Bad data complicates the Ebola outbreak

Tuesday, December 9, 2014

As an epidemiologist and someone who got my start in disease surveillance, this topic is of particular importance to me. While it is obviously terrible that the havoc wreaked by the Ebola outbreak has also hampered data collection and surveillance, I am at least encouraged to see that this is widely recognized as a problem (and making headlines, at least in the disaster news arena). From the AP:
Having accurate numbers about an outbreak is essential not only to provide a realistic picture of the epidemic, but to determine effective control strategies. Dr. Bruce Aylward, who is leading the World Health Organization's Ebola response, said it's crucial to track every single Ebola patient in West Africa to stop the outbreak and that serious gaps remain in their data.
"Decisions about prevention and treatment should be data-driven, but we really don't have the data," agreed Irwin Redlener, director of the National Center for Disaster Preparedness at Columbia University.

A week ago, the World Health Organization insisted at a media briefing it had mostly met targets to isolate 70 percent of Ebola patients and bury 70 percent of victims safely in Guinea, Liberia and Sierra Leone. But two days later, WHO backtracked and said that data inconsistencies meant they really didn't know how many patients were being isolated. Then the U.N. health agency also conceded that many of the safe burials were of people not actually killed by Ebola.
In West Africa, where health systems were already broken before Ebola struck, collecting data amid a raging outbreak has been challenging.

"Suddenly you have all these different sources of data that have to be compiled" from different aid agencies, said Ray Ransom, a data expert at the U.S. Centers for Disease Control and Prevention. "The ability to actually collect information is a different challenge than responding to the outbreak, and the energy has been focused on the response."

He said local officials are good at tracking known or suspected Ebola cases and their contacts but not as reliable relaying that information to national authorities.

The software built to track Ebola outbreaks was initially designed by the CDC to have one person entering data into a computer. That "was perfectly fine since the dawn of time up until" the outbreak exploded this summer, said Armand Sprecher, a public health specialist with Doctors Without Borders.
On a more critical note, I am pretty disappointed that the WHO would have been so confident in their initial numbers, only to have to backtrack later. Any aid worker worth their salt knows that data collection is a mess in the outbreak areas. They certainly were not going to fool anybody.