Read on Twitter
1/ New paper! "Social distancing to slow the U.S. #Covid_19 epidemic", led by Mark Siedner ( @MGHMedicine @MonganInstitute @GlobalHealthMGH @mgh_id @AHRI_News) & @harlingg ( @UCLGlobalHealth & @AHRI_News), w/ Zahra Reynolds, Rebecca Gilbert, & @atheendar ( @PennMEHP @PennLDI)
2/ Social distancing is a mainstay of nonpharmaceutical public health intervention in epidemic settings. Most of you recognize this #FlattentheCurve figure from the @PNASNews paper by Richard Hatchett, Carter Mecher and Mark Lipsitch ( @mlipsitch): https://www.pnas.org/content/104/18/7582
3/ For more about social distancing, see this thread (among many, many informative threads on #Covid_19) by @NAChristakis: https://twitter.com/NAChristakis/status/1234815825430949891?s=20
4/ If you lift social distancing too early, then infections start coming back. Another figure you probably recognize is this @JAMA_current paper by @HowardMarkel et al.: https://jamanetwork.com/journals/jama/fullarticle/208354
5/ Given that @realDonaldTrump is itching to get the U.S. "opened up and just raring to go", we wanted to assess the extent to which social distancing has restrained the growth of the #Covid_19 epidemic. (We started shortly after this disastrous interview) https://www.politico.com/news/2020/03/24/trump-wants-to-restart-economy-by-mid-april-146398
6/ Astoundingly-- Zahra Reynolds and Rebecca Gilbert collected the social distancing data in a few DAYS. That's a lot of trawling of state govt web sites. During March 10–27, all 50 states and the District of Columbia implemented at least one statewide social distancing measure.
7/ These measures included executive orders, state of emergency declarations, restrictions on travel, school closures, etc. Binding, statewide measures were included. Non-binding measures (eg, "recommendations") & lower-level jurisdiction actions (eg, county) were _not_ included.
8/ We merged our data on social distancing policies with the @NYTimes #Covid_19 data, which they helpfully posted on @github: https://github.com/nytimes/covid-19-data
9/ The first social distancing measures were implemented when the median size of the epidemic, across 50 states and D.C., was 36 cases (IQR, 17-72). The mean daily case growth rate at these times was 30.6%, corresponding to a doubling of total cases every 3.3 days.
10/ Here is our primary results slide: mean #Covid_19 daily case growth rate mapped against the date of the first statewide distancing measures:
11/ This figure shows that before implementation, mean daily case growth was relatively constant. Beginning 4 days after implementation (ie, accounting for incubation period), the growth rate declined by 0.8%/day. By day 14, doubling time had increased (from 3.3 days) to 5 days.
12/ Our estimate of the reduction in mean daily case growth corresponds to a reduction in total #Covid_19 cases from 26,356 to 23,266 at 7d after implementation of social distancing, and from 156,360 to 88,105 at 14d. (Obviously we can't be that precise)
.
COMPOUND INTEREST
.
COMPOUND INTEREST
13/ Fewer jurisdictions (n=30 as of Mar 30) have implemented restrictions on internal movement (ie, "lockdowns"). Generally these were enacted later, a median of 10d after the first social distancing measures (IQR 8-12), and at a median epidemic size of 415 cases (IQR, 147-1172)
14/ The mean daily case growth rate had _already_ been declining at this point. There was no additional decline in mean daily case growth after implementation of statewide restrictions on internal movement ("lockdowns"):
15/ We ********** do not ********** think that the take-home message of the previous figure is that "lockdowns don't work". It is very difficult to identify the independent effect of lockdowns. Probably some people who are smarter than us can figure that out.
16/ The effects we estimate are consistent with what can theoretically be expected to result from social distancing and other behavioral interventions-- such as described in this @medRxiv preprint by @thimui_pham and a team from @juliuscenter @UMCUtrecht: https://www.medrxiv.org/content/10.1101/2020.03.12.20034827v2
17/ Our estimates are also consistent with what has historically been observed (cf. the Hatchett/Mecher/ @mlipsitch paper referenced at the beginning of this thread).
18/ LOTS of limitations: social distancing intensified as epidemic worsened, some counties initiated social distancing before state govts, states initiated lockdowns after other social distancing measures, SUTVA violations, incr'd surveillance & testing, less "post" obs time...
19/ I'll conclude with a word from @NIAIDNews director Dr Anthony Fauci: "If you look at what's going on in this country, I just don’t understand why we're not doing that [ie., implementing a federally mandated stay at home order]" https://twitter.com/AC360/status/1245875533113671685?s=20
20/ That's all for now. I'll wrap up the thread and post the medRxiv link once it's live!
