Using Time Series Measures to Explore Family Planning Survey Data and Model-based Estimates

TL;DR

How well FPEM estimates align with survey data is evaluated using time series diagnostic indices from the wdiexplorer R package, which account for countries nested within sub-regions.

Abstract

Family planning is a global development priority and a key indicator of reproductive health. Monitoring progress is challenged by gaps in survey data across countries. The United Nations Population Division addresses this with the Family Planning Estimation Model (FPEM), a Bayesian hierarchical time series model producing annual estimates of modern contraceptive use while sharing information across countries and regions. This paper evaluates how well FPEM estimates align with survey data using time series diagnostic indices from the wdiexplorer R package, which account for countries nested within sub-regions. Visualisation of survey data, modelled trajectories, and diagnostics enables assessment of model performance, highlighting where trends align and where discrepancies occur.

Figures (9)

• Figure 1: Stacked bar plot of contraceptive use showing both modern and traditional methods across countries. Countries are grouped by sub-region, with sub-regions ordered by their average total (modern $+$ traditional) contraceptive use in the most recent year and countries within sub-regions are also ordered according to their total contraceptive use in the most recent years. Central America sub-region has the highest (modern $+$ traditional) contraceptive use with Nicaragua having the highest total contraceptive use in its most recent survey year (2011) in Central America sub-region.
• Figure 2: Comparison of trend strengths in survey data and model-based estimates across countries. Each point represents the ratio of model-based to survey trend strength plotted against the survey trend strength for a country. All countries have ratios above 1, indicating that the model-based trends are stronger than the survey trends. Labelled points correspond to countries with extreme ratios of model-based to survey trends strength. Hovering any point in the https://oluwayomi-olaitan.github.io/Family-Planning-Exploratory-Analysis-Interactive-Plots/interactive-plots/trendstrength_plot.html reveals the corresponding country name, its survey and model trend strength.
• Figure 3: Survey data trajectories and model-based trends for the top 10 countries with the largest ratios of model-based to survey trend strength. Each panel presents the survey observations and corresponding model trajectory of a country, ordered by their ratios. Belize shows the highest ratio, reflecting a dip in its 2006 smoothed out by the model. Similarly, Comoros exhibits a sharp spike in 2000, with an observed proportion of modern contraceptive use far higher than other years.
• Figure 4: Silhouette widths of the survey data across the FP2030 focus countries, grouped by sub-region excluding sub-regions with only one country. Each bar represents the silhouette width of its correspondence country, while the light-shaded rectangular bars indicate the average silhouette width for each sub-region. Sub-regions are ordered by their average silhouette width, and countries within the same group are ordered accordingly with the same colour to facilitate visual distinction within groups. All sub-region except Melanesia exhibit negative group average silhouette width.
• Figure 5: Silhouette widths of the model-based estimates across the 85 FP2030 focus countries, grouped by sub-region excluding sub-regions with only a single country. Sub-regions are ordered by their average silhouette width. Countries within the same sub-region share the same bar and axis text colour to facilitate visual distinction within groups. All sub-regions have negative average silhouette widths, except Melanesia and Southern Africa with positive average silhouette widths.