Covid-19: ensuring equality of access to testing for ethnic minorities

Dear Editor

The editorial by Dodds and Facoya[1] – especially the discussion of the ways complexity of instructions for self-tests and mistrust of institutions can contribute to comparatively low Covid-19 testing rates for persons whose first language is not English and other racial/ethnic minorities ¬– highlights the importance of the disparities measurement issued I discussed in rapid responses of 28 April,[2] 1 May,[3] and 13 May 2020 [4] to a 17 April article by Rimmer.[5]

The responses explained that generally increasing favorable outcomes like the receipt of screening tends to reduce relative racial/ethnic differences in screening rates but increase such differences in rates of failing to be screened. The 13 May response also discussed the anomaly in the US where the National Center for Health Statistics tends to find general increases in screening and vaccination rates associated with reduced racial/ethnic disparities (measured in terms of relative differences in receipt of the procedures) while the Agency for Healthcare Research and Quality in its annual National Healthcare Disparities Reports tends to find the general increases associated with increased racial/ethnic disparities (measured in terms of relative differences in nonreceipt of the procedures).

The responses did not discuss the measurement of health or healthcare disparities in terms of absolute (percentage point) differences between rates, which are unaffected by whether one examines the favorable or the corresponding adverse outcome and which some may well use to measure Covid-19 testing disparities. But the first three references of the 28 April response (reference 6-8 below) discuss the way that absolute differences tend also to be affected by the prevalence of an outcome, though in a more complicated way than the two relative differences. Roughly, and assuming normal underlying risk distributions, when rates are below 50% for both groups being compared, general increases tend to increase absolute differences between rates (so long as rates remain below 50%); when rates are above 50% for both groups, general increases tend to reduce absolute differences. The matter is more complicated when initially the rate is below 50% for one group and above 50% for the other or when either group’s rate crosses 50% during the period examined.

Covid-19 testing rates are probably going to be in ranges where general increases tend to increase absolute differences. But among especially vulnerable subgroups, the rates may initially or eventually be high enough that further increases will tend to reduce absolute difference or make the prevalence-related pattern difficult to predict.

Policies can reduce the forces causing testing or other outcome rates of advantaged and disadvantaged groups to differ and such policies will reduce all measures of disparity. They will thus counter partly or entirely the tendency for general increases in favorable outcomes to increase certain measures of disparity and further reduce the measures of disparity that the general increases tend to reduce. But one can determine the effects of the policies only when one understands the patterns by which the measures employed tend to change as the prevalence of an outcome changes. One can then can explore the extent to which observed patterns of changes in measures are solely function of changes in the prevalence of an outcome and the extent to which they reflect something else (or, and preferably, as discussed in the 28 April and 13 May responses, employ a measure that is unaffected by the prevalence of an outcome)..

References 6 to 8 address the way some researchers discuss value judgments in the choice of measure in circumstances where the relative difference they happen to be examining (that is, in the favorable or the corresponding adverse outcome) and the absolute difference support opposite conclusions about such things as whether disparities are increasing or decreasing. It should be evident, however, that determinations of whether something like cultural competence training reduced the forces causing the outcome rates of two groups to differ, while not uncomplicated, are in no way aided by value judgments.

Effects of misunderstanding of measurement issues on mistrust of institutions among racial/ethnic minorities warrants special mention. Whereas most discussion of disparities issues reflects no recognition that measures tend to be affected by the prevalence of an outcome, a notable exception may be found in the implied assumption that reducing an adverse outcome ought to reduce relative differences in rates of experiencing the outcome. It is most often reflected in statements that a relative difference in an adverse outcome persisted or increased “despite” a general decrease in the outcome. But it is because of, not despite, general reductions in adverse outcomes that we commonly observe increasing relative demographic differences in the outcomes (along with decreasing relative differences in the corresponding favorable outcomes).

This mistaken assumption can have serious consequences in circumstances where distrust within minorities communities is an important issue. For, when measures of disparity increase in the face of actions that are supposed to be reducing them, observers will tend to believe that any racism underlying the disparities must be increasing. See reference 9 with regard to the way that the mistaken assumption in the 2017 Lammy Review of racial/ethnic disparities in UK criminal justice outcomes will tend to increase racial/ethnic minorities’ mistrust of the criminal justice system.

As discussed in the 28 April response, improvements in Covid-19 care will tend to increase relative racial/ethnic differences in adverse Covid-19 outcomes at the same time that the improvements reduce relative racial/ethnic differences in the corresponding favorable outcomes. It would be unfortunate if those observing the former pattern discuss it in a way to suggest that racial/ethnic disparities are increasing notwithstanding development that should be reducing them.

References:
1. Dodds C, Facoya I. Covid: ensuring equality of access to testing for ethnic minorities. BMJ 2020:369;m212. https://doi.org/10.1136/bmj.m2122
2. https://www.bmj.com/content/369/bmj.m1562/rr-2
3. https://www.bmj.com/content/369/bmj.m1562/rr-5
4. https://www.bmj.com/content/369/bmj.m1562/rr-7
5. Rimmer A. Covid-19: Disproportionate impact on ethnic minority healthcare workers will be explored by government. BMJ 2020;369:m1562. 10.1136/bmj.m1562 32303494). https://www.bmj.com/content/369/bmj.m1562
6. Scanlan JP. The monitoring of health inequalities has never been sound. BMJ May 9, 2016 (responding to Mackenbach JP, Kulhánová I, Artnik B, et al. Changes in mortality inequalities over two decades: register based study of European countries. BMJ 2016;353:i1732). http://www.bmj.com/content/353/bmj.i1732/rr
7. Scanlan JP. Race and mortality revisited. Society 2014;51:327-346
http://link.springer.com/article/10.1007%2Fs12115-014-9790-1#page-1
8. Scanlan JP. The mismeasure of health disparities. J Public Health Manag Pract 2016;22(4):416-19.
https://journals.lww.com/jphmp/Fulltext/2016/07000/The_Mismeasure_of_Hea...
9. “Usual, But Wholly Misunderstood, Effects of Policies on Measures of Racial Disparity Now Being Seen in Ferguson and the UK and Soon to Be Seen in Baltimore,” Federalist Society Blog (Dec. 4, 2019).
https://fedsoc.org/commentary/blog-posts/usual-but-wholly-misunderstood-...