Teenage pregnancy and social disadvantage: systematic review integrating controlled trials and qualitative studiesBMJ 2009; 339 doi: http://dx.doi.org/10.1136/bmj.b4254 (Published 13 November 2009) Cite this as: BMJ 2009;339:b4254
- Angela Harden, professor of community and family health1,
- Ginny Brunton, research officer2,
- Adam Fletcher, lecturer in young people’s health3,
- Ann Oakley, professor of sociology and social policy2
- 1Institute of Health and Human Development, University of East London, London, E15 4LZ
- 2Social Science Research Unit, Institute of Education, University of London, London WC1H 0NR
- 3Department of Public Health & Policy, London School of Hygiene & Tropical Medicine, London, WC1E 7H
- Correspondence to: A Harden
- Accepted 12 July 2009
Objectives To determine the impact on teenage pregnancy of interventions that address the social disadvantage associated with early parenthood and to assess the appropriateness of such interventions for young people in the United Kingdom.
Design Systematic review, including a statistical meta-analysis of controlled trials on interventions for early parenthood and a thematic synthesis of qualitative studies that investigated the views on early parenthood of young people living in the UK.
Data sources 12 electronic bibliographic databases, five key journals, reference lists of relevant studies, study authors, and experts in the field.
Review methods Two independent reviewers assessed the methodological quality of studies and abstracted data.
Results Ten controlled trials and five qualitative studies were included. Controlled trials evaluated either early childhood interventions or youth development programmes. The overall pooled effect size showed that teenage pregnancy rates were 39% lower among individuals receiving an intervention than in those receiving standard practice or no intervention (relative risk 0.61; 95% confidence interval 0.48 to 0.77). Three main themes associated with early parenthood emerged from the qualitative studies: dislike of school; poor material circumstances and unhappy childhood; and low expectations for the future. Comparison of these factors related to teenage pregnancy with the content of the programmes used in the controlled trials indicated that both early childhood interventions and youth development programmes are appropriate strategies for reducing unintended teenage pregnancies. The programmes aim to promote engagement with school through learning support, ameliorate unhappy childhood through guidance and social support, and raise aspirations through career development and work experience. However, none of these approaches directly tackles all the societal, community, and family level factors that influence young people’s routes to early parenthood.
Conclusions A small but reliable evidence base supports the effectiveness and appropriateness of early childhood interventions and youth development programmes for reducing unintended teenage pregnancy. Combining the findings from both controlled trials and qualitative studies provides a strong evidence base for informing effective public policy.
Countries such as the United Kingdom and the United States have high teenage pregnancy rates relative to other countries.1 2 3 Although teenage pregnancy can be a positive experience, particularly in the later teenage years,4 5 it is associated with a wide range of subsequent adverse health and social outcomes.6 7 These associations remain after adjusting for pre-existing social, economic, and health problems.8 Despite the establishment of a national teenage pregnancy strategy in 1999,9 teenage birth rates in the UK are the highest in western Europe10 and conceptions among girls under 16 years of age in England and Wales have increased since 2006.11
Recent research evidence shows that traditional approaches to reducing teenage pregnancy rates—such as sex education and better sexual health services—are not effective on their own.12 13 This evidence has generated increased interest in the effects of interventions that target the social disadvantage associated with early pregnancy and parenthood.14 15 16 17 18 19 Social disadvantage refers to a range of social and economic difficulties an individual can face—such as unemployment, poverty, and discrimination—and is distributed unequally on the basis of sociodemographic characteristics such as ethnicity, socioeconomic position, educational level, and place of residence.20 21
The objectives of this study were to determine on the basis of evidence in qualitative and quantitative research the impact on teenage conceptions of interventions that address the social disadvantage associated with early parenthood and to assess the appropriateness of such interventions for young people in the UK.
We undertook a three part systematic review of the research evidence on social disadvantage and pregnancy in young people by using an innovative method we developed previously for integrating qualitative and quantitative research.22 23 The first part of the review focused on quantitative controlled trials and was designed to assess the impact on teenage conceptions of interventions that address the social determinants of teenage pregnancy. The second part focused on qualitative research and examined intervention need and appropriateness on the basis of the perspectives and experiences of young people. In the third part of the review, we integrated the two sets of findings to assess the extent to which existing evaluated interventions do in fact address the social disadvantage associated with early pregnancy and parenthood as determined by the needs and concerns of young people.
The inclusion of qualitative research in systematic reviews facilitates the incorporation of “real life” experiences into evidence based policy making.24 An ability to unpack the worldview of participants at a particular time and location has been highlighted as a key strength of qualitative research.25 26 Although we included trials conducted in any country, we drew only on qualitative studies conducted in the UK to help assess the applicability of interventions to reduce teenage pregnancy within this country in particular.
Our literature searches covered seven major databases and five specialist registers (table 1⇓). Highly sensitive topic based search strategies were designed for each database. We did not use study type search filters and identified controlled trials and qualitative studies using the same strategy.
We included randomised and non-randomised controlled trials that evaluated interventions designed to target social disadvantage and that reported teenage conceptions or births as an outcome measure. The inclusion of trials was not restricted according to language, publication date, or country. We included any qualitative study published between 1994 and 2004 that focused on teenage pregnancy and social disadvantage among young people aged less than 20 years old living in the UK.
Relevant interventions were those that aimed to improve young people’s life opportunities and financial circumstances; for example, through educational or income support. Relevant interventions could be targeted at children, young people, or their families. Controlled trials of sex education or sexual health services and qualitative studies focusing solely on attitudes to and knowledge of sexual health or sex education were excluded.
We hand searched American Journal of Public Health (from January 1999 to January 2004), Journal of Adolescent Health (from January 1999 to February 2004), Journal of Adolescence (from February 1999 to April 2004), and Perspectives on Sexual and Reproductive Health (from issue 1, 1999, to issue 1, 2004). We also reviewed the reference lists of all studies that met our inclusion criteria and contacted experts in the field who suggested further studies to pursue.
We assessed the extent to which controlled trials had minimised bias and error in their findings by using a set of criteria developed in previous health promotion reviews.27 28 29 “Sound” trials were those that reported data on each outcome measure indicated in the study aims; used a control or comparison group equivalent to the intervention group on relevant sociodemographic measures (or, in cases with non-equivalent groups, adjusted for differences in the analysis); provided pre-intervention data for all individuals in each group; and provided post-intervention data for all individuals in each group.
The criteria we used to assess the methodological quality of the qualitative studies were built on those suggested in the literature on qualitative research.26 30 31 32 33 Each study was assessed according to 12 criteria designed to aid judgment on the extent to which study findings were an accurate representation of young people’s perspectives and experiences (box). A final assessment sorted studies into one of three categories on the basis of quality: high quality (those meeting 10 or more criteria), medium quality (those meeting between seven and nine criteria), and low quality (those meeting fewer than seven criteria).
Criteria used to assess the quality of qualitative studies
Quality of reporting
Were the aims and objectives clearly reported?
Was there an adequate description of the context in which the research was carried out?
Was there an adequate description of the sample and the methods by which the sample was identified and recruited?
Was there an adequate description of the methods used to collect data?
Was there an adequate description of the methods used to analyse data?
Use of strategies to increase reliability and validity
Were there attempts to establish the reliability of the data collection tools (for example, by use of interview topic guides)?
Were there attempts to establish the validity of the data collection tools (for example, with pilot interviews)?
Were there attempts to establish the reliability of the data analysis methods (for example, by use of independent coders)?
Were there attempts to establish the validity of data analysis methods (for example, by searching for negative cases)?
Extent to which study findings reflected young people’s perspectives and experiences
Did the study use appropriate data collection methods for helping young people to express their views?
Did the study use appropriate methods for ensuring the data analysis was grounded in the views of young people?
Did the study actively involve young people in its design and conduct?
We used a standardised tool to extract from “sound” controlled trials information on the development and content of the intervention evaluated, the population involved, and the trial design and methods.34 Data to calculate effect sizes for pregnancy and birth rates were identified from study reports and via contact with study authors if data were incomplete or not in an appropriate form.
Data on the development, design, methods, and the populations involved were extracted from the qualitative studies in a standardised way by using an established tool designed for a broad range of study types.35 The findings of the qualitative studies were identified within the “findings” or “results” sections of study reports and exported verbatim into NVivo (version 2; QSR, Victoria, Australia), a qualitative data analysis software package.
The data synthesis was conducted in three stages according to the model described by Thomas and colleagues.22 Firstly, we used statistical meta-analysis techniques to assess the effectiveness of the interventions in the controlled trials. Chi square statistical tests were used to test for heterogeneity (“Q statistic”) between controlled trials; when there was no significant heterogeneity, we combined effect sizes in a random effects statistical meta-analysis using Evidence for Policy and Practice Information Centre reviewer software.36 Relative risk (RR) was used to calculate both individual study and combined effect sizes. Our procedures for meta-analysis followed standard practice in the field37 38 39 and were similar to those used in previous reviews by the Evidence for Policy and Practice Information Centre.29 40
Secondly, we conducted a thematic synthesis of the findings from the qualitative studies,41 42 following established principles developed for the analysis of qualitative data.25 43 44 Study findings were coded line by line to characterise the content of each line or sentence (for example, “frustration with rules and regulations at school,” “expectations for the future”). Codes were compared and contrasted, refined, and grouped into higher order themes (for example, “dislike of school”). The review team then drew out the implications for appropriate interventions suggested by each theme.
Thirdly, we constructed a methodological and conceptual matrix to integrate the findings of the two syntheses. The potential implications of young people’s views for interventions to prevent teenage pregnancy were laid out alongside the content and findings of the soundly evaluated interventions.
Screening of full reports against inclusion criteria, quality assessment, data extraction, and data synthesis were all carried out by pairs of reviewers working independently at first and then together. Initial screening of titles and abstracts was done by single reviewers after a period of double screening to ensure consistency across reviewers.
Study characteristics and quality
Ten controlled trialsw1-w10 and five qualitative studiesw11-w15 met our inclusion criteria. Six controlled trials were judged to be of sufficient methodological quality to provide reliable evidence about the impact of interventions on teenage pregnancy rates.w1-w3 w6 w7 w9 All these trials were conducted in the US and targeted disadvantaged groups of children and young people (tables 2⇓ and 3⇓).
Each of the methodologically sound controlled trials evaluated one of two intervention types: (a) an early childhood intervention, or (b) a youth development programme. Three studies evaluated early childhood interventions that aimed to promote cognitive and social development through preschool education, parent training, and social skills training.w2 w3 w7 Two of these studies—the Perry Preschool Programw2 and the Abecedarian Projectw3—evaluated the long term effects of preschool education and parenting support interventions; the third—the Seattle Social Development Project—evaluated the long term effects of a school based social skills development intervention for children and their parents.w7
A further three studies evaluated youth development programmes that aimed to promote self esteem, positive aspirations, and a sense of purpose through vocational, educational, volunteering, and life skills work.w1 w6 w10 Two of these studies—Teen Outreachw1 and the Quantum Opportunities Programw6—evaluated after school programmes based on the principle of “serve and learn,” in which community service is combined with student learning and educational support; the third—the Children’s Aid Society Carrera-Model Program—evaluated a comprehensive academic and social development intervention delivered in youth centres, which included work experience, careers advice, academic support, sex education, arts workshops, sports, and other activities.w10
In each trial, the control group received no intervention or standard education. The four controlled trials that were deemed not to be of sufficient quality also evaluated youth development programmes in the US.w4 w5 w8 w9All five qualitative studies were judged to be of medium or high quality.w11-w15 These studies included participants from a range of areas throughout the UK and used individual interviews, focus groups, and self completion questionnaires to collect data (table 4⇓). Four studies focused on, or included, the views of young parents,w11 w12 w14 w15 but only two of these studies included the views of young fathers as well as young mothers.w14 w15
Quantitative studies of the effects of interventions on teenage pregnancy rates
Of the six controlled trials deemed to be of sufficient methodological quality, four measured pregnancy rates reported by young women,w1 w2 w7 w10 three measured partner pregnancy rates reported by young men,w1 w7 w10 and two measured birth rates reported by young men and young women separatelyw3 or together.w6 The four controlled trials measuring pregnancy rates reported by young women or young menw1 w2 w7 w10 were included in two random effects meta-analyses: one that assessed the effects of interventions on teenage pregnancies reported by young women and a second that measured the effects of interventions on teenage pregnancies reported by young men. The findings of the two controlled trials that measured birth ratesw3 w6 were not subject to meta-analysis, but their findings are summarised after each meta-analysis. Tests revealed no statistical heterogeneity between the studies, suggesting that it would be appropriate to pool the effect sizes. However, effect sizes for youth development interventions and early childhood education interventions were pooled separately in recognition of the differences between these two types of intervention.
The pooled effect size from the first meta-analysis showed that early childhood interventions and youth development programmes reduced teenage pregnancy rates among young women (RR 0.61, 95% CI 0.48 to 0.77; fig 1⇓). The effect of an early childhood intervention on birth rates reported by young women was similar in the study by Campbell and colleaguesw3 (0.56, 0.42 to 0.75).
The effect of these interventions on pregnancies reported by young men is less clear (fig 2⇓). The pooled effect size from the second meta-analysis showed that young men who had received an early childhood or youth development intervention reported fewer partner pregnancies than those who had not, but this result was not statistically significant (RR 0.59, 95% CI 0.34 to 1.02).
Hahn and colleaguesw6 evaluated a youth development programme and measured birth rates reported by both young women and young men. The intervention reduced the birth rate by 36%, although this result was of borderline statistical significance (RR 0.64, 95% CI 0.40 to 1.03).
Qualitative studies of the views and experiences of young people
Three major themes relating to teenage pregnancy emerged from the findings of the five qualitative studies: dislike of school; poor material circumstances and unhappy childhood; and low expectations and aspirations for the future (fig 3⇓).
Dislike of school was a key aspect of young parents’ accounts of their lives before becoming parents and of young people identified as “at risk” of becoming teenage parents (for example, “Still be at school? I’d rather have a baby than that. I just didn’t like school, it was hard, it was horrible”w14). The reasons young people gave for disliking school varied (fig 3). Some related to the subject matter taught in school, which was seen as boring or irrelevant, especially for young women who had difficult or unhappy home lives and caring responsibilities (for example, “what on earth is this going to do for me?”w15). Other reasons related to insufficient or inappropriate support when falling behind with school work or experiencing bullying by teachers and peers (for example, “I got bullied so I just stopped going”w12). Some young people were frustrated with the inflexibility of “institutional life,” with all its rules and regulations (for example, “You can’t sit with your friends, which I found the best way of learning”w11).
Young parents reported unhappiness, rather than poverty in itself, as the most significant aspect of their childhood experiences that related to becoming a parent, although unhappiness went hand in hand with adversity and material disadvantage in their accounts. Common experiences included family conflict and breakdown, sometimes caused by violence, which could lead to living in care (fig 3). Young fathers reported violent fathers and a lack of suitable role models. Young parents noted how they had to “grow up faster” in order to survive, and also reported a lack of confidence, low self esteem, and high anxiety levels.w11 Some young women saw having a baby at an early age as a way to change their circumstances and ameliorate the effects of adversity. It is important to note, however, that not all the teenage mothers who participated in these studies had grown up unhappy or experienced personal adversity. Regardless of circumstances, some women had wanted to have a baby when they were young and looked forward to still being young when their children were older.
There were differences in the expectations and aspirations of young people who had, or wanted to have, a baby early in life and young people who had or wanted to have a baby later in life. For example, mothers who had children when they were teenagers wanted to leave school as soon as possible and get a job. In contrast, those who became pregnant later in life expected to go to university and travel. Both young mothers and young fathers believed that few opportunities were open to them apart from poorly paid, temporary work in jobs that they disliked (for example, “There are so many jobs out there that I didn’t even know existed . . . I probably could have done something but I just didn’t even think of these high paid jobs I could have done”w14). Young mothers described how having a baby was a more attractive option than entering the workforce, further education, or training. Young men’s lack of ambition was compounded by the low expectations their parents and peers held for them. Young people who wanted children later in life had long term plans and a more positive outlook for the future, and they described how participating in out of school activities such as sports, music, and arts improved their self esteem and motivation.
Do current interventions address the needs and concerns reported by young people?
The themes in our synthesis of qualitative studies suggest areas that should be addressed in preventive interventions, but measures to target these areas have not all been soundly evaluated for their effect on teenage pregnancy rates (table 5⇓).
Youth development programmes and early childhood interventions both go some way to addressing young people’s dislike of school. Two of the three youth development programmes in the controlled trials we reviewed included components designed to promote young people’s academic achievement, such as tutoring and homework assistance,w6 w10 whereas the third aimed to improve young people’s interpersonal skills so they could develop good relationships with their peers and others.w1 One early childhood intervention both taught children conflict resolution skills and trained parents to create a home environment supportive of learning.w7 We did not find any research that had tested the impact on teenage pregnancy rates of interventions designed to change the school culture and environment, such as antibullying strategies, teacher training, or involving young people in making decisions about what happens in the school.
All the youth development programmes aimed to prevent teenage pregnancy by broadening young people’s expectations and aspirations for the future. These programmes offered young people work experience in their local communities, careers advice, group work to stimulate active reflection, and discussion of future careers and employment opportunities. Two of the three soundly evaluated youth development programmes also provided out of school sports or arts activities.w6 w10
Summary of principal findings
This review sought to improve our understanding of the link between social disadvantage and teenage pregnancy by integrating evidence from qualitative studies and quantitative trials.
The evidence from the six controlled trials we looked at showed that early childhood interventions and youth development programmes can significantly lower teenage pregnancy rates. Both types of intervention target the social determinants of early parenthood but are very different in content and timing. Preschool education and support appear to exert a long term positive influence on the risk of teenage pregnancy, as well as on other outcomes associated with social and economic disadvantage such as unemployment and criminal behaviour.45 Programmes of social support, educational support, and skills training delivered to young people have a much more immediate impact.
Our review of five qualitative studies of young people in the UK indicated that happiness, enjoying school, and positive expectations for the future can all help to delay early parenthood. Young people who have grown up unhappy, in poor material circumstances, do not enjoy school, and are despondent about their future may be more likely to take risks when having sex or to choose to have a baby.
The findings of our review are especially important in the light of evidence that sex education and sexual health services are not on their own effective strategies for encouraging teenagers to defer parenthood12; they need to be complemented by early childhood and youth development interventions that tackle social disadvantage.13 18 46 Early childhood interventions and youth development programmes provide enhanced educational and social support in the early years of life and engage young people in developing career aspirations, respectively, thus addressing some of the key themes identified within our qualitative synthesis. However, important gaps exist in the evidence on how effectively current interventions address these themes (table 5). Structural and systemic issues such as housing, employment opportunities, community networks, bullying, and domestic violence were all important issues in young people’s accounts, but these factors have yet to be addressed in appropriate interventions and evaluated as wider determinants of teenage pregnancy.
Comparison with other studies
Our review adds to a growing body of research identifying factors that may explain the association between social disadvantage and teenage pregnancy. Dislike of school, an unhappy childhood, and a lack of opportunities for jobs and education have all emerged as explanatory factors in large scale national and international epidemiological analyses.3 9 17 18 47 48 49 Dislike of school appears to have an independent effect on the risk of teenage pregnancy.49 Our analysis of qualitative research provides additional insight into how factors that increase the risk of teenage pregnancy may operate. For example, a dislike of school was frequently the result of bullying, frustration with rules and regulations, lack of curriculum relevance, boredom, and inadequate support.
As well as developing and testing interventions to modify these antecedents, future research on teenage pregnancy and social disadvantage needs to consider strategies that counter the stigmatisation and discrimination faced by young parents. Some of the social exclusion experienced by young parents is the result of negative societal reaction. However, there is evidence to suggest that teenage parenting can under certain circumstances be a route to social inclusion rather than exclusion.50
Like many other systematic reviews in health promotion and public health, we found few trials conducted in the UK.27 29 40 This raises questions about the generalisability of the trial evidence. Our inclusion of qualitative evidence permitted us to examine the appropriateness of interventions evaluated in US trials from the perspective of young people in the UK. The appropriateness of interventions is an important aspect of generalisability to consider.51 Our inclusion of qualitative evidence does not, however, replace the need for further trials in the UK and elsewhere to address the impact of interventions designed to ameliorate the wider determinants of teenage pregnancy.
A recent study carried out in England evaluated the effects of the Young People’s Development Programme—an intensive, multicomponent youth development intervention based on the Children’s Aid Society Carrera Model Program.w10 52 In contrast to the findings of this review, the quasi-experimental study found that young women in the intervention group were more likely to report pregnancy than those in the comparison group. This finding may be the result of the potentially stigmatising effect of targeting and labelling young people as “high risk” or of introducing participants to other “high risk” young people in alternative educational settings. In comparison with the Young People’s Development Programme, the youth development programmes evaluated by the controlled trials in our review used after school programmes or interventions delivered in community settings rather than the approach of keeping young people out of mainstream schools and working with them in alternative educational settings. This difference in approach may explain the difference in the findings of the two studies and highlights the need to evaluate a revised youth development programme in the UK.
Strengths and limitations of the study
The strengths of our review include the comprehensiveness of our searches, the exclusion of methodologically weak studies, the rigorous synthesis methods used, and the inclusion of qualitative research alongside controlled trials to establish not only “what works” but also appropriate and promising intervention strategies on the basis of young people’s views on the factors associated with teenage pregnancy. Including only studies that evaluated interventions relative to control conditions over the same period of time avoids missing temporal differences between groups. Such changes include the relaxing of abortion laws and the increasing acceptability of abortion over time, which may affect self reported pregnancy rates.
The small numbers of studies we found are a limitation of the available body of research, as is the dominance of controlled trials conducted in the US (although this is a common feature of many health promotion and public health reviews). Our search strategies would have under-represented non-English language studies. As with any systematic review, we cannot be certain that we identified all relevant studies; in particular we may not have identified all unpublished studies, which are more likely to report negative findings than are published studies. We are only aware of one relevant study published since the searches for this review were carried out: the evaluation of the Young People’s Development Programme.52 Whether this study would meet the quality criteria for our review is unclear, but it should be considered in any update.
Conclusion and policy implications
This review provides a small but reliable evidence base that early childhood interventions and youth development programmes are effective and appropriate strategies for reducing unintended teenage pregnancy rates. Our findings on the effects of early childhood interventions highlight the importance of investing in early care and support in order to reduce the socioeconomic disadvantage associated with teenage pregnancy later in life.53 Both the early childhood interventions and the youth development programmes combined structural level and individual levels components, which is in line with many current recommendations in health promotion and public health.54 55 A policy move to invest in youth programmes should complement rather than replace high quality sex education and contraceptive services, and should aim to improve enjoyment of school, raise expectations and ambitions for the future, and provide young people with relevant social support and skills.
What is already known on this topic
Evidence suggests that sex education and better sexual health services do not reduce teenage pregnancy rates
A number of controlled trials have tested the effects of interventions that target the social disadvantage associated with early pregnancy and parenthood, and a number of qualitative studies have considered young people’s views of the factors associated with teenage pregnancy
No systematic review has brought these quantitative trials and qualitative studies together to determine intervention effectiveness and appropriateness
What this study adds
Early childhood interventions and youth development programmes that combine individual level and structural level measures to tackle social disadvantage can lower teenage pregnancy rates
Such interventions are likely to be appropriate for children and young people in the UK because they improve enjoyment of school, raise expectations and ambitions for the future, and ameliorate the effect of an unhappy childhood in poor material circumstances
A policy move to invest in interventions that target social disadvantage should complement rather than replace high quality sex education and contraceptive services
Cite this as: BMJ 2009;339:b4254
Contributors: AH, AO, and GB designed the study and obtained funding. AH, AO, and GB wrote the review protocol. AF, GB, and AH conducted the searches, screened titles and full papers, assessed study quality, extracted data, and undertook the statistical and qualitative syntheses. All authors contributed to the drafting of the paper and approved the final submitted version. AH, AO, and GB are the guarantors. All authors had full access to all the data in the study, including statistical reports and tables, and can take responsibility for the integrity of the data and the accuracy of the data analysis.
Funding: The review was funded by the Department of Health. AH was funded by a senior level research scientist in evidence synthesis award from the Department of Health. The researchers operated independently from the funders and the views expressed in this paper are those of the authors and not necessarily those of the Department of Health.
Competing interests: None declared.
Data sharing: Technical appendix available at http://eppi.ioe.ac.uk/cms/Default.aspx?tabid=674.
This is an open-access article distributed under the terms of the Creative Commons Attribution Non-commercial License, which permits use, distribution, and reproduction in any medium, provided the original work is properly cited, the use is non commercial and is otherwise in compliance with the license. See: http://creativecommons.org/licenses/by-nc/2.0/ and http://creativecommons.org/licenses/by-nc/2.0/legalcode.