didnotalter risky sexual behavior among those par-
ticipants who testednegative;however, C&T was associated
with risk reduction among those who tested positive and with
sero-discordant couples (i.e., couples in which one partner is
infected but the other partner is not). Thus, HIV C&T provides
an effective behavior change strategy for HIV-positive indi-
viduals and sero-discordant couples. A criticism of many of
the early HIV C&T studies is the C&T was not guided by a
sophisticated model of behavior change. The implicit model
seemed to be based on the notion that knowing more about
HIV would lead to behavior change, a purely educational
approach. Since the completion of our meta-analysis, how-
ever, HIV C&T has been in”uenced more by psychological
theory. In addition, recent interventions have recognized that
a single counseling session may not be suf“cient to prompt
behavior change among individuals who test negative.
The •Voluntary HIV-1 Counseling and Testing Ef“cacy
StudyŽ (2000) was conducted in Kenya, Tanzania, and
Trinidad. This randomized controlled trial (RCT) enrolled
3,120 individuals and 586 couples and assigned these partici-
pants to either a counseling group or to a health information
(control) group. In contrast to earlier approaches that relied on
education and persuasive presentations, the intervention used
a client-centered approach, including a personalized risk as-
sessment and the development of a personal risk reduction
plan. This approach was designed to be sensitive to each
client•s emotional reactions, interpersonal situation, social
and cultural context, speci“c risk behavior, and readiness-
to-change risk behavior, consistent with a more psychological
(rather than purely educational) approach. Evaluations
occurred 7 and 14 months after the counseling. At the initial
(7-month) follow-up, STDs were diagnosed and treated, and
participants inbothgroups were retested for HIV and received
the client-centered counseling. At the second (14-month)
follow-up, risk behavior was assessed and additional client-
centered counseling and testing were provided. The results
indicated that the proportion of individuals reporting unpro-
tected intercourse with nonprimary partners declined more
for those receiving C&T than for controls. These results
were maintained at the second follow-up. Consistent with
Weinhardt et al.•s meta-analysis, HIV-positive men were
more likely than HIV-negative men to reduce unprotected in-
tercourse with primary and nonprimary partners, whereas
infected women were more likely than uninfected women to
reduce unprotected intercourse but only with primary part-
ners. These results among HIV-positive patients were repli-
cated among those who tested positive at the “rst follow-up
session. Couples assigned C&T reduced unprotected inter-
course with their primary partners more than control couples,
but no differences were found in unprotected intercourse with
other partners. For those who are interested in using the inter-
vention manual or assessment measures from this study, these
materials are available to download from http://www.caps
.ucsf.edu/capsweb/projects/c&tindex.html.
In the United States, Project Respect compared the
ef“cacy of two C&T interventions guided more explicitly
by social-cognitive theory, and using the CDC•s revised