Title: Using Time Series Measures to Explore Family Planning Survey Data and Model-based Estimates
ArXiv ID: 2602.17034
Date: 2026-02-19
Authors: ** 논문에 명시된 저자 정보가 제공되지 않았습니다. (예: Oluwayomi Olaitan 등) **
📝 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.
💡 Deep Analysis
📄 Full Content
The right of any individual to decide freely and responsibly on the number and spacing of their children has long been recognised as a fundamental human right [1] and is central to improving lives and building equitable societies [2]. This positions improvements to family planning as a cornerstone of progress as reflected in Sustainable Development Goal indicator 3.7.1: "Proportion of women who have their need for family planning satisfied with modern methods". Family planning is therefore a key global development priority, integral to reproductive health services and broader socio-economic development [1]. A range of family planning programmes operate at both global and national levels to expand access to reproductive health services, strengthen health systems, and promote informed contraceptive choice. Among these initiatives is Family Planning 2030 (FP2030) [3], which supports the global commitment to achieving universal access to family planning by 2030. FP2030 evolved from the FP2020 initiative, launched at the 2012 London Summit on family planning with a vision tagged "120 by 20" numeric vision, which sought to enable 120 million additional women and girls access to modern contraception by 2020 [4]. As the FP2020 initiative concluded in 2020, it gave way to FP2030 with a broader vision to promote the right of women and girls to lead healthy lives and make informed choices about contraception [3]. This is an initiative that is closely aligned with the global commitment of universal access to sexual and reproductive health under SDG 3.7.
Despite the global commitment to family planning, progress varies widely across countries and regions, partly due to differences in access, program implementation, and socio-demographic factors [5]. To monitor family planning progress, data are collected through household surveys such as, the Demographic and Health Survey(DHS), Performance Monitoring for Actions(PMA), UNICEF Multiple Indicator Cluster Surveys (MICS), national surveys, and other surveys [2]. These household surveys are typically conducted every 3 to 5 years, resulting in temporal gaps of up to 6 years for certain countries. In some rare cases, national surveys are conducted annually or semi-annually [6].
Temporal gaps in large-scale survey data make consistent monitoring of family planning progress challenging. To address this limitation, the Family Planning Estimation Model (FPEM) was developed [7,8]. FPEM produces annual survey-informed estimates and projections for key family planning indicators using Bayesian hierarchical time series inference. Where survey data are available, the model interpolates estimates and it projects estimates for data gaps between infrequent survey observations. FPEM employs Bayesian hierarchical inference to share information across hierarchical structures, countries nested in subregions, subregions in regions, and regions within the global level. The model estimates are smoothed by design, because they tend to reflect the hierarchical overall behaviour, which makes it important to explore how these modelled trends align with the underlying survey observations. This paper examines the extent to which modelled estimates from FPEM align with the underlying survey data, using selected diagnostic indices from the wdiexplorer package [9]. These indices are time series measures designed to quantify variation, trend, shape and sequential temporal features of country-level panel data. By using wdiexplorer to jointly analyse the survey observations and the corresponding modelled trajectories, as well as evaluating the residuals, the analysis identifies where the model aligns with the observed data and where notable differences arise. This facilitates the identification and comparison of temporal behaviours of the data and the associated model trends, enhancing understanding of patterns in modern contraceptive use across countries. All analyses and visualisations presented in this paper are fully reproducible and available on GitHub at https://github.com/Oluwayomi-Olaitan/Family-Planning-Exploratory-Analysis
.
The remaining sections of this paper are organised as follows. Section 2 provides background to the study, outlining the family planning survey data, the model-based estimates, and the wdiexplorer methodology. Section 3 describes the data processing steps for the survey dataset and the dataset of estimates and presents an initial visualisation of the survey data. Section 4 introduces the exploratory analysis of the survey data and model-based estimates using selected wdiexplorer diagnostic indices. Section 5 presents the residual-based exploratory analysis. Finally, Section 6 concludes with the findings, benefits of assessing how well the model-based trends align with the underlying survey observations, the limitations of the diagnostic indices in the context of family planning data, and directions for future work.
In this section, we discuss the family planning survey data,