Correlating mitigation policy onset/lifting w/ #coronavirus transmission leads to misleading conclusions if underlying consumer behavior & societal changes aren't considered. Where's the "second wave" some like to ask? Let's take Georgia as an example:
🧵 1/ 👇
🧵 1/ 👇
Testing has been a fiasco (one of the States mixing PCR & AB results: theatlantic.com/health/archive…). Even then, as testing ramped, new cases finding has increased and positive rate has at best stayed flat recently.
2/
2/
The increase in new cases is happening despite depressed mobility even after shelter-in-place orders were lifted. Workplace, retail & recreation, transit stations activities still not back at pre-epidemic baseline. % at home also remains elevated.
3/
3/
Total time spent away from home has increased slightly from the lows in April but remains depressed, matching national trend, & slightly better than New York state.
4/
4/
Even more salient, indicator of employment rate at small businesses (number of hourly employees) remains drastically reduced (-59.5%), significantly lower than national average (-39.4%). Note, changes in Georgia were more dramatic than New York.
5/
5/
As expected, number of small businesses open & consumer spending are increasing but both remain well below pre-epidemic baseline.
6/
6/
What does it all mean? Consumer behavior changes endure although restrictions are lifted. This has significant implication for virus transmission. Therefore, more nuanced policy effectiveness analysis is needed than simple correlation of implementation dates w/ outcomes.
7/
7/
We see this in evolution of effective reproduction number (Rt). Much like mobility & consumer behavior changes preceded mitigation policies, Rt started declining soon after Emergency decree. School & business closures likely accelerated decline & kept Rt at or below ~1.
8/
8/
Also noteworthy, testing ramped much after epidemic peaked & Rt trended to ~1. What's unfortunately commonplace these days is denial of interrelated impact of changes in individual behavior & community interactions before dismissing effectiveness of mitigation measures.
9/
9/
Highly unlikely reversals in mobility & community interactions trends won't result in new risks of spread esp. when % pop. infected is low (eg, <4% for GA). The prudent choice remains taking the virus seriously & implementing precautionary measures (balancing trade offs)
10/10
10/10
