IUNI Affiliate YY Ahn (IUB Informatics), along with his postdoctoral research associate Sadamori Kojaku (IUB Computer Science) and their research team, found a theoretical explanation for why “backward” tracing is more effective than “forward” tracing. The team also proposed a potential method to leverage backward tracing in practice.
Ahn, Kojaku, and their team published their recent findings in Nature Physics.
In the paper, researchers show that backward contact tracing, or tracing from whom disease spreads, is highly effective at identifying super-spreading events, which can potentially prevent a large number of transmissions if the tracing is strategically executed. Forward tracing, on the other hand, requires tracing who the disease spreads to.
Findings also indicate that, although it is difficult to identify the direction of spreading in practice, backward-aiming contact tracing can be done effectively by prioritizing those who appear multiple times during the tracing.
Controlling an epidemic such as COVID-19 requires finding and isolating infected individuals as quickly as possible—which is typically aided by contract tracing efforts.
The team says a revision of current contact-tracing strategies to leverage all forms of bias may improve the effectiveness of contact tracing. They also point out that incorporating backward and deep tracing—while ensuring privacy of individuals—may be crucial for digital contact-tracing programs to succeed.
“I think this is a good example that demonstrates the power of network perspective,” Ahn said. “By formulating the contact tracing problem in terms of contact network structure, we could formulate a clear mathematical explanation that teases out two directions of contact tracing.”
The researchers focused on the most effective contract tracing measures, finding backward tracing is far more effective in preventing mass spread of a disease than forward tracing. Backward tracing allows for leveraging the biases in the heterogeneity of the contact network, as well as a simple bias: The more infections an index case produces, the more frequently they will show up as a common contact in tracing.
Ahn and his team say effective (digital) contract tracing that can appropriately leverage backward contact tracing may be crucial for controlling any future epidemic.