The Jump in Covid Test-Positives: Two Alternative Hypotheses

During the 20-day interval between June 15 and July 5 the daily rate of confirmed covid diagnoses in the US more than doubled. How much of that increase is an artifact of more testing, and how much is a real increase in infections?

According to trend data from Johns Hopkins, the daily number of covid tests conducted in the US has been increasing at a steady rate since the beginning of the epidemic. The percentage of tests confirming infection with coronavirus peaked in early April, then declined steadily, bottoming out at 4.4% in early June, before going back up to the present level of 7.6%. During the most recent 20-day interval the testing rate is up about 30 percent while the test-positive rate has increased 70 percent. It would be reasonable to discount the increase in testing rate as an artifact of expanding technical capacity, focusing specifically on the test-positive rate as an accurate measure of the incidence of viral infection in the population. Tentatively then we’d infer that the real rate of covid increased 70 percent over the past 20 days.

But there are other factors to consider. Between June 15 and July 5 the covid death rate decreased by 30 percent. Death is a lagging indicator: all else equal, a declining death rate results from a declining infection rate. It’s possible that the spike in test-positives will eventually be mirrored by a jump in deaths. But how long is the lag? Go back up the timeline to the preceding 20-day interval, from late May to mid-June: the test-positive diagnosis rate was flat, while the death rate dropped 35 percent. Go back another 20 days, early May to late May: test-positives flat, deaths down 17 percent.

Over the past two months then, the coronavirus death rate has decreased by two-thirds, a trend that’s been either independent of or contrary to varying trends in test-positive diagnostic rate. Death may be a lagging indicator of infection, but based on the data it’s not a lagging indicator of official test-positive diagnoses.

So which is it: are infections going up, as the test-positive data strongly imply, or are infections going down, as the death counts indicate? It seems most likely that the upward trend in test-positives reflects not an increase in viral prevalence in the US, but rather an increasing level of accuracy in measuring the actual spread of the virus through the population.

It’s been widely acknowledged that, since the beginning of the epidemic, the official daily tally of covid diagnoses has grossly underestimated the true rate of contagion. Recently the CDC announced that the number of people who have been infected by the virus is at least ten times the total number of test-positives recorded in the official tallies. Together, the increases in testing rates and test-positives are closing the gap, achieving a closer correspondence between official tallies and the actual spread of the virus.

Death is a hard outcome of infection: all else equal, a declining death rate points back upstream to a declining infection rate. But what if all else isn’t equal? Deaths could drop because more young and otherwise healthy people are now getting infected, whereas the old and infirm who previously accounted for most of the deaths are successfully avoiding contagion. The data don’t support this interpretation, as discussed in the preceding post: the age demographics of covid deaths have not shifted downward since the peak of the epidemic. Deaths could also drop due to improvements in medical treatment. That’s possible, maybe even likely, but for better care to have reduced death rates by two-thirds over a matter of a month and a half, absent any new treatment that dramatically reduces symptoms or that cures the disease? Not likely.

If death rates are an accurate lagging indicator of infection rates, then the spread of the virus has been reduced substantially over the past two months. So, as the test-positive rates are going up, the actual infection rates are going down. Will the twain ever meet, with the number of test-positives finally catching up to the actual number of new infections in the population?

It might be happening.

Based on the small number of well-designed random serology surveys, I’ve been using an age-adjusted covid mortality rate of 0.6 percent for the US. Death is a lagging indicator of infection: based on the mortality rate estimate, each death results from 1/.006 = 167 infections. For July 4 and 5 the death counts were 250 per day — the lowest levels since the beginning of the epidemic in late March. 250 x 167 = 42 thousand: that’s the number of new infections per day that would result in 250 deaths per day. For July 4 and 5 the test-positive counts were 45 thousand per day — just about equal to the infection rate estimate based on deaths. The daily test-positives likely won’t climb any higher unless/until the daily deaths go up. Or unless the 0.6% mortality rate is too high…


2 thoughts on “The Jump in Covid Test-Positives: Two Alternative Hypotheses

  1. Anne sent me this WaPo article that touches on many of the issues I address in this post. Points addressed, and my responses:

    “President Trump’s former Food and Drug Administration chief Scott Gottlieb said on CBS’s “Face the Nation” that “the total number of deaths is going to start going up again as the number of hospitalizations starts to spike again.”

    Despite local surges, vovid hospitalizations nationwide have decreased by two-thirds over the past two months, though of course the numbers could go back up. Hospital business models rely on running near capacity, since empty beds don’t bring in revenue. Because of infection fears, doctors were discouraging so-called elective surgeries while individuals were reluctant to go in for treatment. As things open back up, the backlog of suppressed demand for care is being alleviated, increasing hospitalizations for non-covid reasons. Matter of fact, I just received a mailing from my health plan suggesting that I ask my doctor if the time is right to schedule an elective procedure.

    “Infections appear to be spreading fastest among younger people as they start mingling more.”

    The US median age is 38, so the majority of Americans are “younger people.” It’s never been the case that older people are more likely to be infected than younger people.

    “And as testing becomes even more widespread — between 600,000 and 700,000 tests are now being conducted in the United States every day — more people with only mild cases of covid-19 or no symptoms at all are being included in the testing tallies.”

    Likely true, but that’s an artifact of testing, not an indicator of age downshifting in infections.

    ““We know death is a lagging indicator,” said Abraar Karan, a physician at Brigham and Women’s Hospital. “You have cases show up first, hospitalizations next and deaths show up after that.””

    Agree about the lagging indicator. But as I documented in this post and the subsequent one, death rate trends nationwide haven’t been lagging test-positive rates, going back at least two months.

    “It’s been clear for more than a month that infections are rising rapidly.”

    Not so. All along, infection rates have been much higher than diagnostic test-positives would suggest. Test-positive rates have been rising rapidly, but only over the past two and a half weeks. The point of contention is whether test-positive increases correspond with infection increases.

    “There are emerging signs that deaths are already trending upward in some states.”

    That’s how it always goes: up in some states, down in others.

    “Doctors have figured out better ways of treating covid-19 patients. Experts say this could be a small factor in lower death rates. Physicians have come a long way in developing a standardized way of caring for covid-19 patients, compared to the pandemic’s outset. They’ve learned — and research has confirmed — that dexamethasone cuts the risk of death for patients on a ventilator by a third and reduces the risk of death for patients on oxygen by a fifth.”

    That’s really interesting. Deaths can decrease even if infections stay steady if treatment improves markedly. I didn’t know that dexamethasone was having such a big impact. How widespread is its use, and since when? I’ll have to investigate further. Still, it’s worth noting that dexamethasone’s impact is on ventilated and oxygenated patients, who would already be in the hospital. But hospitalizations, like deaths, have also decreased by two-thirds, which would suggest that possible impacts of better care on deaths would have been happening before patients got into the hospital.


  2. More on improved inpatient treatment from this article:

    “Anytime you go on a ventilator, there’s a risk you won’t improve and will spend your last days or weeks unable to speak, and heavily sedated. Early in the pandemic, it looked like going on a ventilator was a longshot in terms of survival. Studies showed a minority came off the machines alive. But more recent evidence has been somewhat more promising. For example, a study in The Lancet showed that of 203 critically ill COVID-19 patients who were put on ventilators in New York hospitals, less than half (about 41%) had died a month after follow up. A study conducted in several ICUs in Atlanta, Ga., found that of 165 ventilated patients, 35.7% died, with fewer than 5% still on ventilation at the end of the study.”


Leave a Reply

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

You are commenting using your 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