Question: So how well done was this study that is garnering headlines? I have reservations. (Couldn’t resist the pun.)
Answer: You’re punny! Hehehehe…. le sigh. Good question too! Three days ago, CDC released results from a small study, “Community and Close Contact Exposures Associated with COVID-19 Among Symptomatic Adults ≥18 Years in 11 Outpatient Health Care Facilities — United States, July 2020.” Two of the study’s findings have made a lot of waves — dining at restaurants and going indoors to bars/coffee shops are both associated with increased odds of COVID infection. As with all studies, this one has strengths and limitations. In my opinion, it’s still too small and too limited to make definitive conclusions, so you’re right to have reservations. That said, we have evidence of transmission events occurring in both indoor restaurant and bar settings (see Q&A of 6/3), and this study adds to the evidence base. In fact, I think it’s the first (or one of the first) studies to look at exposures by outcome, which is a major strength of this study. My main take-away is to keep getting take-away (hahahaha!). Seriously, please keep minimizing risk.
Now a bit more on a few of the study’s limitations — Because this study did not differentiate between indoor/outdoor dining, it’s unclear what the level of risk is for the type of dining. We expect that indoor dining carries higher risk than outdoor dining (see Q&A of 5/20). Second, because the study lumped indoor visits to bars/coffee shops together in the questions asked, and did not differentiate by service delivery method, we cannot know whether the risk identified pertains only to bars, only to coffee shops, to both, and for what type of service delivery method. Furthermore, the number of individuals included in the bars/coffee shops analysis is so small (only 13 cases and 8 controls said yes to the question); really too small to draw conclusions. The same can be said for several other community exposures reported for which no association was found (e.g. gyms, church, public transport) — with so few cases/controls reporting these exposures, the study does not have the power to identify elevated risks. Finally, I wonder whether people who report dining at restaurants and/or going to indoor bars/coffee shops have other behaviors that are more risky and not measured in this study. Perhaps these behaviors are proxies for other risky behaviors — like more comfort/more exposure to sitting for prolonged periods with groups of people in myriad types of settings?
Study Overview
- Study question: Do community and close contact exposures differ by infection status (infected vs uninfected)?
- Methods: Symptomatic adults who recently tested positive for COVID-19 (cases) and symptomatic adults who recently tested negative for COVID-19 (controls) were recruited from 11 health care facility settings across 10 states to participate in the study. Both cases and controls were asked a series of questions about their activities in the 14 days before receiving a COVID-19 test (e.g. their community and close contact exposures). Researchers then compared exposures between cases and controls, adjusting for age, sex, race/ethnicity, and presence of one or more underlying chronic medical conditions.
- Results: Cases were far more likely to have been in close contact with a COVID-19 infected individuals, most of whom were infected family members (42% of cases vs. 14% of controls). For all measured community exposures, only one was statistically significantly higher among all cases — dining at a restaurant. Indeed, restaurant dining was found to be associated with an approximate two to three times increased odds of COVID infection (Figure 1). The study authors state, “case-patients were more likely to have reported dining at a restaurant ([adjusted odds ratio (aOR)] = 2.4, 95% CI = 1.5–3.8) in the 2 weeks before illness onset than were control-participants. Further, when the analysis was restricted to the 225 participants who did not report recent close contact with a person with known COVID-19, case-patients were more likely than were control-participants to have reported dining at a restaurant (aOR = 2.8, 95% CI = 1.9–4.3) or going to a bar/coffee shop (aOR = 3.9, 95% CI = 1.5–10.1).
- Limitations: The study authors list five limitations of their study: 1) The controls may have had similar exposures to the cases since the controls were a group of symptomatic adults who tested negative (e.g. different from the general population), and a number of potential participants did not respond or refused to participate and they may be systematically different from those who agreed to participate; 2) Unmeasured confounding — reported behaviors might represent activities/behaviors not represented/asked about in the study. Especially limiting is that the study did not distinguish between indoor/outdoor dining; 3) Study participants were recruited from 11 health facilities and may not be representative of the broader US population; 4) Participants were aware of their COVID test result, which may have biased their reporting; 5) Case/control status may be misclassified because of imperfect sensitivity/specificity of PCR tests.
Figure 1. Adjusted Odds Ratio for Community Exposures (from CDC)
