Analysis of national and subnational prevalence of adolescent pregnancy and changes in the associated sexual behaviours and sociodemographic determinants across three decades in Ghana, 1988-2019.

objectiveUnderstanding the determinants of adolescent pregnancy and how they have changed over time is essential for measuring progress and developing strategies to improve adolescent reproductive health. This study examined changes over time in the prevalence and determinants of adolescent pregnancy in Ghana.

methodsA total of 11 nationally representative surveys from the Ghana Demographic and Health Survey (1988, 1993, 1998, 2003, 2008, 2014), Multiple Indicator Cluster Survey (2006, 2011, 2017-2018) and Malaria Indicator Survey (2016 and 2019) provided data on 14 556 adolescent girls aged 15-19 for this analysis. A random-effect meta-analysis, time trends and multivariable logistic regression models were used to track the prevalence and determinants of adolescent pregnancy.

resultsThe pooled prevalence of adolescent pregnancy in Ghana was 15.4% (95% CI=13.49% to 17.30%). Rural areas (19.5%) had a higher prevalence of adolescent pregnancy than urban areas (10.6%). In the overall sample, middle adolescents (15-17 years) (aOR=0.30, 95% CI=0.23 to 0.39), adolescents in urban areas (aOR=0.56, 95% CI=0.43 to 0.74), large households (aOR=0.62, 95% CI=0.49 to 0.78), not working (aOR=0.62, 95% CI=0.43 to 0.90) and those unaware of contraceptive methods (aOR=0.49, 95% CI=0.27 to 0.90) were less likely to become pregnant. Adolescents from middle-income (aOR=0.91, 95% CI=0.67 to 1.24) or high-income (aOR=0.59, 95%CI=0.36 to 0.94) households, those who were semiliterate (aOR=0.56, 95%CI=0.39 to 0.82) or literate (aOR=0.28, 95%CI=0.21 to 0.37) and those with fewer previous sex partners were less likely to become pregnant. Not all determinants in the overall sample were consistently associated with adolescent pregnancy in the last three decades. Between 1988 and 1998, determinants of adolescent pregnancy were age, literacy, employment, household size and whether the mother was alive. Between 2003 and 2008, age, literacy, household size, income, age of last sexual partner, number of previous partners and contraception knowledge determined adolescent pregnancy. From 2011 to 2019, age, residence, literacy and menstrual cycle knowledge were determinants of adolescent pregnancy.

conclusionInterventions and policies to prevent adolescent pregnancy should prioritise adolescents from disadvantaged backgrounds.