Skip to main content
NCBI home page
eClinicalMedicine logoLink to eClinicalMedicine
. 2025 Aug 26;88:103453. doi: 10.1016/j.eclinm.2025.103453

Adolescent experiences and exposures: a systematic review of life course outcomes

Prerna Banati a,∗, Maria Vedechkina b, Vanessa De Rubeis a,c, Shiri Kalai d, Lamin Massaquoi a, Ritu Sadana a, Valentina Baltag a, Sarah Baird e
PMCID: PMC12572815  PMID: 41181850

Summary

Background

Adolescence (ages 10–19) is a critical developmental stage that influences outcomes in adulthood and old age. There is growing policy interest in adolescent interventions that are efficiently targeted, demonstrate long-term returns, and impact various domains of well-being. More data is needed to identify which adolescent experiences and exposures can serve as effective levers to shift developmental trajectories and achieve long-term benefits.

Methods

This review examines evidence on the predictive outcomes of adolescent experiences and exposures across the life course. The review is registered in PROSPERO (CRD42023407980). Peer reviewed articles published from March 2010 to July 2025 were included. A framework was adapted to track associations across six core domains of well-being—Agency and Resilience, Connectedness, Family Dynamics, Health and Nutrition, Learning Competence, and Safety—from adolescence through the life course. A systematic search identified relevant studies, followed by a structured descriptive review using the Synthesis Without Meta-Analysis (SWiM) methodology to assess the evidence across domains. Quality was evaluated with an adapted Critical Appraisal Skills Program Checklist.

Findings

Of the 244 included papers, 2231 analyses tested associations between adolescent exposures and life course outcomes, covering data from over 2.1 million adolescents. The Health and Nutrition domain was the most studied domain accounting for 60% of all tested associations, of which 42% were statistically significant. In addition to predicting health outcomes, health in adolescence predicts violence exposure, education, and employment in later life. Safety and Supportive Environments, and Agency and Resilience, though less studied, were also robust predictors of a range of adult outcomes with 35% and 48% of tested associations being statistically significant, respectively. Learning Competence and Safety in adolescence strongly predicts economic success and stability in adulthood. No studies reported on intergenerational effects of exposures during adolescence.

Interpretation

This review uses a systematic approach to demonstrate that adolescent experiences predict adulthood outcomes beyond the original domain of exposure, with significant cross–domain associations. However, it finds that the evidence is unevenly distributed across domains of well-being. In addition, the results shed light on gaps in research on lower-income settings and across demographic variables including sex and socio-economic status.

Funding

This work was supported by Velux Stiftung, Zurich.

Keywords: Adolescent, Life course, Risk and protective factors, Health, Well-being, Development

Introduction

Adolescence, defined here as ages 10–19, is a developmentally sensitive period during which interventions can have significant long-term returns, shaping trajectories into adulthood and beyond.1,2 It is the period of most rapid physical development, second only to infancy.3,4 Social and character development are also central parts of the adolescent experience—the roles of family, peers and communities contribute to values, beliefs and behaviors expressed during adolescence and in later life.1,5,6

Evidence suggests that intervening during adolescence can not only yield immediate benefits but may drive positive changes that endure throughout the life course and across generations.1,5 Central to this is the potential to build on or buffer against early childhood experiences by redirecting behavioral trajectories that send young people on divergent pathways into adulthood.7,8 For example, health-promoting behaviors adopted in adolescence, such as adequate exercise and sleep, or refraining from smoking and binge drinking and are robustly associated with higher levels of healthy behaviors during adulthood9,10 Additionally, experiences in adolescence are profoundly shaped by gender and socioeconomic status (SES) and evidence suggests these determinants play an important role in long term trajectories of health and well-being.1,11

There is growing policy interest in advancing a life course approach to adolescent programming through interventions that demonstrate value for money, are efficiently targeted and have long term returns on well-being.1,12,13 In promoting healthy development, a life course approach aims to amplify the cumulative benefits across individuals' lives and across generations, in a way that responds to emerging health trends, evidence, and societal expectations.14 However, to date, much policy and program development has advanced in the absence of a rigorous assessment of longitudinal relationships across multiple domains of function. To effectively operationalize a life course approach to programming,13,15,16 more information is needed on which individual risk and protective factors, behaviors, experiences, and conditions during adolescence—hereafter collectively referred to as “adolescent exposures”—matter most for lifetime health and well-being. Once identified, these can serve as possible levers for early intervention, ensuring benefits that extend into later stages of life. This systematic review seeks to address two questions.

  • a.

    How do adolescent (ages 10–19) exposures relate to outcomes during later life (ages 20+), and intergenerationally (i.e., of their offspring)?

  • b.

    How do associations differ by age, sex, and socioeconomic status (SES) of adolescents?

To ground our approach in both developmental theory and empirical evidence, we adopted a life course conceptual framework that views adolescence (ages 10–19) as a pivotal stage of growth and transition.1,4 Within this framework, adolescent exposures—ranging from protective resources such as supportive parenting and quality education to risk factors like violence, substance use, or social exclusion—collectively shape adult outcomes through complex and interdependent pathways.1 We distilled these diverse exposures into six overarching domains (Agency & Resilience; Connectedness & Contribution to Society; Health & Nutrition; Learning Competence, Education Skills & Employability; Safety & Supportive Environments; and Family Dynamics & Relationships) based on existing models of adolescent well-being and developmental needs.17,18 By employing a domain-based framework, our review captures the breadth of adolescent experiences while allowing us to identify which domains are most frequently studied and which show the strongest evidence of long term implications—insights essential for informing interventions and policy decisions aimed at optimizing adolescent development and sustained well-being.

Methods

Search strategy and selection criteria

Studies were identified through a literature search of Medline, PsycINFO, Embase, Web of Science and ERIC. The search strings can be found in Annex 1. The search was implemented by a team of three independent reviewers and covered the period between January 2010 and July 2025. Inclusion and exclusion criteria are shown in Table 1. All studies required a case and control group (or equivalent binary classification), with data collected longitudinally and a minimum 5-year interval between exposure and outcome. Exposures must have been experienced during adolescence (10–19 years old), and outcomes had to occur during adulthood (over the age of 20) or intergenerationally. This study relied on the PRISMA checklist guidelines for reporting systematic reviews.20

Table 1.

Inclusion and exclusion criteria.

Inclusion criteria Exclusion criteria
  • •

    Publication between March 2010 and July 2025.

  • •

    Full-text in English language.

  • •

    Must be a case–control design or cohort study with a group comparison measure (i.e., exposure vs. no exposure in adolescence).

  • •

    Longitudinal prospective observational data with at least two time points of data should be available, with a mean follow up at least 5 years later.

  • •

    Baseline measures must be taken between age 10–19 and follow-up measures at age 20 years or more.

  • •

    Since experiences of adversity are typically assessed using retrospective report,7,19 studies using retrospective measures of adversity at baseline were included as long as the measure was taken between the ages of 10–19.

  • •

    Our search was not limited by geographic boundaries.
  • •

    Experimental, diagnostic or intervention studies.

  • •

    Genetics studies.

  • •

    Modelling studies.

  • •

    Cross-sectional studies.

  • •

    Association studies (i.e., studies with no control group or zero exposure comparison).

  • •

    Non-systematic reviews and scoping reviews.

  • •

    Studies that focus on non-demographic subpopulations that limit generalizability (i.e., individuals with a specific health condition).

  • •

    Mediation studies where direct effects are not reported.

  • •

    Case reports, protocols, measurement evaluation or validation studies, modeling studies, genetic studies, systematic reviews or meta analyses.

  • •

    Conference proceedings, dissertations and pre-prints.

  • •

    Retrospective assessments of exposure (except for adversity if assessed between ages 10–19)
Open in a new tab

Study selection

All identified articles were uploaded to Covidence systematic review software (Veritas Health Innovation, Melbourne, Australia. Available at www.covidence.org), which was used for all levels of screening, including duplicate removal, title and abstract and full text screening. A total of 90,902 articles were retrieved during the search, 43,283 of which were identified as duplicates. The remaining studies (n = 47,619) were assessed against the inclusion/exclusion criteria. Given the large initial number of abstracts, abstract level screening was conducted by a single reviewer, after a double screening for 10% of abstracts reached adequate agreement between two reviewers (76% agreement; kappa = 0.61). A total of 45,893 articles were excluded at title/abstract, and 1726 were excluded at the full-text screen (Fig. 1). Two reviewers independently reviewed all articles that were included in the full-text screen (n = 1704) to determine final eligibility. A third member of the team resolved disagreements. Common reasons for exclusion included: studies using inadequate study design (n = 288), data was not disaggregated to confirm exposure during the adolescent age10, 11, 12, 13, 14, 15, 16, 17, 18,20 (n = 349) or insufficient follow up period (<5 years) (n = 141).

Fig. 1.

Fig. 1

Open in a new tab

PRISMA flow chart.

Data extraction, appraisal and synthesis

The following data were extracted from included studies: author, year of publication, publication title, country and region of study, sample size, study objective, sampling characteristics, study design, participant characteristics (age, sex, SES), exposure age, exposure description, outcome age, outcome description, effect measure, effect size, directionality of effect and statistical significance.

Exposures and outcomes were then coded into six domains (Table 2), based on five domains identified by Ross et al. (2020)17: Agency and Resilience; Connectedness and Contribution to Society; Health and Nutrition; Learning Competence Education Skills and Employability; and Safety and Supportive Environments. We added a sixth domain called Family Dynamics and Relationships, as Baltag et al. (2020)18 notes the importance of family formation in the transition to the third decade of life. Our review included studies that track the evolution of a single domain (e.g., adolescent Health and Nutrition and adult Health and Nutrition), and those that track relationships between different domains of function across different life stages (e.g., adolescent Learning, Competence Education Skills and Employability and adult Health and Nutrition). We defined these life stages as1 Adolescence [age 10–19]; and2 Adulthood [age 20+] which includes Young Adults [20–39]; Middle aged adults [40–59] and Older adults [60+].

Table 2.

Domains of well-being with corresponding measures.

Domain Describes domain Examples (Adolescent exposure) Examples (Life course outcome)
Health and nutrition Physical health and capacities; mental health and capacities; optimal nutritional status, diet and exercise
  • •

    Obesi

  • •

    Depression

  • •

    Regular breakfast consumption
  • •

    Alcohol dependence

  • •

    Chronic widespread pain

  • •

    Cardiometabolic risk
Connectedness and contribution to society Valued and respected as part of a community, integrity, morality and character, cultural and civic engagement
  • •

    Activism

  • •

    Volunteering

  • •

    Screen time
  • •

    Sense of well-being

  • •

    Relationship quality

  • •

    Social class
Safety and a supportive environment Physical and emotional safety, built environment, equality and fairness, social and financial protection, non-discrimination
  • •

    Violence

  • •

    Bullying/teasing

  • •

    Foster care
  • •

    Home ownership

  • •

    Attachment with child

  • •

    Discrimination
Learning, competence, education, skills, and employability Motivation for learning, education, skills, competencies, employability, employment, income security and financial protections
  • •

    Education level

  • •

    Education/employment/training status

  • •

    Successful transition to work
  • •

    Education attainment

  • •

    Welfare dependence

  • •

    Employment status
Agency and resilience Identity, purpose, resilience, fulfilment
  • •

    Spirituality

  • •

    Self-awareness

  • •

    Decision-making
  • •

    Life satisfaction

  • •

    Optimism

  • •

    Control over resources

  • •

    Perceived survival expectation
Family dynamics and relationships Marriage, caregiving, parenting and family formation, peer relationships
  • •

    Quality peer relationships

  • •

    Early marriage
  • •

    Sole parenting

  • •

    Childlessness

  • •

    Marital status
Open in a new tab

When multiple analyses were conducted in a single study, each exposure/outcome relationship was extracted as a separate entry. Our data therefore consisted of the total number of estimates across all extracted studies, each of which represents a unique relationship between a given exposure and outcome variable.

Data were extracted by one reviewer and verified by a second reviewer. Exposure and outcome data was grouped by domain and characteristics of the participants at baseline and follow up (e.g., sex, age, SES), and duration between baseline and follow up. Given the diversity of exposures and outcomes among the studies included, we employed a narrative synthesis approach, following recommended guidance for systematic reviews when meta-analysis is infeasible. The heterogeneity in study designs, measures, and follow-up periods was substantial, so pooling effect sizes quantitatively was not appropriate.21 Instead, we synthesized findings using a structured descriptive review based on SWiM (Synthesis WIthout Meta-nalysis) guidelines by identifying patterns and themes across studies while maintaining transparency in how conclusions were drawn.21 This approach enabled us to explore how adolescent exposures relate to later life course outcomes without combining disparate data into a single summary effect. Diverse exposures and outcomes in each domain meant there were a large number of covariates which differed across studies. Where multiple effect estimates were reported, we extracted the most adjusted estimate, to ensure the most robust findings were considered in the analysis. Only 5% of all tested associations did not adjust for any covariates, with most studies adjusting for age, gender and some SES measures when relevant.

A structured descriptive review was employed to analyze patterns and trends across the different lines of evidence.22 This was done by codifying and analyzing data using a ‘best-evidence synthesis’ approach23,24 to describe frequencies of results extracted from the representative sample of published literature. Vote counting methodology was used to assess the directionality of an effect.25

For each study, extensive information was extracted on the population under study and where exclusions were applied. As indicated in our protocol, we did not include studies with specific sub-populations that limit generalizability (i.e., non-demographic sub-populations), for example adolescents living in state care or juvenile detention facilities. Very few longitudinal studies have sample sizes that are nationally representative.

Quality assessment (QA) was conducted using an adapted Critical Appraisal Skills Program Checklist26 with input from the CHARMS Checklist27 across four areas1: data completeness (attrition and exclusion rates, reasons for attrition/exclusions reported)2; study power (sample size reported and/or justified, significance levels and/or confidence intervals reported)3; model development (adequate methods, modeling assumptions, selection of predictors, validation/adjustment, model performance described, and included covariates); and4 other sources of bias (any additional concerns not otherwise addressed). Papers were coded as either high quality (most information provided) or poor quality (most information not provided).

This systematic review was registered with the International Prospective Register of Systematic Reviews (PROSPERO), under the identifier CRD42023407980, prior to conducting the literature search.

Role of funding sources

WHO acknowledges financial support from Velux Stiftung, Zurich, including support to conduct research and advance metrics and evidence on healthy ageing including across all stages of the life course.

Results

State of the evidence and gaps in the literature

The full-text screening generated 244 articles that met all inclusion criteria, capturing a total of 2231 estimates testing relationships between adolescent exposures and life course outcomes. The articles represent data from over 2.1 million adolescents, and over 16 million unique adolescent data points with an average follow up period of 20 years (and range of 5–83 years of follow up, follow up period can be from earlier in childhood). The mean age of adolescent participants was 14.9 years old with a range of 10–19, and the mean age at follow up was 30.6 years, with a range of 21–85. Over 45% (n = 108) of studies recruited adolescents in schools, 20% (n = 42) were based on national registries, 14% (n = 31) recruited from clinics or hospitals, and the remainder of studies recruited from households, institutions, or other venues.

Few studies originated outside of high-income settings. Approximately 94% (225 studies) were based on populations in the US, Europe, Australia or New Zealand. A total of 8 studies were conducted on populations based in the Americas outside of the US, and 6 were based on populations residing in Asia and Middle East. The remaining 4 studies operated in multi-country settings. Over 72% (n = 170) studies did not disaggregate analysis by age. Approximately one-fifth (22%) of studies disaggregated analysis by gender. The majority of included studies were of high quality (Table 3).

Table 3.

Quality assessment of extracted articles (n = 244).

Assessment Description Standards met (high quality)
Data completeness Assesses if attrition and exclusions were reported, reasons for attrition/exclusions identified, sensitivity analysis of missing data conducted where required 161 (66%)
Study power Assess if sample size calculations were reported and justified, assesses significance levels and/or confidence intervals were reported 214 (88%)
Model development Appropriateness of model evaluated, assumptions checked, sensitivity analysis to assess robustness of estimates 174 (71%)
Open in a new tab

The identified literature was unevenly distributed across domains of well-being (Fig. 2). Health and Nutrition was the most studied domain of exposure in adolescence, with a total of 1343 tested associations (representing 60% of total tested associations) across 121 individual studies (Table 4). Safety and a Supportive Environment was the second most studied domain of exposure, with 338 (15%) total associations tested, followed by Connectedness and Contribution to Society (n = 202, 9%), Learning Competence, Education Skills and Employability (n = 161, 7%), Family Dynamics and Relationships (n = 137, 6%), and Agency and Resilience (n = 50, 2%). Notably, Agency and Resilience was not studied in relation to Agency and Resilience in adulthood, but rather as a predictor of health, safety, and employability outcomes in adulthood.

Fig. 2.

Fig. 2

Open in a new tab

Graphical representation of relationships between adolescent predictors and life course outcomes. Note: Dark colour and ∗% indicates significant findings, and lighter colour is nonsignificant at p ≤ 0.05. Thickness of the lines represents the proportion of the total number of associations tested across all included studies.

Table 4.

Number of tested associations identified between adolescent and life course well-being domains (percentage significant at p < 0.05 in brackets).

Adolescent predictor domains
Agency & Resilience Connectedness & Contribution to society Family dynamics & relationships Health & Nutrition Learning competence education skills & Employability Safety & Supportive environment Total
Life course outcome domains
 Agency & Resilience 0 3 (33%) 2 (0%) 31 (32%) 4 (0%) 9 (22%) 49 (27%)
 Connectedness & Contribution to society 0 9 (33%) 4 (100%) 38 (39%) 15 (640%) 22 (5%) 88 (33%)
 Family dynamics & relationships 0 12 (33%) 10 (20%) 32 (28%) 3 (0%) 13 (38%) 70 (29%)
 Health & Nutrition 20 (35%) 120 (17%) 82 (37%) 1061 (40%) 106 (21%) 222 (34%) 1611 (36%)
 Learning competence education skills & Employability 20 (55%) 43 (44%) 33 (55%) 115 (60%) 27 (67%) 29 (66%) 267 (58%)
 Safety & Supportive environment 10 (60%) 15 (87%) 6 (33%) 66 (53%) 6 (83%) 43 (35%) 146 (52%)
 Total 50 (48%) 202 (30%) 137 (41%) 1343 (42%) 161 (32%) 338 (35%) 2231 (39%)
Open in a new tab

In terms of adult outcomes, the bulk of the tested associations focused on Health and Nutrition outcomes (n = 1611, 72%), followed by Learning Competence, Education Skills and Employability (n = 267, 12%), Safety and a Supportive Environment (n = 146, 7%), Connectedness and Contribution to Society (n = 88, 4%), Family Dynamics and Relationships (n = 70, 3%), and Agency and Resilience (n = 49, 2%).

Adolescent predictors of life course outcomes

Of the 2231 associations tested across the included studies, 870 (38%) were statistically significant at p < 0.05 (Table 4; Fig. 2). In the Health and Nutrition domain, 42% (n = 564) of tested associations were statistically significant. In addition to predicting health outcomes in adulthood, there is strong evidence that health status in adolescence predicts employment, violence exposure and perpetration, and education in later life (Table 5).

Table 5.

Number of tested associations between adolescent ‘Health and Nutrition’ domain and life course outcomes (percentage significant at p < 0.05 in brackets).

Adolescent health and nutrition predictors
Physical health Mental health BMI Physical activity Nutrition Substance use Screen use Sleep hygiene Sexual health Oral health Total
Life course outcomes
 Physical health 42 (33%) 20 (55%) 19 (36%) 56 (77%) 57 (81%) 72 (54%) 34 (82%) 18 (89%) 317
 Mental health 3 (33%) 98 (29%) 2 (100%) 84 (74%) 14 (71%) 25 (40%) 10 (60%) 1 (0%) 1 (0%) 248
 BMI 6 (67%) 36 (61%) 19 (5%) 12 (25%) 19 (68%) 2 (50%) 5 (80%) 99
 Physical activity 4 (100%) 5 (100%) 17 (76%) 2 (100%) 28
 Nutrition 1 (100%) 3 (100%) 3 (100%) 58 (67%) 65
 Substance use 66 (74%) 2 (50%) 22 (50%) 145 (39%) 3 (33%) 1 (100%) 2 (0%) 241
 Screen use 1 (100%) 1 (100%) 6 (100%) 1 (0%) 9
 Cognitive health 1 (100%) 2 (20%) 9 (44%) 12
 Healthy ageing 1 (100%) 1 (0%) 1 (0%) 3 (67%) 6
 Sleep health 1 (0%) 8 (0%) 8 (75%) 1 (0%) 4 (100%) 19 (37%) 41
 Mortality 2 (0%) 3 (100%) 4 (100%) 9
 Social engagement 2 (0%) 15 (87%) 4 (75%) 3 (0%) 24
 Family relationships 1 (0%) 8 (100%) 15 (53%) 3 (100%) 27
 Education 25 (56%) 12 (42%) 10 (30%) 6 (0%) 21 (76%) 1 (0%) 4 (0%) 81
 Employment and Income 12 (83%) 34 (76%) 28 (64%) 6 (67%) 13 (54%) 93
 Violence 33 (33%) 9 (0%) 42
 Health costs 9 (56%)
 Total 53 (49%) 327 (54%) 104 (51%) 206 (69%) 150 (73%) 280 (49%) 51 (76%) 46 (61%) 3 (33%) 6 (0%) 1343
Open in a new tab

To better understand these relationships, directionality analysis was conducted. Positive associations were defined as instances where an adverse exposure significantly predicts an adverse outcome, or a favorable exposure predicts a favorable outcome, while negative associations indicate the inverse. This analysis revealed that adolescent health has predominantly positive significant associations across domains: 65 positive and 4 negative associations for Learning, Education Skills, and Employability; 35 positive associations for Safety and a Supportive Environment; 15 positive associations for Connectedness and Contribution to Society; 8 positive and 2 negative associations for Agency and Resilience; and 9 positive associations for Family Dynamics and Relationships. These findings underscore the broad and largely beneficial impact of adolescent health on various aspects of life in adulthood.

In the Safety and a Supportive Environment domain, 35% (n = 112) were significant. Connectedness and Contribution to Society had 30% (n = 60) significant associations, while Learning, Education Skills, and Employability had 32% (n = 54), and Family Dynamics and Relationships had 41% (n = 56). Although Agency and Resilience was a relatively understudied domain of exposure in adolescence, it had the largest proportion of statistically significant findings (48%; n = 24). This indicates that despite being less frequently studied in the life course literature, agency and resilience in adolescence is a robust predictor of health, employability, and safety in adulthood.

The proportion of significant findings also varied across life course outcome domains (Fig. 2; Table 4). In the Health and Nutrition domain, 36% (n = 580) of associations were significant, and for Learning Competence, Education Skills, and Employability, 58% (n = 154) were significant. Safety and a Supportive Environment had 52% (n = 76) significant associations, while Connectedness and Contribution to Society had 33% (n = 29), Family Dynamics and Relationships had 29% (n = 20), and Agency and Resilience had 27% (n = 13). Notably, Health and Nutrition; Learning, Education Skills, and Employability and Safety and a Supportive Environment in adulthood are significantly predicted by all six adolescent domains of exposure (Fig. 2; Table 4). While several studies focused on predicting later life fertility and fecundability, none of these directly tested intergenerational outcomes (defined as outcomes taking place in the children of participants).

For illustration, further analysis was conducted in two exposure areas of contemporary interest—High BMI and Physical Activity—and where outcome measures in the studies lent themselves to comparative analysis. This analysis is not intended to replace a meta-analysis, but rather, provides further insight on clusters of outcomes, age sequencing, and possible pathways for future interrogation.

High adolescent BMI, overweight or obesity predicts negative health and social consequences in later life

When drilling down on adolescent BMI, 113 estimates of BMI, overweight or obese status in adolescence were identified, of which 51% were significant (Table 5). Of these, 23 associations (6 studies) were observed for high BMI (Fig. 3). Overweight or obesity have implications for adult outcomes in young adulthood (20–39 years) and mid-adulthood (40–59 years). High adolescent BMI is significantly associated with adult outcomes at both individual (obesity, smoking, cardiovascular health) and family (marriage, earnings) socio-ecological levels. Later life significant associations were seen in Health and Nutrition, Family Dynamics and Relationships, and Learning, Competence, Education, Skills and Employability domains. Missing areas for investigation include associations of adolescent BMI with adult mental health, social connectedness, and measures of agency and resilience.

Fig. 3.

Fig. 3

Open in a new tab

Associations between adolescent high BMI, overweight or obesity and life course outcomes (brackets indicate number of associations). ∗reference is normal BMI or normal weight.

Adolescent physical activity predicts positive health and social function in later life

Adolescent physical activity was captured in 226 associations, of which 68% were significant (see Table 5). Of these, 19 (10 studies) were observed for physical activity that meets recommended or higher levels. Most outcomes were measured in the 20–39 age range, and included obesity, a diverse set of mental health outcomes, substance use, and education. Associations with health and mental health outcomes are also seen in mid-adulthood (40–59) (Fig. 4). This evidence suggests that adolescent physical activity has implications for physical and mental health in later life, though apart from education and employment, no evidence was available on its relationship to other domains of well-being.

Fig. 4.

Fig. 4

Open in a new tab

Associations between adolescent physical activity and life course outcomes (brackets indicate number of associations). ∗reference is low or no physical activity.

Discussion

The goal of this review was to assess the evidence for the relationship between a wide set of adolescent risk and protective factors, behaviors, experiences, and conditions and life course outcomes using a systematic approach. To our knowledge, this review is the first to systematically map and synthesize predictive relationships between different domains of function from adolescence into adulthood.

We have presented a description of adolescent exposures for which evidence substantiates a longitudinal association with later life outcomes. This analysis brings together evidence from diverse disciplinary backgrounds, including psychology, economics, education and health research. Despite this, we find that the evidence is unevenly distributed across domains of well-being: Health and Nutrition in adolescence is the most frequently studied predictor of lifelong outcomes, while predictors in other domains are fewer. Predictors in the domain of Agency and Resilience, which include self-efficacy, goal setting, decision making, autonomy, adaptability, problem solving and optimism in adolescence are relatively understudied, perhaps reflecting the absence of validated measures and the infancy of quantitative estimation literature in the field.28 Additionally, the outcomes of Family Dynamics and Relationships appear to be understudied, despite wide consensus of the usefulness of parenting programs in adolescence.13 Further research focusing on intergenerational outcomes and the quality of family interactions could uncover the role of supportive home environments in adolescence shaping adult relationships and caregiving practices.

Our findings substantiate the notion that adolescence is a developmentally sensitive period in which interventions can exert long-term influences on adult health and well-being.1 From a life course perspective, these results align with conceptual models suggesting that early life experiences “accumulate” or “cascade” to shape future trajectories.2 Specifically, we found that while the Health and Nutrition domain was the most frequently studied, the Agency and Resilience domain—which includes constructs like self-efficacy, goal setting, and optimism—showed the most robust predictive strength, despite being underrepresented in the literature. For instance, only 42% of the analyses in the domain of Health and Nutrition significantly predicted later life outcomes, including health, violence exposure, education and employment. This discrepancy highlights a crucial insight: the quantity of research in a domain does not necessarily equate to the quality or strength of the relationship with future outcomes.

In contrast, although relatively understudied, Agency & Resilience in adolescence was a robust predictor of adulthood outcomes that may warrant further research focus. This observation reflects theories within the positive youth development (PYD) framework, emphasizing that fostering intrinsic assets (e.g., autonomy, problem-solving) in adolescence can produce durable benefits extending into adulthood.29 Our findings thus support the hypothesis that strengthening internal competencies in adolescence has ripple effects on later health, economic, and social outcomes.30 Given the limited but promising evidence, researchers should prioritize developing standardized, validated instruments to measure constructs such as self-efficacy and optimism across diverse cultures, investigating how these factors buffer against adverse outcomes.

Despite the strong conceptual rationale for linking adolescent exposures to adult outcomes, the evidence across multiple domains proved inconsistent. For instance, Safety and a Supportive Environment and Connectedness and Contribution to Society were extensively studied yet had relatively modest proportions of significant associations (36% and 30% of significant findings, respectively). This may indicate either genuine variability across contexts or methodological heterogeneity in study design and measurement approaches, including inconsistencies in definitions of “support” or “connectedness.” Such measurement challenges are echoed in other systematic reviews, which have similarly noted that constructs related to social environments are highly sensitive to local definitions and cultural norms.31,32 To reduce these inconsistencies and enhance cross-study comparability, researchers and funding agencies could undertake structured research-priority exercises—similar to the Child Health and Nutrition Research Initiative (CHNRI) approach33—to identify common well-defined indicators for critical exposures and outcomes, standardized measurement tools, and explicit research questions. By aligning study designs and measurement methods across multiple investigations, future work can more reliably compare findings, clarify causal pathways, and ultimately inform evidence-based interventions targeted at adolescent exposures with the greatest potential impact on later-life outcomes. Indeed, more research on specific components of these domains—such as neighborhood cohesion, civic engagement, peer networks, school climates, or community resources—may provide more consistent evidence on how these factors can mitigate the long-term effects of adverse events in adolescence.

The proportion of significant findings also varied across adult outcome domains, providing further insights into which later-life outcomes may be most closely tied to experiences and exposures in adolescence. While Family Dynamics and Connectedness and Contribution to Society (being active, socially engaged) are considered important in older age,34 there seems to be little interest in this domain as a target for longitudinal research in adulthood, warranting further investigation within the context of population ageing and longevity. Our review also underscores the importance of adolescent exposures for later life Learning Competence, Education Skills, and Employability (54% significant) and Safety and a Supportive Environment (52% significant), highlighting adolescence as a critical period for shaping economic success and stability in adulthood. This aligns with a robust body of research demonstrating the role of early experiences in shaping long-term educational and professional outcomes.35,36

Beyond these domain-specific insights, our work underscores the need for a holistic approach that integrates multiple dimensions of adolescent development. Notably, our findings add to the existing literature by demonstrating that socio-economic and health-related outcomes in adulthood are significantly predicted by exposures across all six adolescent domains, beyond early education and environmental factors alone. This highlights the interconnectedness and cascading potential of adolescent experiences to shift developmental trajectories across a range of functional outcomes into adulthood. Policies and program interventions during adolescence which consider multiple domains of well-being are likely to be most effective in fostering positive long-term health and socio-economic outcomes. It also further underscores the importance of linking multiple developmental stages—spanning infancy, childhood, adolescence, and youth—to adult well-being and healthy aging.14 While our frequency-based approach and use of domain-level data should be interpreted with caution— as it does not allow for precise effect estimation or a thorough investigation of potential confounders and effect modifiers—the findings from this review may serve as a foundation for more targeted, detailed future analyses.

Despite the strength of some domains in predicting life course outcomes, our review highlights notable gaps and biases in the literature. While many studies were of high quality, a sizeable proportion (over a third) lacked adequate assessment of data completeness. This is a critical consideration in the analysis of longitudinal studies where differential attrition can weaken study findings. Similarly, approximately a third of studies lacked appropriate descriptions of model development (choice of the model, underlying assumptions, covariates included, and sensitivity analyses). A sizeable portion of studies did not disaggregate data further by age (72%) or gender (78%) obscuring potential differences in longitudinal relationships across these important demographic variables.

Furthermore, an overwhelming majority of studies were conducted in high-income settings, particularly the US, Europe, Australia, and New Zealand. Only a small fraction of studies (7%) focused on populations from lower-income regions. Conducting longitudinal research in low- and middle-income country (LMIC) settings often requires substantial resources, both financially and in terms of infrastructure for sustained follow-up.37 Consequently, many of the high-quality longitudinal data sets originate from high-income contexts with more robust funding mechanisms. The relative absence of research studies accounting for geographic variations and other important determinants limits the specificity of findings and hampers our ability to strengthen interventions that reach underrepresented groups. Indeed, trajectories for adolescents depend strongly on structural determinants4 such as poverty.

This study highlights the need for more comprehensive longitudinal research in a range of global contexts, where socioeconomic and infrastructural constraints may uniquely shape adolescent trajectories. Additionally, further work should systematically disaggregate findings by age (e.g., early vs. late adolescence), sex/gender, and other potentially important factors such as race, ethnicity, and disability status. Such subgroup analyses are essential for identifying vulnerable populations and tailoring interventions that account for the interplay between social determinants and adolescent development.1,4 By broadening geographic coverage and demographic detail, future research can offer a more robust, nuanced understanding of how adolescent exposures influence adult outcomes across varied social and economic settings.

Given the breadth of our review and the large initial number of abstracts, abstract level screening had adequate, but not perfect inter-rater reliability.38 Our analysis did not account for cluster effects when multiple outcomes were drawn from a single study. The focus on statistical significance (p < 0.05), rather than clinical significance, may limit the applicability of findings to public health interventions. Relatedly, although this review identifies the breadth of statistically significant associations at the domain level, we do not systematically infer directionality which can be difficult to parse when exposures and outcomes vary in valence (e.g., comparing high BMI to healthy eating), and when conceptual distinctions between ‘positive’ and ‘negative’ factors are not always clear. A more granular analysis at the individual exposure–outcome level would be required to fully capture directional nuances—a valuable endeavor for future research seeking to disentangle the complex pathways linking adolescent risk/protective factors to adult outcomes. Finally, we cannot make claims of causal inference and were unable to shed light on chronicity, severity and timing of exposures, highlighting the need for more detailed longitudinal data, including intergenerational studies.

This study substantiates the importance of optimizing trajectories, the value of developmental and longitudinally focused study, and the role of life-course models of analysis. Adolescent experiences shape adult outcomes beyond their original domain of exposure, revealing the cross-domain influence adolescent factors exert on future health, well-being, and socio-economic success. These findings have important implications for the development of meaningful targets for intervention that address adolescent risk and protective factors across the life course. However, more comprehensive and context-specific research that considers diverse populations and demographic variables is essential. In addition, this review advocates for a greater focus on understudied domains such as Agency and Resilience, which appear to be powerful predictors of adult outcomes despite the limited attention they have received to date. By addressing the identified gaps and expanding the scope of study, researchers can better understand the adolescent factors that underlie and promote positive youth development trajectories, and lay the groundwork for healthier, more productive lives.

Contributors

PB and VB conceived and designed the study. MV, VDR, SK, LM and PB collected and curated the data. MV, PB and VDR conducted the analyses with input from LM and SK. PB drafted the initial manuscript, and VB, RS, and SB provided critical revisions. All authors reviewed and approved the final version of the manuscript. PB and MV had full access to all the data and had final responsibility for the decision to submit for publication.

Data sharing statement

No new data was generated during this study. All data extracted for the umbrella review is made available in detail in Appendix 1. Any further details are contained in the original publications, the references for which are contained either in the paper or in Appendix 1. The protocol for the study has been published and can be found at International Prospective Register of Systematic Reviews (PROSPERO), under the identifier CRD42023407980.

Declaration of interests

The authors affirm that they have no conflicts relevant to this manuscript.

Acknowledgements

PB, VB and RS are members of the World Health Organization. The authors alone are responsible for the views expressed in this publication and they do not necessarily represent the decisions, policy, or views of the World Health Organization.

Footnotes

Appendix A

Supplementary data related to this article can be found at https://doi.org/10.1016/j.eclinm.2025.103453.

Appendix A. Supplementary data

Supplement Table 1
mmc1.xlsx (193.5KB, xlsx)
Web Annex 1
mmc2.docx (16.6KB, docx)

References

  • 1.Patton G.C., Sawyer S.M., Santelli J.S., et al. Our future: a lancet commission on adolescent health and wellbeing. Lancet. 2016;387(10036):2423–2478. doi: 10.1016/S0140-6736(16)00579-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.Masten A.S., Cicchetti D. Developmental cascades [Editorial] Dev Psychopathol. 2010;22(3):491–495. doi: 10.1017/S0954579410000222. [DOI] [PubMed] [Google Scholar]
  • 3.Dahl R., Allen N., Wilbrecht L., Suleiman A.B. Importance of investing in adolescence from a developmental science perspective. Nature. 2018;554:441–450. doi: 10.1038/nature25770. [DOI] [PubMed] [Google Scholar]
  • 4.Viner R.M., Ozer E.M., Denny S., et al. Adolescence and the social determinants of health. Lancet. 2012;379(9826):1641–1652. doi: 10.1016/S0140-6736(12)60149-4. [DOI] [PubMed] [Google Scholar]
  • 5.Banati P., Lansford J.E. Oxford University Press; 2018. Handbook of adolescent development research and its impact on global policy. [Google Scholar]
  • 6.Lerner R.M. In: Handbook of child psychology: volume I. Theoretical models of human development. Damon W., Lerner R.M., editors. John Wiley & Sons; 2006. Developmental science, developmental systems, and contemporary theories of human development. [Google Scholar]
  • 7.Johnson M.K., Crosnoe R., Elder G.H., Jr. Insights on adolescence from a life course perspective. J Res Adolesc. 2011;21(1):273–280. doi: 10.1111/j.1532-7795.2010.00728.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Dragastin S., Elder G.H., Jr. Hemisphere; Washington, DC: 1975. Adolescence in the life cycle. [Google Scholar]
  • 9.Baltag, Sidaner E., Bundy D., et al. Realising the potential of schools to improve adolescent nutrition. BMJ. 2022;379 doi: 10.1136/bmj-2021-067678. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Neufeld L.M., Andrade E.B., Ballonoff Suleiman A., et al. Food choice in transition: adolescent autonomy, agency, and the food environment. Lancet. 2022;399:97. doi: 10.1016/S0140-6736(21)01687-1. [DOI] [PubMed] [Google Scholar]
  • 11.Cyranowski J.M., Frank E., Young E., Shear M.K. Adolescent onset of the gender difference in lifetime rates of major depression: a theoretical model. Arch Gen Psychiatry. 2000;57(1):21–27. doi: 10.1001/archpsyc.57.1.21. [DOI] [PubMed] [Google Scholar]
  • 12.Banati P. In: Developmental science and sustainable development goals for children and youth. Verma S., Petersen A.C., editors. Springer; 2018. Bringing life course theory to the sustainable development goals; pp. 313–328. [Google Scholar]
  • 13.WHO The life course approach: from theory to practice. 2021. https://www.who.int/europe/publications/i/item/9789289053266 Downloaded from.
  • 14.WHO . 2024. WHO framework on operationalizing a life course approach. Forthcoming. [Google Scholar]
  • 15.WHO (2020) World Health Organization; Geneva: 2020. Decade of healthy ageing: Baseline report.https://www.who.int/publications/i/item/9789240017900 [Google Scholar]
  • 16.UNICEF . 2018. Life course.https://unicef-irc.org/research/life-course/#:∼:text=A%20better%20understanding%20of%20how,and%20effectiveness%20of%20development%20programming Downloaded from. [Google Scholar]
  • 17.Ross D.A., Hinton R., Melles-Brewer M., et al. Adolescent well-being: a definition and conceptual framework. J Adolesc Health. 2020;67:472. doi: 10.1016/j.jadohealth.2020.06.042. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18.Baltag V., Guthold R., Sawyer S.M. 2020. Global health in young adulthood: risk factor burden, vulnerabilities, and health determinants. Handbook of Global Health; pp. 1043–1087. [Google Scholar]
  • 19.Felitti V.J., Anda R.F., Nordenberg D., et al. Relationship of childhood abuse and household dysfunction to many of the leading causes of death in adults. The adverse childhood experiences (ACE) study. Am J Prev Med. 1998;14(4):245–258. doi: 10.1016/s0749-3797(98)00017-8. [DOI] [PubMed] [Google Scholar]
  • 20.Page M.J., McKenzie J.E., Bossuyt P.M., et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372 doi: 10.1136/bmj.n71. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Campbell M., McKenzie J.E., Sowden A., et al. Synthesis without meta-analysis (SWiM) in systematic reviews: reporting guideline. BMJ. 2020;368 doi: 10.1136/bmj.l6890. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Paré G., Trudel M.C., Jaana M., Kitsiou S. Synthesizing information systems knowledge: a typology of literature reviews. Inf Manage. 2015;52(2):183–199. [Google Scholar]
  • 23.Slavin R.E. Best evidence synthesis: an intelligent alternative to meta-analysis. J Clin Epidemiol. 1995;48(1):9–18. doi: 10.1016/0895-4356(94)00097-a. [DOI] [PubMed] [Google Scholar]
  • 24.Green B.N., Johnson C.D., Adams A. Writing narrative literature reviews for peer-reviewed journals: secrets of the trade. J chiropractic Med. 2006;5(3):101–117. doi: 10.1016/S0899-3467(07)60142-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.McKenzie J.E., Brennan S.E. Report to the Australian National Health and Medical Research Council; 2014. Complex reviews: methods and considerations for summarising and synthesising results in systematic reviews with complexity. [Google Scholar]
  • 26.Critical Appraisal Skills Programme CASP case control and cohort checklists. https://casp-uk.net/casp-tools-checklists/ Available at:
  • 27.Moons K.G., de Groot J.A., Bouwmeester W., et al. Critical appraisal and data extraction for systematic reviews of prediction modelling studies: the CHARMS checklist. PLoS Med. 2014;11(10) doi: 10.1371/journal.pmed.1001744. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Windle G., Bennett K.M., Noyes J. A methodological review of resilience measurement scales. Health Qual Life Outcomes. 2011;9:8. doi: 10.1186/1477-7525-9-8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Lerner R.M., Lerner J.V., Bowers E.P., Geldhof G.J. In: Handbook of child psychology and developmental science. 7th ed. Vol. 1. Lerner R.M., editor. Vol. 1. Wiley; New York: 2015. Positive youth development and relational developmental systems; pp. 1–45. [Google Scholar]
  • 30.Masten A.S. Global perspectives on resilience in children and youth. Child Dev. 2014;85(1):6–20. doi: 10.1111/cdev.12205. [DOI] [PubMed] [Google Scholar]
  • 31.Cordier R., Milbourn B., Martin R., Buchanan A., Chung D., Speyer R. A systematic review evaluating the psychometric properties of measures of social inclusion. PLoS One. 2017;12(6) doi: 10.1371/journal.pone.0179109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.Plackett R., Hulin J., Mukuria C., et al. Measures of social connectedness in adult populations: a systematic review. BMC Public Health. 2024;24:3384. doi: 10.1186/s12889-024-20779-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Rudan I., Gibson J.L., Ameratunga S., et al. Setting priorities in global child health research investments: guidelines for implementation of CHNRI method. Croat Med J. 2008;49(6):720–733. doi: 10.3325/cmj.2008.49.720. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.WHO (2024b) WHO commission on social connection. https://www.who.int/news/item/15-11-2023-who-launches-commission-to-foster-social-connection
  • 35.Bakken L., Brown N., Downing B. Early childhood education: the long-term benefits. J Res Child Educ. 2017;31(2):255–269. [Google Scholar]
  • 36.Chetty R., Friedman J.N., Hilger N., Saez E., Schanzenbach D.W., Yagan D. How does your kindergarten classroom affect your earnings? Evidence from project STAR. Q J Econ. 2011;126(4):1593–1660. doi: 10.1093/qje/qjr041. [DOI] [PubMed] [Google Scholar]
  • 37.Imam A., Wariri O., Dibbasey T., et al. Conducting clinical research in a resource-constrained setting: lessons from a longitudinal cohort study in the Gambia. BMJ Glob Health. 2021;6(8) doi: 10.1136/bmjgh-2021-006419. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Belur J., Tompson L., Thornton A., Simon M. Interrater reliability in systematic review methodology: exploring variation in coder decision-making. Sociol Methods Res. 2021;50(2):837–865. [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplement Table 1
mmc1.xlsx (193.5KB, xlsx)
Web Annex 1
mmc2.docx (16.6KB, docx)

Articles from eClinicalMedicine are provided here courtesy of Elsevier

RESOURCES