Investigating the fall in teenage pregnancy rates in the UK from 1999 onwards using systematic review and natural experimental methods

Teenage pregnancy is associated with numerous health risks, both to young women and to infants, and may contribute to entrenched social inequalities. Reducing high rates of teenage pregnancy may be an effective target for action to reduce inequalities. Between 1990 and 2016, UK countries saw relatively high rates of teenage pregnancy compared with other European high-income countries. A prominent intervention in response to these observations was England’s Teenage Pregnancy Strategy (henceforth referred to as the Strategy), which was launched in 1999 and cost £280m across 10 years of implementation. Across the Strategy period there was a 51% reduction in under-18 pregnancy rates between 1998 and 2014. The Strategy was hailed as a success and as a model for other countries to follow. However, researchers have questioned this conclusion of effectiveness. My analyses first examine the range of plausible hypothesised causes of the UK’s falling rates. I construct an initial causal tree of interventions and changes in social determinants presented in prominent sexual health literature and hypothesised to be causative of falling pregnancy rates. I use this tree to inform a systematic literature search of prior reviews of teenage pregnancy prevention. I synthesise the evidence presented of both the effectiveness of each cause and the exposure of the UK population using modified Bradford Hill criteria. I use this evidence to construct an updated causal tree. I present this causal tree as a summary of potentially testable causal pathways to determine the contribution of each to the observed falling rates in the UK. I then conduct an evaluation of the effect of the Strategy on teenage pregnancy rates. I use two natural experimental methods to address potential biases and triangulate between datasets, outcomes and methods. Using controlled interrupted time series models, I compare England under-18 pregnancy rates with Scotland and Wales from 1992-2016. Changes in trends in England at 1999 closely match those in Scotland (0.08 fewer pregnancies per 1000 women per year in England; -0.74 to 0.59) and Wales (0.14 more pregnancies per 1000 women per year in England; -0.48 to 0.76). In my second analysis, I use synthetic control models to construct predicted under-18 birth rate and under-20 pregnancy rate trends in England from data for 15 comparable countries. Across all models and sensitivity analyses, controls saw similarly falling teenage birth and pregnancy rates in the absence of the Strategy. I conclude that the Strategy had little or no effect on England’s teenage pregnancy rates. Future policy approaches to preventing teenage pregnancy should not rely on reproduction of the Strategy to achieve similar results. Further investigation should seek to understand the true causes of falling pregnancy rates. This thesis demonstrates the value of robust public health policy evaluation and the importance of achieving accurate estimates of the effects of a policy on its target outcomes