New preprint: Here we report quantitative prospective longitudinal testing from @nba players, staff and vendors & consider how these data can be used to improve diagnostic and surveillance testing algorithms. https://www.medrxiv.org/content/10.1101/2020.10.21.20217042v1 1/n

Testing is a key to managing the COVID-19 pandemic. But not all positive tests are the same. For example, a positive early in an infection means a greater risk of transmission than late in an infection. 2/n

Ct values from PCR tests can, in theory, reveal a person’s progress through the stages of infection https://academic.oup.com/cid/advance-article/doi/10.1093/cid/ciaa619/5841456 @michaelrtom @michaelmina_lab 3/n

And knowing the viral dynamics of an infection can provide support for smart surveillance strategies that make use of frequent, low sensitivity testing -- see https://www.nejm.org/doi/pdf/10.1056/NEJMp2025631. @danlarremore 4/n

But there’s been a problem. Tests are usually administered after a person develops symptoms, and by then viral shedding has already peaked and begun to decline. We haven’t been able to directly observe how Ct values spike during the critical initial stage of infection. 5/n

Here we address this gap. With an all-star team including @StephenKissler @HarvardChanSPH @NathanGrubaugh @JosephFauver @YaleSPH, members of @IQVIA_US, @BioReference, @QuestDX & the NBA, we collected samples from players, staff, & vendors... 6/n

... who were monitored for SARS-CoV-2 infection as part of the resumption of the 2019-20 NBA season 7/n

Due to the great work by the NBA’s occupational health team, there were no players or team staff who tested positive in the NBA’s “bubble” phase of the season’s restart. 8/n

For those with acute infection, we found that Ct values peaked within about 3 days and viral RNA shedding took longer to return to the limit of detection in those reporting symptoms compared to those who did not (~11 days vs 7 days). 9/n

Low Ct values (=high viral RNA concentrations) reliably distinguish acute infections from persistent shedding, but a single test can’t tell if viral RNA concentrations are rising or falling. 10/n

A second test within two days does a much better job, however 11/n

These findings underscore 2 main points. First, the benefit of frequent SARS-CoV-2 screening, regardless of symptoms, since infections are likely to progress from detectable to peak infectiousness within just a few days. 12/n

Second, modestly more sophisticated decision algorithms can help us extract much greater value from tests we already have available, helping us chart a patient’s course through infection, thereby improving both clinical and public health management. 13/n

Assay reliability and reproducibility are clearly important issues and more work is needed on this, but our results suggest that testing informed by understanding of virus trajectory can improve diagnostic and surveillance algorithms. 14/n

This article is a preprint and so comes with the usual caveats. The study population was predominately male and not representative of the broader population; larger studies will help confirm the findings and assess the generalizability. 15/n

Still, we anticipate that these findings will help by providing data on early infection and clarifying the link between test results and the stages of infection. 16/n

Feedback, constructive criticism, and questions are all very welcome - we want to make sure that this work is maximally useful! 17/n

What a fantastic collaboration! Many thanks again to the co-authors & all those who made this project happen. Particular thanks to the @NBA for its science-driven & careful approach and its encouragement and enthusiasm to share the findings from its amazing efforts! 18/18