the Harvard test-trace-isolate plan

This WaPo story is how I became aware of this release of this Pandemic Resilience: Getting it Done report from May 12th.  As it's a 30+ page PDF, I have extracted all of the best bits (in my personal opinion) in <30 quotes below. Hope this helps!

"Low prevalence areas (“green zones”) should focus on surveillance and therapeutic testing. Moderate-prevalence areas (“yellow zones”) should aggressively use TTSI to suppress COVID-19 over the next one to two months and can do this even with the economy fully open. High-prevalence areas (“red zones”) should focus over the next two months on mitigation testing and tracing strategies, targeted especially at their most vulnerable populations, while also maintaining stay-at-home orders and advisories."

"The policy-making landscape has converged around the view that massively scaled up testing, tracing, and supported isolation programs (TTSI), as well as local infection control, should be at the core of the nation’s pandemic response prior to the arrival of a vaccine or major therapeutic breakthrough. The goal of ramping up TTSI programs is to reopen the economy safely and sustainably and to stay open, building an infrastructure of resilience that can withstand repeated further waves of COVID."

"A national strategy of massively scaled up TTSI would replace blanket stay-at-home orders with targeted isolation and would save both lives and livelihoods."

Based on South Korea's data, "the necessary number of tests to detect positive cases and suppress all transmissions ... suggests about ~25 tests conducted to detect one positive case" -- so, 4% positive rate

"Green zone communities have few active infections (a current infection prevalence of less than roughly one per 36,000). Yellow zone communities have known active infections at a small scale (less than 1% prevalence of active virus in the population). Red zone communities have an outbreak which public health authorities have not been able to suppress (1% or higher prevalence of virus in the population)."

"Our core recommendation is that the nation use an ambitious surge of testing and tracing this summer to pursue a near-term (2 to 3 months) national target of bringing the reproduction number of the virus down to .75 (R = .75) and a long-term target (12 to 18 months) of consistent suppression to produce a context in which outbreaks will be easily containable."

Test and trace goals: "GREEN ZONES: one test per day for every 10,000 people; one five-person contact tracing team for every 100,000 people. YELLOW AND RED ZONES: testing capacity at the level of 2500 tests x (deaths/day), 60 teams of five tracers for every death/day in the locale.
These are aggressive targets whose goals are to clear out the disease as quickly as possible during a surge period of testing."

"A basic lesson of TTSI programs is that the earlier the investment, the less expensive it will be."

Leaders in green zone: "I need to make sure my state decision-makers know that keeping all green areas green is the #1 priority and that doing this will take only a small percentage of our testing capacity."

Leaders in yellow zone: "If my testing allocation is lower than <described above>, I should keep less of my economy open, pretty much in straight percentage terms to my testing allocation. I need to make sure my state decision-makers know that turning all yellow zones green is their number #2 priority. The more aggressive I can be up front, the less I will have to invest in aggregate."

Leaders in red zone: "I should keep only essential services open ... If I can maintain <described above> testing levels and contact tracing teams, I can open my economy fully when I get to yellow or prevalence below 1%.”

"If total national capacity hits 5 million tests a day, then we could cover the testing needs of all three categories of locales simultaneously." (otherwise Red takes third priority, as tests are least helpful there)

"If you live in a low prevalence community, your community needs to be testing only 1 out of every 10,000 people per day.
If you live in a moderate or high prevalence locale, your community wants to find as many cases as possible. To do this, public health officials can use contact tracing. This means finding the contacts of infected person so those contacts can get a test."

"If and only if your community is testing broadly and equitably, then it is also useful to look at the positivity rate. This is the percentage of tests administered each day that return a positive result. The lower the positivity rate, the better the testing program is doing. "

"If my town has equitable access to testing but the positivity rate is above 12%, public authorities may not be seeing enough cases to know if there is a fresh outbreak or new surge emerging or they may be missing the spread of the disease in parts of our community.
If my town has equitable access to testing but the positivity rate positivity rate is above 10%, public authorities will have less time than would be ideal to respond to a surge.
If my town has equitable access to testing but the positivity rate above is 4%, public authorities will probably have to use stay-at-home orders again to control a new surge.
If my town has equitable access to testing and the positivity rate is only 4%, we are using testing in the most robust way possible to suppress covid-19."

"broad quarantine ... is insufficiently targeted to interrupt the actual chain of transmission. Nonetheless, it is a valuable tool for mitigation in contexts of scarcity of testing resources. "

"the approximate rule “5 contact tracers per new case per day.” The reason for this is that 1 tracer needs to do an interview (as soon as possible), and then a team of 5 people should spend the next 12 hours to try and find all those at risk and discuss with them how to take appropriate steps. ... A trace time beyond three days has little value in controlling the epidemic while a trace time of one day is plausibly effective."

"Green zones primarily need sentinel testing and modest amounts of TTSI testing. Yellow zones need sentinel testing, a full-scale TTSI program, and some critical context testing. Red zones need robust critical context testing and, to the degree possible, full scale TTSI tracing programs."

(in yellow:) "If all contacts are tested immediately using a fast lane approach, isolated until they stop shedding virus if they are positive, and quarantined for the full fourteen days and/or receive a second confirmatory test if negative, then the number of active infections should rapidly decline in yellow zone communities. Having sufficient capacity to execute TTSI with a trace time of less than one day enables safe re-openings of the economy beyond the essential workforce. Re-openings undertaken without this capacity in place will be accompanied by increases in disease prevalence."

I'm going to omit most of the "slow lane, fast lane, carpool lane" sections regarding building the testing infrastructural capacity and the several different speeds/targets of testing.

Voluntary self-isolation: "we will need to offer individuals the opportunity to self-isolate for up to two weeks if they are unable to do so in their homes. There is substantial evidence that providing a voluntary option to safely isolate will help to dramatically reduce spread of infection to one’s family and therefore the spread of infection overall. We believe 14 percent of infected individuals and exposed contacts will need such an option. Hotels are largely sitting idle at present and can provide local options for such voluntary self-isolation sites."

Income support for voluntary self-isolation: "offer income support to individuals for whom loss of income during 14 days of voluntary self-isolation represents a prohibitive barrier to being able to self-isolate. We estimate 40 percent of individuals will need a measure of income support in order to be able to afford to voluntarily self-isolate. Paying these individuals a stipend of $50 per day (analogous to federal jury duty) will greatly increase the success of containment efforts"

"The total investment we recommend comes to $74 billion, most of which would be spent over 12 months, with some being spent over 18 months. This contrasts to a monthly cost for collective stay-at-home orders of $350 billion." (I'm am unclear here whether "cost" of stay-at-home means "GDP impact" or "stimulus bills")

"Risk communication messaging from government officials should incorporate the concept that TTSI is a strategy for disease control designed to put the disease on a constant downward deceleration, to achieve its near elimination, and to leave the country well-positioned to respond rapidly to new outbreaks by containing them"

"Risk communication messaging from government officials should incorporate the concept that residents of locales in the red zones should avoid inessential travel to yellow or green zones and should ideally self-quarantine for fourteen days upon arrival in a green or yellow zone."

And it concludes on page 36, with the rest being appendices: numbers from around the country, describing possible interstate compact models for administering testing, local level model for small city response