PCR-pooling

Estimating the efficacy of PCR pooling in boosting detection of Covid-19 infectious.

Interactive Dashboard implements algorithm from (Hanel and Thurner, 2020, preprint) (https://arxiv.org/pdf/2003.09944.pdf).

Problem statement:

• Extensive population screening to find Covid-19 carriers is an asset against the epidemic.
• Testing capacity is currently a bottleneck.
• Pooling samples for PCR has been suggested to boost testing capacity.
• Is there an optimal sample size to minimise the tests necessary to identify a person as positive? What does it depend on?

Proposed solution:

• (Hanel and Thurner, 2020, preprint) have suggested a statistical method to gauge the optimal sample size and to quantify the number of missed infectious.
• We implement an interactive dashboard to tune the method to different regional cohorts.

Variables:

• Single test false positives rate.
• Single test false negatives rate.
• Infected fraction of the population. In principle, this should also consider a rough estimation of not-yet-found carriers. As a rule of thumb, this number may be 5 to 10 times bigger than official reports.

Outputs:

• Best sample size per test.
• Gain (persons per test).
• Estimated maximal number of missed infectious, per test.

Default values are for Luxembourg.

Author: Daniele Proverbio; Date: 31/03/2020; Affiliation: LCSB, University of Luxembourg