In this article, I’ll try to translate what key parts of a report from the Centers for Disease Control and Prevention (CDC) actually said. I’m posting links to this column on social media like Facebook, so before I go further, let me note that EVERY STATISTIC AND QUOTE CAME DIRECTLY FROM THE CDC’S SEPTEMBER 11, 2020 WEEKLY REPORT on COVID—19. Hopefully the censors will not flag this column and eager beaver lockdown fans won’t falsely report this as fake news—that has happened to me in the past.
Start with the headline run at NPR:
Adults With COVID-19 Twice As Likely To Have Eaten At Restaurants, CDC Study Finds
Compare this to the title of the report at the CDC:
Community and Close Contact Exposures Associated with COVID-19 Among Symptomatic Adults ≥18 Years in 11 Outpatient Health Care Facilities — United States, July 2020
If you’re a restaurant or bar owner, there’s actually good information here, and the CDC report could help guide policy to keep patrons as safe as is (knowingly) possible. If you’re a media worker who copied the headline from various legacy media, take yourself to a corner and give yourself a solid head thump.
As soon as I saw the article at NPR and read it, I asked myself why, when almost all other data was reported as a percentage of the control group or the case patients group, that wasn’t the case with the statistic on participants who reported eating at restaurants. That made no sense to me. Shouldn’t some scribe at NPR have asked himself or herself what percentage of the case patients reported it?
Better yet, shouldn’t some scribe somewhere have actually read the full CDC report?
So much was omitted from the reportage on the data the CDC released.
For starters, I have no idea where they got the “twice as likely” figure unless they’re stretching an estimate. That stretch is evident in the table appended to the end of the CDC report on 314 symptomatic adults who were outpatients in 11 academic health care facilities and who received positive and negative SARS-CoV-2 test results. There were 154 in the Case Patients Group and there were 160 in the Control Participants Group.
In the discussion section, the CDC did say this:
“Adults with confirmed COVID-19 (case-patients) were approximately twice as likely as were control-participants to have reported dining at a restaurant in the 14 days before becoming ill.”
The table data show that 40.9 percent of the Case Patients Group (63 people) reported eating at a restaurant 14 days before the illness occurred. The number in the Control Patients Group who reported eating out during the same time period totaled 27.7 percent (44 people). How does that come to “twice as likely”? That’s a big “approximately”, right?
Can we assume that 91 people in the Case Patients Group and 116 people in the Control Participants Group did not report eating at a restaurant during that time period? It would seem so.
The closest estimate I see that justifies as “twice as likely” result would be work colleagues. In the Case Patients Group, 11 people reported close contact with a work colleague with known COVID-19. In the Control Participants Group, only 6 reported such contact.
Another example has to do with previous close contact with known COVID. In the Case Patients Group, 65 people had such contact while only 23 in the Control Patients Group reported the same. That’s almost three times as many who had close contact with known COVID. Why wasn’t that the clickbait header?
You could go on and on.
The CDC in the report did post caveats about this study. I guess most media workers were too busy to report on those caveats. Also, bear in mind this study is based on self-reportage of participants; boldface has been added by me:
The findings in this report are subject to at least five limitations. First, the sample included 314 symptomatic patients who actively sought testing during July 1–29, 2020 at 11 health care facilities. Symptomatic adults with negative SARS-CoV-2 test results might have been infected with other respiratory viruses and had similar exposures to persons with cases of such illnesses. Persons who did not respond, or refused to participate, could be systematically different from those who were interviewed for this investigation. Efforts to age- and sex-match participating case-patients and control-participants were not maintained because of participants not meeting the eligibility criteria, refusing to participate, or not responding, and this was accounted for in the analytic approach. Second, unmeasured confounding is possible, such that reported behaviors might represent factors, including concurrently participating in activities where possible exposures could have taken place, that were not included in the analysis or measured in the survey. Of note, the question assessing dining at a restaurant did not distinguish between indoor and outdoor options. In addition, the question about going to a bar or coffee shop did not distinguish between the venues or service delivery methods, which might represent different exposures. Third, adults in the study were from one of 11 participating health care facilities and might not be representative of the United States population. Fourth, participants were aware of their SARS-CoV-2 test results, which could have influenced their responses to questions about community exposures and close contacts. Finally, case or control status might be subject to misclassification because of imperfect sensitivity or specificity of PCR-based testing (9,10).
Most media carelessly report information the government provides. Any media outlet reporting on sensitive or critical matters should not be taken seriously if they do not require “reporters” to read the actual information they are presenting to you. NPR editors should, not for the first time, hang their heads in shame.
I know a lot of small business owners who depend on bars and restaurants for their livelihood. The lack of regard shown this constituency by media is appalling.
~~Featured photo: Transmission electron microscopic image of an isolate from the first U.S. case of COVID-19, formerly known as 2019-nCoV. The spherical viral particles, colorized blue, contain cross-sections through the viral genome, seen as black dots. (Image by Hannah A Bullock; Azaibi Tamin)
(Kay B. Day/Sept. 14, 2020)
The new Web is weighted towards big partisan media outlets, not small publishers. The only way we can stay online is by relying on the support of our readers.