UC Berkeley Press Release
Small groups of superspreaders lead to most infections, new study says
BERKELEY – A detailed analysis of eight deadly diseases shows that a small subset of particularly infectious people can exert a powerful influence over how outbreaks progress.
A research team led by the University of California, Berkeley, found that diseases such as Severe Acute Respiratory Syndrome (SARS) and measles are prone to "superspreading events" in which a few people, given the right conditions, can ignite explosive epidemics. However, the researchers say that such volatility also means that outbreaks are more likely to fizzle out relatively quickly.
The findings will be published Thursday, Nov. 17, in the journal Nature.
(Image courtesy of James Lloyd-Smith, UC Berkeley)
"From Typhoid Mary to SARS, it has long been known that some people spread disease more than others," said James Lloyd-Smith, a post-doctoral researcher at UC Berkeley's Department of Environmental Science, Policy and Management and lead author of the study. "However, we lacked a way to measure these differences reliably for diseases spread by casual contact, so the impact of individual variation in infectiousness hadn't been clear. What this study provides is a first look at the extent of this variation for a handful of important diseases, and a model that helps us understand how this affects disease outbreaks."
There were too many uncertainties in available data for avian flu to be included in this study, but the findings could still have direct bearing on the unfolding avian flu outbreaks worldwide, the researchers said.
"For many diseases, we found that the proportion of infected people that do not infect anyone else is higher than previously expected, suggesting that health officials should not be lulled into complacency by an absence of flu transmission events," said Wayne Getz, UC Berkeley professor in the Department of Environmental Science, Policy and Management and principal investigator of the paper.
As for what characterizes a superspreader, the researchers say it depends upon the disease and other factors. The researchers said an individual is at greater risk of being a superspreader if her or she has a job that brings him or her into frequent, close contact with a large number of people.
Health professionals involved in direct patient care, for instance, are at greater risk for both contracting and spreading certain diseases. This was illustrated all too clearly with the 2003 SARS outbreak that was traced to a Hong Kong hotel where an infected physician from China stayed.
"In that case, the index patient was a physician who had contracted the infection from a patient in China, and then traveled to a hotel that brought him into contact with many other travelers," said Lloyd-Smith. "He infected at least 10 people from six different countries, but what exactly made him so infectious is unknown."
In addition to occupation, people with compromised immune systems or who have concurrent infections - particularly ones that make a person cough or sneeze - may help transmit diseases more readily. Another common theme involves patients who are misdiagnosed or who do not seek treatment, so they go on infecting others for a longer time.
Methods used to treat an infected patient can also create a superspreading event. With the example of SARS, procedures such as inserting airway tubes or administering medication in vapor form through a nebulizer may have inadvertently helped the disease to spread by aerosolizing the virus.
"The factors that make an individual a superspreader are going to vary from one disease to another, and may have both genetic and behavioral components," said Getz. "Behavioral components can be quite obvious, such as promiscuity in the case of sexually transmitted diseases, while genetic components are more subtle and could relate, for example, to the morphology of an infected person's respiratory tract in the spread of the common cold or influenza."
Prior models of epidemics had assumed a steady rate of growth averaged across a population, precisely because individual differences had been difficult to measure. To help resolve this, the researchers developed statistical methods that take into account variations in infectious histories of individuals, properties of the pathogen and environmental circumstances.
They analyzed outbreak data from eight infectious diseases, ranging from SARS and smallpox to measles and pneumonic plague. The outbreaks they analyzed all occurred within the past 60 years. Instead of looking at the population curve, the researchers worked with data at the individual level to obtain a more detailed view of the infection's progress.
They noticed "asymmetrical" infection patterns in which the majority of infected people do not go on to infect anyone else, and the disease often dies out on its own without signaling public health surveillance systems.
"For diseases like SARS, major outbreaks occur when the disease hits the jackpot by infecting a superspreader," said Lloyd-Smith. "A superspreader could go on to infect 10, 20 or even more people if the conditions are right. In one extraordinary case, a sailor transmitted measles to about 250 other people in Greenland."
The researchers say the study, which quantifies what many health professionals have suspected, has implications for emerging disease surveillance and control.
"During the start of an epidemic, public health agencies need to pay the same attention to estimating the proportion of patients that are not transmitting the disease as they do to counting the cases of those that transmit to many," said Getz. "This will facilitate early identification of factors associated with superspreaders with the efforts focused on eliminating superspreading events."
"This study shows that public health agencies need surge capacity," Lloyd-Smith added. "They need to be equipped to ramp up their services quickly for times when a rapid flare-up occurs."
Other co-authors of the paper are Sebastian Schreiber, associate professor of mathematics at the College of William and Mary in Virginia; and P. Ekkehard Kopp, professor of mathematics at the University of Hull in Great Britain.
This research was supported by the National Institutes of Health, the James S. McDonnell Foundation, the National Science Foundation and the South African Centre for Epidemiological Modelling and Analysis.