Poster award | May 15, 2012
How influenza pandemics spread
© CDC / Cynthia Goldsmith
MPIDR-Researchers Carlo Giovanni Camarda and Jim Oeppen have been awarded a prize at the annual conference of the Population Association of America (PAA) for their poster showing a new approach to the analysis of influenza pandemics.
In the scientific project they presented at the conference, which took place from May 3-5 in San Francisco, the two scientists from the Laboratory of Survival and Longevity described a new technique to estimate the “Reproductive Number” of a pandemic. This number is an important parameter for epidemiologists as it indicates how many new cases will be caused by a person already infected. When the number is greater than one the pandemic spreads through the population like a tree growing many new branches. Intervention measures are designed to get the number below one which makes the pandemic die out. If the number is low, closing schools might be enough to get it below one by lowering the contact rate. Health authorities can also use the number in a formula to calculate the percentage of a population that would need to be vaccinated to prevent a future pandemic.
The novelty of their method is that it allows for the fact that new influenza cases that occur on weekends and holidays are frequently reported on the following days, leading to a weekly cycle of under- and over-counting that hides the true pattern. The new approach simultaneously corrects the data and updates the estimate of the Reproductive Number.
Applying the method to daily data from the Russian influenza pandemic of 1889/90 in England and Wales and the 1918 Spanish influenza pandemic in Chicago, the scientists show that failure to allow for the undercount on weekends leads to inaccurate estimates of the Reproductive Number. For Chicago in 1918 the raw data lead to an estimate of 4.3, which falls to 3.3 when the data are corrected. The researchers also showed that the convention of using a single number to describe the growth phase of a pandemic is misleading because the parameter changes before the pandemic reaches its peak.
The influenza pandemic that exploded in the autumn of 1918 is thought to have killed more than 50 million people worldwide – perhaps as many as 100 million. This research is part of the international scientific effort to understand past pandemics and try to mitigate the effects of future ones.
"Estimating the Reproductive Number of an Influenza Epidemic from data with Day-of-the-Week reporting biases", Poster, PAA 2012, Carlo Giovanni Camarda and Jim Oeppen.