Phylodynamic inference of the contribution of transmission routes in infectious disease outbreaks
Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Understanding which types of contact drive pathogen transmission is critical for outbreak control. Traditionally, inference of transmission events with contact tracing or network analyses is used to identify high-risk contacts. Inference based on genetic data might offer higher accuracy on identifying the true events, but is rarely used to evaluate the contribution of different types of contact to transmission. We extended the Bayesian phylodynamic model phybreak to incorporate data on different types of contacts between cases. The model estimates the fraction of transmission events attributable to different types of contact by combining pathogen genetic sequences, sampling times, and contact data. We evaluated performance through simulations across various scenarios — different prevalences of contact types, independent versus correlated types — and applied the method to the 2020 SARS-CoV-2 outbreak in Dutch mink farms. Simulations showed that when contacts of a specific type were sparse but important for transmission, the model accurately estimated the number of transmissions attributable to that type of contact. Performance declined with more prevalent or more positively correlated contact types. Contact data improved transmission tree inference, particularly under conditions where genetic data alone lacked resolution. When applying the method to the SARS-CoV-2 outbreak, we estimated that in 76% of all transmission events between farms linked via shared personnel, transmission occurred through a link of this type. Veterinary service providers and feed suppliers were less strongly associated with transmission. This method allows for simultaneous inference of transmission trees and quantification of the importance of contact types using both genetic and structured contact data. It improves accuracy of inferring who infected whom under realistic conditions and is applicable to a range of outbreak settings. These findings can inform intervention and surveillance strategies in both human and animal populations.
Author summary
In outbreaks of infectious diseases, there are often many possible ways a pathogen can be transmitted — through direct contact, e.g. skin-to-skin contact, or indirect contact, e.g. shared health-care workers, service providers or environments. However, it is hard to know which of these connections actually matter most for transmission, although this can be crucial to stop the spread of the disease. In this study, we created a method that uses both genetic data from a pathogen and records of contacts between cases to investigate which connections contributed most to spreading a disease.
We tested our approach using simulated outbreaks and applied it to the spread of COVID-19 among Dutch mink farms. We found that our method worked well on simulated outbreaks, when the contacts were not too common. In the example of COVID-19 in mink farms, we found that shared personnel was the most risky type of contact, more so than shared feed suppliers or veterinary service providers. This insight can help authorities to focus their intervention efforts on the most risky contacts in future outbreaks. Although we have shown an example of an outbreak among animals, our method can also be used for human infectious disease outbreaks.
