RR:C19 Evidence Scale rating by reviewer:
Potentially informative. The main claims made are not strongly justified by the methods and data, but may yield some insight. The results and conclusions of the study may resemble those from the hypothetical ideal study, but there is substantial room for doubt. Decision-makers should consider this evidence only with a thorough understanding of its weaknesses, alongside other evidence and theory. Decision-makers should not consider this actionable, unless the weaknesses are clearly understood and there is other theory and evidence to further support it.
The authors are to be congratulated to proving further empirical evidence about the racial differences in COVID-19 morbidity and mortality. However, I am not sure that the presented data analysis was as careful (line 61) as it could be? Given the nature of the variables (some being individual, and others being community level), I wondered why something like multi-level (mixed effects) models were not employed? I feel there is a danger that the findings may be subject to the ecological fallacy. I also urge the researchers to be informed by the REporting of studies Conducted using Observational Routinely-collected Data (RECORD) guidelines (see: https://www.record-statement.org/) in reporting their findings. I could not help but think of all the important unmeasured confounders that might be influential – and explain some of the differences noted. “Race” may simply be a marker for these unmeasured confounders. I got no sense of the quality of the variables/data in this study, how they were linked (and the various diagnostics around that), and the completeness of records. Many statistical tests were undertaken, yet no consideration of test multiplicity was evident. And new analyses/results appeared in the Discussion.
Line 29. Replace bivariate and multivariate with bivariable and multivariable, respectively.
Line 35. Insert ‘nearly’ or ‘almost’ in front of ‘twice’.
Line 37. Please include 95% CI.
Line 49. Inset (US) after the United States.
Line 52. Spell out CDC.
Line 54. ‘Latinx’ is new to me, and not widely used by Hispanic or Latino populations (see: https://www.pewresearch.org/hispanic/2020/08/11/about-one-in-four-u-s-hispanics-have-heard-of-latinx-but-just-3-use-it/)
Lines 64-65. The authors wish to “explore[s] whether there is a causal relationship between race, neighborhood factors, and COVID-19 in Cook County, Illinois.” The study design and analysis limit any causal assertions.
Results section: inconsistent precision of reporting percentages and other summary statistics.
Line 122: Why is “Unknown” included in Table 1 for sex – when it is empty?
Line 124. Table 2 is not referenced in the text. It should be at line 124.
Line 126. ‘Univariable’ analysis is referred to – but it is bivariate – consistent with this described in the Abstract.
Line 132. The columns need labelling within the Table; ensure formatting is consistent (ie. “1,234”)
Line 135. I could not see where ** was used?
Line 146. A p-value of 0.0691 is non-significant. It is not ‘marginally significant’.
Line 147. The table title lacks description.
Line 148. I’m not sure I understand what ‘<0.00*’ means? Also, the format of this table is unclear. For Model 4, which IRR refers to poverty. I feel the table construction/layout deserves more attention.
Line 152 and elsewhere. I wonder why multi-level models were not employed? There are individual and community level variables. Individuals are nested within communities (as I understand these data). How will the potential effects of the ecological fallacy be avoided?
Line 165. Again the title is very short.
Line 165. Table 4. Why is it important to report the Estimate, SE, and Wald Chi-square – when the other details are included?
Line 176-178. What is the difference in age of death for minorities compare to Whites not dying of COVID-19? It may be that they generally die younger? This may be insightful – when speaking to unmeasured confounders and the accumulated life experience of disadvantage.
Line 194. As I understand it ‘poverty’ in a community level factor, yet ‘COVID-19 related death’ is an individual level factor. People’s SES will vary (potentially considerably) with those areas designated as being subject to poverty. So this might be an ecological fallacy affect.
Line 198. New analyses are presented within the Discussion. This is contrary to usual/accepted practice.
Line 267. I would be very cautious making the statement that “…that race, more than SES, is the only…”
Limitations: Variable coverage and quality of data? Validity of ‘Race’ variable. Data linkages? Emigration/immigration effects – how are they captured within the dataset.