One epidemiologist’s take on the week ahead (March 22)
For the week starting Monday March 23, 2020
In the big picture, watch for the narrative this week to shift. Sometime this week, the US may surpass Italy in number of cases, a psychological blow and further incentive to enact quarantine measures (shelter-in-place is quarantine by another name, shedding dubious legal precedent). I expect the emphasis on epidemiologic modeling will take a back seat to concerns about the provision of clinical care, mask shortages, hospital stories. The emergence of New York as an epicenter will serve as a stark warning to other cities to get medical systems ready. Financial markets will be a big story as stimulus legislation goes through gyrations.
At the same time, there will be more and more stories of people having recovered with mild cases, maybe weakening public resolve. The drumbeat of people who think we should end quarantine in short order will increase, indeed it has already started.
Case Counts
As expected, the US took the number 3 spot in the world last week, behind China and Italy, in terms of number of confirmed-cases-among-the-tested. While we use confirmed-cases-among-the-tested to mentally keep score, it’s worth remembering that it is like looking into the past. You know how the light from stars takes millions of light years to reach us on earth, making our current night sky a tableau of past astronomic cataclysms? Same goes for “The Count” — you can assume 10 to 14 days from the actual transmission event when you see the counts. (Hence, beware of intervention graphs that suggest that policies have immediate effects using the complete case count.) Instead, a physician friend in Raleigh is keeping an eye on the number of days it takes for hospital deaths to double.
A proper epi curve would show not the daily confirmed-cases-among-the-tested, but instead the horizontal axis would be the date of onset-of-symptoms-among-the-tested. Even if the data are delayed, this is more useful information for deciding on public health interventions. Here’s more on why this happens from CDC (and a quick training lesson).
Will other states enact shelter-in-place orders?
The data on which to make this decision is not going to come from surveillance. Cases-among-the-tested will go up this week (starting at approximately 32,000 on Sunday night), due to testing and community-based spread, but we will not know how many infections have been acquired this week. This is the biggest question of the week. Expect to see increasing restrictions on what businesses can be open, and the experience in early mandatory quarantine states will inform those not under official lockdown. If the milestone of more “cases” than Italy happens this week, the pressure for official lockdowns will increase exponentially.
To Test or Not to Test?
We were bemoaning the lack of test kits last week. However, for those who can weather the sickness at home, the message shifted to not come in for testing this weekend. What gives? I think this is a classic example of something we see routinely in health services research and epidemiology called the “healthy” user effect or health-seeking bias. When a new treatment or test becomes available, there is a rush for getting it among people who are otherwise healthier than the general population. In this example, as testing capacity increased and the message for more testing was promoted, the “worried well” (e.g., mild to moderate infections) got into line to get tested first. Epidemiologist Nick Reich at UMass Amherst sees this effect in the NYC data.
This isn’t to say that last week’s demands for more testing were misplaced —we still need more testing to be easier and more available. An alternate approach to the first-come-first-served approach to American healthcare would be to deploy testing systematically, contact trace, isolate. This example from the Italian village of Vò shows the difference that could be made with an approach that is population-based. Instead, we are relying on information filtered through our healthcare system’s existing (structural, racial, financial, etc.) inequities. These filters determine who gets medical care first, and who we see represented in the surveillance data.
This also means that the definition for who is eligible for getting a test has shifted. As the days go by, data from this weekend onwards will not be perfectly comparable to earlier weeks.
15 Days of Pain
Last week, we saw a glimpse of the possible endgame. The “15 Days to Slow the Spread” campaign from the White House got higher visibility with Fauci holding up a sign from the podium. This photo was taken by Alex Wong on March 20. But when did the clock start on these 15 days??
“15 Days to Slow the Spread” appeared on the evening of March 16 as a temporary story on the White House’s “1600 Daily”, and was pushed via Twitter on March 18, 2020 . The next day came the “hammer and dance” theory, proscribing a few weeks of quarantine (“hammer”) followed by months of infection tracing (“dance”). And finally the picture of Fauci holding up the idea of 15 days on the 20th.
No real new data has come around in the last week to suddenly make the 15 days an empirically informed strategy. This is made up, with no scientific support that I have seen. (West Bengal, India is trying to get it done in 4 days!) Possibly it represents the best guess for non-biologic phenomenon? Like the possible earliest that checks could go out to Americans for stimulus? Or psychology experiments? April Fool’s Day (March 16+15)? Who knows, but worth keeping an eye on how this specific number evolves this week.
Contact Tracing and Forced Isolation… y’all ready?
I see calls for relaxing quarantine in time to enjoy the Spring. Part of the rationale is that we would have gotten ahead of the virus and then could do contact tracing among real cases, with isolation (like has been done in many other countries). In normal times, contact tracing and forced isolation are public health measures reserved for the poor, marginalized, and racial minorities. In the United States we use these tools to glean hidden social network information from people with HIV and hepatitis, people who use drugs, men who have sex with men, Black men with higher incidence of some STDs, people living in crowded conditions with tuberculosis, and measles in secretive off-the-grid groups. The public health belief, with origins in paternalism, is that these people cannot be trusted to manage prevention on their own, so the State must step in. Well, contact tracing does work in these situations, and in church lunch picnic disasters. However, it comes at a cost to personal privacy and freedom of movement, and requires considerable public health personnel. In order for contact tracing and forced isolation to become a national strategy, we have to acknowledge our past and make the case that the sacrifice of these values is for the greater good.
Cynically, then, watch for an increase in elderly patients being portrayed as not having the agency to make their own decisions, Baby Boomers not heeding voluntary social spacing, or shock stories of identifiable groups of people not heeding the voluntary call or now deemed to be at higher risk (e.g., asthma). Young spring break partiers are a start, but that kind of group shaming will emerge, and start to create the foundation for establishing contact tracing. Please understand I am not advocating for this, but just putting out my expectations.
WASH YOUR HANDS. DON’T TOUCH YOUR FACE. DO GROCERY ONCE A WEEK. STAY HOME AS MUCH AS YOU CAN…. BUT GET EXERCISE BY TAKING WALKS.
