A quick check of the ever more reassuring University of Washington IHME model shows that America reached the peak number of deaths three days ago and is on it’s way to improvement. Which is great. Except for the part where that didn’t happen. In between the previous update on April 13 and the latest update, the IHME equation consistently under-predicted deaths by 400 or more. Then, after taking that into account, and following an update on Thursday, it went on to … under-predict deaths by 400 on Friday. Which makes the way the model also cut the overall number of deaths by 8,000 in the latest update seem more than a little suspect.
Discounting additions made after New York rolled in thousands of previously unreported deaths earlier in the week, Friday marked the highest single day of reported deaths in the United States with over 2,500. It’s also becoming increasingly obvious that New York is far from alone in failing to account for cases and deaths that are happening outside of hospitals. Across the nation, extended care facilities have become overwhelmed by cases among both patients and staff, and the rate of people dying in their own homes is up an order of magnitude.
All the evidence indicates that the number of cases of COVID-19 in the United States has not plateaued. But the rate of testing has.
For weeks now, the COVID-19 Tracking Project has recorded not just the daily statistics on new cases and deaths across the nation, but the number of tests conducted state by state. Some states have tested at relatively high rates. For example, both Rhode Island and New York have tested at a rate that works out to almost 30 people out of 1000. But over half the states have tested at rates less than 10 people out of 1000—sometimes much less. Puerto Rico, which has presumably been shipped supplies with the same reluctance that Trump has demonstrated in helping the island withe every other disaster, has tested at a rate of only 3 out of 1000.
But the pure rate of testing doesn’t tell the whole story. The over 500,000 tests administered in New York are still showing a rate of over 40% positive—demonstrating that it’s still phenomenally difficult to get a test there. Meanwhile, Vermont has conducted only 12,000 tests, but the rate of positives was just 6%, which shows that Vermont … still isn’t doing enough tests.
What does genuinely adequate testing look like? Well, it’s not bad to consider the Iceland approach. The isolated country of 364,000 has taken a radical approach to dealing with COVID-19. Rather than imposing social distancing across the board, they’re simply testing everyone. Literally everyone. Currently, Iceland has tested just under half the population, and while the 1,750 cases confirmed may seem relatively high for a nation whose population is roughly that of Wyoming (412 cases), the very low death rate shows that Iceland is doing a good job at finding the cases with mild or no symptoms — the kind of cases that aren’t even being considered for testing in most U.S. states.
The United States isn’t just far behind Iceland when it comes to rate of testing, it’s behind New Zealand, and Singapore, and Canada, and Hong Kong, and Switzerland, and … basically every nation which has been pointed out as a model for how the epidemic should be handled. The United States has tested at almost the exact same rate as the successful program in South Korea, except for the little fact where South Korea went into testing at a high rate when the number of cases was tiny, allowing it to trace cases and effectively isolate those affected. The U.S. began significant testing only after cases had exploded, meaning that the testing has mostly involved proving what was already clear — yes, this already hospitalized patient displaying severe respiratory distress does indeed have COVID-19.
And that remains an enormous problem. Because every study has demonstrated that not those with very light symptoms spread COVID-19, but those with few symptoms are responsible for by far the majority of all cases. In fact, 80% of cases in some studies began through contact with someone not displaying clear symptoms for COVID-19. That’s not just because about half of all cases have light symptoms, but because even those who later develop a more severe form of the disease are most contagious in the early days of the infection before those symptoms appear. There’s also a very basic human factor in this—people are less likely to hang around with those who are sweating and coughing in the midst of a pandemic, while those who are sick are less likely to socialize.
With all that said, let’s look at where things stand …
With Italy on the chart simply as means of demonstrating the raw scale of the epidemic in the United States, it’s clear that the number of new cases in the United States has not begun to diminish. While the eye tends to generate some good news on looking at these values, I invite you to pick up a piece of paper, a nearby letter, or anything else with a straight edge and lay it against the screen. What you’ll see is that the angle of the line for the United States has barely changed since the end of March. Really. The genuine rate of growth has changed, but only slightly. Here’s what it looks like in terms of new cases each day.
Since the first of the month, the number of new cases each day has wavered around in the same range between 25,000 and 35,000 cases a day. There is no discernible trend in these numbers, certainly not one that signals a prolonged period of decline. Instead, there are a series of dips and troughs that seem to correspond closely to … weekends. The large dip centered on April 12 is pinned directly to Easter Sunday. The decline around that date, and the dip a week before, is much more likely to be connected with reduced testing around those dates than it is on a genuine reduction in cases coming into testing facilities.
And, in fact, that whole period of 25,000-35,000 cases a day also represents something else: a flattening not of COVID-19 cases, but of America’s ability to test for COVID-19 cases. Throughout the same period, the range of tests conducted on a daily basis was 130,000 to 160,000. That means that, nationwide, about 20% of all tests conducted have been positive. Again, this is a number that shows people are only getting tested when displaying symptoms or with known contact. In some states and areas, it takes both to warrant a test.
Even worse, the 160,000 rate came near the start of this period. In the last two weeks, the number of tests conducted each day has actually trended slightly downward. Even as the number of known cases has increased, the number of tests hasn’t begun to keep up. With 700,000 known cases, the United States is now testing at a rate below that it was employing when there were 300,000 known cases — which was, quick reminder, just two weeks ago. In fact, this week the United States hit 600,000 cases on Tuesday, then passed 700,000 on Friday, just three days later.
In the best possible view of the existing data, the United States has managed to straighten the line of growth. That’s a precursor to genuinely flattening the curve, but it’s only a precursor.
However, it can be argued that testing over the last weeks hasn’t just failed to improve, it’s now much worse than it was at the beginning of the month when it comes to probing the defining the epidemic in a way that allows the nation to take appropriate action. Yes, we’ve leveled the number of new cases being logged each day, but every indicator says we’ve done so only because we’ve hit the limits of current testing.
- Tests are still impossible to obtain for most people.
- Tests are not being used to sample the population at large.
- Tests are not being involved in a system of case tracking and quarantine.
Testing was America’s original failure in response to COVID-19. It continues to be the primary failure. The testing is not just inadequate, but so inadequate that it generates a false sense of progress. And that’s deadly.
Posted with permission from Daily Kos.