Declining COVID-19 mortality risk ratios must be interpreted with caution

Abstract

In their recent letter in J Epidemiol Community Health, Schultes et al examined patterns of declining race/ ethnic disparities in COVID19 mortality in Connecticut from March 2020 to December 2021. This work is a helpful contribution to the expanding literature on race/ethnic disparities in SARSCoV2 mortality, and due to its attention to the differential toll of SARSCoV2 in congregate versus noncongregate settings by race/ethnicity is particularly important. However, motivated by our shared concern about the inequities highlighted by Schultes et al, we feel compelled to address the authors’ contention that their ‘findings suggest that attenuation of racial and ethnic disparities is an achievable public health goal’. We believe that this statement reflects a broadly held misapprehension about the meaningfulness of declining mortality rate ratios (MRRs) as evidence of progress against inequity in pandemicrelated mortality. While changes in the MRR for nonwhites as compared with whites—the main indicator of disparity used in the paper—may reflect some progress against race/ethnic and classbased disparities SARSCoV2 in infection and mortality, a narrow focus on declining groupspecific MRRs may paint a far more optimistic picture of this progress than is warranted. On a methodological level, ratiobased measures of SARSCoV2 inequality are made difficult to interpret because of their susceptibility to distortion from change in the denominator: Diminishing MRRs by race may reflect declines in the numerator (the minoritised group in question) or increases in the denominator (an advantaged group, typically whites). In fact, Lawton et al found that much of the declines in countylevel SARSCoV2 MRRs by race were better explained by increasing overall prevalence reflecting increased infection and death among whites than precipitous declines in infection among nonwhites. From a health justice perspective, it is critical that withinpandemic successes in attenuating disparities not be conflated with success in combating racial capitalism and other manifestations of structural racism which contributed to differential participation in hazardous ‘essential work’ and other risks that made irreversible earlypandemic mortality disparities inevitable. We suggest that, at a minimum, researchers and policymakers contemplate a simple thought experiment before concluding that secular declines in race/ ethnic MRRs for COVID19 suggest progress against infection inequity: If a novel pandemic characterised by similar lethality and transmissibility to SARSCoV2 was to emerge in the coming months, would the factors that led to declining MRRs during this pandemic translate into dramatically decreased disparities in death at the beginning of the next one? We doubt it and advocate for caution in interpretation of these declines as a result.