Charting a Path Forward for the Demographic and Health Surveys Program

On May 6, researchers, statisticians, and policy advocates gathered at the Population Association of America’s annual meeting in St. Louis to assess the health of the Demographic and Health Surveys (DHS) Program, a critical source of population and health data for low- and middle-income countries. In 2025, the program suffered a devastating blow when the U.S. government abruptly ended funding after over 40 years of support.

The session was organized by Madeleine Short Fabic and Trevor Croft of ICF. Panelists included Funmilola OlaOlorun (University of Ibadan, Nigeria), John Wilmoth (UN Population Division), Jacob Adetunji (Gates Foundation), and Mark Mather (PRB).

The Program Continues—But Funding Gaps Remain

When the U.S. government withdrew funding in February 2025, the immediate fears—that decades of data would disappear overnight—did not materialize. Thanks largely to a $39 million grant from the Gates Foundation (plus $4 million from Bloomberg Philanthropies), ongoing surveys were able to continue, and near-final datasets were preserved and shared with countries.

Six new datasets have been released in the last six months, and 15 surveys are underway. But Short Fabic was direct: “Funding is slowing. Gaps remain.” The program is narrower in scope than before 2025. Countries that never relied on USAID funding continued without interruption; others are still looking for financial support.

Short Fabic described the DHS as having two distinct components: the individual country surveys, funded by countries and donors including the Global Fund, World Bank, UNICEF, UNFPA, and the Gates Foundation; and the DHS platform, or the technical standards, quality assurance processes, comparability tools, and data accessibility infrastructure that help turn the surveys into a global public good. The platform is currently supported by the Gates Foundation and a few other donors.

In May 2026, ICF issued a Request for Information seeking partners to build mirrors or backups for the DHS archive website, recognizing that reliance on a single host creates fragility for a global public good. The request is open through June 19.

New AI Tools Support Data Processing and Analysis

Short Fabic reported that three AI-driven innovations—all intended to be non-proprietary—are under development to support countries’ survey efforts:

Automated data processing helps country teams move from raw survey data to publicly available datasets more quickly and with less outside support.

AI-assisted report writing feeds standardized DHS tables into a model trained on 40 years of country survey final reports. In testing, the tool generated a first draft of a 16-chapter final report that reached 86% accuracy against the Lesotho final report, which had been withheld from the model during training. Real-world country testing is imminent.

Natural language data analysis allows users to query DHS data in plain language and receive answers, tables, and visualizations. Short Fabic suggested this could evolve toward something closer to a DHS research assistant, capable of answering questions that go beyond the program’s 4,000 standard indicators.

New Research Shows the Critical Importance of DHS Data

Wilmoth, Director of the UN Population Division, offered the session’s most eye-opening evidence of just how critical DHS data are to global population and health measurement.

According to a recent analysis, DHS data accounted for 16% of under-five mortality data, 9% of adult mortality data, and 10% of total fertility rate data in World Population Prospects 2024, the UN’s authoritative estimates and projections of the global population. For the least developed countries, those figures rise to 57%, 48%, and 24%, respectively. In some countries, 80% to 90% of all input data rely on the DHS.

“We would really have almost no data for certain countries without the DHS,” Wilmoth said. “It’s a critically important data source.”

Wilmoth also shared findings from a survey, distributed through IPUMS and IUSSP, on what data users value most about DHS. The responses—internationally comparable indicators, geographic coverage, public-use microdata, harmonized recode datasets, the range of topics, and quality assurance—map closely to what ICF says makes DHS a global good, though users did not rank country ownership among their top priorities. (Country ownership might be more important for sustainability than usability, Wilmoth noted.)

Countries Need Ownership and Resources to Lead

OlaOlorun offered a practitioner’s perspective on what DHS data have meant in Nigeria. As a longtime DHS user, a co-principal investigator of the Performance Monitoring for Action (PMA) project and professor at the University of Ibadan, she described using DHS data across research, teaching, and stakeholder engagement, and as a benchmark for comparing PMA survey estimates.

“It serves as a gold standard for my students,” she said.

OlaOlorun also acknowledged structural challenges. Nigeria and other countries can no longer rely on donor funding to pay for DHS surveys, and political will is needed so that governments invest in them.

She also shared practical lessons on survey sustainability. Using the same surveyors across annual survey cycles reduces training time and builds institutional knowledge, and involving policymakers and ministries from the planning stage helps build ownership of data.

“I think this is a good opportunity to reset and think about how to move forward in the absence of donor support,” she said.

Stakeholders Explore Pathways for Long-Term Governance

Wilmoth presented the UN Statistical Commission process as a possible pathway toward sustainable long-term governance of demographic and health surveys. In March 2025, the UN Statistical Commission established a Task Force on Sustainable Demographic and Health Statistics Surveys, operating under the Inter-Secretariat Working Group on Household Surveys (ISWGHS). The task force has largely met two of its objectives—ensuring archive access and completing ongoing surveys—and plans to issue a report on long-term solutions at its March 2027 session, Wilmoth said.

Wilmoth shared this process as a possible pathway toward sustainable DHS governance. With multilateral endorsement and oversight, the Statistical Commission could set standards and facilitate support for country data collection, he said.

“The Statistical Commission essentially is an annual gathering of all the chief statisticians of all the countries in the world,” Wilmoth said. “That’s a powerful group if they get mobilized around this topic.”

But pursuing such a path comes with major challenges, including the tension between country ownership of surveys and the need for data standardization and comparability. The indicator framework for the UN’s Sustainable Development Goals, or SDGs, endorsed by 191 countries, offers an example of how to resolve this tension, Wilmoth said.

“We need to emphasize the value of an agreed indicator framework as a way to keep standardization as part of this process going forward,” he said.

Wilmoth acknowledged that what form any such mechanism might take remains unclear and that the process is still at the consultation stage. Other pathways—including regional, bilateral, and independent consortium models—remain live options as the global community continues its deliberations.

The Community Must Act on the Gates Foundation’s Investment

Adetunji described the Gates Foundation’s decision to step in as straightforward in principle: “The overall summary of why it was done was that it was the right thing to do.” The foundation— itself caught off guard by the abrupt termination—stepped in with goals to stabilize the survey infrastructure, complete in-process surveys, buy time for planning, and test options for archive preservation. Of the $39 million Gates committed, $13 million came directly from the CEO’s central office, reflecting the urgency of the moment.

Adetunji was candid about progress. The current effort has focused on household surveys, not yet addressing service provision assessments, questionnaire updates, capacity-building tools, and other DHS components, and has had limited success attracting additional donors. Time is also running out—the foundation’s grant ends in 2028.

“We are giving ourselves and other stakeholders three years to find a solution for the long term,” he said. “I don’t really know what the result will be.”

Four Lessons for Global Supporters

PRB’s Mark Mather drew on efforts to support the U.S. federal statistical system to offer four lessons for global advocates:

Focus on users and use cases, not just data systems. In the U.S., the most effective pushback on proposed changes to the American Community Survey has come directly from data users—from local planners, health departments, and nonprofits sharing stories of the impact on their work. For DHS, the case is built the same way: from Ethiopia’s health extension program to child marriage legislation in Nigeria, the data have a documented record of changing policy.

Build data user communities. Broad coalitions that say “data matter” rarely move the needle. What works is organized users with specific concerns who can act quickly when threats emerge. The “Friends of” model—Friends of BLS, Friends of NICHD—has proven effective for U.S. federal data programs, inspiring PRB’s Federal Data Forum. A similar community could do the same for the DHS.

Protect the least resourced. In the U.S., if the National Center for Health Statistics is hollowed out, states with robust health data infrastructure, like California and Massachusetts, will adapt, while rural and under-resourced states will struggle. The same pattern holds globally: If DHS fragments into country-by-country efforts, better-resourced countries will continue while others fall away, eroding data equity.

Keep the door open for government reinvestment. U.S. government funding is most likely to return under three conditions: documented cost, ready infrastructure, and political will. “It’s not just about sustaining surveys,” Mather said. “It’s about preserving the conditions that make it possible for funding to be restored.”

Using This Moment to ‘Build Something Better’

The session closed with a consensus that the future of DHS—its scope, governance, and funding model—remains unknown. “Everything could change in a moment,” Short Fabic said.

What is clear is that the current stability represents progress, not the end of the work.

“My hope and intention is to ensure that those best parts of the DHS continue, and that ideally we use this period to work together to build something better,” Short Fabic said.