Carefree Mis‐​Reporting of Virus Deaths From Alarmist Media

This article appeared in the Real Clear Markets on July 15, 2020.
  • Related Content

Virus Deaths Climb as Cases Hit New High” was the original title of a recent front‐​page Wall Street Journal feature, later changed online to “Deaths Begin Trending Higher.” Either title portends important news, but the only evidence is gibberish: “The average daily death toll in the U.S. rose to 599 in the seven days through July 8, up from 510 deaths a day as of July 4, according to a Wall Street Journal analysis of data from Johns Hopkins University.”

July 4 is obviously one of the “seven days through July 8” so the alleged “rise” makes no sense. A correct apples‐​to‐​apples comparison would have said, “The average daily death toll in the U.S. fell to 599 in the seven days through July 8, down from 786 in the seven days through July 1.” Those averages (599 and 786) are from the Wall Street Journal’s own graph of “Daily reported Covid‐​19 deaths.” But revealing the actual 35% drop in deaths over back‐​to‐​back seven‐​day periods would require rewriting the headline.

In a remarkable understatement, the article admits “deaths haven’t surged in the same way the infections have.” In fact, nationwide deaths have fallen according to all of their own figures.

The graph accompanying the article, reproduced here, shows a 7‐​day rolling average of daily cases and deaths expressed as an index number, where April 1=100. The death index jumped from 100 on April 1 to a peak of 284 on April 18. But it then fell steadily to 76 by June 23—which was nearly two months after many states reopened for business from April 24 to May 4. There was then a brief uptick in the death index above 100 from June 25 to July 1, which appears consistent with a multi‐​week lag between infection and death following the late May Memorial Day activities and mass protests. Since July 4, however, the 7‐​day average of deaths has again fallen well below 100—even as the number of positive tests keeps moving in the opposite direction. The 7‐​day average was 65 from July 4 to July 6, a record low. There were only 217 recorded coronavirus deaths on July 5—the lowest since March 24, when the epidemic was barely started. What happened over the following four days was scarcely a new trend.

As the authors reluctantly concede, “soaring case counts are partly attributable to expanded testing that is detecting asymptomatic or less severe cases, often among younger people.” Quite right. But that makes it deceitful to keep equating more positive tests with “more infections”—as they do eight times.

Soaring case counts in a few Southern states are also partly attributable to a belated but essential emphasis on testing workers in hospitals and nursing homes — jobs with a high risk of testing positive. A later July 13 Wall Street Journal report found 800‑1000% increases in nursing home cases from the last week of May to the week ended June 28 in Phoenix, Tampa, Houston, San Antonio, and Kansas City. But there was no related surge in nursing homes deaths in those cities (and very few total deaths in Missouri or Kansas). So, this June discovery of a few hundred more cases needing to be quarantined was the predictable and intended result of fast‐​tracked intensive testing at nursing homes.

The July 11 article about climbing deaths claims nameless “epidemiologists caution that deaths typically lag behind other indicators.” Ignore what such anonymous experts supposedly said and look at the graph. Deaths first rose long before tested cases because few cases were tested. Then cases rose rapidly when testing finally accelerated, but deaths fell. Cases don’t predict deaths—with or without a lag.

A currently obligatory media narrative is that coronavirus deaths are rising in the Sunbelt states because Republican Governors supposedly reopened too much and too early. But the states that opened first and most aggressively were Oklahoma (April 24), Alaska (April 24) Georgia (April 28), Tennessee (April 30), Idaho (May 1) and Utah (May 1), yet recent daily deaths in those states averaged only 3, 0, 20, 14, 1 and 4, respectively. Southern California has seen some increased deaths recently, but that is hard to blame on incautious reopening or Republicans. There have been increased deaths in Texas, but also Arizona (which delayed reopening until May 15) and Florida (May 18). Oher sunbelt states still have average daily deaths below 20, most below 10. 

At the national level, the week ending July 4 was the 11 th week in a row the CDC reported declining deaths from COVID-19, flu and pneumonia. The Wall Street Journal graph likewise shows COVID-19 deaths trending lower, not higher, as do the 7‐​day averages the authors mentioned. Yet their two headlines scream the opposite.

Searching for evidence that their graph was wrong, the authors noted that “Arizona, South Carolina and Alaska joined Texas as the states with the biggest percentage increases in deaths, with fatalities rising more than 20% in each state since the beginning of the month.” But that was very few states and very few days. And the percentage increase is meaningless if the number was unusually low on July 1 or unusually high on the last day being compared with the first. Besides, Alaska has had only 17 COVID-19 deaths all year, and it’s not a Southern state.

The headline and article are extremely misleading, but the graph is not.

usconfirmed-covid-19.jpg

Alan Reynolds

Alan Reynolds, an economist, is a Senior Fellow with the Cato Institute.