Increased access to contraceptives and family planning resources have empowered women to make informed decisions about their reproductive health. However, approximately one in three women who start a modern method of contraception abandons use within a year, a fact that has puzzled researchers, policymakers, health providers and advocates.

For the past two years, the Big Data for Reproductive Health Bass Connections team and two Data+ teams have explored this conundrum using years of data collected through the United States Agency for International Development’s Demographic and Health Surveys Program (DHS).

Covering over 80 countries, this data has the potential to provide a detailed picture of contraceptive behavior, but, according to team leader Megan Huchko, it can be “quite hard to understand [the data] without pretty significant training and analytic skills.” The team’s goal, therefore, has been to develop user-friendly tools that can convert this raw data into visualizations to help the policy and healthcare communities see and understand trends in contraceptive use over time.

The team spent much of 2018-2019 identifying, building and testing new tools for data visualization, including an interactive chord diagram tool designed to show how contraceptive users switch between methods (or abandon contraceptive use entirely) over a specific, six-month time period. Team members also shared their work with key stakeholders to discuss opportunities for collaboration and encourage policymakers and ground-level health practitioners to use their tools to inform their own work.

This year, the team took a deeper dive into the DHS data to consider whether it might be possible to predict patterns in women’s contraceptive use and discontinuation through the application of machine learning and predictive modeling. This kind of predictive tool, team members inferred, could enable a healthcare worker to provide customized recommendations for family planning methods based on a person’s demographic and geographic characteristics.

Breaking into two main subgroups focused on machine learning and stakeholder engagement, the team reviewed and tested algorithms for their predictive tool at the same time as they assessed the digital family planning landscape and interviewed stakeholders to determine if and how a predictive tool might be of use in the field.

Next year, the team plans to shift focus to examine social media data to analyze prospective trends in contraceptive use. Team members will partner with IntraHealth’s digital health team to investigate whether social media data sources, such as Google Trends and Twitter, can be used to measure contraceptive use and its discontinuation, design public health interventions and conduct disease surveillance. Ultimately, the goal is to make big data useful for reproductive health and continue to increase access to informed family planning worldwide.

Read more from team leader Megan Huchko, check out the team’s year-end profile in the Fortin Foundation Bass Connections Virtual Showcase and browse the team’s website to learn more about their research.

Learn More

• See other faculty perspectives.
• Browse the Fortin Foundation Bass Connections Virtual Showcase.
• Read about this project team’s plans for 2020-2021.