Ricochet is the best place on the internet to discuss the issues of the day, either through commenting on posts or writing your own for our active and dynamic community in a fully moderated environment. In addition, the Ricochet Audio Network offers over 50 original podcasts with new episodes released every day.
This morning’s data sort is by “Recoveries” as I am trying to “look up” a little even as I ask the question: Who/what do you trust? Yesterday I posted a link to a University of Washington affiliated think tank that put out some wonderful graphics attempting to make predictions, state by state (and aggregated nationally) about the bell curve for this epidemic. Or as New York Governor Cuomo said the other day, they were saving stored ventilators for the “apex” of their epidemic. The Washington think tank purported to give an answer as to when that apex might be. Within hours a number of the predictions were falsified which calls the model into question. Not just that the predicted number was wrong, but that the range was wrong as well.
I’ll get back to that in a moment. The Worldometers table pictured above has a new column: 1st Case. We can now look at country data and see it within the context of how long their particular epidemic has been going on. The circled date for the US is “January 20.” This was a man traveling through Wuhan, China early in January and returned to the US on January 15. He fell ill with pneumonia on January 16. The CDC confirmed that he was suffering from COVID-19 on January 20. The reports at the time said that “CDC officials now say there is evidence that ‘limited person-to-person spread is happening.‘”
Well, that turned out to be an understatement. (And there is some belief that the virus may have been circulating in the country undetected as far back as November.)
Helmuth von Moltke the Elder famously said: “Battle plans never survive contact with the enemy.” Which is another way of saying that things don’t go as planned, information turns out not just to be wrong, but serially wrong. That is, we don’t just get it wrong, make corrections, and get it right. We get it wrong, then wrong, and wrong again until we finally get it right.
And that is our dilemma. As we watch our systems struggle, and life and death are in the balance, how do we maintain our confidence in leadership and authority? It doesn’t help that not only do they get it wrong, but the media is crafting a narrative designed to undermine our confidence in the cohesion of our national response. We could turn off our televisions and close down our web browsers, but where would that get us?
So let’s go back to the IHME COVID-19 Projections. They are a great set of graphics. They are data-rich as you pull the sliders across looking into each state’s future. They convey important information about possible deaths and demands on health care. And, so far, they are wrong. The question that IHME has to figure out is why are they wrong? Some might say that they are not wrong because the experience may yet come back into the predicted range. And for some locations that may be true. But why are they wrong now? If there is a single key factor it may be China. From the IHME paper:
Local government, national government, and WHO websites [21–25] were used to identify data on confirmed COVID-19 deaths by day at the first administrative level (state or province, hereafter “admin 1”). Government declarations were used to identify the day different jurisdictions implemented various social distancing policies (stay-at-home or shelter-in-place orders, school closures, closures of non-essential services focused on bars and restaurants, and the deployment of severe travel restrictions) following the New Zealand government schema.  Data on timings of interventions were compiled by checking national and state governmental websites, executive orders, and newly initiated COVID-19 laws. Data on licensed bed and ICU capacity and average annual utilization by state were obtained from the American Hospital Association. We estimated ICU utilization rates by multiplying total bed utilization rates by the ratio of ICU bed utilization rates over total bed utilization rates from a published study. Observed COVID-19 utilization data were obtained for Italy  and the United States,  providing information on inpatient and ICU use. Data from China  were used to approximate inpatient and ICU use by assuming that severe patients were hospitalized and critical patients required an ICU stay. Other parameters were sourced from the scientific literature and an analysis of available patient data.  Age-specific data on the relative population death rate by age are available from China,  Italy,  Korea,  and the US  and show a strong relationship with age (Figure 1).
Using the average observed relationship between the population death rate and age, data from different locations can be standardized to the age structure using indirect standardization. For the estimation of statistical models for the population death rate, only admin 1 locations with an observed death rate greater than 0.31 per million (e-15) were used. This threshold was selected by testing which threshold minimized the variance of the slope of the death rate across locations in subsequent days.
From the projected death rates, we estimated hospital service utilization using an individual-level microsimulation model. We simulated deaths by age using the average age pattern from Italy, China, South Korea, and the US (Figure 1) due to the relatively small number of deaths included for the US (n = 46) and the fact that the US age pattern is likely biased toward older-age deaths due to the early nursing home outbreak in Washington state. For each simulated death, we estimated the date of admission using the median length of stay for deaths estimated from the global line list (10 days <75 years; 8 days 75+ years). Simulated individuals requiring admission who were discharged alive were generated using the age-specific ratio of admissions to death (Figure 3), based again on the average across Italy, China, and the US. The age-specific fraction of admissions requiring ICU care was based on data from the US (122 total ICU admissions over 509 total admissions). The fraction of ICU admissions requiring invasive ventilation was estimated as 54% (total n = 104) based on 2 studies from China. [36,37] To determine daily bed and ICU occupancy and ventilator use, we applied median lengths of stay of 12 days based on the analysis of available unit record data and 8 days for those admissions with ICU care. 
No doubt that IHME will be updating their model outputs as new data comes in. But in the initial formulation, they needed to draw data from places with significant case experience. And that was China, and to a lesser extent Italy, in the formulation. That puts a lot faith in the data from China. I do not want to dip into “yellow peril” territory, but the conduct of the Chinese Communist Party does not inspire confidence. No one believes the Chinese gross data reported on Worldometers. Maybe the experts with personal relations with Chinese epidemiologists can decode whatever careful language they must be using in describing what is going on there. But even as China-published data was trending downward, various reports would come out contradicting the official counts. Just the other day Report: Thousands Of Urns Shipped To Wuhan, Where The Virus Is Supposedly Under Control suggests that the virus may still be raging in some parts of China. This, even as China is lending aid to its One Belt One Road (OBOR) partner in Italy. Do we doubt that Chines Community Party would deny some segment of its own people critical supplies in order to promote its international objectives? We do not.
And there are US officials. The CDC has gone from “limited person-to-person spread” to highly contagious community spread. The President has to cut through red tape in order to create a “right to try” approach to the epidemic, not just terminally ill patients from other causes. The federal government bureaucrats sought early on to centrally control the response to the outbreak through its siloed expertise. The President has demanded a “whole of government” and private enterprise approach. The media has been as dismissive of the President in his leadership in the epidemic as they have been of his legitimacy to be President at all.
As a nation, we are being asked to make great sacrifices. The worker relief program is critical as the consumer demand has been totally cut off from some goods and services and redirected to others. Our politicians have larded it up with irrelevant stuff that we will be paying for in the coming decades as well. Our governors have restricted our liberty in the name of safety, and over time that will be the greatest sacrifice we have made if liberty is not fully restored.
But there are two kinds of sacrifices: willing and unwilling. With confidence in leadership, the sacrifices are mostly willing. Without confidence in leadership, our sacrifices are coerced. Who/what do you trust?
[Note: Links to all my COVID-19 posts can be found here.]Published in