Spatial Clustering in Temporal Trends of Female Genital Mutilation Risk: Leveraging Sparse Data in Ethiopia, Kenya, and Somalia.

Female genital mutilation (FGM) is a harmful practice rooted in gender inequality. Its elimination is part of national and international agendas including the Sustainable Development Goals of the United Nations. Understanding its geographical evolution is crucial for targeted programming. However, due to sparse data, it is challenging to establish international comparability and statistical reliability. Data on FGM is observed at different points in time and periodicity across countries and in contexts with varying age-risk patterns, all of which can be a source of inaccurate and biased estimates. We perform an exemplary analysis, drawing on survival and complex survey analysis in Ethiopia, Kenya, and Somalia. This novel approach addresses measurement challenges specific to FGM data and produces an internationally comparable indicator-the probability of not experiencing FGM by age 20. We pinpoint the onset of statistically significant FGM decline at the subnational level from cohorts born in the 1970s until the 1990s. In the same period, we observe no decline in FGM risk across regions clustered around international borders and increasing subnational inequalities within countries. Our methods thus provide crucial insights into the geographical pattern of temporal trends in FGM risk.