Thread by @joshuasweitz

This thread will take a few steps, and begins with a premise:

Testing for viral shedding (via PCR) is critical and we need more of it (far more), but it's not enough.

Instead, we need to implement targeted and large-scale serological surveys to test for antibodies...

1/15

Serology can tell us who has been sick & recovered, which can help (i) Measure the 'denominator' by estimating the prevalence of COVID-19 (refining case fatality rates); (ii) Serve as a resource for convalescent plasma to treat patients (see more ccpp19.org)

2/15

Serology can also identify recovered individuals who can (safely) return to work, often in critical services, as @mlipsitch, @trvrb, @NAChristakis, and others have noted, e.g.:

nytimes.com/2020/03/28/opi…)

Our model takes a step further.

3/15

We propose and analyze an epidemic intervention model in which recovered individuals serve as focal points for sustaining safer interactions via interaction substitution, i.e., to develop what we term ‘shield immunity’ at the population scale.

4/15

To do so, we develop a nonlinear population model in which the infection rate is diluted by substituting interactions between individuals of unknown status with recovered individuals (e.g., thereby preferentially replacing S-I with S-R or I-R).

5/15

As a result, the deployment of recovered individuals can lead to an amplifying, negative feedback on transmission can then be used to help curtail the outbreak, while reducing effects of social distancing.

6/15

We evaluate this effect in a simplified SIR model, finding epidemic size is reduced and peak time is shortened (where the strength of the shield immunity scales with both the number of recovered individuals and the extent to which are identified/deployed).

7/15

Unlike herd immunity (which would lead to disastrous consequences given the severity of COVID-19), the intent of shield immunity is to dilute transmissable interactions intentionally through elevating interactions with recovered, rather than depletion of susceptibles.

8/15

We then extended the basic model to include critical features of #COVID19, include age-stratified risk and both a symptomatic and asymptomatic route.

9/15

We show that shield immunity can potentially reduce fatalities and hospitalization burden, and can be used synergistically with social distancing (albeit potentially reducing the burden/level of social distancing needed)

10/15

For the full details, see here:

Code:
github.com/WeitzGroup/cov…

Paper:
github.com/WeitzGroup/cov…

SI:
github.com/WeitzGroup/cov…

11/15

We recognize there are significant challenges to implementing interventions that aim to develop population-scale shield immunity.

12/15

Nonetheless, the magnitude of the current public health and economic crisis demands action, including new approaches to reduce transmission and identify a path to move forward from shelter-in-place orders.

13/15

Serological assays have been tested for SARS-CoV-2 (see @florian_krammer & @MarionKoopmans & @bart_haagmans work - and others) prompting public calls to leverage serology to enable recovered individuals to return to the workplace, e.g.:

telegraph.co.uk/news/2020/03/2…

14/15

Our manuscript provides theory- and modeling-based evidence that we – collectively – can do more, helping not just ourselves, but each other.

For more information, see:
github.com/WeitzGroup/cov…

15/15