224 AIDS/HIV
signi“cantly over time, and ar gued that these results mark a
•declining phaseŽ in the HIV epidemic among IDUs, at least
in New York. Replication of these “ndings in other major
urban settings is needed; nevertheless, the “ndings are quite
encouraging regarding the modi“cation of risk behavior
among what many people would assume is the most dif“cult
and hard-to-reach population of IDUs.
Midstream Approaches
Interventions targeted toward de“ned populations for the pur-
pose of reducing risk behavior, typically in natural environ-
ments or preexisting organizations, constitute midstream
approaches. Midstream approaches target people who may be
engaging in the risk behavior as well as people who have not
yet begun the risk behavior in an effort to prevent the behavior.
Midstream approaches are often implemented in schools, pri-
mary care health clinics, and community-based settings.
School-Based Programs
Reviews of the school-based education programs indicate
that one-half of these programs effectively reduce unpro-
tected intercourse (Kirby & DiClemente, 1994). Effective
programs tend to be guided by social learning theories, focus
on speci“c strategies to reduce risk behavior, use active
learning methods of instruction, address social and media in-
”uences and pressures to have sex, reinforce values against
unprotected sex, and provide modeling and practice of com-
munication or negotiation skills.
Before the AIDS pandemic, preventionists often found it
challenging to secure the support of parent organizations and
school boards for programs involving discussion of sexual or
drug use behavior. Recently, however, a number of school
districts have approved HIV prevention programs, especially
in urban settings where HIV is more widely seen as a health
threat. For example, in an inner-city school system in New
York City, Walter and Vaughan (1993) evaluated a teacher-
delivered curriculum with high school students. Their pro-
gram, based on social cognitive theory, resulted in modest
improvements in HIV-related knowledge and self-ef“cacy,
and reduced sexual risk behavior. The effectiveness of this
curriculum, which was accepted following meetings with all
key constituent groups, demonstrates the feasibility of such
an approach in school-based settings.
Increasingly, acceptance of school-based programs is oc-
curring in nonepicenters as well. For example, in California
and Texas, K. Coyle et al. (1999) are currently evaluating the
effectiveness of •Safer Choices,Ž a multicomponent HIV,
STD, and pregnancy prevention program for 9th and 10th
grade students. The intervention draws upon social-cognitive
theory (Bandura, 1986), social in”uence theory (Fisher,
1988), and models of school change, and includes a skills-
based classroom curriculum for students, formation of a team
of peer educators to conduct out-of-classroom activities,
parent education activities and newsletters, and improvement
of the schools• linkages with community resources. This
innovative study is a RCT involving 3,869 students from
20 schools. In an initial report, using data from students
who completed both the baseline and the initial follow-up (at
7 months), the results have been promising: The intervention
enhanced 9 of 13 psychosocial variables including knowl-
edge, self-ef“cacy for condom use, normative beliefs and
attitudes regarding condom use, perceived barriers to con-
dom use, risk perceptions, and parent-child communication.
The program also reduced the frequency of intercourse with-
out a condom in the three months prior to the survey and
increased use of condoms at last intercourse and the use of
contraceptives at last intercourse. (The latter outcome has
less value for HIV prevention but does help to reduce unin-
tended pregnancies.)
Although increased acceptance of sexual health programs
in U.S. schools promises to reduce HIV transmission domes-
tically, it can be argued that the greatest need for education
and risk reduction programs is in countries where HIV and
AIDS are most prevalent. Klepp, Ndeki, Leshabari, Hannan,
and Lyimo (1997) described a study that tested the effects
of the •NgaoŽ (Swahili for •shieldŽ) education program in
Tanzania. The program, in”uenced by the theory of reasoned
action (Fishbein & Middlestadt, 1989) and social learning
theory (Bandura, 1986), was designed to reduce children•s
risk of infection and to improve their tolerance of and care
for people with AIDS. This RCT (baseline, intervention,
12-month follow-up) was conducted through public primary
schools in two regions of Tanzania. The schools were strati-
“ed according to location and randomly assigned to interven-
tion or comparison conditions. The results revealed that the
intervention group exhibited improved knowledge and more
tolerant attitudes toward people with AIDS. Fewer students
from the intervention schools (7%) than from the comparison
schools (17%) had had their sexual debut during the follow-
up interval.
Schools provide an ideal venue for the promotion of
healthy behavior because they allow access to young people
before they establish risky habits. Moreover, in countries
where education is mandatory, schools can reach the vast
majority of older children and adolescents without additional
outreach expense. Continuing controversy exists regarding