Thread by @nicolamlow

People are talking about *forward* and *backward* #ContactTracing #COVID19

What do they mean?

I made a case study for the @GOARN @WHO contact tracing group

A fictional contact tracer’s tale

My slides are here:
ispmbern.github.io/covid-19/Forwa…

Short thread (1/7)

Most #ContactTracing seeks contacts from 2 days before the infected person developed symptoms

#SARSCoV2 can be transmitted before symptoms emerge.
That is looking *forwards*

Who might the index case have infected?

Contacts in quarantine soon enough won’t infect others

2/7

Forward contact tracing can break onward transmission chains.

But it doesn’t find the source who infected the index case

To find the source, we need to go *backwards*

Index cases are more likely to have been infected by a source, who also infected others in a cluster

3/7

An important characteristic of #SARSCoV2 shows why going *backwards makes sense

It doesn’t transmit randomly. Best shown here, by @jburnmordoch

https://twitter.com/jburnmurdoch/status/1311660310156066816?s=20

And shown v early on ncbi.nlm.nih.gov/pubmed/32019669

4/7

So, #BackwardContactTracing focuses on high-risk potential #superspreading events

Look at our index case's garden party

After finding a cluster, you need to trace *forward again
Needs speed, effort, communication, cross-checking records, #TestTestTest and quarantine

5/7

Mathematical modelling finds #BackwardContactTracing identifies more infected contacts per index case and reduces effective reproduction number more than #ForwardContactTracing alone: @medrxivpreprints
doi.org/10.1101/2020.0…
doi.org/10.1101/2020.0…
6/7

When #COVID19 cases are increasing rapidly, @c_drosten says «Looking back is more important than looking forward»

We should try #ClusterBusting in countries as well as Japan and South Korea

Let's evaluate it too

zeit.de/2020/33/corona…

7/7