Transitions between contraceptive states among women aged 15-49 Years in Guinea

This study examined transitions between contraceptive states within women's contraceptive trajectories and identified the factors associated with contraceptive continuation, method switching, and discontinuation.

The study used contraceptive calendar data from the 2018 DHS Guinea. A woman-month event dataset was constructed to reconstruct contraceptive and reproductive histories over the five years preceding the survey.

Transitions between contraceptive states - non-use of contraception, use of traditional methods, use of modern methods, pregnancy, and pregnancy outcomes - were modelled as recurrent events. To account for multiple transitions experienced by the same woman over time, the Prentice-Williams-Peterson gap-time survival model was applied.

The probability of transitioning to another contraceptive state was estimated at 1.0% among women not using contraception, 1.6% among users of traditional methods, and 2.7% among users of modern methods. The probability increased to 7.5% during pregnancy and reached 61.1% following pregnancy outcomes. Transition probabilities varied only marginally by age, ranging from 23.6% among women aged 15-24 years to 23.0% among those aged 25-34 years and 22.3% among women aged 35-49 years. Similar patterns were observed across educational attainment and household wealth categories. By contrast, exposure to family planning messages and place of residence were not significantly associated with contraceptive transitions.

Contraceptive transitions are primarily driven by reproductive events rather than by socioeconomic characteristics. These findings emphasise the importance of strengthening family planning interventions at key stages of the reproductive life course. However, the findings should be interpreted with caution, as the analysis did not account for sociocultural norms, the availability of family planning services, and the characteristics of local health systems. In addition, the retrospective contraceptive calendar data are subject to recall bias.