This article is more than 5 years old. Information may no longer be current.
News reports on Ebola outbreak correspond with surveillance data
Click Here to Manage Email Alerts
A recent study showed that news reports on the Ebola outbreak corresponded with records from epidemiological surveillance data, suggesting that authoritative sources can reliably predict outbreak patterns that may otherwise be hard to assess in real-time.
Gerardo Chowell, PhD, associate professor of epidemiology and biostatistics at Georgia State University School of Public Health, and colleagues reviewed news reports on 286 suspected, probable and confirmed cases of Ebola in Guinea, Sierra Leone and Liberia that were published by The New York Times (21 articles, two videos), The Washington Post (six articles), EbolaDeeply (six articles) — an independent digital media project focused on Ebola — and WHO (11 articles, five situation reports) from March 2014 to March 2015.
Gerardo Chowell
A higher proportion of cases were reported for Liberia (49%) compared with Sierra Leone (28.3%) and Guinea (22.7%), reflecting an American-centric bias toward a country strongly tied to the United States, the researchers wrote. They extracted demographic, epidemiologic and clinical data on each patient in these reports and the nature of their exposure.
The analysis included 79 Ebola virus disease clusters that ranged from one to 33 cases (mean cluster size = 4.3; 95% CI, 3.2-5.3). The estimated case fatality rate was 74.4% (95% CI, 68.3-79.8), which persisted throughout the epidemic and corresponded with reports from the WHO Ebola Response Team.
Based on news reports, a majority of patients were exposed to the Ebola virus through family members (57.3%) and health care settings (18.2%). Hospital transmission was greatest in Liberia (24.5%), followed by Sierra Leone (13.6%) and Guinea (9.1%; P < .04). Funeral transmission was more frequent in Sierra Leone (27.3%), followed by Guinea (18.2%) and Liberia (1.8%; P < .0001).
Overall, the most frequent source of transmission during the epidemic was family exposure (57%), which was similar across the regions.
A breakdown of monthly cases in 2014 correlated with WHO’s reported outbreak trajectory (P = .026) and showed that funeral exposure (60%) was the main source of transmission in April, but was replaced by hospital transmission (70%) in June and July. By August, both funeral and hospital transmission rates declined, likely from enhanced infection control protocols in health care settings and local educational campaigns, according to the researchers.
Thirty-nine individual reproduction numbers were identified based on transmission chains. A negative binomial distribution model yielded a mean reproduction number of 2.3 (95% CI, 1.8-2.7).
“Our mean estimate of the reproduction number based on cluster data lies toward the higher end of the range of previously published estimates around 1.5-2.5,” the researchers wrote. “This is likely explained as a reporting bias, that news reports favor larger clusters.”
In addition, the sample of data collected from news stories had a higher proportion of infected health care workers compared with official WHO statistics (16.1% vs. 3.5%).
“Other potential biases might exist toward writing survivor stories or sensational articles,” the researchers wrote.
However, news reports, epidemiological surveillance data and data from previous outbreaks were in “close agreement” regarding the regional estimate of case fatality rates and the predominance of hospital and funeral transmission prior to enhanced interventions, which led the researchers to conclude that unstructured news reports from trustworthy sources may be capable of assessing epidemiological patterns of emerging infectious disease threats. – by Stephanie Viguers
Disclosure: The researchers report no relevant financial disclosures.
Click Here to Manage Email Alerts