RR:C19 Evidence Scale rating by reviewer:
Reliable. The main study claims are generally justified by its methods and data. The results and conclusions are likely to be similar to the hypothetical ideal study. There are some minor caveats or limitations, but they would/do not change the major claims of the study. The study provides sufficient strength of evidence on its own that its main claims should be considered actionable, with some room for future revision.
This study makes reliable, well substantiated, and generally supported claims that essential occupational sectors represent a key site of intervention for preventing excess morbidity and especially mortality from SARS-CoV-2 infection and COVID-19. The main suggestions would be to consider the intersection of sector-ethnicity-occupation slightly more systematically throughout the study and to consider the role of non-work/workplace factors that might further spread infection among these groups of workers. This study provides some much-needed insights on contextual (e.g., not just demographic, or biomedical) factors related to SARS-CoV-2 infection to better highlight some of the most vulnerable groups and the sites or environments in which socio-behavioral and/or industrial and environmental interventions can be employed to reduce spread. In this sense, the study substantially advances understandings of who is most affected by the pandemic and why.
There are a few points regarding how the data are presented that could be improved, largely in the abstract and discussion sections. Essentially, the authors have identified both excess mortality in (1) specific sectors, (2) specific ethnic/racial groups, and then (3) specific combinations of certain ethnic minority workers working in specific sectors, adding (4) general data on excess mortality in specific occupations to further contextualize the ethnic and sectoral differences. I would like to see the discussion reflect the excellent data in a more systematic way. For example, discuss first which sectors have the greatest excess mortality and why; second, which ethnic groups have the greatest excess mortality and then, (3) the ethnicity/race-sectoral combinations which the excess mortality (in terms of RR) exceeds, say, both the general RR for the ethnic/racial group and for the sectoral group by at least .20 (or a threshold of the authors’ choosing).
There is brief reference to some group-specific patterns (e.g., high excess mortality among Asian healthcare workers and Latino agricultural and production workers), but the explanation is sometimes a bit muddled (e.g., referring to demographics rather than specific explanations). These potential explanations could be fleshed out and presented more systematically. In terms of general ethnic/racial and sectoral disparities, for example, why might excess mortality for Latinos be so high in general and even in non-essential occupations? The extant literature or the occupational data featured within the study—perhaps also stratified by race/ethnicity (as with sectors)—could be used to better contextualize these specific patterns more systematically. For example, whether the predominance of a specific occupation held among a specific ethnic group working in that sector (e.g., prevalence of certain ethnic groups employed in food service compared with other types of retail) or in combination with demographic factors (e.g., the health care worker group for some ethnic groups being more comprised of medical doctors who are likely to be older individuals). The abstract should be a summation of the most important points from this systematic analysis—currently, the sum of the results being highlighted in the abstract seem cherry-picked with limited context.
Relatedly, at least some brief attention could be given to the non-workplace factors contributing to elevated excess mortality for some groups. Some of this may have to do with where various groups live and the compounding influence of place/neighborhood on spread, in addition to actual housing stock (which is mentioned briefly). It would be helpful to—at least in a few sentences—to consider the broader intersection of factors associated with the nexus of work and living arrangement (e.g., carpooling, socializing after work in the same places, shared housing or neighborhoods among multiple individuals working at the same facility or set of facilities). Consider the work by Bui et al. (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7439983/) and Lewis et al. (https://pubmed.ncbi.nlm.nih.gov/32970656/) in Utah as a close complement to this California work and as a way to extend the discussion on excess mortality for these groups. A few last technical considerations would be as follows:
(1) State upfront which years of data (2018 and 2019) are being used as the comparators for 2020 rather than referring to historical periods generally—currently, the years are not identified until the end of the methods section.
(2) It may be helpful to use death rates (e.g., per 1,000 or 100,000) rather than raw death counts for context and comparability to other states.
(3) It would be helpful to stick to just one mode of reporting excess mortality, e.g., RR rather than % increased risk.