Congo's Ministry of Health counted more than 6,250 confirmed Ebola infections and 3,039 deaths on Thursday, according to government data reported by Reuters from Dakar. [1] That is 64 cases and 32 deaths above the ministry table this paper carried Wednesday. The outbreak, Congo's 17th, is Bundibugyo. There is still no approved vaccine or specific treatment. [1] Death-toll copy will sell 3,000 again. The operational fact is a method: anonymised mobile-phone records, not used before in an Ebola response, to map where people move when the tracing boards cannot see most new infections.
Wednesday's file held 6,186 cases and 3,007 deaths and put the CDC fraction first: most new confirmed cases were coming from only 15 to 20 percent of known contacts. About four in five sat outside the chains responders already had listed. Thursday moves that ministry table by about 30 deaths. It does not reconcile it with Monday's WHO table of 6,041 and 2,911. Keep those ledgers labeled. Do not add Africa CDC's "possibly three times larger" claim onto 6,250. That remains an undercount warning, not a third census.
Olivier Le Polain, head of epidemiology and analytics for response at the WHO Health Emergencies Programme, said population movements have always mattered in outbreaks and matter more in this one. [1] Traditional risk maps assume the virus will spread into the next district. Mobility records can flag a place that is geographically distant and tightly connected by travel. "It gives us a more nuanced understanding of risk," he said, pointing to Ituri, where mining and trade drive large movements and where the virus is believed to have started circulating. [1]
The analysis is produced by the Swedish nonprofit Flowminder from records generated when phones connect to different antennas. Flowminder founder Linus Bengtsson told Reuters the group estimates how people move, "which in turn is a predictor of how infectious people move." [1] It receives aggregated, anonymised records and does not identify subscribers. Vodacom, Congo's largest operator, provides the data free. Vodacom did not respond to Reuters. The company has more than 26 million active subscriptions and nearly 35 percent of the market. [1]
Flowminder's first analysis, in early June, followed people who had spent time in Bunia, Mongbwalu, and Rwampara between April 3 and April 23. The largest flows went to nearby areas in Ituri and North Kivu. By the end of June, all 10 destinations receiving the largest flows from those original outbreak areas had confirmed Ebola cases. [1] That is a map catching up with a virus, not a new headcount.
Kisangani, a Congo River hub between east and west, drew particular attention. Le Polain said the risk was not a surprise even though the city did not border the first affected areas. "By mid-June, the data showed the risk was significant and that the outbreak could spread there faster than we initially thought. We strengthened operations there before the first cases were detected." [1] Later work tied Kisangani to Kinshasa, to river settlements, and to areas near the Central African Republic and South Sudan borders. Fifteen of the 50 areas most connected to Kisangani sit in Kinshasa province. A WHO modelling exercise presented in late June put a 70 percent probability on a Kinshasa detection by the end of September. [1] That is a probability on a calendar. It is not a case in the capital today.
The coverage gaps are the method's honest limit. The analysis uses Congolese networks only, so international borders are a blind spot. It also uses a single operator. Rival networks are invisible. Flowminder and U.N. officials said they are trying to add other operators. [1] A phone map built on 35 percent of subscriptions is a proxy. It is not a census of the four-fifths of new infections that still sit off the contact lists.
WHO's Africa office, filing from Nyankunde on Thursday, described a different instrument: talk. At least 32 community dialogues have brought together more than 1,500 leaders in that Ituri health zone. Community alerts that once stayed at about 11 a day now reach 60 to 70, said Dieudonné Ramazani, a WHO risk-communication expert there. [2] No attacks on response teams have been reported since the Ebola treatment centre near the general hospital was assaulted on July 15 and 16. [2] Alerts per day are not incidence. They are a measure of whether people will pick up the phone. The phone-mobility map exists because the tracing fraction still cannot see most of the epidemic. The ministry table moved. The method is the news.
-- KENJI NAKAMURA, Tokyo