COVID-19 Model Shows Travel Restrictions Work, But Testing and Isolation Key
A mathematical model based on 4 publicly available data sets finds travel restrictions cut the rate of transmission in Wuhan, China. It also found that it likely takes several cases in a given location before the virus sparks an outbreak.
New research based on mathematical modeling of SARS-CoV-2 transmission rates finds transmission declined by approximately half when China’s Wuhan province introduced travel restrictions. The study also highlights the importance of testing and isolation as a defense against the spread of the virus.
Investigators from the London School of Hygiene and Tropical Medicine wanted to estimate transmission rates of the novel coronavirus, as well as the variance of transmission rates over time. To do so, they used a stochastic transmission dynamic model to process 4 publicly available data sets from Wuhan province, as well as data from internationally exported cases of COVID-19. The results were published last week in The Lancet.
The team found that the median daily reproduction number of the virus dropped from 2.35 down to 1.05 within the week following the introduction of strict travel restrictions in Wuhan on January 23rd.
They also found that when the virus is introduced to new locations, the first case or 2 don’t necessarily lead to outbreaks. Using existing SARS data related to the variability of transmission, they calculated that once there are 4 independently introduced cases in a given location there is a 50% likelihood of an outbreak.
“Even if the reproduction number is as high as in Wuhan in early January, it could take several introductions for an outbreak to establish, because high individual-level variation in transmission makes new chains of transmission more fragile, and hence it becomes less likely that a single infection will generate an outbreak,” writes corresponding author Adam J. Kucharski, PhD, of the school’s Centre for Mathematical Modelling of Infectious Diseases, and colleagues.
They say this suggests early identification of patients with the illness, followed by careful isolation, can be a strong strategy against transmission.
The authors note that there is a high degree of difficulty in predicting a novel virus in real time.
Data might be incomplete or inaccurate, and there can be a significant lag between the time of infection and the onset of symptoms. However, the mathematical models use varying techniques to attempt to offset such uncertainties.
As more data have come out, Kucharski and colleagues say the number of cases in Wuhan in February was much lower than the model predicted, though that might have to do with a lack of data. Conversely, while the authors say most of their estimates of international transmission have been relatively accurate, the model appears to have under-estimated the number of cases that would be exported to the United States, France, and Australia.
“Once extensive travel restrictions are introduced, as they were in Wuhan, the signal from such data gets substantially weaker,” Kucharski and colleagues note. “If restrictions and subsequent delays in detection of cases are not accounted for, this could lead to artificially low estimates of [median daily reproduction] or inferred case totals from the apparently declining numbers of exported cases.”
The authors conclude that transmission rates likely vary substantially over time with SARS-CoV-2. As the virus spreads, they write that data and modeling will be highly important to understand which control measures work, and how well they work, in different settings.