By late September 2026, the Democratic Republic of Congo had recorded 7,773 confirmed Ebola cases and 3,759 deaths β and health officials were using mobile phone records to try to stay one step ahead of the virus.
It is the first time anonymised phone data has been deployed in an Ebola response. The approach lets public health teams map population movement and prepare high-risk areas before cases are detected there.
The Fastest-Growing Ebola Outbreak on Record
The WHO declared the outbreak on 15 May 2026 in Ituri province. By 21 September, confirmed cases had spread across seven provinces. The WHO has described it as the fastest-growing Ebola outbreak ever recorded.
The strain driving it is the Bundibugyo virus β a variant with no approved vaccine or treatment. That makes early detection not just helpful, but essential. And that is exactly where the phone data comes in.
How the Phone Data Actually Works
Every time a mobile phone connects to a network, the antenna it links to logs a record. As people travel, their phones ping different antennas across different locations. Researchers stitch those records together to estimate how populations move between areas.
This is not individual tracking. The system reveals how groups of people move β not any single person's location. Swedish non-profit Flowminder, which specialises in population mobility data, is the organisation processing the records for the DRC response.
Vodacom β the DRC's largest mobile operator with more than 26 million active subscriptions and a market share of nearly 35% β is providing the data free of charge, according to Reuters. Flowminder says it receives only aggregated, anonymised records with no access to any information that could identify individual subscribers.
What the Data Found β and When
Flowminder's first analysis came out in early June 2026. It focused on Bunia, Mongbwalu, and Rwampara β the areas believed to be the original epicentre. Researchers tracked where people who had been in those zones between 3 and 23 April subsequently travelled.
The results were striking. By the end of June, every single one of the 10 destinations receiving the highest volumes of travellers from those zones had reported confirmed Ebola cases.
"Estimates of how people move help predict how infectious people move," said Linus Bengtsson, founder of Flowminder.
Traditional outbreak risk models assume a virus spreads mainly to areas physically adjacent to a hotspot. Mobility data challenges that assumption. Olivier Le Polain of the WHO Health Emergencies Programme points to Ituri as a clear example β the province's mining and trade industries drive large-scale population movements, meaning travel links can be stronger between distant towns than between nearby ones.
The Coverage Gap Problem
The method is not without limits. The entire analysis currently rests on data from a single operator. Vodacom's 35% market share means the mobility picture being built is incomplete β anyone on a rival network is invisible to the model. In a country as vast and sparsely connected as the DRC, that is a significant blind spot.
Vodacom is working with Orange on solar-powered base stations for rural parts of the DRC, which could eventually extend coverage β and improve the quality of future analyses. But right now, those gaps remain a real constraint on what the data can reliably show.
Building on What Came Before
Mobile tools have featured in Ebola responses before. During the West African epidemic of 2014β2016, digital health technology helped responders through mobile reporting apps and free public information services. The DRC's current approach adds network-level movement data to that toolkit β a meaningful upgrade in predictive capability.
The WHO's own figures from 23 September show a 26% drop in cases in Ituri over a 21-day period, even as North Kivu recorded a 73% rise over the same window. Whether mobility data helped contain Ituri or flag North Kivu's risk early is still being assessed. But the fact that all 10 predicted high-risk destinations confirmed cases suggests the method is doing something right.