Read on Twitter
ALOHA may be in trouble
Start. Aug 1 you can go to Hawaii w/out 14d quarantine
Just $150 to get #COVID19 test 72 hrs before flight & if negative & no fever you're good or $200 in Hawaii & wait until neg result
A challenge & a partial solution for Hawaii https://www.staradvertiser.com/2020/06/24/breaking-news/watch-live-gov-david-ige-discusses-trans-pacific-travel-to-hawaii/
Start. Aug 1 you can go to Hawaii w/out 14d quarantine
Just $150 to get #COVID19 test 72 hrs before flight & if negative & no fever you're good or $200 in Hawaii & wait until neg result
A challenge & a partial solution for Hawaii https://www.staradvertiser.com/2020/06/24/breaking-news/watch-live-gov-david-ige-discusses-trans-pacific-travel-to-hawaii/
Background: #COVID19 has crushed tourism to Hawaii (visitors down 99.5%) and with it a huge part of their economy, despite being relatively lucky so far in terms of impact (although it's not looking good recently).
https://www.nytimes.com/interactive/2020/us/hawaii-coronavirus-cases.html
https://www.nytimes.com/interactive/2020/us/hawaii-coronavirus-cases.html
To stop COVID-infected travelers Hawaii previously instituted a mandatory 14d quarantine in your hotel for visitors. This understandably dissuaded most travelers. Why go to Hawaii and sit in a room for 14d before getting to enjoy the ocean, beach, luaus etc.?
To encourage more tourism and revive the economy Hawaii will try to use testing & fever screening upon arrival to allow visitors to skip the 14d quarantine.
Why is that likely to miss a substantial fraction of infected travelers?
Why is that likely to miss a substantial fraction of infected travelers?
Most people following me will already know the answer: a COVID test w/out any symptoms can't rule out infection:
https://twitter.com/DiseaseEcology/status/1270093309126438912
Best example of this issue in action was failed attempt to screen 640 new army recruits leading to huge outbreak:
https://connectingvets.radio.com/articles/fort-benning-confirms-142-covid-19-cases-in-2-battalions
https://twitter.com/DiseaseEcology/status/1270093309126438912
Best example of this issue in action was failed attempt to screen 640 new army recruits leading to huge outbreak:
https://connectingvets.radio.com/articles/fort-benning-confirms-142-covid-19-cases-in-2-battalions
What is (partial) solution?
One reasonable option that would increase cost/hassle a bit but would be worth it from a public health point of view:
A 2nd test to catch those missed by 1st test;
However, exact timing of 2nd test isn't simple.
2 scenarios:
One reasonable option that would increase cost/hassle a bit but would be worth it from a public health point of view:
A 2nd test to catch those missed by 1st test;
However, exact timing of 2nd test isn't simple.
2 scenarios:
To explain, I've made figures (avg incub period 5.5d & infectiousness starts 2.3d before symptoms; sources https://twitter.com/DiseaseEcology/status/1271281847754846211). I assume that if infectiousness is >0, you will test positive by swab b/c high viral load. Thread provides solid evidence for this assumption.
(We *might* be able to find RNA w/ swab test before you are infectious (b/w days 0-3.2 after infection in figure) but this is huge unknown - no data exist to support this assumption yet, and it might be a *very* short period during initial rapid viral growth.)
Sc1) Infection occurs b/w 1st test&flight: an obvious challenge. Testing day of flight would miss infection (viral load too low), but 3d after flight could catch infection & early (blue Test 2)
10d after flight (green Test 3) would also be + but it'd be too late to stop spreading
10d after flight (green Test 3) would also be + but it'd be too late to stop spreading
Sc 2: Infection occurs BEFORE test but is missed by test due to low virus amount in respiratory system. Test 2, 0-1d after flight, would catch people w/ avg symptom onset & infectiousness timing (fig: top). But what about people w/ longer incubation periods (13d: fig, bottom)?
If viral load is low for 10d (A BIG UNKNOWN) (bot. fig) then Test 2 would miss them. Testing again a week later (Test 3) would catch these people, but testing everyone this late is too late to catch people w/ early or avg incub period (top).
Luckily the incubation period is not uniformly distributed, and many people have incubation periods near average, so long periods are relatively rare (but do occur!).
https://www.medrxiv.org/content/10.1101/2020.06.20.20134387v1
https://www.medrxiv.org/content/10.1101/2020.06.20.20134387v1
Conclusion: Hawaii's current plan to test 72 hrs before flight will miss substantial fraction of infected travelers. Adding 2nd test 2-4 days after flight would catch many of these missed infections. Can't catch them all w/out 3+ tests, but a 2nd test would help a bunch!
Note: all the scenarios here apply to any attempt to use screening to find infected people before or after events (start of classes, sporting event, protest).
Data on sensitivity of testing before symptom onset, *especially* for people w/ longer incubation periods would help!
Data on sensitivity of testing before symptom onset, *especially* for people w/ longer incubation periods would help!
Thanks to many people I've discussed this broader issue with, including @sarahzhang @jLewnard @CT_Bergstrom @joannamasel @mlipsitch @nataliexdean @aetiology @bencowling88 @bennyborremans @jlloydsmith
A follow-up: It looks like one of the islands in Hawaii, Kauai is going to use 2 tests separated by 6d (arrival and +6d) w/ quarantine b/w them in designated hotel, & acknowledges this will still miss a few cases. Great stuff Kauai! https://www.civilbeat.org/2020/06/could-new-kauai-proposal-be-a-model-for-bringing-back-tourism/
