School, community, family, clinic, and technology-based approaches to adolescent pregnancy prevention: a comprehensive systematic review

Abstract

Adolescent pregnancy remains a global public health issue associated with adverse maternal and neonatal outcomes, educational disruption, and long-term socioeconomic consequences. This systematic review synthesised evidence on school, community, family, clinic, and technology-based interventions for adolescent pregnancy. A comprehensive literature search was conducted in Embase, ProQuest, Scopus, and PubMed to identify relevant articles published between 2018 and 2023 The screened articles were full-text, and the selected articles were open access. The search strategy combined MeSH terms and keywords using the Boolean operators Reviews and Meta-Analyses guidelines, 1352 records were identified, and 25 studies (randomised controlled trials [RCTs], mixed-methods, quasi-experimental) were included. Study quality and risk of bias were assessed using structured appraisal tools: the Joanna Briggs Institute Guidelines Tools for RCTs and quasi-experimental studies, and the Mixed Methods Appraisal Tools for mixed-methods studies. The findings were analysed through a design-stratified narrative synthesis to account for methodological heterogeneity. A meta-analysis was not conducted because the included studies were methodologically and clinically heterogeneous, thereby limiting the validity of pooling effect sizes. School- and technology-based interventions improved sexual health knowledge and attitudes but showed limited behavioural change. Community programs demonstrated moderate improvements in contraceptive use and delayed sexual initiation at 6–12 months. Clinic-based models increased communication and service utilisation but rarely measured pregnancy incidence. Only intensive family-based interventions with structured reinforcement reduced pregnancy rates at ≥12 months. Knowledge improvement does not guarantee a reduction in pregnancy. Sustained impact depends on intervention intensity, reinforcement, and long-term follow-up, supporting integrated multicomponent strategies with standardised pregnancy indicators.

Introduction

Adolescent pregnancy remains a global public health challenge, adversely affecting the health of both adolescents and their children.[1,2] It is closely associated with social consequences, including community stigma, risk of school dropout, and limited economic opportunities.[3] Despite ongoing global intervention programs, declines in adolescent pregnancy rates have been inconsistent across regions and socioeconomic groups. This pattern highlights the persistence of structural and contextual barriers.[4]

Multiple determinants influence adolescent pregnancy.[5,6] The Socio-Ecological Model offers a comprehensive framework for analyzing the multifaceted nature of this issue.[7] At the individual level, limited reproductive health literacy and inadequate refusal and decision-making skills elevate risk among adolescents.[8] Interpersonal factors, such as family and peer communication patterns, significantly affect sexual behavior. Institutional factors, including the roles of schools and health systems, determine access to education and health services.[9] Community norms regarding gender, early marriage, and sexuality further contribute to risk, while national policies and health systems influence the availability of youth-friendly services.[10] Interventions that address only a single level are unlikely to achieve sustainable behavioral change if reinforcing mechanisms at other levels persist. Consequently, effective prevention strategies necessitate coordinated efforts across all levels of the socioecological system to ensure lasting outcomes.[11]

Consistent with this multilevel framework, an increasing number of studies have evaluated interventions implemented in schools, communities, families, and clinics, and through digital platforms. School-based programs have demonstrated improvements in specific knowledge and behaviors.[12] Community-based interventions changed gender norms, strengthened life skills, and promoted community awareness of preventing adolescent pregnancy, which can be achieved through local rules, policies, and cultural norms.[13,14,15] . Family-based interventions enhanced communication between parents and adolescents^.[16] Clinic-based interventions increased the use of contraceptives and made them easily accessible^.[17] Technology-based interventions have been shown to expand reach and privacy, although reduced engagement and digital inequality may limit their sustained behavioural impact.[18] While these interventions correspond to different ecological levels within the socioecological model, they are often evaluated in isolation.

Most existing systematic reviews focus on a single type of intervention, without conducting cross-setting comparative analyses.[19,20] Furthermore, few syntheses address the considerable methodological heterogeneity present in the literature. Consequently, policymakers and practitioners lack a comprehensive assessment of how intervention models function across different ecological contexts and how methodological differences influence conclusions regarding effectiveness. Grounded in the Socio-Ecological Model and synthesizing multiple contexts, this framework enables the evaluation of effectiveness patterns across ecological levels. It examines how methodological design and risk profiles affect measured outcomes. It assesses whether intervention impacts vary by delivery context and framework constraints, while systematically accounting for methodological heterogeneity and study quality. This review aimed to provide a contextual and theory-driven synthesis of adolescent pregnancy prevention strategies. Specifically, the systematic review (1) identified and categorized intervention models across ecological levels, (2) evaluated effectiveness through a narrative synthesis structured by risk of bias, and (3) analyzed contextual implementation factors influencing program sustainability. By adopting an integrated ecological perspective, this study aimed to deepen understanding of adolescent pregnancy prevention across diverse sociocultural and health system contexts.

Materials and Methods

Study Design and Setting

This systematic literature review employed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) approach to select articles from previous studies. The review protocol was developed before data collection and comprised four main phases: identification, screening, eligibility assessment, and inclusion.

Data Sources and Search Strategy

Research articles were identified by searching the Embase, ProQuest, Scopus, and PubMed databases for publications from 2018 to 2023 to capture recent developments in adolescent pregnancy prevention interventions. The screened articles were full-text, and the selected articles were open access. The search strategy combined MeSH terms and keywords using the Boolean operators. The search strategy and keywords used in this review are presented in Table 1.

Table 1.

Search strategy and keywords

Database		Keywords
Embase		‘adolescent’/exp OR teen*:ab,ti OR youth*:*:ab,ti OR adolescen: *:ab,ti
		‘health education’/exp OR ‘sexual education’/exp OR ‘health promotion’:ab,ti OR sex education’:ab,ti OR ‘interventionstudy’:ab,ti
		‘adolescent pregnancy’/exp OR ‘pregnancy prevention’:ab,ti OR ‘adolescent pregnancy’:ab,ti OR ‘unintended pregnancy’: :ab,ti OR ‘teen pregnancy’:ab,ti OR ‘sex behavior’:ab,ti OR ‘sexual behavior’:ab,ti OR ‘sexual reproductive health’:ab,ti
PubMed		MAINSUBJECT.EXACT(“Teenagers”) OR abstract(teen* OR adolescen* OR youth*)
		(MAINSUBJECT.EXACT(“Health education”) OR MAINSUBJECT.EXACT(“Sex education”) OR MAINSUBJECT.EXACT(“Health promotion”)) OR (“SEX EDUCATION” OR “HEALTH PROMOTION” OR “HEALTH EDUCATION” OR INTERVENTION STUDY)
		(MAINSUBJECT.EXACT(Adolescent sexuality”) OR MAINSUBJECT.EXACT(“Teenage pregnancy”)) OR abstract(“pregnancy prevention” OR “adolescent pregnancy” OR “unintended pregnancy” OR “sexual behavior”)
ProQuest		MAINSUBJECT.EXACT(“Teenagers”) OR abstract(teen* OR adolescen* OR youth*
		(MAINSUBJECT.EXACT(“Health education”) OR MAINSUBJECT.EXACT(“Sex education”) OR MAINSUBJECT.EXACT(“Health promotion”)) OR abstract(“SEX EDUCATION” OR “HEALTH PROMOTION” OR “HEALTH EDUCATION” OR INTERVENTION STUDY)
		MAINSUBJECT.EXACT(“Teenage pregnancy”) OR abstract(“pregnancy prevention” OR “adolescent pregnancy” OR “unintended pregnancy” OR “sexual behavior”)
Scopus		TITLE-ABS-KEY (teen* OR adolescen* OR youth*)
		TITLE-ABS-KEY (health AND promotion OR sex AND education OR health AND education OR intervention AND study)
		TITLE-ABS-KEY (pregnancy AND prevention OR teen AND pregnancy OR adolescent AND pregnancy OR unintended AND pregnancy)

Eligibility criteria

Studies were included if they met the following criteria.

1. Population: Adolescents or youth aged 10–19 years.

2. Intervention: Programs designed to prevent adolescent pregnancy implemented in schools, communities, clinics, families, or delivered through technology.

3. Comparison: Comparison groups included those receiving alternative interventions or those observed without intervention.

4. Study design: Randomized controlled trials (RCTs), mixed-methods, experimental, and quasi-experimental designs.

5. Outcome measures: Reduction in adolescent pregnancy rates, changes in age at sexual initiation, improvement in life skills, and increased knowledge and positive attitudes regarding adolescent pregnancy prevention.

6. Language: Published in English.

Exclusion Criteria

The following were excluded:

1. Studies focusing exclusively on adult populations (aged ≥19 years).

2. Narrative reviews, editorials, commentaries, and conference abstracts without full texts.

3. Studies examining interventions addressing adolescent health in general, without a specific focus on pregnancy prevention.

We manually searched the reference lists of the eligible studies for additional studies. Before starting this review, we searched for existing systematic reviews and meta-analyses on this topic. We manually searched the reference lists of those systematic reviews using the same screening process.

Data Analysis

The primary outcome of interest was the prevention of adolescent pregnancy, defined as a reduction in adolescent pregnancy rates, changes in the age of sexual initiation, improvement in life skills, and increased knowledge and positive attitudes regarding adolescent pregnancy prevention. All articles were reviewed during data analysis. Data were extracted using the PICOS (Population, Intervention, Comparison, Outcome, and Study Design) framework from all relevant articles, following reading and summarisation. Articles were screened based on their titles, abstracts, and full texts, and the study quality was assessed using the Joanna Briggs Institute (JBI) Guidelines Tools and the Mixed Methods Appraisal Tools (MMAT). A meta-analysis was not conducted because the included studies were methodologically and clinically heterogeneous, thereby limiting the validity of pooling effect sizes.

Risk of Bias Assessment

Risk of bias was assessed using the JBI critical appraisal tool for RCTs and quasi-experimental studies. The mixed methods studies were assessed using MMAT. Each study was evaluated based on the criteria of these tools. There are 13 questions for RCTSdesigns and 9 for quasi-experimental designs, with answer options of Yes, No, Unclear, or NA. Meanwhile, 15 questions are used to assess the quality of mixed methods studies, with answers Yes, No, or Can’t Tell. The score (%) is calculated as the number of Yes answers divided by the total number of questions. Scores are divided into three categories: Poor, Fair, and Good. Based on the assessment results, all articles were categorized as GOOD QUALITY and eligible for analysis.

Data Synthesis

A narrative synthesis approach was employed to summarize the findings, given the heterogeneity in study designs, intervention types, and outcome measures. Interventions were categorized into five models: school-based, family-based, community-based, clinic-based, and technology-based. Details of the data synthesis are presented ini Tables 2 and 3.

Table 2.

School-based analysis

Authors, year, country		Population/sample size		Intervention (type, method, duration)		Study design		Outcomes						Sustainability
								Knowledge		Behavior		Pregnancy incidence
Yakubu et al., 2019 Ghana		Students aged 13–19 years n=367		Comprehensive Sexuality Education Program Methods: face-to-face classroom discussions, demonstrations, role-playing, and problem-solving techniques Duration: 1 month (1–2 sessions per week). Post-test: 3 months.		Cluster-RCTs		Significant increase in knowledge		Significant improvement in sexual abstinence		Not measured		The measurement period was effective for short-term outcomes, but longer-term effects were not evaluated.
Lohan et al., 2018 UK		Students aged 14–16 years n=831		“If I Were Jack” (interactive video drama, IVD) + classroom materials Method: teacher-delivered (trained) IVD on a computer/tablet Duration: 4 weeks (4 lessons); 12-month trial, follow-up at 5 and 9 months.		RCTs		Increase in knowledge		A slight positive impact of delaying or reducing unprotected sex		Not measured		The intervention showed positive short-term results, but sustainability remains uncertain due to the lack of long-term follow-up data
Rohrbach et al., 2019 US		Students aged 12–14 years n=4562		It’s Your Game: Keep It Real (IYG) (HIV/STI pregnancy curriculum-based and interactive computer program) Method: combining classroom-based, individual journaling, and computer-based activities Duration: 2 academic years, 24 lessons. Follow-ups at 1 year after intervention.		Quasi-experimental		Significant increase in knowledge		Positive decrease in sexual activity and presexual behaviors		Not measured		Positive short-term impact, but long-term sustainability needs further study
Lohan et al., 2022 UK countries		Students aged 14–15 years in secondary schools n=6561		“If I Were Jack” program Method: 90 min of face-to-face training. Duration: 12 weeks (1 session per week). Follow-up for 12–14 months.		Cluster-RCTs (cRCTs)		Significant increase in knowledge		Positive increase in contraceptive use but no change in sexual abstinence		Not measured		Positive short-term impact, sustainability remains uncertain
Houck et al., 2018 US		Students aged 12–14 years n=420		Talking About Risk and Adolescent Choices Program (emotional regulation intervention and SRH) Method: small-group education (4–8 adolescents) Duration: 12 sessions (twice a week, 1 h per session) + 2 boosters. Follow-up for 30 months.		Cluster-RCTs		Significant increase in knowledge		Significant decrease in delayed sexual debut and fewer condomless sex acts		Not measured		Positive short-term impact, but the long-term sustainability remains uncertain
Rotz et al., 2018 US		Students aged 14–15 years n=942		Teen Prevention Education Program (PEP) (comprehensive sexual education—peer-to-peer) Method: training peer educators through daily elective classes Duration: 5 90-min workshops, 1 academic year. Follow-up in 6 months.		Quasi-experimental		Significant increase in knowledge		No significant change in sexual activity or condom use		Not measured		Positive short-term impacts on knowledge, but no significant long-term behavioral changes
Manlove et al., 2021 US		Students aged 13–17 years n=626		Re: MIX Program (comprehensive sexual education) Method: face-to-face in class (co-facilitation by professional + young parent educators Duration: 10 weeks (1 h per session per week) Follow-up: 12 months.		Cluster-RCTs		Increased significantly, with long-term retention of information		No significant change in sexual activity or condom use		Not measured		Positive short-term effects were sustained over the long term, particularly in knowledge and self-efficacy
Nik Farid et al., 2018 Malaysia		Junior high school students, 12 years old n=209		MyCAP (Malaysian Care for Adolescent Project)—internet-based Method: interactive website Duration: 1 session (1.5 h for online intervention). Follow-up for 12 months		Quasi-Experimental		Significant increase in knowledge		No significant change in sexual behaviors or contraceptive use		Not measured		The intervention’s positive effects on knowledge were sustained over time, but behavioral changes were not achieved
Daley et al., 2019 US		Students aged 14 years n=4789		Teen Outreach Program (TOP)—Positive Youth Development (PYD) model (CSE) Method: weekly curriculum sessions (≥25 sessions), Community Service Learning (CSL) (≥20 h), face-to-face with trained adult facilitators Duration: 9-month and 1-year follow-up.		RCTs		Significant increase in knowledge		Improved immediately after the program, but the effects diminished over time		Not measured		Positive short-term effects, but the impacts were not sustained in the long term
Kenyon et al., 2019 US		Students aged 11–14 years n=45		My Journey (culturally attuned school-based sexual health curriculum) Method: facilitated by two local facilitators, classroom teachers; delivery in class Duration: 28 lessons. Implementation over one semester.		Quasi-experimental (single-group pretest–posttest)		Increased significantly in terms of sexual health and cultural understanding		Improved, particularly in terms of decision-making and sex refusal self-efficacy		Not measured		Positive short-term effects; long-term sustainability is uncertain without ongoing reinforcement
Gómez-Lugo et al., 2022 Columbia		Girls aged 12–19 years n=2047		COMPAS Program (Competencies for Adolescents with a Healthy Sexuality) Methods: delivery—face-to-face at school Duration: 5 sessions—1 h, once per week. Follow-up in 6 months		Cluster-RCTs		Increased significantly, with sustained improvements		Improved in terms of self-reported intentions and assertiveness, but no significant change in actual behavior or condom use over time.		Not measured		Positive short-term effects, but long-term sustainability was limited
Thongnopakun et al., 2022 Thailand		Students aged 13–15 years n=71		“Our Love, Our Control” (SRH Edutainment Program) Method: online media via a closed Facebook group Duration: 20 weeks. Follow-up: week 8 and week 20.		Quasi-experimental		Increased significantly, with the effect remaining positive but slightly reduced over time		Improved in the short term, but declined by the 20th week, indicating short-term change		Not measured		Positive in terms of knowledge, but behavioral sustainability was limited without continued reinforcement
Millanzi et al., 2022 Tanzania		Adolescents aged 12–19 years n=660		Integrated reproductive health (RH) lesson materials with problem-based pedagogy (PBP) Method: face-to-face, small groups. Duration: 4 weeks, 4 sessions (1–2 times per week, 30–90 min per session).		Cluster-RCTs		Increased significantly, with improvements in sexual health literacy.		Improved in terms of soft skills (self-esteem, negotiation, refusal), but the direct impact on sexual behavior was not clearly measured.		Not measured		Integrating reproductive health, the positive effects were sustained in the short term (3 months), but some decline was observed at 6 months

IVD=interactive video drama, RCTS=randomized controlled trial, IYG=It’s Your Game: Keep It Real, HIV=human immunodeficiency virus, STI=sexually transmitted infection, cRCT=cluster-RCT, SRH=sexual and reproductive health, PEP=Prevention Education Program, Re: MIX=Maximize their strengths, Imagine a healthy future, and eXplore their identities. MyCAP=Malaysian Care for Adolescent Project, TOP=Teen Outreach Program, PYD=Positive Youth Development, CSE=Comprehensive Sexuality Education, CSL=Community Service Learning, COMPAS=Competencies for Adolescents with a Healthy Sexuality, RH=reproductive health, PBP=Problem-Based Pedagogy

Table 3.

Community-based, family-based, clinic-based, and technology-based analysis

Authors, year, country		Population/sample size		Intervention (type, delivery methods, duration)		Study design		Outcomes						Sustainability
								Knowledge		Behavior		Pregnancy incidence
Community Based
Chinman et al., 2017 US		Adolescents aged 11–13 years n=419		Making Proud Choice Supported by Getting to Outcomes (MPC + GTO) Methods: manual, face-to-face training, and onsite technical assistance (TA) Duration: MPC: 8 weeks; delivered once per year, GTO: 2 years. Follow-up: 66 months, year 1, year 2.		Cluster-RCTs		Significant increase in knowledge		No significant changes in sexual activity, condom use, and sex without a condom		Not measured		Significantly increased condom attitudes/intentions and knowledge, but not in behavior, for a 2-year follow-up
Manlove et al., 2022 US		Male adolescents aged 15–18 n=110		Manhood 2.0 (social messages about masculinity and gender norms) Method: gender norms discussion, trained facilitator, group-based reflective discussion. Duration: 8 sessions (1–2 h per session), twice a week for 4 weeks. Follow-up: last session (4 weeks).		Mixed-methods		Increased overall (information exposure and self-reported knowledge), but no change in some objective “percent correct” knowledge scores		Cannot be concluded; insufficient behavioral evaluation		Not measured		Without long-term follow-up, the sustainability of effects cannot be evaluated
Oman et al., 2018 US		Adolescents aged 13–18 years n=1036		“Power Through Choice (PTC)—Comprehensive Sexuality Education” Methods: interactive approach Duration: 10 sessions (90 min/session), twice a week. Follow-up at 6 and 12 months.		RCTs		Not measured		Sex without birth control, past 3 months, decreased at 6 months, no change at 12 months		No change at 6 months, decreased at 12 months		Behavior sustainability decreases by 12 months (effect fades), and pregnancy outcome improvement emerges by 12 months
Campero et al., 2021 Mexico		Students aged 11–19 years n=747		“Yo importo, yo aprendo, yo decido”—CSE Method: face-to-face in school and the community Duration: 6 months, with activities every 3 weeks (2 h per session). Follow-up: 6 months.		Quasi-experimental		Increased SRH knowledge/self-efficacy and SRHR knowledge		Not measured		Not measured		There is only one follow-up point, the study cannot establish whether effects persist beyond the immediate post-intervention period
Family-based
Brown et al., 2021 US		Adolescents aged 13–19 years n=786		LiFT Program (Linking Families and Teens) Method: delivered by a trained facilitator. Additional: text messages, booster call for parents (3–5 weeks after the workshop) Duration: 5-h family workshop + post-workshop support (text message and phone call): Follow-up: 3 and 12 months		Cluster-RCTs		Not measured. The main outcome is parent–child communication		No change at both 3 and 12 months		Decreased at 3 months (significant); decreased at 12 months (favorable direction, marginal)		Increased/sustained for communication and perceived competence, while pregnancy reduction is partially sustained
Clinic-based
Ford et al., 2019 US		Adolescents aged 14–15 years and their parents/guardians n=118 dyads		Parent-targeted brief interventions with a written handbook and brochures Method: delivered by a nonclinical health coach + doctor/nurse practitioner Duration: 10 min during a well-care visit + follow-up call 2 weeks later. Follow-up analysis: 4 months.		RCTs		Increase adolescent frequency of communication about sex and alcohol. Knowledge not measured		Not measured		Not measured		Short-term improvement sustained at 4 months. Long-term improvement unknown
Jenner et al., 2023 US		Female adolescents aged 18–19 years n=1770		Entertainment–education video (“Plan A”) Method: viewed via Chromebook in clinic waiting room Duration: 1 session (23 min). Follow-up: 3 months		RCTs		Significantly increase contraceptive knowledge		Partially improved (subgroup LARC uptake, no overall change in condomless sex)		Not measured		Positive effect in 3 months, long-term effect not measured
Technology-based
Ybarra et al., 2021 Tanzania		Adolescent girls aged 14–19 years n=703		Girl2Girl (text messaging) Method: regular text message—SMS (2–3 times per week), with optional interactive quizzes and counseling via hotline Duration: 20 weeks. Measurement at baseline, 3 months, and 6 months.		RCTs		Knowledge increased, particularly in contraceptive awareness		Increased protective contraceptive use, while sexual activity rates showed no significant change		Not measured		Short-term maintenance of knowledge and partial behavioral improvement
Martínez-García et al., 2022 US		Female adolescents aged 14–18 years n=1210		Mobile application “Crush.” Method: Access by mobile phone. Participants also receive multimedia messages every 3 days (motivation, content boosters, reminders) Duration: Full access for 6 months. Baseline measurements at 3 and 6 months.		RCTs		Increased at 6 months, no significant change at 3 months		No significant change at both 3 and 6 months		Not measured		Psychosocial improvements sustained and strengthened through 6 months; behavioral outcomes not sustained because they did not significantly change at any point
Narkarat et al., 2021 Thailand		Female students, with an average age of 14 years n=128		Mobile health education based on LINE messages and Facebook animations Methods: Delivery by LINE and Facebook Duration: 24 weeks (3 messages per week, total 72 LINE messages). Measurements at baseline and after 24 weeks.		Quasi -experimental—pre–post- design		Significantly increased knowledge (sexual health literacy)		Not measured		Not measured		Demonstrated short-term improvement (24 weeks); long-term sustainability not assessed
Kissinger et al., 2022 US		Black/African-American females aged 18–19 years n=637		BuTifull (web-based online program) Method: 8 interactive sessions (videos, texts, activities, discussion forums) Duration: 4 weeks (website access, each session ±30 min)—measurements at baseline, 6 months, and 12 months.		RCTs		Significantly increased and sustained		Modest improvement in protective behaviors; no major change in sexual activity frequency		No significant change		Knowledge improvements sustained; behavioral gains partially sustained; long-term effects beyond the study window unknown
Turah et al., 2019 Indonesia		Adolescents aged 15–17 years n=106		Gapin Apps (Mobile Apps–Android) about Sexual and Reproductive Health Education Method: Presented through the Gapin application on smartphones Duration: 7 days. Follow-up at day 7.		Quasi-experimental		Significantly increased in knowledge and attitudes		Not measured		Not measured		Short-term improvement only (7 days); long-term improvement not measured

MPC + GTO=Making Proud Choice Supported by Getting to Outcomes, TA=technical assistance, PTC=Power Through Choice, RCTs=randomized controlled trials, CSE=Comprehensive Sexuality Education, SRH=sexual and reproductive health, SRHR=Sexual and Reproductive Health and Rights, LiFT=Linking Families and Teens, LARC=Long-Acting Reversible Contraception, LINE=Communication app which allows us to make FREE voice calls and send FREE messages

Results

Study Selection

This study identified 1352 articles from four databases (Embase, ProQuest, Scopus, and PubMed) published between 2018 and 2023. After removing 352 duplicates, 1000 articles remained for title screening. Of these, 63 proceeded to abstract screening, and 35 were excluded for failing to meet the eligibility criteria. This left 28 studies for full-text screening, of which 25 were eligible for analysis. Details of the selection process are presented in Figure 1.

Figure 1.

PRISMA flow diagram of article selection process. PRISMA: Preferred Reporting Items for Systematic Reviews and Meta-Analyses

Discussion

School-based Intervention

Schools are strategic environments for interventions to prevent adolescent pregnancy because teenagers spend most of their time there. Schools provide academic education and influence social norms, attitudes, and behaviors related to sexual and reproductive health (SRH).[18] They are also ideal for comprehensive SRH education, which can increase knowledge and reduce adolescent pregnancies and sexually transmitted infections.[21,22,23] This study identified 13 articles on school-based interventions. The study designs used included RCTs and cluster-RCT[12,24,25,26,27,28,29,30] and quasi-experimental.[27,31,32,33,34]. This variation is important because the strength of causal inference and the risk of bias differ between designs, and this was reflected in the variation in the consistency of behavioral outcomes.

The delivery methods used also varied widely, including face-to-face classroom discussions, demonstrations, role-play, problem-solving, a combination of classroom and journaling activities, computer activities, edutainment through Facebook, and website-based SRH modules. It means, “school-based” is not limited to a single type of intervention, but rather an umbrella of interventions that includes pedagogical approaches, facilitators (trained teachers vs. local facilitators vs. professional educators/young parent educators), and media (offline–online).

The duration of the interventions ranged from one session (Malaysian Care for Adolescent Project [MyCAP]) to two academic years/24 lessons (It’s Your Game: Keep It Real) and one academic year (Prevention Education Program [PEP]), as well as a medium duration of 4–12 weeks. These differences in duration/intensity also affected the amount of learning participants received, which influenced consistent changes in knowledge, but they differed slightly in behavioral variables. The results of the analysis of most school-based studies (3 studies) showed a significant increase in knowledge. Almost all interventions, whether curriculum-based (comprehensive sexual education), interactive video, problem-based pedagogy, or online edutainment, showed a consistent increase in knowledge. However, the behavioral outcomes were inconsistent.

Positive behavioural changes, such as increased abstinence and contraceptive use, and decreased sexual activity, were observed in some studies; however, others showed insignificant behavioural changes (e.g. PEP, Re: MIX, MyCAP) or decreases at follow-up. Pregnancy incidence was not measured in all school-based studies. Regarding the timing of measurement, follow-up began at 3, 6, or 12–14 months, and extended up to 30 months . Some studies reported weakening behavioral effects over time (e.g., Teen Outreach Program: behavioral effects declined; Facebook edutainment: behavioral improvement declined at week 20). This suggests that longer follow-up is often associated with challenges in sustaining behavioral outcomes. School-based interventions are highly effective in improving adolescent sexual literacy and understanding in the short term. Still, the effects of increased knowledge on consistent and sustained behavioral change remain limited. Without sustained and long-term reinforcement, the impact on behavior tends to decline. The success of school-based interventions is also influenced by collaboration between schools, parents, and communities. Some interventions do not actively involve families or communities. However, social support from parents and communities is an important determinant of changes in adolescent behavior.[24,25]. Family and community engagement strategies are vital for promoting healthy behaviors among students. These strategies include communicating with families, providing support or education, and fostering collaborations between schools and communities to support students and their families.[22] Integrated interventions with communities in Nigeria have shown positive changes in attitudes and beliefs related to adolescent reproductive health, demonstrating the importance of aligning school-based programs with community norms and values.[35]

Community-based Intervention

Community-based interventions are key to addressing adolescent pregnancies. These approaches extend beyond formal education by engaging adolescents in their social environments. Engaging all community members helps reduce adolescent pregnancy by mobilizing residents and educating stakeholders. These interventions utilize evidence-based adolescent reproductive health practices, with training and support from the government and local organizations.[36] Community-based programs empower adolescents, their families, and stakeholders. They provide sexual education, life skills, and access to reproductive health services, often through youth clubs, faith-based organizations, and nongovernmental organizations.[13,14,37,38] This study identified four community-based articles. The study design included cluster-RCT, RCT, quasi-experimental, and mixed-methods. The delivery methods used varied, ranging from manual training and onsite technical assistance (Making Proud Choice Supported by Getting to Outcomes), gender-norm-based reflective discussions (Manhood 2.0), interactive approaches (Power Through Choice), and face-to-face learning in schools and communities (“Yo import, yo aprendo, yo decido”). Community interventions are not merely individual interventions; they strengthen community support. The wide range of durations, from 4 to 5 weeks to 2 years of technical assistance, an variable follow-ups led to inconsistent changes in the identified studies.

The results of the analysis showed an increase in knowledge in several studies, but inconsistent behavioral changes. One study showed a decrease in risky behavior at 6 months, but the effect did not last until 12 months. The incidence of pregnancy was measured in only one study, and the results showed a decrease in pregnancy rates at 12 months.[38] The results also showed that knowledge/attitude can be sustained, while behavior measurement results were inconsistent in longer follow-ups. The community approach has the potential to reach adolescents in a broader social context, but behavioral effects tend to be unstable in the long term. The sustainability of the effects appears to depend on the duration of mentoring and the intensity of technical support.

Family-based Intervention

Family-based interventions are vital for shaping adolescent sexual health outcomes. They achieve this by enhancing communication, supervision, and emotional connections between parents and adolescents. Effective communication is a protective factor against risky sexual behaviors. In sub-Saharan Africa, improved communication has been shown to reduce the risks of early sexual intercourse and adolescent pregnancy.[39,40] Parents provide key information about adolescents’ SRH. Their communication is often reactive, one-sided, and authoritarian. This limits open discussions on romantic relationships and sexual health. Many parents are unprepared to discuss SRH topics with their children.[41] Research shows that effective communication between parents and adolescents and a supportive family environment can reduce risky sexual behaviors among adolescents.[16]

O ne study assessed a family-based intervention called Linking Families and Teens (LiFT). LiFT is a workshop that strengthens communication skills and increases knowledge about SRH. It starts with separate workshops for adolescents and parents, then brings them together for sessions on shared values, communication, and creating a family action plan. The parent training component focuses on enhancing parents’ SRH knowledge and skills, thereby enabling more effective communication with their adolescents.[16] The design used in the study was cluster-RCT. The results showed no significant change in sexual activity. A significant decrease in pregnancy incidence occurred at 3 months, with a positive change at 12 months’ follow-up. The study also showed increases in communication between parents and adolescents and in perceived competence. Family-based interventions have a direct impact on pregnancy incidence compared to other settings. Although sexual behavior does not change significantly, strengthening family communication can be a protective mechanism in preventing adolescent pregnancy.

Clinic-based intervention

Clinic-based interventions aim to increase adolescents’ access to reproductive health services. By enhancing availability, these programs help prevent adolescent pregnancies and support the overall well-being of women. Clinic-based intervention programs for reducing risky sexual behavior and teenage pregnancy emphasize adolescent-oriented services, protect confidentiality, and ensure affordability.[41]

This study identified two articles discussing clinic-based interventions. Clinic-based interventions are usually carried out during visits to health facilities. Both identified studies used an RCTSdesign, but differed in their delivery methods. The approaches used were short intervention programs for parents, involving adolescents and parents who visited clinics.[17] In one study, a brief parent-targeted intervention was conducted during the visit (10 min and a follow-up call 2 weeks later) by a health coach and doctor/nurse, while in the other a 23 minutes entertainment education video (“Plan A”) was shown in the waiting room before patients consulted with the doctor.[42] Both studies shared the same characteristics. That is, brief interventions delivered during clinic visits. The relatively short duration limited the study’s ability to produce outcomes that require longer interventions

The results showed that Plan A increased knowledge and improved some behavioral changes (Long-Acting Reversible Contraception – LARC uptake). In contrast, parent-targeted brief interventions increased the frequency of communication about sex and alcohol but did not measure knowledge, behavior, or pregnancy incidence. Both studies had relatively short intervention and follow-up periods (3–4 months). In general, clinic-based education is effective for improving knowledge and communication in the short term. However, the limited duration of the intervention and follow-up hinders the evaluation of sustainability and impact on pregnancy incidence.

Technology-based intervention

Technology-based interventions represent an innovative and effective approach to preventing adolescent pregnancies. By leveraging digital platforms, these interventions provide users with easily accessible, private, and engaging health information.[43] Digital technology is particularly valuable for reaching adolescents whom traditional school-based programs may not serve well. Multiple technology-integrated sexual health education efforts have been linked to increased clinical service use and improved sexual health knowledge among the participants.[44]

This study identified five technology-based interventions for preventing adolescent pregnancy. The designs used were RCT[43,45,46] and quasi-experimental.[18,47] Delivery methods varied widely, including text messages, mobile applications, social media education, and web-based education. This variation affected engagement and exposure sustainability, which was reflected in the sustainability of outcomes. The duration of the interventions ranged from 7 days to 6 months, and from 4 to 24 weeks. Follow-up was generally conducted for <6 months, except for BuTifull (web education), which conducted measurements for up to 12 months. The relatively short follow-up period limited the effect on pregnancy incidence. Based on the research findings, all studies reported consistent increases in knowledge and partial behavioral changes (e.g. increased use of protective contraception, but no reduction in sexual activity frequency). Pregnancy incidence was generally not measured, but in BuTifull, there was no significant change. Technology approaches are effective in improving sexual literacy and contraceptive awareness. However, their impact on behavior and pregnancy incidence remains limited. Long-term effects require further evaluation.

Comparative heterogeneity analysis

A comparative analysis of the five intervention bases revealed heterogeneity across study designs, delivery methods, intervention durations, follow-up durations, measured outcomes, and sustainability.

Study design

School-based and community-based interventions exhibited the widest variety of design types, including RCT, cluster RCT, quasi-experimental, and mixed methods. Clinic-based interventions were relatively homogeneous because all studies used an RCT design, whereas family-based interventions were represented by only one design: cluster RCT. Technology-based interventions combined RCT and quasi-experimental designs. This variation affected the final results, with patterns showing that even strong experimental designs do not always yield consistent changes in behavior or pregnancy incidence. The implementation context and the complexity of factors influencing sexual behavior often determine the effect of an intervention more than the methodological design.[20,48]

Delivery methods

Delivery methods were diverse. School-based programs used interactive pedagogy (role-play, discussion, problem-based learning). Community-based programs focused on social norms and technical assistance. Clinic-based programs integrated brief interventions into services. Family-based programs offered workshops with digital boosters. Digital-based programs included short message service, applications, and interactive websites. This pattern shows that interventions with diverse, interactive models consistently increase knowledge, whereas passive digital approaches yield more limited behavioral outcomes. This is supported by research showing that interventions based on direct interaction or relational components have a stronger behavioral effect than single approaches.[49]

Intervention duration

The duration of interventions ranged from 1 session (clinic-based, partially digital) to 2 academic years (school-based). Community-based interventions even included two years of mentoring. Compared with shorter durations, longer durations were more consistent in sustaining knowledge gains, but did not guarantee long-term behavioural change. Very brief interventions were effective at raising awareness but were insufficient to influence complex outcomes such as pregnancy incidence. The relationship between dose and effectiveness is not always linear; the quality of interaction and integration of context is more decisive than the length of the program.[50]

Outcome measurement

Across all intervention settings, knowledge was the most responsive outcome, improving in almost all cases. Behavior showed high heterogeneity, with partial improvement, no significant change, or a decrease in longer-term follow-up. Pregnancy incidence was rarely measured, especially in school-based and clinic-based settings, so direct evidence of pregnancy prevention was limited. Family-based interventions focused on improving parent–child communication and were the only interventions to show a decrease in pregnancy incidence over 12 months. Supporting literature also confirms that changing sexual behavior requires multilevel and long-term interventions, while increasing knowledge is easier to achieve.[19,51]

Follow-up length

From the analysis results, follow-up ranged from 7 days to >30 months. Short follow-ups showed strong positive effects, whereas evaluations ≥12 months often showed a decline in behavioral effects. Community-based studies with long follow-ups showed that pregnancy outcomes could manifest over a longer period. Methodological evidence shows that assessing pregnancy incidence requires at least 12 months of observation to capture meaningful changes.[16,46]

Sustainability

The most sustainable results were observed in the knowledge variable, whereas behavioral changes often did not persist without a reinforcement component. School- and community-based programs reported weakening short-term effects; digital-based programs maintained literacy but did not consistently affect behavior; clinic-based programs were limited to short-term effects; and family-based programs showed sustainability in communication and positive signals in pregnancy incidence. International literature shows that multicomponent interventions with periodic boosters are more likely to maintain behavioral change.[19,26]

Location

The review results revealed a significant imbalance in the geographical distribution of evidence: of the 25 studies, 14 came from the United States (56%), while only 2 studies originated from the United Kingdom, Tanzania, and Thailand each, and 1 study from Ghana, Malaysia, Colombia, Mexico, and Indonesia each. This concentration indicates that the synthesis primarily reflects the context of education systems, health services, and social norms in high-income countries, thereby limiting the external validity for low- to middle-income regions. Because the effectiveness of interventions depends heavily on implementation capacity and local ecology, these findings cannot be interpreted as global conclusions. Furthermore, many studies did not measure pregnancy outcomes, and follow-up durations varied, reinforcing the limitations of generalization.

The geographic concentration of evidence not only limits external validity but also raises equity implications because the reach and effectiveness of interventions for adolescents are heavily determined by structural barriers (access to services, stigma, privacy, and resource limitations), which are more severe among hard-to-reach groups such as out-of-school youth, highly mobile youth, and youth living in remote areas.[52] Under these conditions, the intervention design needs to be low-threshold through multichannel delivery (community and primary services connected to school components, where feasible) to avoid coverage bias that only benefits easily reached groups.[52] For digital components, interventions should provide low-bandwidth/offline options and explicitly equity-oriented implementation approaches so that digital transformation does not widen the access gap.[53] Finally, evaluations need to report stratified reach and uptake according to equity markers (e.g., school status, economic disadvantage, digital access) so that effectiveness is not solely achieved among the populations that are easiest to recruit and retain.[54]

Conclusion

Comparatively, heterogeneity analysis showed that all intervention types are consistently effective in increasing knowledge but vary significantly in promoting behavioural changes, especially in pregnancy incidence. School-based and technology-based interventions excel in improving sexual literacy, but behavioural impacts are often partial or decline in longer follow-ups. Community-based interventions show potential in medium-term outcomes, but behavioural impacts are not always sustained. Clinic-based interventions are effective in increasing short-term awareness and communication, while family-based interventions are the only approach that shows a reduction in pregnancy incidence up to 12 months. Overall, the sustainability of the impact is greatly influenced by the duration, intensity, and presence of reinforcement components.

This study has several limitations. First, varying designs and outcome measurements limit direct comparison between interventions. Second, few studies measured pregnancy incidence, so claims about prevention effectiveness remain limited. Third, different follow-up periods make sustainability estimates inconsistent. Fourth, the review excluded meta-analysis and considered only English-language studies. However, the ecological analysis is a key strength. This approach helps design effective interventions by showing how factors at different ecological levels work together to prevent adolescent pregnancy. Future research should use strong experimental designs with at least 12 months of follow-up, standardized pregnancy incidence indicators, and multicomponent interventions involving schools, families, communities, and technology to increase long-term impact. An integrated ecological perspective will help future studies better understand pregnancy prevention among adolescents in different sociocultural and healthcare settings and guide policymakers as they plan effective, evidence-based programs.

Future research needs to integrate robust experimental designs with follow-up ≥12 months, use consistent pregnancy incidence indicators, and develop multicomponent interventions that combine school, family, community, and technology elements to increase long-term impact. By adopting an integrated ecological perspective, the results of future research are expected to develop a comprehensive understanding of pregnancy prevention among adolescents in diverse sociocultural and health care settings and serve as a reference for policymakers in evaluating and planning effective, evidence-based intervention programs.

Artificial intelligence policy

Artificial intelligence tools were used solely to translate the manuscript into English and to support language proofreading (grammar, clarity, and style), without generating, modifying, or interpreting the study’s scientific content, analyses, or conclusions.

Ethical approval

Ethics approval was not required because this study was based exclusively on published literature and did not involve human participants, identifiable personal information, or primary data collection. Nevertheless, methodological rigor was maintained throughout the review process by applying predefined eligibility criteria, conducting systematic literature searches, using standardized study selection procedures, and critically appraising the methodological quality of the included studies to reduce the risk of bias.

Author’s contributions

PEM conceived and designed the study, developed the review protocol, conducted the literature search, screened studies, extracted and analyzed the data, performed the methodological quality assessment, interpreted the findings, drafted and revised the manuscript, managed correspondence with the journal, and approved the final version. MW, SM, KHAD, WN, MAT, EEFK and VYL contributed to the literature search, study screening, data extraction, methodological quality assessment, data verification, interpretation of findings, and critical review of the manuscript. TK provided methodological guidance, supervised the review process, validated the study selection and data synthesis, critically reviewed the manuscript for important intellectual content, and contributed to manuscript revision. KNS contributed to the study conceptualization, provided overall supervision and scientific oversight, critically revised the manuscript, approved the final version for publication, and supervised the research throughout all stages. All authors read and approved the final manuscript and agreed to be accountable for all aspects of the work.

Data availability statement

All data analyzed in this study were obtained from previously published articles cited in the reference list. No new datasets were generated during this study.

Conflicts of interest

There are no conflicts of interest.

Funding Statement

Indonesia Endowment Fund for Education Agency (Lembaga Pengelola Dana Pendidikan Republik Indonesia).