1/ The Skagit Choir outbreak was dominated by aerosol transmission

52 out of 60 infected after 2.5 hr practice

Our Skagit Choir Paper has been accepted after peer review, available at: https://twitter.com/ShellyMBoulder/status/1306352867213631488
2/ Not isolated event. Other choir superspreading events in: Netherlands, Austria, Canada, Germany, England, South Korea, Spain, and France

We started investigating this event shortly after the LA times report by @RichReadReports
3/ Submitted in June. Aerosol transmission seemed pretty obvious, didn’t dwell on demonstrating it

We were surprised to find skepticism that aerosols led to the outbreak (including @WHO), so in revised version we added extensive description of event (in paper and Supp. Info.)
4/ In this thread we review some of the information why we think aerosols dominated

County Public Health & choir members concur that choir practice was were infection happened

Choir spokesperson: “It is not a highly social group. It is pretty seriously about the music”
5/ Known index case

“No one had contact with the index case during that week” [prior to the rehearsal]

Some of the time split into 2 groups in rehearsal, attack rate in room where index case stayed as expected

Prob. of 2nd index case: 0.3%.
6/ 3 modes of transmission: fomites (surfaces), ballistic droplets, and aerosols

Fomites general I: everyone agrees not very efficient, e.g. CDC: “not thought to be the main way the virus spreads” https://www.cdc.gov/coronavirus/2019-ncov/prevent-getting-sick/how-covid-spreads.html).
8/ Fomites in choir I:

- aware of COVID-19, no one reported any direct personal contact

- entrance door propped open, used sanitizer

- index case did not handle chairs, did not have snacks

- ½ had snacks. Many who did not have any snacks got sick
9/ Fomites in choir II:

- index went to specific bathroom that only ~3 other people used

- Many who did not use any bathroom got sick

CONCLUSION for fomites: extremely unlikely to explain more than very few cases
10/ Second mode of transmission: Ballistic droplets I:

- Not too close: 61 ppl where normally there would be 120

- Infections everywhere in the room, w/o clear pattern. Many cases behind index. **Infections up to 44 ft (13.5 m) from (& behind) index case**
11/ Ballistic droplets II:

- index case did not have anyone within the 3 m (10 ft) likely landing area for droplets emitted while singing. Droplets don’t travel backwards

- Many ppl (including index) came in just in time to start and left right at the end
12/ Ballistic droplets III:

- Only 1 x 10 min. break. Rapid transitions otherwise, not much talking

- Index case talking to others was minimal during break (& entire event). Others talked to 2-3 ppl on average during 10 min. break
13/ Ballistic droplets IV:

- CDC: need 15 min. Of close contact to get infected.

CONCLUSION for droplets: seems impossible to infect 52 people by impacting droplets on their eyes, nostrils, or mouth in this event
14/ 3rd pathway: aerosols (float in air, infect by inhalation)

- aerosol emission much higher when singing

- air was well mixed within the room because of convection due to heat emitted by ppl

- Ventilation was low

- CONCLUSION: aerosols can easily explain the outbreak
15/ Summary of transmission in Skagit Choir rehearsal:

"Per Occam’s razor, explanation most probable: inhalation of infectious respiratory aerosol from “shared air” was leading mode of transmission"

Unfortunate combo: high occupancy & duration, loud vocalization, low vent
17/ We then model event w/ box-Wells-Riley model (similar to http://tinyurl.com/covid-estimator ), conclude index case was exhaling ~1000 infect. doses / hr

Quite high: 6-16x higher restaurant or bus

Quant. consistent w/ known much higher aerosols singing loudly vs talking intermittently
18/ What could have helped limit outbreak?

- Rehearsal outdoors would have been most effective

- Masks, ventilation, portable filters, shorter duration: all help, none silver bullet. All together: 52 cases → 5 cases
19/ CONCLUSION: Skagit Choir superspreading event (52 cases) was dominated by aerosols

- Choir practices risky during COVID-19

- Mitigation: http://tinyurl.com/faqs-aerosol 

- Any feedback is welcome as always
20/ For ppl investigating outbreaks:

Most outbreak reports (eg German meat p.) *do not include* information needed to investigate aerosol spread quantitatively:
- ventilation rate, easily measurable by: https://twitter.com/jljcolorado/status/1289360404271861760
- volume of the room

Pls include in future reports
21/ Authors: @shellyMBoulder, W. Nazaroff, A. Boerstra, G. Buonanno, S. Dancer, J. Kurnitski, @linseymarr, L. Morawska, @CathNoakes & me

We thank choir & *especially* spokesperson Carolynn Comstock, & Skagit County Public Health Dept. for extensive support of our investigation
You can follow @jljcolorado.
Tip: mention @twtextapp on a Twitter thread with the keyword “unroll” to get a link to it.

Latest Threads Unrolled: