See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/26710273

Peer-led interventions to reduce HIV risk of youth: A review

Article in Evaluation and Program Planning · August 2009

DOI: 10.1016/j.evalprogplan.2009.07.001 · Source: PubMed

CITATIONS READS

172 580

2 authors:

Eleanor Maticka-Tyndale Jessica Barnett

University of Windsor Wright State University
136 PUBLICATIONS 6,715 CITATIONS 8 PUBLICATIONS 369 CITATIONS

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Eleanor Maticka-Tyndale on 22 April 2020.

The user has requested enhancement of the downloaded file.

This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/copyright
 Author's personal copy

Evaluation and Program Planning 33 (2010) 98–112

Contents lists available at ScienceDirect

Evaluation and Program Planning

journal homepage: www.elsevier.com/locate/evalprogplan

Peer-led interventions to reduce HIV risk of youth: A review

Eleanor Maticka-Tyndale *, Jessica Penwell Barnett
Dept. of Sociology and Anthropology, University of Windsor, 401 Sunset Ave, Windsor, Ontario, Canada N9B 3P4

A R T I C L E I N F O A B S T R A C T

Article history: One approach in HIV prevention programming targeting youth is to use peer leaders in what is referred
Received 6 December 2008 to as peer education programming. This paper critically reviews and synthesizes the results and lessons
Received in revised form 13 April 2009 learned from 24 evaluated peer-led programs with an HIV/AIDS risk reduction component that target
Accepted 4 July 2009
youth in the communities where they live and are delivered in low- and middle-income countries.
Interventions were identified through a comprehensive search of the peer reviewed AIDS-related
Keywords: literature as well as publication lists of major organizations in the UN family that address HIV and AIDS.
Peer education
Our synthesis of study results finds that these programs have demonstrated success in effecting positive
HIV prevention
Evaluation review
change in knowledge and condom use and have demonstrated some success in changing community
attitudes and norms. Effects on other sexual behaviors and STI rates were equivocal. We include an
overview of characteristics of successful programs, a review of program limitations, and recommenda-
tions for the development and implementation of successful community-based peer-led programs in
low-income countries.
ß 2009 Elsevier Ltd. All rights reserved.

One approach in HIV prevention programming targeting youth
worldwide is to use peer, rather than professional, leaders in what
have been referred to as peer education programs. Peer-led
programs have been delivered in schools, clinics, community
centers, workplaces, and in informal settings where members of
target populations congregate. They build on the natural exchange
of information between people of similar age or status (Turner &
Shepherd, 1999). Peer education and peer-led interventions
typically target peer groups and communities rather than
individuals as the unit of change, with agents of change coming
from within the group or community (i.e. peers) rather than
brought in from outside. The approach is based on the assumption
that, especially among adolescents, peers learn from each other,
are important influences on each other, and that norms and
behaviors are most likely to change when liked and trusted group
members take the lead in change (Aggleton & Campbell, 2000;
Campbell, 2004; Fee & Youssef, 1996; Shiner, 1999; Turner &
Shepherd, 1999).
According to Gerber and Kauffman (1981), peer education has
its roots in the ‘‘monitorial system’’ set up by Joseph Lancaster in
London, England, in the early 1800s that was designed to reduce
teacher workloads. Teachers taught lessons to a select group of
student ‘‘monitors’’ who then passed these on to their classmates.
Helm, Knipmeyer, and Martin (1972) identify the influenza
outbreak at the University of Nebraska in the United States as
* Corresponding author. Tel.: +1 519 253 3000x2200; fax: +1 519 971 3621.
E-mail address: Maticka@uwindsor.ca (E. Maticka-Tyndale).
one of the earliest uses of peer educators in the health sphere with
trained students providing prevention and care information to
fellow students. By the 1990s, peer education was one of the most
widely used approaches in HIV prevention initiatives targeting
youth (Bernert & Mouzon, 2001; Horizons, 2000). Today, peer
education is included as a component in a number of large-scale
initiatives designed to reduce the spread of HIV among youth,
including the 100 million pound initiative funded by the
Department for International Development, UK, in Nigeria
(www.dfid.gov.uk) and South Africa’s National HIV Prevention
Program for Youth, LoveLife (www.kff.org/about/lovelife.cfm).
Following a model that networked peer educators across 14
countries in the European Union called EUROPEER, both national
and international organizations have been formed in other world
regions. NOPE (National Organization of Peer Educators; website
www.nope.or.ke), for example, mobilizes peer-led community
interventions and trains and networks peer educators across
Kenya. YPEER (Youth Peer Education Network; website
www.youthpeer.net), with chapters in 27 countries in Eastern
Europe, Central Asia, the Arab States and Africa, networks and
trains peer educators and expands peer-led programming within
and across regions.
With the spread of peer education, including the development
of national and international organizations to promote peer
education, it becomes increasingly important to synthesize the
evidence from existing programs to better guide decision-making
and program planning. Kim and Free (2008) recently published a
systematic review and meta-analysis of 13 peer-led adolescent
sexual health education interventions. Using several indicators of

0149-7189/$ – see front matter ß 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.evalprogplan.2009.07.001
 Author's personal copy
E. Maticka-Tyndale, J.P. Barnett / Evaluation and Program Planning 33 (2010) 98–112
condom use and sexual behavior, they found no evidence of a
positive program effect. They surmised that this poor showing for
peer-led interventions was potentially due to the paucity of
evaluation studies using rigorous designs (which excluded the vast
majority from consideration in their synthesis) and the haste with
which many programs are implemented, especially in regions
facing the crisis of high and rising rates of HIV spread. The
methodological standards required for their meta-analysis espe-
cially excluded interventions and evaluation studies from low-
income countries. Sub-Saharan Africa (SSA), for example, is
estimated to account for 59% of peer-led programs (Kelly et al.,
2006), but provided only 4 of the 13 studies to Kim and Free’s
review.
Kelly et al. (2006), in their recent review of programming
delivered by non-government organizations (NGOs) in low-
income countries stress the very limited resources with which
these organizations work. This often places NGOs in the position of
balancing between delivering more sophisticated (and costly)
programs and expanding the reach of more basic programs as well
as between investing in expanded program delivery and more
sophisticated evaluation. The shortage of rigorous evaluation
studies is exacerbated by the reluctance of donor agencies to invest
in them (Pettifor, MacPhail, Bertozzi, & Rees, 2007). Besides the
resource constraints, some have questioned the relevance of the
criteria applied in meta-analyses and syntheses such as Kim and
Free’s to interventions targeting groups and communities—the
primary target groups for peer-led interventions (Auerbach, 2008;
Rapkin & Trickett, 2005). These interventions require flexibility of
delivery and consideration of a wide array of interconnected
factors whose influence on outcomes is expected to be highly
variable depending on the positioning of individuals (Auerbach,
2008; Pettifor et al., 2007; Rapkin & Trickett, 2005), a situation that
does not readily fit the requirements of the most rigorous
evaluation designs (DiClemente, Crosby, & Wingood, 2005; Rapkin
& Trickett, 2005).
The dearth of syntheses of peer-led interventions limits our
knowledge about ‘‘what works’’ in low-income countries. This
paper sets out to help fill that gap in knowledge. We critically
review and synthesize the results and lessons learned from
evaluated peer-led programs with an HIV/AIDS risk reduction
component that target youth in geographical communities (i.e.
communities where they live) and are delivered in low-income
countries. We have focused on low-income countries because this
is where HIV prevalence is highest and where there are the greatest
resource challenges for program delivery. The focus of our
attention is on programs in geographical communities because
this is where vulnerabilities are grounded and where most risky
behaviors occur (Campbell, 2004) and because programs in
communities have the potential of reaching the largest number
of youth (Maticka-Tyndale & Brouillard-Coyle, 2006). We follow
the approach used in a paper that synthesized school-based
programs in sub-Saharan Africa (Gallant & Maticka-Tyndale, 2004)
together with the framework for assessing evidence of program
impact provided by Habicht, Victora, and Vaughan (1999) and used
in the recent WHO review and synthesis of HIV prevention
programming targeting youth in low-income countries (Ross, Dick,
& Ferguson, 2006). This allows us to broaden the methodological
inclusion criteria to capture a larger number of programs with
more diverse evaluation designs while still maintaining a
structured and critical approach to assess the quality of evaluation
results.
1. Method
Evaluated prevention programs were located by searching
literature databases such as Psychological Literature (PsychLit),
99
Population Information Program (POPLINE), Sociological Abstracts,
and MEDLINE; the tables of contents of journals that published
articles evaluating interventions with an HIV/AIDS content
between 1994 and 2008; and publication lists from international
organizations such as the United Nations Joint Commission on HIV/
AIDS (UNAIDS) and the World Health Organization Global Program
on AIDS (WHO/GPA).
To be included in this review, an intervention had to meet the
following criteria: (a) youth (as culturally defined) were included
in the target population (most targeted those 15–24); (b) it was
delivered in a geographical community; (c) at least some content
dealt with knowledge, attitudes, norms, and/or behaviors relevant
to the prevention of HIV/AIDS; (d) it was designed to be delivered
primarily by youth peers; (e) it was delivered in a low- or middle-
income country; (f) it was evaluated (both quantitative and
qualitative evaluations were included) and information about the
evaluation methods and results was provided; (g) the content and
delivery methods were described; (h) the report or paper
describing the intervention and its evaluation was available in
English or French and published between January 1994 and
November 2008. This time period encompassed the dates used in
the WHO publication (Ross et al., 2006) where similar syntheses
were reported and the last date on which we accessed the
literature in preparing this paper. Interventions were excluded if
they did not meet these criteria or if they were delivered primarily
in a school, workplace, or health facility.
Several steps were taken to synthesize materials. First,
descriptions of each program were examined and charted based
on their theoretical framework, targeted outcomes, intervention
content, implementation strategies, duration, local community
input and/or cultural modifications, program monitoring, and
discussion of any issues salient to program delivery or evaluation.
A condensed version of these charts is included here as Table 1.
Second, descriptions of evaluation procedures were reviewed
and charted based on research design, sampling frame and size,
data collection, threats to internal validity, validation of measures
and appropriateness of analytic techniques. Condensed versions of
these charts are included here as Tables 2 and 3. Using Habicht
et al.’s (1999) continuum of evidence for an intervention’s effect,
interventions were categorized as providing evidence of an effect
that ranged from adequate to plausible to probable. Demonstration
that an effect was related to the intervention was considered
adequate if the design of the evaluation study could show that the
expected changes occurred. This was often in the form of
performance or process indicators (e.g. programs were held, target
populations were reached, condoms were distributed) and a
change in the desired direction for health, knowledge, attitude or
behavioral indicators or community activities. There was plausible
evidence that a demonstrated effect was related to the interven-
tion if, in addition to demonstrating change, study design and data
analysis allowed for alternative explanations of change to be ruled
out through use of a control group and/or statistical controls for
potentially confounding variables. Plausibility evidence ranged
from weak to strong depending on the design, analysis techniques
and handling of threats to internal and external validity. Finally, to
draw conclusions about an intervention’s effects on the grounds of
probability required the exclusion of explanations for the effect
other than the intervention, typically through a randomized
control design. The reported outcomes were examined from within
this framework.
Comparison of the effectiveness of interventions was drawn
within outcome categories (e.g. knowledge changes across
interventions) as well as within the intervention and across
outcome categories. Conditional results (e.g. only girls, only rural
youth) and results of tests for confounding factors were also noted.
The outcomes from evaluation studies with the strongest
 Author's personal copy

100
Table 1
Characteristics of programs.

Evaluated by Country Theory Intervention site Youth or Content Form Implementation

Comm
involved Targeted Main activities Delivery method Total participants Intervention Peer educators Training Monitored
outcomes duration

Mixed-method evaluation designs

Bhuiya et al. Bangladesh Health and youth Yes KAB, use of Information, referrals Print, curricula, Facilitators: 3437, 20 months; Facilitators aged 4–5 day Yes
(2004) facilities RH services to services groups, hotline, PE: 1256 20 sessions 21–28 years plus
anonymous refresher
question box
Brieger et al. Ghana and Youth Service Yes KB Information, referrals 1-on-1, group, 18 months Cascade Yes
(2001) Nigeria Organizations to services, condom print, drama model
provision
Elkins et al. Thailand Self- 77 rural villages A Condom use skills Condom race 5775–11,550 1 day per Health Part of

E. Maticka-Tyndale, J.P. Barnett / Evaluation and Program Planning 33 (2010) 98–112

(1998) efficacy people; 40–49% event volunteers 1 day
of each village
Esu-Williams Zambia Anti-AIDS clubs Yes B PLHA care and Assist PLHAs, 300 PE trained 18 months Members of Yes
et al. (2004) advocacy, 1-on-1, group, anti-AIDS clubs
information, condom drama, print,
provision, referrals to lectures
services
Frontiers Zambia PLA Communities Yes B Information, condom 1-on-1 4000 PE counseling; 20 months
(2001) provision, economic 4–15% of adolescents
development
Mathur et al. Nepal PLA Yes Compare PLA Information, life skills, Seminars, I: 61% rural, 34% 12–24 Two coed groups, Yes
(2004) PE model to challenge social norms community urban C: 38% rural, months younger/unmarried
standard PE fairs, print, 38% urban and older/married
model group, drama populations adolescents
Merati et al. Bali Social Traditional Yes Reduce HIV Information, values Group Single 7 h Nominated by adult 3 days
(1997) influence comm. youth and STDs clarification, activities session leaders and peers
model groups decision-making
and communication
skills
Muyinda Uganda Yes K Information, condom 1-on-1, group 86% of target 12 months Volunteers 1 week + Yes
et al. (2003) provision group, 71 selected 6 months
adolescent by study team refresher
girls; 176
other villagers
Nastasi Sri Lanka Social- Yes K, improve Information, social Group, written 89% attended one+ 1 month; Recruited by 15 h + Yes
et al. (1998) construction sexual norms challenged exercises session, 66% 12 sessions public health 15 h
negotiation attended six+ worker and refresher
sessions university
researchers
Sergeyev Russia Peer On the street, Prevent HIV Information, risk 1-on-1, print, 2 years Volunteer IDUs
et al. (1999) influence storefronts among IDUs assessment, needle risk assessment
exchange, condom interview
provision

Quantitative evaluation designs only

Askew et al. Kenya Clinics, churches, Yes KB Information, life skills, Drama, video, 20% of adolescents 12 months Community PE Cascade
(2004) other spaces referrals to services, songs, debates, recruited by govt. model
not specified condom provision festivals, workers and comm.
curricula, leaders
1-on-1, group,
print
 Author's personal copy

Diop et al. Senegal Determined by Yes K, use of RH Information Curricula, games, 18,358 completed 15 months Youth CBO and Cascade Yes
(2004) youth in each services music, role-play, entire curriculum, clinic staff model
group; e.g. youth information 2317 interacted
centers, homes booths at with clinic PE
festivals, radio
Entre Nous Cameroon Schools, youth Yes KB Information, life skills 1-on-1, group, 5000 19 months Criteria: permanent
Jeunes: associations, talks/conferences, resident, motivation/
IRESCO community soccer matches, attitude, ability
(2002)a events print
Speizer et al. Cameroon KAB; peers Existing clubs KB Information, referrals 1-on-1, group 39% of youth in 18 months Recruited from clubs 1 week
(2001)a influence and to services community and organizations; + quarterly
peers organizations (12,000 in assessed motivation refresher
groups, 5000
1-on-1)
Moyo et al. Zimbabwe Clinics, youth KB, use of Information, condom 1-on-1, group, 9 months Criteria: 16–24 years, Yes
(2000) center, schools, RH services provision, referrals to new RH clinic, fluent English and

E. Maticka-Tyndale, J.P. Barnett / Evaluation and Program Planning 33 (2010) 98–112

churches, on services recreation, print 1 local language,
the street, etc. personality,
motivation and
availability
Özcebe and Turkey Homes, public K Information 1-on-1, group, 60% of adolescents 2 years Investigators chose. 40 h Yes
Akin (2002) spaces print Criteria: 15–24 years, + quarterly
volunteer, permission refreshers
from family;
knowledge

Qualitative evaluation designs or monitoring only

Bagamoyo Tanzania Popular Communities in Yes KB Information, social Popular theatre 30,000 + Youth choose young Cascade
College of Arts theatre 4 Districts norms challenged artists from the model
et al. (2002) community
Brady and Kenya Sports facilities Improve quality Soccer league, Group Youth members of
Khan (2002) in informal of life for youth information, life sports association
settlements skills
Cáceres et al. Peru Yes A, skills, use Information, referrals Print, group, 6 months
(1999) of RH services to services video, rock
concerts, health
fairs
Hughes-d’Aeth Zambia 4 NGO facilities Yes KB, build Life skills, condom Drama, music, Recruited by NGOs Trained
(2002) evaluated community provision, information, games, group, by NGOs
capacity to cope services for PLHAs, print, radio,
with HIV/AIDS referrals to services
Mitchell et al. Uganda On the street Yes KB Information, life skills, 1-on-1, group, 91% of NGO children 2001 to Voted in by peers Semiannual
(2007, 2002) condom provision, drama, puppetry, and 43% of street present refresher
referrals to services sports, festivals/ children training
events
Sharma (2002) Nepal Youth action Yes KA, Information, condom Drama, 1-on-1s, 20 youth groups 6 months Group leader college
groups and junior communication, provision events, seminar, aged, PE secondary
red cross circles use of RH print materials school aged
services
Shuguang and China On the street, Yes B Information, social Safe sex role- 705 trishaw drivers 1 month Criteria: popularity, 4h
Van de Ven drop-in center norms challenged, model stories, >6 years educ., exp.
(2003) develop peer space group w/social work,
and drop-in center meetings, communication skills,
story stickers connectedness,
motivation
UNFPA (2004) Burkina Faso Homes and public Yes KA, life skills Information, non-Rx 1-on-1, group, Nominations; Criteria: 2 sessions Yes
spaces contraceptives sold radio, film aged 18–24 years, had
1–2 children, able to
travel and speak in
public
101
 Author's personal copy

102
Table 1 (Continued )
Evaluated by Country Theory Intervention site Youth or Content Form Implementation
Comm
involved Targeted Main activities Delivery method Total participants Intervention Peer educators Training Monitored
outcomes duration

UNICEF (2002) Ghana Compared Peer KB Information, condom Peer Education 66,208 in Peer 2 years; CENCOSAD: recruited 5 days + Yes
Education Project provision, life skills, Project: group Education Project; 4–2 h from existing 3 day
and CENCOSAD referrals to services activities, talks, 12,000 in sessions community groups refresher
games; CENCOSAD
CENCOSAD also

E. Maticka-Tyndale, J.P. Barnett / Evaluation and Program Planning 33 (2010) 98–112

had 1-on-1,
sports, drama,
debates
Notes: Abbreviations are as follows: K: knowledge; A: attitudes (including norms and self-efficacy); B: behaviors; RH: reproductive health; PLA: participatory learning and action; PE: peer educator; PLHA: person living with HIV/
AIDS; CBO: community-based organization.
a
Entre Nous Jeunes was evaluated at two different sites in the Cameroon, each reported separately. The delivery of this intervention was modified as appropriate to each site and the evaluations assessed some different outcome
goals.

Table 2
Summary of quantitative evaluation methods.

Evaluated by Design Sampling N Pre/post Experimental/control Time to outcome Measures Base Statistical analysis Strength of
contamination measure validated effects evaluation
controlled designa

Askew et al. Pre–post w/2 exp. Population-based survey, I: 322/357 boys; 3% of adolescents in Approximately Yes Yes Chi-square with Moderate
(2004) conditions plus control; multi-stage cluster design 907/881 girls C: control report program 18 months gender and age plausibility
two sites per condition 321/342 boys, 837/ participation controlled
(6 sites total) 910 girls
Bhuiya et al. Pre–post w/2 exp. Population-based survey I: 973/1016 C: Distance between sites 24 months pre–post; Yes Yes Multivariate Strong
(2004) b conditions plus control 1006/1042 100 km+; other NGO 2 months models, data plausibility
RH activities at control post-intervention weighted
site
Brieger et al. Pre–post w/control Modified cluster for I: 911/908 C: Plausible 18 months or more No No Chi-square, analysis Strong
(2001) b each site/neighborhood 873/893 (site variable) of variance, regression plausibility for
knowledge;
moderate for
other
Diop et al. (2004) Pre–post w/2 exp. Population-based survey, I: 1,019/1,032 C: No; other RH activities 27 months Yes Yes Multivariate models, Moderate
conditions plus control two-stage cluster design 961/830 at control site data weighted plausibility
Elkins et al. Pre–post, no control All members of 1 team/ 164/164 n/a <1 day Yes Yes Non-parametric sign Weak
(1998) b village; 29 of 77 villages rank test plausibility
Esu-Williams et al. Pre–post w/control Random sampling from 16 I: 365/496 C: Unlikely 18 months Yes Yes Chi-square, t-test, Strong
(2004) b of 30 clubs in each cond. 431/487 logistic regression plausibility
Frontiers (2001) b Pre–post w/3 exp. Population, participants, PE Population: 1,634/ Unknown Unknown Weak
conditions plus control 1,720 Participants: plausibility
103 PE: 83
 Author's personal copy

Entre Nous Jeunes: Pre–post w/control Two-stage probability I: 1,248/1,238 C: No; other RH activities 19 months pre–post; Yes Weak
IRESCO (2002) design for each 1,256/1,226 at control site 1 month plausibility
neighborhood post-intervention
Speizer et al. Pre–post w/control Clustered randomized by I: 402/405 C: No 17 months pre–post; Yes Chi-square, Strong
(2001) household 400/413 pre 6 months after 3 multivariate plausibility
program initiation, models/DV
post 3 months
post-intervention
Mathur et al. Pre–post w/control 724 adolescents, 965 Plausible 51 months pre–post; Yes Weak
(2004) b households/979 post 1–8 months plausibility
adolescents, 1003 after intervention
households
Merati et al. Pre–post, no control PE in groups selected for 97; 278 lost to n/a Approximately Yes Yes Paired-sample t-tests, Weak
(1997) b high motivation and attrition 6 months Wilcoxin signed-rank plausibility
participation tests (for PE only)
Moyo et al. Pre–post, no control Cluster sampling; post 250/606 n/a 12–21 months Yes No Bi- and multivariate Weak

E. Maticka-Tyndale, J.P. Barnett / Evaluation and Program Planning 33 (2010) 98–112

(2000) sample demographically analysis plausibility
different than pre
Muyinda et al. Pre–post w/control I: girls who lived in village I: 71/71 C: 24/24 Unknown 12 months No Adequacy
(2003) b whole 12 months and
participated C: girls who
lived in village whole
12 months
Nastasi et al. Pre–post, no control Participants who attended 66; 23 lost to n/a 1 month Yes Yes Paired-sample t-tests, Weak
(1998) 50%+ of sessions attrition Wilcoxin signed-rank plausibility
tests
Özcebe and Akin Pre–post w/control; All 15–24 year olds living I: 222/199 C: Unlikely Approximately Yes 1-way analysis of Moderate
(2002) two sites per condition in villages who complete 219/166 2 years variance plausibility
survey
Sergeyev et al. Pre–post (3 posts), Program participants 484/159/83/33 n/a 3 months b/w Yes Paired-sample t-tests Adequacy
(1999) b no control and PE assessments
Notes: Abbreviations: I intervention; C control; PE peer educator.
In evaluation studies that used comparison groups from locations other than geographical communities, only results for youth in geographical community groups are included.
a
Based on Habicht et al.’s (1999) categorizations.
b
For qualitative methods used in these evaluation studies, see Table 3.

103
 Author's personal copy

104 E. Maticka-Tyndale, J.P. Barnett / Evaluation and Program Planning 33 (2010) 98–112

Table 3
Summary of qualitative evaluation methods.

Evaluated by Design Sample N Pre/Post Data collection Strength of

evaluation
designa

Bagamoyo College of Pre–post Community members PE interviewed community members; Adequacy

Arts et al. (2002) documented post-performance discussions
Bhuiya et al. (2004)b Snapshotc Youth; parents; community 31 youth Interviews w/youth; FGD w/parents and Adequacy
leaders/stakeholders community leaders/stakeholders
Brady and Khan (2002) Case study Staff; participants; parents Interviews; FGD Adequacy
Brieger et al. (2001) b Pre–post Youth; comm. members Interviews; FGD w/youth Adequacy
Cáceres et al. (1999) Pre–post Youth; community stakeholders 800 youth Pre: surveys and FGD; post: interviews Adequacy
and FGD/workshops; rapid assessment
of clinics
b
Elkins et al. (1998) Follow-up Convenience sample of N = 160 Interviews Adequacy
community members
aged 15-45
b
Esu-Williams et al. (2004) Process PE; staff; PLHA FGD; interviews; participatory workshops Adequacy
Frontiers (2001) b Pre–post PE N = 118 Interviews and survey Adequacy
Hughes-d’Aeth (2002) Series of Comm. stakeholders Interviews w/stakeholders, review of prog. Adequacy
mini-case documents; observation of PE activities
studies
b
Mathur et al. (2004) Process Youth; comm. stakeholders/ 14 youth and Interviews, FGD, and participatory Adequacy
members 3 comm. activities w/youth; interviews and group
stakeholders assessments w/comm. stakeholders/
via interviews members
b
Merati et al. (1997) Follow-up PE Approximately FGD Adequacy
60–90
Mitchell et al. (2007, 2002) Process PE; youth 47 youth via FGD; PE admin. questionnaire; review of Adequacy
Participatory survey prog. documents
Evaluation
b
Muyinda et al. (2003) Process Youth; comm. stakeholders/ 60 adult Interviews w/youth and adult participants; Adequacy
members; adult participants; PE participants workshops w/youth; FGD w/comm.
stakeholders/members and PE
Nastasi et al. (1998) Process PE; youth Group discussion and questionnaires Adequacy
w/youth; PE activity log and self-evaluation
b
Sergeyev et al. (1999) Snapshot PE; youth 484 Interviews; health knowl. test Adequacy
Sharma (2002) Snapshot PE; comm. stakeholders Interviews, observation AND FGD w/PE; Adequacy
discussions w/stakeholders; 3 indiv. case
studies (2 youth, 1 comm. member)
Shuguang and Van de Snapshot PE; youth 150 PE; Questionnaire Adequacy
Ven (2003) 705 youth
UNFPA (2004) Process PE; married youth; community Talks, FGD, review of programme documents; Adequacy
leaders, stakeholders; 6 case studies
PE husbands; medical staff
UNICEF (2002) Snapshot PE; staff; youth Observation; interviews; FGD; review of Adequacy
programme documents; feedback
workshops; pharmacy visits
a
Based on Habicht et al.’s (1999) categorizations.
b
For quantitative evaluation methods used in these studies see Table 2.
c
Snapshot is one-time only data collection.

evaluation designs were given greater attention than those with
weaker designs in the presentation and discussion of results. These
results are summarized in Tables 4 and 5.
This paper first describes the interventions and their imple-
mentation followed by a description and critique of the evaluation
study designs and then a review of the evaluation results. Finally,
the implications for peer education interventions delivered in
communities in low-income countries are discussed with a
suggested set of guidelines for the design of effective interventions
based on the programs reviewed here.
2. Results
2.1. Interventions
Twenty-five evaluation studies applied to 24 interventions (the
intervention Entre Nous Jeunes was delivered and evaluated
separately in two different sites) were located. The unit of analysis
in this paper is the 24 interventions. Of these, 14 were delivered in
countries of sub-Saharan Africa, 7 in countries in Asia and the
Pacific, and 1 each in Turkey, Russia and South America. Ten used
both quantitative and qualitative evaluation methods, 6 only
quantitative and 9 only qualitative. Table 1 summarizes the
interventions.
The form of the interventions varied widely. An intervention in
Uganda replicated the senga tradition (Muyinda, Nakuya, Pool, &
Whitworth, 2003) in which an older youth or adult mentored and
educated a younger adolescent on matters of gender and sexual
and reproductive health. The Mathare Sports Association operating
in one of Nairobi’s largest informal settlements used soccer as the
main activity and involved youth in peer-led workshops, discus-
sions, and activities that challenged existing gender and sexual
norms and imparted information about HIV and AIDS (Brady &
Khan, 2002). In an intervention in Thailand, youth took part in
condom relay races at community festivals that included peer-led
instruction on condom use, practicing simulated use as part of a
relay race, and open conversation about condoms and their
importance in protecting against HIV (Elkins, Dole, Maticka-
Tyndale, & Stam, 1998). The intervention evaluated by Esu-
Williams, Schenk, Motsepe, Geibel, and Zulu (2004) in Zambia
added to the regular anti-AIDS clubs’ activities on HIV prevention,
the provision of home-based care and support to PLHAs by club
 Author's personal copy

E. Maticka-Tyndale, J.P. Barnett / Evaluation and Program Planning 33 (2010) 98–112 105

Table 4
Outcome results from peer-led interventions using quantitative research designs.

Reference Country Knowledge Abstinence/ Recent sex # partners Condom use/ STD
never had sex self-efficacy symptoms

+ ns + ns + ns + ns + ns + ns

Quasi-experimental
Askew et al. (2004) Kenya    
Bhuiya et al. (2004) Bangladesh     
Brieger et al. (2001) Ghana and Nigeria  a
Diop et al. (2004) Senegal        
Esu-Williams et al. (2004) Zambia     
Frontiers (2001) Zambia     
Entre Nouse Jeunes: Cameroon      
IRESCO (2002)
Speizer et al. (2001) Cameroon   
Mathur et al. (2004) Nepal     
Muyinda et al. (2003) Uganda   
Özcebe and Akin (2002) Turkey 
Sergeyev et al. (1999) Russia 
Total (N = 12) 8 2 2 3 3 3 3 3 0 2 3 0 7 4 0 4 1 0

Pre–post survey, no control

Elkins et al. (1998) Thailand a
Merati et al. (1997) Bali  a
Moyo et al. (2000) Zimbabwe    
Nastasi et al. (1998) Sri Lanka  
Total (N = 4) 2 2 0 0 1 0 0 0 0 0 0 0 3 0 0 0 1 0

Totals for evaluation strength

Strong plausibility 2 1 0 1 1 1 0 2 0 0 1 0 2 1 0 0 0 0
Moderate plausibility 2 0 1 0 2 1 1 1 0 0 2 0 3 2 0 0 1 0
Weak plausibility 5 3 1 2 1 1 2 0 0 2 0 0 4 1 0 3 1 0
Adequate 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 1 0 0

Grand total (N = 16) 10 4 2 3 4 3 3 3 0 2 3 0 10 4 0 4 2 0

Notes: + change in desired direction for at least some participants; ns no significant change for any groups of participants; change in undesired direction for at least some
participants;
In evaluation studies that used comparison groups from locations other than geographical communities, only results for youth in geographical community groups are
included.
a
Condom self-efficacy.

members. Sergeyev et al. (1999) evaluated an intervention that
targeted intravenous drug users in Russia with young IDU-led
storefront and on-the-street risk assessments, needle exchanges,
and provision of condoms. Many programs used youth centers and
pre-existing youth groups to deliver HIV prevention programs or
community-wide festivals and events to convey information.
2.1.1. Theoretical framework
Most interventions were based on field experience from prior
peer and community work, with a specific theoretical framework
articulated in only 8 of the 24 interventions. The targeted outcomes
of knowledge, attitudes, and beliefs, together with the use of
peer educators (PE), participatory learning modalities, and the

Table 5
Outcomes for peer-led interventions using qualitative and/or process evaluation.

Reference Country Change norms and Distribute resources Connect Y to services Provide Y with
mobilize community information

+ Unknown + Unknown + Unknown + Unknown

Bagamoyo College of Tanzania  

Arts et al. (2002)
Brady and Khan (2002) Kenya  
Brieger et al. (2001) Ghana and Nigeria 
Cáceres et al. (1999) Peru   
Elkins et al. (1998) Thailand  
Esu-Williams et al. (2004) Zambia 
Frontiers (2001) Zambia  
Hughes-d’Aeth (2002) Zambia    
Mathur et al. (2004) Nepal 
Mitchell et al. (2007, 2002) Uganda   
Sergeyev et al. (1999) Russia   
Sharma (2002) Nepal   
Shuguang and China   
Van de Ven (2003)
UNICEF (2002) Ghana  
UNFPA (2004) Burkina Faso    

Total (N = 15) 8 3 0 8 0 0 3 1 0 11 1 0
 Author's personal copy
106 E. Maticka-Tyndale, J.P. Barnett / Evaluation and Program Planning 33 (2010) 98–112
involvement of community stakeholders (both youth and adults)
suggest that if theoretical frameworks were identified for the
remainder of the interventions, they would include cognitive
behavioralism, participatory learning, and social influence theories.
2.1.2. Targeted outcomes
In many of the interventions, HIV risk reduction was addressed
within the context of a wider focus on reproductive health (RH).
For this paper, only the outcomes specific to HIV risk reduction are
addressed. Improving knowledge related to HIV transmission,
acquisition and prevention (16 interventions) and altering sexual
behaviors related to risk and risk reduction such primary and
secondary abstinence, age of sexual debut, number of partners and
condom use (12 interventions) were the two most common
targeted outcomes. Among the qualitative studies, improving the
quality of life for participants and changing community norms that
contributed to risk of HIV infection were also common. One
intervention compared different PE approaches (Mathur, Malhotra,
& Mehta, 2004) and another specifically targeted condom use skills
and self-efficacy (Elkins et al., 1998).
2.1.3. Program activities
Programs employed a variety of activities to achieve their
targeted outcomes. The most common included provision of
information; practicing life skills; challenging social norms
through discussion, debate, drama and role play; distributing
condoms; and making referrals to RH services. Program delivery
methods were largely based on practical considerations such as
cultural or community norms and available resources. Popular
delivery methods included the use of drama/role play, 1-on-1
consultations, group activities/discussions, distribution of print
materials, condom distribution, and community events.
2.1.4. Peer educators (PE)
There was considerable variation in methods for selection and
training of PE and in the information provided about these.
Information regarding PE selection ranged from nonexistent (3
reports), to details of specific criteria and selection methods (6
reports). Only 5 studies reported that PE were chosen by fellow
youth. In the remainder, PE were either volunteers from among
group members or were selected by group leaders. In three of the 5
where youth chose the PE, this was identified as critical for
establishing trust and reaching the targeted participants (Mitchell,
Oling, Onen, Nyakake, & Manyindo Kihugur, 2002; Mitchell,
Nyakake, & Oling, 2007; Muyinda et al., 2003; UNFPA, 2004). Of
those providing information on training length, the shortest
reported training was 4 h and the longest 2 weeks. Most authors
identified training content as including training on relevant
information and program content as well as on message delivery
skills (e.g. public speaking, drama, counseling) with a specific
training curriculum often used. A cascade approach was used in
training for 4 interventions and 7 included regular refresher
courses.
Retention of PE was discussed in all publications with authors
noting that multiple factors such as competing time commitments,
pressures from family and peers, ‘‘growing up,’’ and the need for
paid employment work against the retention of PE over a long time
period. Maintaining contact among PE and with adult supervisors
and providing material compensation and rewards were the two
most commonly used methods to increase retention.
2.1.5. Duration, reach, and monitoring
Program duration, reach (number of participants), and mon-
itoring were not consistently reported. In programs that used
informal delivery methods such as community events or naturally
occurring gatherings and group activities it was impossible to
identify the specific number of contacts with specific youth.
Similarly, since a cornerstone of peer-led programming is the
involvement of peer groups or entire communities in activities that
may be repeated over an extended time period, the duration of a
program often could only be reported as the length of time it was
operating in a community. Most interventions included here
continued for many months, several as long as 2 years, and one for
over 4 years. The few that had a fixed-term with a more
curriculum-based content were of a shorter, limited duration
(e.g. a single 7 h session in Balinese age mate groups was evaluated
by Merati, Ekstrand, Hudes, Suarmiartha, & Mandel, 1997).
The intensive nature of programs that targeted a distinct
subgroup of youth and used primarily 1-on-1 delivery kept the
numbers that were reached low (e.g. 71 adolescent girls were
reached in the senga program in Uganda evaluated by Muyinda
et al., 2003). Programs that used community-wide events had a
much wider reach, but often the numbers had to be estimated (e.g.
condom races at community festivals in Thailand were estimated
to reach from 5000 to nearly 12,000 people in 77 rural villages;
Elkins et al., 1998).
Few publications provided information on monitoring. In those
that did, monitoring was most often described as posing a
challenge, most often related to literacy levels, skills, time and
resources. One creative solution to monitoring with minimal
resources and low literacy levels was the use of pictogram forms to
facilitate record keeping by illiterate PEs (UNFPA, 2004). Monitor-
ing is seen by some as an unnecessary, costly and time consuming
exercise done merely to satisfy the requirements of donor agencies
but with little or no benefit to the program itself. However,
evidence of the importance of monitoring to insuring quality
program delivery was provided in three evaluations (Askew et al.,
2004; Diop et al., 2004; Sharma, 2002), with poor monitoring
identified as a potential contributor to weak delivery.
2.2. Evaluation designs
Of the 25 evaluations, 6 used only quantitative methods, 9 only
qualitative and 10 used both. The 16 quantitative designs are
reported in Table 2 and the 19 qualitative in Table 3. All but two of
the 16 studies that used a quantitative evaluation methodology fell
within Habicht et al.’s (1999) plausibility category. Eleven used a
quasi-experimental pre–post design with a control group and 5
used pre–post surveys without a control group. In four studies
(Askew et al., 2004; Bhuiya et al., 2004; Diop et al., 2004; Frontiers,
2001) quasi-experimental designs incorporating two or more
experimental conditions were used. In all cases, one condition was
community-based only and the others incorporated a school-based
component. For this review, results for these four studies are
reported only for the community-based condition.
While the designs of the 11 quasi-experimental studies were
relatively strong, problems encountered during five of the studies
compromised study power. Though most control sites were far
enough away for cross-site contamination to remain unlikely,
Askew et al. (2004) reported that 3% of the adolescents in the
control site claimed they had participated in the intervention
program. Given the size of the sample in this study, this did not
compromise study power. In addition, three authors noted the
initiation of RH-related initiatives by other groups at the control
sites during the evaluation period (Bhuiya et al., 2004; Diop et al.,
2004; IRESCO, 2002). The degree of impact that this contamination
had on results is unknown. However, the effect on evaluation
outcomes is most likely to reduce the effect size. Statistical power
was compromised in one study by small sample size (Muyinda
et al., 2003). Of note is that the low power or effect size in all cases
was due to natural, unavoidable field occurrences and not to flaws
in study design.
 Author's personal copy
E. Maticka-Tyndale, J.P. Barnett / Evaluation and Program Planning 33 (2010) 98–112
Several of the pre–post surveys without controls also encoun-
tered problems. In the study by Moyo, Bond, Williams, and Mueller
(2000), the cross-sectional pre and post samples did not match
demographically, however, statistical controls were applied to
compensate for these differences. Two studies lost sizable portions
of their samples due to attrition (Nastasi et al., 1998; Merati et al.,
1997), compromising both power and generalizability of results.
The qualitative components of evaluation studies consistently
involved individual or focus group interviews/discussions (see
Table 3 for a full description). In the mixed methods studies these
provided more elaborate and detailed information that was used to
supplement and assist in interpreting the results from the
quantitative, survey methods. In eleven evaluation studies inter-
views were conducted with PEs, while in fifteen, interviews were
conducted with participating youth and often also with other
community members. Interviews were conducted both before and
after the intervention in only four studies (Bagamoyo College of
Arts, Tanzania Theatre Centre, Mabala, & Allen, 2002; Brieger,
Delano, Lane, Oladepo, & Oyediran, 2001; Cáceres, Cabezudo,
Jiménez, Valverde, & Pérez Luna, 1999; Frontiers, 2001). In all
others, interviews took place during or after the intervention.
Three studies included surveys as well as interviews, although
these were not statistically analyzed, with only descriptive results
presented. The qualitative evaluation studies fit Habicht et al.’s
(1999) criteria for providing evidence at the level of adequacy, i.e.
documenting that programs were delivered and changes occurred.
2.2.1. Sampling—quantitative
Samples were drawn from three populations, the targeted
youth (whether or not they participated in the program), identified
program participants, and the PE themselves. Population-based
samples of targeted youth reflect the level of knowledge, attitudes
or behavior among these youth regardless of their participation in
or contact with the program. Since evaluation participants do not
necessarily participate in the intervention but relies on a diffusion
of innovation model for intervention content to move from
participants to non-participants, this design typically produces the
weakest results. However, it is the most appropriate for peer-led
programs where the desire is to change communities or entire
populations. Of the seven evaluation studies that used population
samples, all but one (Bhuiya et al., 2004) used multi-stage cluster
designs and only Moyo et al. (2000) used a population-based
sample without a control group.
Six evaluation studies used samples of program participants.
Participant samples specifically address the effect of the program
on those who participated, with a minimum threshold of
participation often set for evaluation studies. Only Nastasi et al.
(1998) documented a participation threshold, restricting partici-
pation in the evaluation study to youth who attended at least half
of the 12 program sessions.
Finally, PE samples assess the impact of being a PE on the PEs
themselves, but do not provide information on the impact of the
intervention on the targeted population or participants. Samples of
PEs are most likely to produce the strongest results since these
youths have had the greatest exposure to the program. Three
studies specifically evaluated program impact on PEs with one of
these (Merati et al., 1997) restricting its sample exclusively to PEs.
Since over 3/4th of the original group of nearly 400 PEs were lost to
attrition in this latter study, results from this evaluation study
must be interpreted with caution.
2.2.2. Data collection and measurement in quantitative designs
Given the diverse intervention approaches, time from pre- to
post-measurement varied widely, ranging from a few hours (for
interventions delivered in discrete time periods) to 51 months (for
interventions that were more amorphous, on-going, and long
107
term). Two studies collected data at more than one post-
intervention period. Muyinda et al. (2003), with an intervention
targeting a small, discrete population and activities (1-on-1
mentoring and tutelage) collected evaluation data at 6 and 12
months post-initiation. Sergeyev et al. (1999), targeting a
population that is prone to ‘disappear,’ reported evaluation results
over three 3-month intervals with the sample decreasing from 484
at pre-assessment to 159 at the first and down to 33 at the final
assessment. Data collection required extended time periods with
large community samples (e.g. Mathur et al., 2004 reported
collecting baseline data over the course of 15 months). Speizer,
Oleko Tambashe, and Tegang (2001) report that baseline data were
collected 6 months after program initiation raises concerns that
the comparison was actually between change that occurred early
in the program to change that occurred later in the program rather
than between pre- and post-program.
Information regarding measurement tools was limited. A few
authors indicated modification of standardized tools, but most
reported measure creation. Several authors indicated piloting their
measures but, often due to restrictions on publication length, did
not provide details. Very few authors provided psychometric
assessments of their measures, and three provided no information
regarding the measurement tools at all. Potentially weak
measurement is a limitation of the studies reviewed.
2.2.3. Data collection and sampling—qualitative designs
A wider range of populations was sampled in studies with a
qualitative component. Commonly, some combinations of the
targeted youth, PEs, program staff, and community members or
stakeholders were sampled. This is consistent with the intent to
provide a more context-rich examination of program implementa-
tion and results. Sample size was not provided in 9 of 21 studies
(see Table 3). Of the remaining 12, several provided a partial
sample size (e.g. for the interviews, but not the focus group
discussions). The content covered during interviews and focus
group discussions was detailed in seven studies, with one (UNFPA,
2004) providing interview guidelines specific to each type of actor
(i.e. PE vs. youth vs. community member, etc.).
2.2.4. Data analysis
Of the 18 studies that included quantitative analyses, only 9
conducted and reported statistical analyses that fully used the
strength of their research design (Askew et al., 2004; Bhuiya et al.,
2004; Brieger et al., 2001; Diop et al., 2004; Elkins et al., 1998; Esu-
Williams et al., 2004; Moyo et al., 2000; Nastasi et al., 1998; Speizer
et al., 2001). Most often missing were full comparisons of
intervention and control groups across both the pre and post
data collection. Instead, 5 of the 11 studies with quasi-experi-
mental designs used only within group, pre–post comparisons
(Frontiers, 2001; IRESCO, 2002; Mathur et al., 2004; Muyinda et al.,
2003; Özcebe & Akin, 2002). This approach lost the advantage
provided by the research design to produce impact information
based on statistical intervention-control comparisons.
2.2.5. Strength of evaluation studies
The last columns in Tables 2 and 3 provide information on the
strength of each evaluation study based on Habicht et al.’s (1999)
criteria. Use of controls, sample size, intervention/control con-
tamination, measurement validation, data collection, and data
analysis techniques were considered in arriving at these categor-
izations. Four evaluation studies used designs that produced
conclusions about program effects at the level of strong
plausibility; however, Brieger et al. (2001) reached this level of
strength only for the knowledge outcome and was otherwise at the
level of moderate plausibility. Three studies produced conclusions
exclusively at the level of moderate plausibility and seven had
 Author's personal copy
108 E. Maticka-Tyndale, J.P. Barnett / Evaluation and Program Planning 33 (2010) 98–112
weak plausibility. Two that included quantitative methods and all
qualitative evaluation designs produced conclusions at the level of
adequacy.
2.3. Evaluation results
Table 4 presents results from quantitative evaluations and
Table 5 for qualitative evaluations. Studies reporting results from
both quantitative and qualitative analyses are included in both
tables. A recurrent issue in interpreting the evaluation studies was
the absence of comparability between targeted outcomes and
reported results. In most cases ‘‘extra’’ results (e.g. reporting
results regarding STI symptomology when reduction of STI rates
was not a targeted outcome) or assumed determinants of targeted
outcomes rather than direct measurement of the outcome itself
(e.g. condom use as a determinant of HIV incidence) were reported.
2.3.1. Knowledge
Knowledge was the most widely assessed outcome. Of the 13
studies in which knowledge results were reported, 10 (including
all but one of the studies with strong plausibility designs) reported
positive change in at least one sample subgroup. One study with a
strong plausibility design and two studies with weaker designs
reported no significant change. For the two studies reporting
negative change, this was only for some subgroups or types of
knowledge, with positive change obtained for others. Although
Nastasi et al. (1998) reported a clean gender split, with girls
experiencing positive change and boys nonsignificant, this study
had only a weak plausibility design and no clear gender
differences were evidenced in studies with stronger designs. It
does appear that certain types of knowledge are easier to learn
than others. Across all evaluation study designs, there was a trend
toward nonsignificant or negative change in knowledge of birth
control and the body (e.g. puberty, pregnancy, fertility). On the
other hand, knowledge related directly to HIV (e.g. prevention
methods and transmission routes) was most likely to exhibit
positive change.
Twelve qualitative analyses included reports about provision of
relevant information to youth. Eleven reported success, while
Sharma (2002) did not report clear results, potentially because this
intervention was unable to connect youth to services where
information would be provided. In addition, all eight of the studies
that assessed distribution of resources (e.g. educational print
materials and condoms) were successful in doing so.
2.3.2. Norms, attitudes, and skills
Eight of the 11 studies that reported qualitative results
regarding the transformation of community norms provided
evidence of success. Examples of community change included
modifying harmful practices such as reusing razor blades in
traditional scarring (the community returning to disposable,
sharpened wood chips) (Hughes-d’Aeth, 2002) and modifying
aspects of initiation rites that encouraged sexual risk behaviors
(Bagamoyo College of Arts et al., 2002). In addition, many authors
reported increased social capital for the PE, who frequently lacked
social power and resources due to being young, or female, or
unemployed. This is particularly relevant given the current debate
over social capital and risk for negative reproductive health
outcomes (Pronyk et al., 2008).
In the quantitative studies, the only ‘attitude’ related to HIV
prevention that was assessed was condom self-efficacy, and this in
only three studies. Elkins et al. (1998) and Merati et al. (1997), two
studies with pre–post designs, but the latter only assessing the PEs,
reported positive change; whereas, Brieger et al. (2001) using a
design that supported results at the level of moderate plausibility
found no significant change.
2.3.3. Sexual activity
Results regarding primary and secondary abstinence (the latter
operationalized as whether or not respondents had sex in the last
3–6 months) and number of partners were mixed. Three studies
measuring primary abstinence, two with designs supporting
conclusions at the level of strong plausibility and one with a
weak plausibility design reported a desirable effect for at least one
subgroup. Two of these studies (Esu-Williams et al., 2004; Mathur
et al., 2004) also reported undesirable effects on one subgroup.
Three studies with moderate and one with weak plausibility
designs, however, reported nonsignificant results for primary
abstinence.
Of the six studies reporting on secondary abstinence, one at the
level of moderate plausibility (Diop et al., 2004), and two with
weak plausibility (Frontiers, 2001; IRESCO, 2002) reported positive
change and three with strong and moderate plausibility reported
no significant change (Bhuiya et al., 2004; Esu-Williams et al.,
2004; Speizer et al., 2001). Of note is that in the two evaluations of
the Entre Nous Jeunes intervention, the stronger design (Speizer
et al., 2001) reported no significant change (recall, however, that
Speizer et al., 2001 actually compared early to later post data),
while the weaker design (IRESCO, 2002), which did compare pre-
to post-intervention, produced a positive change.
The only studies that reported success at reducing the number
of sexual partners had a weak plausibility design (Frontiers, 2001;
IRESCO, 2002). Three studies with strong or moderate plausibility
designs reported no significant effects (Bhuiya et al., 2004; Diop
et al., 2004; Esu-Williams et al., 2004).
Across reports with mixed results on changes in sexual
behavior, the negative or nonsignificant results were most likely
to occur among males or a particular subgroup of males. However,
across all reports as a whole, results were most often similar for
females and males.
2.3.4. Condom use
There were primarily positive gains for condom use. Of the ten
interventions that measured use, eight obtained positive results for
at least one subgroup (including all but one with a strong
plausibility design), two only nonsignificant results (one, Bhuiya
et al., 2004, with a moderate plausibility design) and none reported
negative results. Evaluation studies in which participant sub-
groups were assessed separately did find differences in results for
condom use based on gender, marital status and rural/urban
residence. Since few studies performed such breakdowns, we
cannot be certain of the reliability of these results. In addition, all
qualitative studies that targeted condom distributed provided
evidence of successful distribution.
2.3.5. STD symptoms
STD symptom measures were all self-report. Of the 6 studies
that asked about STD symptoms, none had a strong plausibility
design. One study with a moderate plausibility design (Askew
et al., 2004) obtained nonsignificant results, as did another with a
weaker design (Moyo et al., 2000). The 4 studies that obtained
positive results for at least some subgroups all were at the level of
weak plausibility or adequacy.
2.3.6. Results summary
PE programs were generally successful at increasing knowl-
edge, increasing condom use, and providing youth with informa-
tion and condoms. They may be successful at increasing condom
use self-efficacy and decreasing self-reports of STD symptoms,
although the weak designs of evaluation studies that assessed
these outcomes precludes drawing definitive conclusions. The
effect of programs on primary and/or secondary abstinence, and
number of sexual partners was less encouraging. PE programs may
 Author's personal copy
E. Maticka-Tyndale, J.P. Barnett / Evaluation and Program Planning 33 (2010) 98–112
produce delays in first sexual intercourse (primary abstinence) for
some groups of youth, but are unlikely to change the sexual
behaviors of youth who are already sexually active. The most
successful programs were in Senegal (Diop et al., 2004), Zambia
(Esu-Williams et al., 2004) and Cameroon (IRESCO, 2002; Speizer
et al., 2001). All three shared intervention designs that included
primarily group activities in a club-like atmosphere. The Zambian
intervention also involved youth in caring for people with HIV.
Each produced primarily positive gains with negative results
always restricted to specific subgroups or unique outcomes that
were balanced by positive results for others. All three produced
results at the level of strong plausibility. Least successful was a
program in Ghana and Nigeria (Brieger et al., 2001) which included
similar activities to those in the successful programs but produced
only nonsignificant results and had a strong plausibility design.
What must be noted about this program, West African Youth
Initiative (WAYI), is that it produced positive results when
delivered by PE in a school setting, suggesting that it was better
suited to delivery in this latter environment. Among interventions
using a qualitative evaluation design (as well as some with mixed
designs), when the community was targeted through participation
in program development, festivals, and other large-scale events,
there was a demonstrated shift in community norms toward an
increased recognition of local factors that make youth vulnerable
to HIV infection and a stated willingness to work to change these.
3. Discussion
The purpose of this review was to synthesize what is known
about peer-led community based interventions in low-income
countries with the intent of informing on-going and future
programming decisions. As discussed above, our review of 24
programs indicates that peer-led interventions of diverse forms
and content can produce positive change in areas related to HIV
prevention, especially knowledge, community norms, and condom
use. The mixed results for changing sexual behavior suggest that
peer educator led programs may not be the most effective in
changing sexual behavior. What must be noted, however, is that
few of the interventions reviewed here had strong evaluation
designs, many encountered field challenges that compromised
their statistical power, all evaluated results over a relatively short-
term, and none permitted identification of the specific components
that produced particular outcomes. As noted by several con-
tributors to a recent volume on community interventions to reduce
the spread of AIDS, such interventions are particularly difficult to
evaluate, there are multiple direct and indirect causal pathways of
influence, and evidence suggests that some interventions of this
type take a longer time to demonstrate an effect (Trickett &
Pequegnat, 2005).
The programs reviewed here were all delivered in natural
community settings with minimal or no control of the setting. This
reflected the ‘‘real life’’ situation in which interventions are
delivered; however, it periodically introduced serious challenges
for the evaluation studies. Although the evaluation studies set out
to assess program efficacy, given the natural settings, they
simultaneously were assessing program effectiveness. The most
serious challenges included the introduction of additional pro-
gramming in control sites and attrition of peer leaders, both of
which reduced the power of the studies, making it more difficult to
obtain significant results for the program of interest and exposing
conclusions to Type II errors (i.e. concluding no significant impact
when there was one). None of the programs were evaluated with
the strongest of study designs, the randomized control trial, and
only four of the studies used a design that supported conclusions at
what Habicht et al. (1999) refer to as a level of strong plausibility.
Of note is that the greatest successes were reported in studies with
109
the weakest designs. These must be interpreted with caution, given
the weakness of the designs and call attention to the need for
attention to high quality evaluation. In interpreting and reporting
results we took both the challenges faced by the interventions and
evaluation studies as well as the strength of study designs into
consideration, placing more emphasis on results from studies with
stronger designs. However, this may have inadequately repre-
sented the gains made by interventions evaluated with weaker
designs and, with Kim and Free (2008), we recommend caution in
excluding a particular type of intervention or activity from
consideration merely because it has not been adequately
evaluated.
3.1. Factors contributing to program success
Douglas Kirby’s extensive work with curriculum-based inter-
ventions has led the way in identifying factors that contribute to
program success (Kirby, Laris, & Rolleri, 2005). We follow Kirby’s
example in identifying factors that contribute to program success
(measured as at least one positive outcome and only conditional
negative outcomes) in the peer-led interventions included in this
review. Twelve of the 16 quantitative evaluations report at least 1
positive and no negative outcomes. Three of the remaining report
both positive and negative outcomes depending on the subpopu-
lation of interest and one reports no statistically significant results.
Of the 9 interventions evaluated using only qualitative methods, all
reported at least 1 positive and no negative outcomes.
3.1.1. Community needs assessment
Several evaluators identified the importance of a community
needs assessment prior to designing the program. Unless prior
work in a particular region is available to shed light on current
conditions, it is necessary to first find out what youth know (and
how they talk about it), what they are doing, what cultural norms/
practices shape their sexual landscape, and what they would like to
know. Without a needs assessment, programmers run the risk of
designing an intervention that uses inaccessible language or
concepts, culturally inappropriate messages, and/or fails to
provide useful information or services. In addition, a needs
assessment can be the first step in a broader scheme of community
involvement.
3.1.2. Community involvement
Active input from the target community (youth, as well as
community stakeholders) was used in the selection of program
content and activities, as well as implementation in all of the most
successful programs. Input from youth helps to ensure the message
and methods are effective, while acceptance and support from
those in power paves the way for a successful (and sustainable)
program. In several cases (Mathur et al., 2004; Muyinda et al.,
2003; UNFPA, 2004), programmers utilized the traditional social
structure to mobilize their intervention, resulting in the success of
program activities (including changing harmful community
norms) that would likely have otherwise been inhibited. In two
cases (Bagamoyo College of Arts et al., 2002; Mitchell et al., 2002,
2007), the specialized knowledge and community connections of
the PE were critical to program success.
Mathur et al. (2004) actually compared a standard PE program
with a PE program based on participatory learning and action
(PLA). They found that the PLA PE contacted more youth and
yielded better overall outcomes. In addition, the authors noted that
only the PLA intervention changed community norms by garnering
the youth increased self-confidence and respect from their elders.
Brady and Khan (2002) noted that investment from the broader
community facilitated the genesis of social space for girls and a
more even workload distribution between males and females.
 Author's personal copy
110 E. Maticka-Tyndale, J.P. Barnett / Evaluation and Program Planning 33 (2010) 98–112
Personal empowerment, positive alteration of community norms,
and investment from powerful community allies are noted as
important factors for sustained change.
3.1.3. Training
Training needs for PE were noted in several programs as being
greater or different than those for adults. This is not surprising
since we are talking about young people, often teenagers, who are
likely to have more limited education and life and work experience
than adults. Obviously training is needed for both adult and peer
leaders in the content that is relevant to the intervention, such as
information about HIV transmission and prevention. It is also
critical that the training cover community-specific issues related to
sexuality (such as moral belief systems, cultural practices, etc.),
message delivery and communication skills (e.g. group, individual,
drama, counseling), literacy (both verbal and textual), and self-
care. Several authors discussed the pressure that was placed on
peer leaders by friends and participants, as well as the difficulties
some peer leaders faced in their community. These factors make it
important that PE learn how to take care of themselves
emotionally when facing pressure, community ostracism, or
suspicion. Other topics it may be useful to cover include: gender
and sexuality issues, legal and ethical issues, concrete skills such as
proper condom use (and practice in required skills), and activity
organization.
Training time varied widely across the studies reviewed with no
clear evidence to suggest an ideal time. Two themes emerged in
examining training implementation across studies. First, inter-
ventions that used a cascade model, training local adult leaders or
lead PE who then trained the main PE, experienced more difficulty
with program implementation (e.g. PE had poor grasp of knowl-
edge, were not comfortable discussing issues/organizing events,
did not keep records, etc.) than programs where the PE were
trained directly by staff (for an example see Diop et al., 2004).
Second, refresher trainings appeared useful for those programs
with an extended timeframe. Refresher trainings served multiple
purposes. They allowed for training of new PE, rekindled
enthusiasm for the program, helped ensure delivery of accurate
information, and allowed for addressing or adjusting issues based
on program or PE experience.
3.1.4. PE supervision
There is an image of the peer leader as someone who is trained
and then sent out to find peers and deliver the intervention. In the
few instances where this has been evaluated, it has most often
faired poorly. In Nepal (Sharma, 2002) young adult leaders from
rural communities were brought to a central location for training
and then sent back to their local communities to pass on their
training and conduct a variety of activities. At the post-evaluation
period, there was little evidence of any activities. Many of the
trained youth had left their communities or moved on to other
activities. In Mathur et al.’s (2004) comparison of a PLA model to a
standard PE program the higher level of interaction between PEs
and adults in the former was identified as a major contributing
factor to its greater success. In contrast, the IDU peer-educators in
the intervention evaluated by Sergeyev et al. (1999) produced
desirable outcomes without on-going supervision or structure.
Clearly there is no definitive answer to the question of how
much structure and supervision peer leaders need. In fact, these
results raise additional questions. For example, we could ask
whether the training or selection of the peer leaders is the key
difference across these cases. However, if we want to stay ‘on the
safe side,’ it is advisable in planning a peer-led program to build in
a structured method for regular contact with supervisory staff. On-
going support of the PE is multi-purpose. It allows supervision to
help ensure the intervention is delivered with integrity. It also
helps PEs problem-solve, gain confidence, and enhance their social
investment in the program, aiding in retention.
3.1.5. PE retention
Retention of peer leaders was a concern for all PE programs.
Youth, especially in low-income countries, are a mobile popula-
tion. Youth grow up. They are also likely to have other priorities
demanding their attention, such as school, work, family respon-
sibilities, etc. The successful studies reviewed here demonstrated a
variety of strategies to enhance PE retention. These included:
reimbursement for expenses, incentives (pay, microcredit,
bicycles, ability to sell condoms for a small profit, etc.),
development of professional and job related skills, manageable
scope of work, acknowledgement, participation in project devel-
opment and decision-making, and support or supervision.
3.1.6. Sustainability
In an ideal situation, programs successful at effecting positive
change would be sustained long after the research team has left the
scene. As pointed out by Kelly et al. (2006), sustainability is a
challenge for community-level interventions. Some of the reviewed
studies sought to meet this challenge by partnering with govern-
ment. Governments have preexisting structure, authority, and
resources to maintain a program they support. For one example of
success with this strategy see Diop and Diagne (2007). However,
governments are not always willing to partner with NGOs or
research teams. Mathur et al. (2004) provided some recommenda-
tions for program sustainability without government partnership.
First, empower the youth and community through participatory
techniques (this also helps develop a sense of ownership). Second,
develop structures and mechanisms of delivery that are not reliant
on research team staff. Third, cede pieces of the program to well-
established organizations (such as NGO’s).
3.1.7. Summary of elements for success
As we have discussed above, certain program design and
implementation choices have a demonstrated correlation with
program success. Based on our review, we find the elements for a
successful PE program to include: a community needs assessment
(unless current data are already available from another source),
well-thought out PE selection (preferably with input from youth/
community stakeholders), adequate PE training, PE monitoring/
supervision, involvement of youth and community stakeholders in
program development and implementation, a structure for
program delivery, PE retention efforts, a system to locate and
train replacement PE, and a system for sustainability. With the
exception of PE retention and replacement and a system for
sustainability, which are unique to peer-led community-based
programs which may not have an identifiable sponsor, these
parallel the components for successful curriculum-based programs
identified by Kirby and his colleagues (2006). Although every
program will not need every one of these elements to effect
positive change, they may serve as a guideline for program
development based on the ways in which past programs have
achieved (or failed to achieve) success.
3.2. Limitations of peer-led programs
Although PE programs have demonstrated effectiveness in
producing positive change across several outcomes, there are
limitations to this approach. First, certain program requirements
must be met in order to yield these positive changes. As with all
types of programs, a PE program must be implemented well in
order to work.
Another potential limitation for PE programs is that adults may
be better at conveying factual information (although trained,
 Author's personal copy
E. Maticka-Tyndale, J.P. Barnett / Evaluation and Program Planning 33 (2010) 98–112
supervised peers are able); while peers may be better at engaging
youth in conversation about norms, attitudes, and behaviors.
Across the studies reviewed here there were several where youth
preferred to get factual (especially medical) information from
adults, or young adults/older adolescents with more schooling,
since they trusted them to know more than they believed their
peers knew. Adults and more educated older adolescents also
demonstrated more confidence in their ability to communicate
this information. However, there are also interventions where
peers have successfully conveyed information about sexual and
reproductive health.
In addition, in some communities, close peers may be ‘too close
for comfort’ and a sense of confidentiality. An intervention in Bali
(Merati et al., 1997) used pre-existing age-graded peer groups and
trained one member of the group as a peer educator to deliver
sexual and reproductive health information. Several participants
commented that they felt uncomfortable discussing their own
intimate matters or concerns with them because they were too
close to the peer educator and had an on-going relationship with
this person. They preferred someone who was more distant.
Researchers have experienced this same phenomenon in qualita-
tive work in closely knit communities or communities where there
is a strong shame culture and participants are uncomfortable
talking about sexual and reproductive matters with members of
their own community since they will have to face them after the
research is over. On the other hand, in several studies reviewed
here the success of the intervention seemed to hinge on the
established relationship between the peers and target community
(Mitchell et al., 2002, 2007; Muyinda et al., 2003). Again, it is
important to conduct a needs assessment and involve the
community in order to understand what will be most acceptable
and effective.
A final limitation revolves not around PE programs themselves,
but their evaluation. As evidenced here, high quality evaluation
studies are rare. Implementation of community-based programs is
time consuming and demanding of scarce resources. Resources for
quality evaluations are even more scarce, costly and donor
agencies are not as in press with funds for evaluations as they
are for programming itself. Yet, without monitoring and evalua-
tion, funds may well be misdirected, opportunities lost, and
mistakes or weak, ineffective programs perpetuated. What is
evidenced here, as well, is the difficulty of evaluating programs
targeting community-level and behavior change. Focus group
discussions with youth in low-income countries confirmed
previous findings that, when responding to surveys/interviews,
male youth tend to exaggerate their sexual experience while
female youth tend to minimize theirs. In addition, the youth
indicated that they tended to report more honestly during follow-
up due to their increased level of comfort with the issues and staff
(Merati et al., 1997). These factors conspire to cloud any evaluation
of youths’ sexual behaviors. Community-level change is even more
difficult to assess, with no standard indicators available.
4. Conclusions
These results suggest that peer-led interventions are able to
reach large numbers of youth when they are delivered at a scale
designed for such reach. They can be effective in changing
community norms, connecting youth to services, and distributing
HIV prevention resources. Those that evaluated specific changes
among youth demonstrated success at increasing knowledge and
the use of condoms. Changes in other areas of sexual behavior were
not as consistent across programs. The ability of programs to effect
abstinence – either through delaying first intercourse or through
having already experienced youth stop engaging in intercourse –
appears limited. Effects were more evident among young women
111
than young men, but even here the results were not consistent
across all interventions. Programs targeting community norms
also appeared to have met with some success, although the
evaluation study designs in these programs were less rigorous and
consequently, conclusions are more tentative. This synthesis paper
demonstrates that peer-led interventions can effect change, but
must be developed with attention to the requirements of such
interventions. It also demonstrates the challenges faced when
evaluating such interventions in natural settings—precisely the
settings in which they are ultimately delivered.
References
Aggleton, P., & Campbell, C. (2000). Working with young people—Towards and agenda
for sexual health. Sexual and Relationship Therapy, 15, 283–296.
Askew, I., Chege, J., Njue, C., Radeny, S., Kenyan Ministry of Health, & Kenyan Ministry of
Education, Science, and Technology, et al. (2004). A multi-sectoral approach to
providing reproductive health information and services to young people in Western
Kenya: Kenya adolescent reproductive health project. New York: Population Council.
Auerbach, J. (2008). Confronting the ‘‘evidence’’ in evidence-based HIV prevention:
Summary report. San Francisco AIDS Foundation, Science and Public Policy Depart-
ment.
Bagamoyo College of Arts, Tanzania Theatre Centre, Mabala, R., & Allen, K. B. (2002).
Participatory action research on HIV/AIDS through a popular theatre approach in
Tanzania. Evaluation and Program Planning, 25, 333–339.
Bernert, D. J., & Mouzon, L. D. (2001). Peer education in the 90s: A literature review of
utility and effectiveness. The Health Educator, Spring 2001 Edition, 31–37.
Bhuiya, I., Rob, U., Chowdhury, A. H., Rahman, L., Haque, N., Adamchak, S., et al. (2004).
Improving adolescent reproductive health in Bangladesh. New York: Population
Council.
Brady, M., & Khan, A. B. (2002). Letting girls play: The Mathare Youth Sports Association’s
football program for girls. New York: Population Council.
Brieger, W. R., Delano, G. E., Lane, C. G., Oladepo, O., & Oyediran, K. A. (2001). West
African Youth Initiative: Outcome of a reproductive health education program.
Journal of Adolescent Health, 29, 436–446.
Cáceres, C. F., Cabezudo, C., Jiménez, O., Valverde, R., & Pérez Luna, G. (1999). Sexual
health in a young city in Peru: A community-based intervention. Sexual Health
Exchange 2.
Campbell, C. (2004). Creating environments that support peer education: Experiences
from HIV/AIDS-prevention in South Africa. Health Education, 104(4), 197–200.
DiClemente, R., Crosby, R., & Wingood, G. (2005). Community HIV prevention inter-
ventions: Theoretical and methodological considerations. In E. Trickett & W.
Pequegnat (Eds.), Community Interventions and AIDS (pp. 222–248). NY: Oxford
University Press.
Diop, N. J., Bathidja, H., Toure, I. D., Dieng, T., Mane, B., RamaRao, S., et al. (2004).
Improving the reproductive health of adolescents in Senegal. New York: Population
Council.
Diop, N. J., & Diagne, A. F. (2007). Mainstreaming adolescent reproductive health in
Senegal: Enhancing utilization of the findings from the Youth Reproductive Health
Project. New York: Population Council.
Elkins, D. B., Dole, L. R., Maticka-Tyndale, E., & Stam, K. R. (1998). Relaying the message
of safer sex: Condom races for community-based skills training. Health Education
Research, 13, 357–370.
Esu-Williams, E., Schenk, K., Motsepe, J., Geibel, S., & Zulu, A. (2004). Involving young
people in the care and support of people living with HIV and AIDS in Zambia. Horizons
Final Report. Washington, DC: Population Council.
Fee, N., & Youssef, M. (1996). Young people, AIDS, and STIs: Peer approaches in
developing countries. In J. Mann & D. Tarantola (Eds.), AIDS in the World II: Global
dimensions, social roots and responses (pp. 247–248). New York: Oxford University
Press.
Frontiers in Reproductive Health. (2001). Peer educators can promote safer sex
behaviors. Zambia Youth Reproductive Health OR Summary 17. Washington, DC:
Population Council.
Gallant, M., & Maticka-Tyndale, E. (2004). School-based HIV prevention programmes
for African youth. Social Science and Medicine, 58, 1337–1351.
Gerber, M., & Kauffman, J. M. (1981). Peer tutoring in academic settings. In P. S. Strain
(Ed.), The utilization of classroom peers as behaviour change agents. New York:
Plenum Press.
Habicht, J., Victora, C., & Vaughan, J. (1999). Evaluation designs for adequacy, plausi-
bility and probability of public health programme performance and impact.
International Journal of Epidemiology, 28, 10–18.
Helm, C. J., Knipmeyer, C., & Martin, M. R. (1972). Health aides: Student involvement in
a university health center program. Journal of American College Health Association,
20, 246–251.
Horizons. (2000). Peer education and HIV/AIDS: Past experience, future directions.
Washington, DC: Population Council.
Hughes-d’Aeth, A. (2002). Evaluation of HIV/AIDS peer education projects in Zambia.
Evaluation and Program Planning, 25, 397–407.
Institut de Recherche et des Etudes des Comportements (IRESCO) [Institute of Research
and Studies of Behaviour]. (2002). Peer education as a strategy to increase contra-
ceptive prevalence and reduce the rate of STIs/HIV among adolescents in Cameroon.
Washington, DC: Population Council.
 Author's personal copy
112 E. Maticka-Tyndale, J.P. Barnett / Evaluation and Program Planning 33 (2010) 98–112
Kelly, J., Somlai, M., Benotsch, E., Amirkhanian, Y., Fernandez, M., Stevenson, L., et al.
(2006). Programmes, resources, and needs of HIV-prevention nongovernmental
organizations (NGOs) in Africa, Central/Eastern Europe and Central Asia, Latin
America and the Caribbean. AIDS Care, 18(1), 12–21.
Kim, C., & Free, C. (2008). Recent evaluations of the peer-led approach in adolescent
sexual heatlh education: A systematic review. International Family Planning Per-
spectives, 34(2), 89–96.
Kirby, D., Laris, B.A., & Rolleri, L. (2005) Impact of Sex and HIV Education Programs on
Sexual Behaviors of Youth in Developing and Developed Countries. Youth Research
Working Paper No. 2. Triangle Park, N.C.: Family Health International.
Mathur, S., Malhotra, A., & Mehta, M. (2004). Youth reproductive health in Nepal: Is
participation the answer? Washington, DC: International Center for Research on
Women EngenderHealth.
Maticka-Tyndale, E., & Brouillard-Coyle, C. (2006). The effectiveness of community
interventions targeting HIV and AIDS prevention at young people in developing
countries. In D. A. Ross, B. Dick, & J. Ferguson (Eds.), Preventing HIV/AIDS in young
people: A systematic review of the evidence from developing countries. Geneva:
WHO.
Merati, T. P., Ekstrand, M. L., Hudes, E. S., Suarmiartha, E., & Mandel, J. S. (1997).
Traditional Balinese youth groups as a venue for prevention of AIDS and other
sexually transmitted diseases. AIDS, 11(Suppl. 1), S111–S119.
Mitchell, K., Nyakake, M., & Oling, J. (2007). How effective are street youth peer
educators? Lessons learned from an HIV/AIDS prevention programme in urban
Uganda. Health Education, 107, 364–376.
Mitchell, K., Oling, J., Onen, T., Nyakake, M., & Manyindo Kihugur, S. (2002). Harnessing
talent: Ugandan street youth using drama. Sexual Health Exchange, 1, 15–16.
Moyo, I., Bond, K., Williams, T., & Mueller, L. (2000). Reproductive health antecedents,
attitudes, and practices among youth in Gweru, Zimbabwe: Findings prior to and
following the introduction of ‘‘youth-friendly’’ services. SEATS family planning
service expansion and technical support: FOCUS on young adults. Washington, DC:
USAID.
Muyinda, H., Nakuya, J., Pool, R., & Whitworth, J. (2003). Harnessing the senga
institution of adolescent sex education for the control of HIV and STDs in rural
Uganda. AIDS Care, 15, 159–167.
Nastasi, B. K., Schensul, J. J., Amarasiri de Silva, M. W., Varjas, K., Silva, K. T., Ratnayake,
P., et al. (1998). Community-based sexual risk prevention program for Sri Lankan
youth: Influencing sexual-risk decision making. International Quarterly of Commu-
nity Health Education, 18, 139–155.
Özcebe, H., & Akin, L. (2002). Peer education approach to young people on reproductive
health as an example from rural area, Turkey. Turkish Journal of Population Studies,
24, 51–64.
Pettifor, A., MacPhail, C., Bertozzi, S., & Rees, H. (2007). Challenge of evaluating a
national HIV prevention programme: The case of loveLife, South Africa. Sexually
Transmitted Infections, 83, i70–i74.
View publication stats
Pronyk, P., Harpham, T., Morison, L., Hargreaves, J., Kim, J., Phetla, G., et al. (2008). Is
social capital associated with HIV risk in rural South Africa? Social Science &
Medicine, 66, 1999–2010.
Rapkin, B., & Trickett, E. (2005). Comprehensive dynamic trial designs for behavioral
prevention research with communities: Overcoming inadequacies of the rando-
mized controlled trial paradigm. In E. Trickett & W. Pequegnat (Eds.), Community
Interventions and AIDS (pp. 249–277). NY: Oxford University Press.
Rapport Evaluation Thematique du projet UNFPA [Thematic Evaluation Report of
UNFPA Project]. (2004). Mobilisation communautaire, participation et renforce-
ment de l’autonomie des adolescent(e)s [Community mobilization, participation
and empowerment of adolescents] (unpublished manuscript). Burkina Faso: Rap-
port Evaluation Thematique du Projet UNFPA.
Ross, D. A., Dick, B., & Ferguson, J. (Eds.). (2006). Preventing HIV/AIDS in young people: A
systematic review of the evidence from developing countries. Geneva: WHO.
Sergeyev, B., Oparina, T., Rumyantseva, T. P., Volkanevskii, V. L., Broadhead, R. S.,
Heckathorn, D. D., et al. (1999). HIV prevention in Yaroslavl, Russia: A peer-driven
intervention and needle exchange. Journal of Drug Issues, 29, 777–803.
Sharma, M. (2002). Youth for each other programme: Rapid impact assessment. Nepal
Red Cross Society Junior/Youth Department HIV/AIDS Prevention Programme. Nepal:
Centre for Development and Population Activities.
Shiner, M. (1999). Defining peer education. Journal of Adolescence, 22, 555–566.
Shuguang, W., & Van de Ven, P. (2003). Peer HIV/AIDS education with volunteer trishaw
drivers in Yann, People’s Republic of China: Process evaluation. AIDS Education and
Prevention, 15, 334–345.
Speizer, I. S., Oleko Tambashe, B., & Tegang, S.-P. (2001). An evaluation of the ‘‘Entre
Nous Jeunes’’ peer-educator program for adolescents in Cameroon. Studies in
Family Planning, 32, 339–351.
Trickett, B.J. & Pequegnat, W. (2005) Community Interventions and AIDS. NY: Oxford
University Press.
Turner, G., & Shepherd, J. (1999). A method in search of a theory: Peer education and
health promotion. Health Education Research, 14(2), 235–247.
UNICEF Ghana. (2002). Evaluation of HIV/AIDS prevention through peer education,
counselling, health care, training and urban refuges in Ghana. Evaluation and
Program Planning, 25, 409–420.
Eleanor Maticka-Tyndale is a University Professor of Sociology and holds a Canada
Research Chair in Social Justice and Sexual Health at the University of Windsor, Canada.
HIV prevention among youth is one of her primary areas of research. She has completed
evaluation studies of interventions in Thailand and Kenya.
Jessica Penwell Barnett is a graduate student in the Sociology program at the University
of Windsor, Canada, and a research assistant in the Social Justice and Sexual Health
Research Lab. Her primary area of research interest is positive aspects of sexual health.