It Is Rising — Oy Vey!

According to the IHME statistical model’s projections, the US reached its pinnacle of daily covid deaths two days ago, on Good Friday. Though the subsequent mortality dip in the curve isn’t expected to be as precipitous as was the ascent, the model still projects that the daily body count will drop to single digits by early June.

During the upswing, actual deaths have come close to the projected numbers. What about the downstroke? It’s too soon to tell, since we’re still only a day or two past the presumed apex.

What about other countries that reached the top of the death curve earlier in the month — have their rates dropped? Italy for example — its early sharp death spike reached a peak of 969 on March 27. The IHME model projected that, due mostly to widespread social isolation, the Italian death rate would decline markedly, down to 335 on April 10 and 281 on April 11. According to Worldometer’s data, however, Italy’s actual death counts the past two days were 570 and 619 — nearly twice the numbers projected by the model. While the death rate did drop during the first week after the peak, it’s now plateaued for the subsequent week.

How about Spain, its situation similar to Italy’s? Spain’s daily deaths topped out at 961 on April 2. For April 10 and 11, the IHME model projected Spanish deaths at 304 and 262; Worldometer reports that actual deaths were 634 and 525 — again, twice the projected numbers.

In order for the death rates to drop, the contagion rate has to drop. COVID-19 is a highly contagious disease: under pre-isolationist societal conditions the average infected person was infecting 2.4 others. Slowing the infection rate keeps the healthcare system from being overwhelmed, but it doesn’t necessarily reduce the total number of people who will eventually contract the disease and die from it. In order actually to bring new infections from this first wave to a halt, the interpersonal contagion rate has to drop below 1.0. If that threshold is achieved and new infections drop to small numbers, then when a new outbreak of the disease shows up it becomes feasible to deploy more precisely targeted interventions of social isolation and contact tracing, keeping the viral spread from building momentum and spreading out of control.

But what if current social isolation measures are able to reduce the contagion rate to some level below 2.4,  but not enough to drop below the 1.0 threshold required for the wave to extinguish itself? The infection rate will slow, and so will the death rate, but gradually, and inexorably, the viral wave will sweep its way across the entire population.

It’s likely that early next week, based on three days’ worth of new data, IHME will revise its mortality estimates upward for Italy and Spain. It’ll take at least another two weeks for the post-apex body count trend to make itself evident. Conceivably the models will have proven overly optimistic, and the first wave will be the only wave, which won’t wash out until it’s passed over the entire population.

Lately the daily count of newly confirmed diagnoses has been holding steady at around 34K per day. Everyone agrees that the real numbers are much higher, maybe ten times as high. Let’s say there are 5 million active cases in the US with 340 thousand new cases added daily. Even if the nation could mobilize massively on short notice — which clearly it cannot — there’s no way that any sort of effective individual isolation and contact tracing could be implemented mid-wave, bringing the numbers down. The wave will have to run its course: maybe 200 million people infected, maybe 1 million dead.

If current levels of social isolation can’t stop the wave, then the policy question will come down to two options: either keep the wave moving slowly, by maintaining continued levels of social isolation; or let it speed up, by relaxing the constraints and returning to “normal.” Maybe issue warnings to the old and infirm who are the most likely to die if infected: for your own safety, continue sheltering at  home. Implement infrared fever-detection protocols at workplaces, schools, and nursing homes to facilitate targeted isolation of individual cases that could slow group contagion. Roll out new treatments that offer some relief of symptoms, possibly reducing severity and bringing down death rates. Get that vaccine invented, tested, manufactured, and rolled out.

4 thoughts on “It Is Rising — Oy Vey!

  1. Yep, the data might not show the precipitous drop that the U of Washington models predicted. But didn’t we think we were just flattening the curve to allow the number of infections to be manageable by the health system? Of course, it would be good to get the numbers low enough to make the “Identify-quarantine-followup” strategy implementable. Is it possible? I just read this article, https://metropole.at/open-letter-yale-epidemologist/?fbclid=IwAR3cQ5hQOzK3t2wQut2cOs8UREOb1058_DwphccV2GYqDFD4pUNpQ_VFKsc (suggested on my NextDoor list) by an infectious disease expert. It makes me wonder how many folks are “cheating” on isolation? I hope everyone is holding the line. I worry more about the decision makers who will try to read the numbers too soon pointing the country toward some out of control roller coaster of illness. Still, the numbers are fascinating…

    Liked by 1 person

  2. Nice article — no cheating! I think we’re doing a good job, to the point where social distancing behaviors that at first demanded a good deal of conscious attention are starting to get compiled into habits, which make them easier.

    Good point about flattening the curve — at first that constituted the primary focus of intervention. Maybe it’s always been a pipe dream that current social isolation measures, installed after the cat was already out of the bag, could drop the curve back down to the X-axis, so we collectively could do better the second time around, getting ahead of the curve with rigorous testing, identification of infected individuals, and social contact tracing/isolation.

    Suppose the curve stays flattened to its current rate of around 350K new cases per day. At that steady rate the new cases coming on line would equal those who’ve previously been infected but who have either recovered or died — steady-state resource use. Let’s further suppose that, on a flattened curve, the virus will infect 50% of the total population: that’s around 160 million people. 160M divided by 350K/day = 457 days for the viral wave to pass through the population. That would take us until July of 2021, maintaining the present level of social isolation in order to keep the curve flat; i.e., no school, few jobs, no large gatherings, etc. Never gonna happen. With relaxed social isolation — going back to work and school and the ball park — the curve will bulge upward again, but the wave will pass through more rapidly, let’s say by the end of this year, doubling the burden on a healthcare system that’s already past the breaking point. Not pretty.

    But let’s assume that the CDC-Homeland Joint Task Force is accurate in estimating that 80% of the fatalities will be concentrated among the 20% of the population that’s 65 years and older. If the vulnerable elderly self-isolate for the next 8 months in the interest of self-preservation, then the coming wave won’t hit them, won’t send them to intensive care, won’t require ventilators to keep them alive, won’t kill them. If that works even halfway successfully, healthcare resource demand could remain flattened close to current levels and mortality could be halved to *only* half a million or so. Then the big wave will have passed through, and since such a big percentage of the population would have acquired immunity, any second wave would be minor or could be gotten in front of. Then comes the vaccine and it’s over.

    Is that optimistic? No. But it might be realistic.

    Like

  3. The Italian death count for April 12 was 431. That’s down a lot from the prior day’s 619, but it’s still nearly twice as high as the 242 projected for April 12 in the IHME model. One wonders whether the count wasn’t quite thorough on Easter Sunday, and whether the numbers will bounce back up again tomorrow.

    Like

  4. This Harvard study forecasts a long right tail to the covid curve. The article refers to R0: the reproduction rate, or the average number of people infected by someone who has been infected — what I’ve been calling the contagion rate. The curve flattens to the extent that R0 is decreased. To make the curve come down the other side — i.e., to extinguish the epidemic — R0 has to drop to under 1.0. From the article:

    For social distancing to have reversed the epidemic in China, the effective reproduction number must have declined by at least 50-60%, assuming a baseline R0 between 2 and 2.5. Through intensive control measures, Shenzhen was able to reduce the effective reproduction number by an estimated 85%. However, it is unclear how well these declines in R0 might generalize to other settings: recent data from Seattle suggests that the basic reproduction number has only declined to about 1.4, or by about 30-45% assuming a baseline R0 between 2 and 2.5. Furthermore, social distancing measures may need to last for months to effectively control transmission and mitigate the possibility of resurgence.

    Because social distancing doesn’t lower the R0 below the critical threshold, the spread of the disease is slowed but not stopped:

    When transmission was not subject to seasonal forcing, one-time social distancing measures reduced the epidemic peak size. Under all scenarios, there was a resurgence of infection when the simulated social distancing measures were lifted. However, longer and more stringent temporary social distancing did not always correlate with greater reductions in epidemic peak size. In the case of a 20-week period of social distancing with 60% reduction in R0, for example, the resurgence peak size was nearly the same as the peak size of the uncontrolled epidemic: the social distancing was so effective that virtually no population immunity was built. The greatest reductions in peak size come from social distancing intensity and duration that divide cases approximately equally between peaks.

    The IHME models expect current social distancing measures to reduce R0 to below 1.0, with only something like 5% of the population having been infected before the wave extinguishes itself, making possible either a series of subsequent waves with distancing or a more proactive approach of testing, quarantine, and contact tracing that keeps new waves from forming. It remains to be seen, however, whether the current distancing measures will actually drop R0 below 1, or whether the the spread will continue through the population until herd immunity is achieved, dropping R0 below 1 because there won’t be enough non-immune people left to infect. That, of course, assumes that immunity persists and that the virus doesn’t start mutating more rapidly, resulting in recurrent outbreaks like the flu…

    Like

Leave a Reply to Anne Doyle Cancel reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s