Exploring the evidence on the effectiveness of interventions to cut rates of teenage pregnancy
David Paton, PhD, is chair of industrial economics, Nottingham University Business School.
Paton, D. (2009) Exploring the evidence on strategies to reduce teenage pregnancy rates. Nursing Times; 105: 42, early online publication.
This article outlines the goals of the Teenage Pregnancy Strategy, its progress in reducing conceptions over the past 10 years and the scientific evidence on the effectiveness of these policies.
Keywords: Teenage pregnancy, Conception rates, Sexual health interventions
- This article has been double-blind peer reviewed
- Some sexual health interventions can have unexpected and unwanted effects on teenage sexual behaviour.
- Providing emergency contraception in school and other settings is unlikely to contribute to lower teenage pregnancy rates.
- Insisting on parental consent before providing sexual health services to minors may have beneficial impacts on teenage sexual health.
- Practitioners should consider the possible impact of interventions on STIs as well as teenage pregnancy.
Despite efforts by successive governments over the past 20 years, teenage pregnancy rates in the UK have remained stubbornly high. In 1999, a Teenage Pregnancy Strategy was launched with specific targets to achieve a 15% reduction in the under-18 conception rate in England by 2004, a 50% reduction by 2010, and to establish a long-term downward trend in the under-16 conception rate (Social Exclusion Unit, 1999). A further objective was to improve the life chances of teenage mothers, particularly by reintegrating them into education.
In a parallel initiative in 2004, the government also adopted a target of reducing the rate of new diagnoses of sexually transmitted infections (STIs) among under-16s and 16-19 year olds.
Responsibility for implementing the Teenage Pregnancy Strategy was given to a new body, the Teenage Pregnancy Unit (TPU), since subsumed within the Department for Children, Schools and Families Every Child Matters (ECM) programme. Each local authority in England was given a specific target for reducing under-18 pregnancy rates and resources to help them achieve this target.
The TPU/ECM remains responsible for national initiatives and for advising on and monitoring delivery at local level. Funds were allocated from the end of 1999 and increased significantly to a peak of £40m in 2003-2004. About 85% of expenditure is allocated for local implementation, with the remainder being taken by central costs.
From the start, the focus has been on reducing conceptions through better and earlier sex education, improving young people’s knowledge of and access to family planning services, and providing easier access to abortion services for young people.
Progress so far
The under-16 conception rate in 2007 was about 1% higher than when the strategy was launched in 1999, while the under-18 rate was about 7% lower (Fig 1). However, most of the decrease in the under-18 rate came between 1999 and 2001, when relatively little had been spent on the strategy. Since then, this downward trend has slowed considerably and there appears to be no prospect of the government meeting its target of a 50% reduction by 2010.
Although the national data provides little evidence that the strategy has had any impact in reducing conception rates, there is evidence to suggest that it may have affected the outcome of teenage conceptions. For example, the percentage of under-18 conceptions ending in abortion increased from 53% in 1999 to just under 62% in 2007.
New diagnoses of STIs among teenagers have continued to increase throughout the period (Fig 1). Although some of this increase may be due to greater awareness and diagnosis of STIs, there are few signs that teenage sexual health is improving.
It is important to note that national trends may be affected by several factors and it does not necessarily follow that the strategy has had no impact on adolescent sexual health. It may be that certain measures have had positive effects that have been negated by other counterproductive measures. Another possibility is that any beneficial impact of the strategy has been masked by wider cultural and socio-economic factors that have also affected pregnancy rates. To help explore these possibilities, it is useful to assess the evidence on how specific interventions of the type promoted by the strategy affect teenage pregnancy rates.
Types of evidence
Evidence on the impact of particular interventions comes from a variety of sources. Within the broad definition of qualitative studies, we can identify at least three key approaches – randomised controlled trials (RCTs), population studies and mathematical modelling, each of which carries advantages and disadvantages.
RCTs attempt to estimate the difference in an outcome between a treatment and control group before and after an intervention, such as exposure to a particular sex education curriculum. A potential problem with RCTs is the possibility that knowledge of being observed may alter participants’ behaviour in either control or intervention groups. Also, RCTs tend to focus on a very specific group of subjects and it may be difficult to use them to predict the impact of a policy on the wider population.
A good example of this problem occurs with RCTs looking at the impact of interventions to increase access to emergency contraception. Such studies typically select both control and intervention groups from women attending family planning clinics. The results are informative about the effect of easier access on those who are sexually active, but it is harder to draw conclusions about the impact of such a policy on the general population.
Population-level studies tend to use secondary data either on individuals or on regional areas. The best studies exploit the advantages of panel data, that is, data on several observations over several periods of time, to estimate the impact of an intervention. This type of approach has the advantage of testing directly for an aggregate impact at the level the policy is targeted. A common criticism of this type of study is that the likelihood of a policy being enacted in a particular area may be directly related to the outcome variable of interest.
For example, it is reasonable to expect that interventions to reduce teenage pregnancy rates will be introduced more intensively in areas with high (or increasing) teenage pregnancies. If these areas are also ones in which it is particularly difficult to reduce teenage pregnancy rates, then it would be unreasonable to infer a causative impact from a positive association between the intervention and pregnancy rates.
A recent refinement to population-level studies that attempts to get around this problem is the use of matching estimators. These try to ensure that areas affected by an intervention are compared only with otherwise similar areas that are unaffected by the intervention.
Mathematical modelling can be used to measure the impact of behavioural changes over time. For example, some recent work combined survey data on sexual activity among young people with estimates of contraceptive failure, to identify the extent to which falls in teenage pregnancy rates in the US could be attributed to abstinence or contraceptive use. However, these studies tend to be very sensitive to the assumptions used.
For example, Santelli et al (2007) concluded that about 77% of the decrease among 15-17 year olds between 1995 and 2002 should be attributed to improved contraceptive use and just 23% to greater abstinence. Using the same data but slightly different assumptions, Mann and Stine (2007) suggested that both contraceptive use and abstinence made equal contributions, while Mohn et al (2003) attributed two-thirds of the decrease to lower sexual activity.
In any case, these studies examine the direct mechanism by which pregnancy rates go down. They are not able to provide evidence on the impact of a policy intervention on, for example, family planning access.
Much - although by no means all - of the existing evidence comes from the US. Although such evidence from the US can provide valuable insights for the UK, cultural and societal differences between the two countries may mean research findings from one cannot automatically be applied to the other.
Access to family planning
Improving access to family planning services, particularly those aimed at young people, has been a high priority from the start of the Teenage Pregnancy Strategy. Despite this, the evidence on the impact of such measures on teenage conception rates is not especially encouraging.
Wilkinson et al (2006) found those local authorities that have been allocated more money have experienced larger reductions in under-18 conception rates (at least in the early years). But further analysis showed that those areas with higher quality contraceptive services and with better access to services experienced lower reductions in conception rates than other areas. Although this association is not necessarily causal, the finding is in line with a range of population-level studies and RCTs from the US and the UK (Paton, 2006; 2002; DiCenso et al, 2002). These found little evidence that access to family planning reduces teenage pregnancy rates, particularly among younger age groups.
There is some evidence that access to family planning may be associated with fewer teenage births once conception has happened (Kearney and Levine, 2009; Wolfe et al, 2001) but more abortions (Wilkinson et al, 2006). These findings are consistent with the increase in the proportion of teenage pregnancies ending in abortion in England since the start of the strategy.
The evidence is perhaps strongest on the impact of increased access to emergency contraception. An exhaustive review of RCT evidence concluded that “to date, no study has shown that increased access to [emergency contraception] reduces unintended pregnancy or abortion rates on a population level” (Raymond et al, 2007). Population-level studies from the UK (for example, Girma and Paton, 2006) have also found that access to emergency contraception does not lead to reductions in teenage pregnancy rates.
That better access to family planning services does not seem to reduce teenage conception rates may seem counter-intuitive at first sight. However, there are a number of reasons why such interventions might be ineffective. One possibility is that increased access to services through, for example, schools, simply substitutes for services already provided by GPs.
It may be that increasing access to family planning does reduce teenage pregnancy rates, but that the effect is too small for studies to pick up. This explanation seems unconvincing in the case of emergency contraception given the range of data that has failed to find any positive impact. A particular issue with emergency contraception may be recent evidence suggesting its effectiveness is lower than previously thought (Stanford, 2008).
Economists such as Levine (2003) suggest that greater access to family planning (or abortion) reduces the perceived risks associated with early sexual activity and, as a result, leads some young people to increase risk-taking behaviour. Combined with relatively high failure rates for many methods of family planning among teenagers (Kost et al, 2008), fewer pregnancies from greater use of birth control are counterbalanced by more pregnancies arising from more sexual activity.
The question of whether access to family planning increases risky sexual behaviour is by no means resolved. A meta-analysis (DiCenso et al, 2002) found no evidence that contraceptive access affects sexual behaviour, whereas Raymond and Weaver (2008) reported that access to emergency contraception is associated with a significant increase in risky behaviour.
Use of diagnoses of STIs as a proxy for sexual risk taking (Klick and Stratmann, 2008; Paton, 2006) may be a fruitful approach to understanding why access to family planning in general and emergency contraception in particular seems to have little effect on teenage pregnancy rates.
The Teenage Pregnancy Unit has consistently emphasised that the role of confidentiality is crucial when providing family planning and abortion services to young people, especially those below the age of consent, which can be a particularly sensitive issue for school nurses.
A common rationale for such a policy is that by assuring young people that parents do not have to be informed, uptake of services will increase and this will in turn contribute to lower underage conception rates. However, if access to services does not in fact reduce conception rates, then the case for guaranteeing confidentiality is considerably weakened.
Very few studies have actually examined the impact of removing (or enforcing) confidentiality for contraception on pregnancy rates. Those that have (for example, Paton, 2002) have failed to find a significant impact on underage conception rates, although there is some evidence of an impact on births relative to abortions.
Because of the Fraser ruling, family planning could not be provided to underage girls without parental involvement in England and Wales for most of 1985 (Fig 2). Take-up at family planning clinics among this age group dropped by about 30% in 1985, yet the underage conception rate in England decreased slightly relative to the rate among older teenagers. Similarly, the rate did not increase relative to the underage conception rate in Scotland where the Fraser ruling did not apply (Paton, 2002).
Several researchers have examined the impact of laws in the US mandating parental involvement (including consent in some cases) before abortions are performed on minors. Difficulties of doing such studies mean the issue is by no means settled, but the majority to date found that parental involvement laws lead to significant decreases in underage abortion rates (Joyce et al, 2006; Levine, 2003); in overall conception rates (Levine, 2003); and to decreases in teenage STIs (Klick and Stratmann, 2006). There is some evidence that parental involvement laws are most beneficial for younger teenagers. Colman et al (2008) found that such a law decreased both abortions and births among girls aged 17 at the time of the birth or abortion. However, in a slightly older cohort (girls aged 17 at the time of conception), abortions decreased by a lower amount while births increased slightly.
There is also the question of whether such laws lead to delays in obtaining abortions and, as a result, increase the number of late ones. There is certainly some evidence that the proportion of late abortions increases after parental consent laws but the reduction in the overall abortion rate seems to be such that the actual rate of late abortions does not significantly increase (for example, Joyce and Kaestner, 2001). The evidence is not strong and this is clearly an issue that merits further research.
School-based sex education
In general, studies that have examined the impact of particular school-based sex and relationships education (SRE) interventions on teenage pregnancy rates are not particularly encouraging. A review of RCTs that tested for an impact of SRE schemes concluded that none were effective in reducing teenage pregnancy rates (DiCenso et al, 2002). Wilkinson et al (2006) also found no association across local authorities in England between the quality of SRE provision and reductions in the under-18 conception rate.
Recent RCTs (Stephenson et al, 2008; Henderson et al, 2007) have reported a similar lack of impact on unwanted pregnancy rates. One exception is Cabezon et al (2005), who found evidence that an abstinence-based programme had a statistically significant impact in reducing both early sexual activity and pregnancy rates.
The majority of studies on this topic examine a particular SRE programme relative to existing SRE models and they are not testing the impact of school sex education relative to no school sex education. The few exceptions to this rule present conflicting evidence.
Oettinger (1999) found that, among some sub-groups, teenagers who were exposed to school-based SRE experienced slightly higher pregnancy rates than those who were not exposed. However, Kohler et al (2008) found that SRE is associated with lower self-reported pregnancy rates among teenagers. Sabia (2006) concluded that SRE has no measurable impact on pregnancy rates. Given that school-based SRE has become nearly universal, there is little prospect that future research will be able to resolve these contradictions satisfactorily.
Care is needed in making firm policy conclusions from the results of these studies. On some issues, such as the impact of abstinence-based education, the evidence on conception rates is limited and more research is needed.
The aim of SRE is not solely to reduce teenage pregnancy rates, and schemes may have benefits that cannot be observed in statistical studies of pregnancy rates. However, it is important not to overstate the impact that innovations to SRE policy are likely to have on teenage pregnancy rates in the UK.
Despite more than £200m being spent on the Teenage Pregnancy Strategy, there has been little discernible impact on conception rates, at least at a national level. Although disappointing, these results should not be surprising.
The evidence that direct interventions such as improved school sex education and confidential access to family planning services help to lower teenage pregnancy rates is, at best, weak.
In the case of access to emergency contraception, a large number of studies using a range of data and methodological approaches have failed to find evidence of any reduction in pregnancy rates. Limitations in the quantity and/or quality of research means there should be caution in drawing general policy conclusions.
Some of the more encouraging evidence is on measures aimed at increasing parents’ involvement in younger teenagers’ abortion decisions. However, while this evidence is consistent with qualitative research emphasising the importance of family influences (Stammers, 2007), points of contention still remain.
These findings have implications for nurses working to implement the Teenage Pregnancy Strategy in schools, in community settings and those involved in policy discussions. The fact that many of the measures currently in place appear unlikely to have an impact on teenage pregnancy rates does not necessarily mean that nurses should abandon them. Interventions may have a number of objectives beyond simply cutting teenage pregnancy rates and these objectives may be sufficient to justify them.
However, it is important that practitioners and policymakers avoid giving the impression to stakeholders, for example parents, that particular measures need to be introduced to cut teenage pregnancy when the evidence does not back this up. Rather, a particular initiative should be considered on its own merits.
Ethical considerations should of course be paramount, especially when considering the provision of information and services to children below the age of consent. Even where a particular course of action is felt to be ethical, policymakers and practitioners need to pay careful attention to the possibility that measures designed to improve sexual health among teenagers will affect behaviour in unintended ways.
Cabezon CP et al (2005) Adolescent pregnancy prevention: an abstinence-centered randomized controlled intervention in a Chilean public high school. Journal of Adolescent Health; 36: 64-69.
Colman S et al (2008) Misclassification bias and the estimated effect of parental involvement laws on adolescents’ reproductive outcomes. American Journal of Public Health; 98: 10, 1881-85.
DiCenso A et al (2002) Interventions to reduce unintended pregnancies among adolescents: systematic review of randomised controlled trials. BMJ; 324: 1426-34.
Girma S, Paton D (2006) Matching estimates of the impact of over-the-counter emergency birth control on teenage pregnancy. HealthEconomics; 15: 1021- 32.
Henderson M et al (2007) Impact of a theoretically based sex education programme (SHARE) delivered by teachers on NHS registered conceptions and terminations: final results of cluster randomised trial. BMJ; 334: 133-37.
Joyce T, Kaestner R (2001) The impact of mandatory waiting periods and parental consent laws on the timing of abortion and state of occurrence among adolescents in Mississippi and South Carolina. Journal of Policy Analysis and Management;20: 2, 263-82.
Joyce T et al (2006) Changes in abortions and births and the Texas Parental Notification Law. New EnglandJournal of Medicine; 354: 10, 1031-38.
Kearney MS, Levine PB (2009) Subsidized contraception, fertility and sexual behavior. Review of Economics and Statistics; 9: 137-51.
Klick J, Stratmann T (2008) Abortion access and risky sex among teens: parental involvement laws and sexually transmitted diseases. Journal of Law, Economics and Organization; 24: 2-21.
Kohler PK et al (2008) Abstinence-only and comprehensive sex education and the initiation of sexual activity and teen pregnancy. Journal ofAdolescent Health; 42: 344-51.
Kost K et al (2008) Estimates of contraceptive failure from the 2002 National Survey of Family Growth. Contraception; 77: 10-21.
Levine PB (2003) Parental involvement laws and fertility behavior. Journal of Health Economics; 22: 861-78.
Mann JR, Stine C (2007) The roles of abstinence and contraception in declining pregnancy rates. American Journal of Public Health; 97: 969.
Mohn JK et al (2003) An analysis of the causes of the decline in non-marital birth and pregnancy rates for teens from 1991 to 1995. Adolescentand Family Health; 3: 39-47.
Oettinger GS (1999) The effects of sex education on teen sexual activity and teen pregnancy. Journal of Political Economy; 107: 3, 606-44.
Paton D (2006) Random behaviour or rational choice? Family planning, teenage pregnancy and STIs. Sex Education: Sexuality, Society and Learning; 6: 281-308 .
Paton D (2002) The economics of abortion, family planning and underage conceptions. Journal of Health Economics;21: 27-45.
Raymond EG et al (2007) Population effect of increased access to emergency contraceptive pills: a systematic review. Obstetrics and Gynecology; 109: 181-88.
Raymond EG, Weaver MA (2008) Effect of an emergency contraceptive pill intervention on pregnancy risk behavior. Contraception; 77: 333-36.
Sabia JJ (2006) Does sex education affect adolescent sexual behaviors and health? Journal of Policy Analysis and Management; 25: 4, 783-802.
Santelli JS et al (2007) Explaining recent declines in adolescent pregnancy in the US: the contribution of abstinence and improved contraceptive use. American Journal of Public Health; 97:150-56.
Social Exclusion Unit (1999) Teenage Pregnancies. London: HMSO.
Stammers T (2007) Sexual health in adolescents. BMJ; 334: 103-4.
Stanford JB (2008) Emergency contraception: overestimated effectiveness and questionable expectations. Clinical Pharmacology and Therapeutics; 83: 19-21.
Stephenson J et al (2008) The long-term effects of a peer-led sex education programme (RIPPLE): a cluster randomised trial in schools in England. PLoS Medicine;5: 1579-90.
Wilkinson P et al (2006) Teenage conceptions, abortions and births in England: 1994-2003, and the national teenage pregnancy strategy. Lancet; 368: 1879-86.
Wolfe B et al (2001) The role of economic incentives in teenage nonmarital childbearing choices. Journal of Public Economics; 81: 473-511.