Article Text
Abstract
Background The Teenage Pregnancy Strategy was published in July 1999 with the aim halving under-18 pregnancy rates in ten years. The strategy came to an end in 2010 with the UK’s change of government. Overall costs of implementation are estimated around £280m. From 1999–2014 rates of pregnancy to women aged under 18 fell by more than 50%. Hailed as a unique, nation-wide, comprehensive, evidence-based intervention, the strategy has been promoted as a reproducible model for other countries with high teenage pregnancy rates. In evaluating policy impact, observational studies alone may not be able to account for background trends and other events. We aimed to evaluate the impact of England’s Teenage Pregnancy Strategy on pregnancy and birth rates using natural experimental methods.
Methods Women aged under-20 and living in England during the intervention period were the target population. UK pregnancy rates were taken from Office of National Statistics and Information Services Division Scotland reports for years 1992–2016. We compared under-18 pregnancy rates in England with Scotland and Wales as controls using interrupted time series methods. Observed level and trend changes in controls at 1999 were used to predict a ‘No Strategy’ England. We compared under-18 birth rates and under-20 pregnancy rates of England with European and English-speaking high-income countries using synthetic control methods. Estimates of births were taken from the Human Fertility Database, estimates of population from the Human Mortality Database and abortions from the European Health Information Gateway, supplemented from national statistics bodies, for years 1990–2013. Background trends were estimated using a weighted mean of control countries’ rates.
Results Although pregnancies and births in England fell during the Strategy period, we found no difference compared with control. In interrupted time series analyses, trends in rates of teenage pregnancy beginning in 1999 in England were similar to Scotland (0.08 fewer pregnancies per 1,000 women per year; -0.74 to 0.60) and Wales (0.14 more pregnancies per 1,000 women per year; -0.48 to 0.76). In synthetic control analyses, yearly under-18 birth rates in England were very similar to synthetic control predictions post intervention, and under-20 pregnancy rates were marginally higher than control across the post intervention time period. Placebo testing and other sensitivity analyses supported the finding of no effect.
Conclusion Our analyses cast doubt on the effectiveness of England’s teenage pregnancy strategy. These results should be factored into decision making if other countries, or England in the future, were to contemplate similar costly strategies.