Half-baked thoughts about superspreading and contact tracing. With the existence of pre-symptomatic spread, the serial interval for this virus is quite short, making forward contact tracing challenging. 1/5

Japan’s cluster tracing approach looks backwards to find clusters. Given overdispersion, cases can often be linked back to clusters. We can then look out from clusters to find more cases. We will be too late to prevent some transmission, but could stop later generations. 2/5

My feeling is that, to be effective, the relevant time interval is not the time between generations of cases, but rather the time between generations of clusters. Control is not about preventing all secondary cases, but rather preventing the next superspreading event. 3/5

Professor Oshitani talks about the forest and the trees. “Even if we tolerate some cases go undetected, as long as we can prevent clusters where one infects many, most chains of transmissions will be dying out.” 4/5 https://www.japanpolicyforum.jp/diplomacy/pt20200605162619.html

Thoughts? Has this already been modeled? I don’t have the bandwidth to study this myself, so I am sharing with the Twitterverse in hopes that others see something here. 5/End