Does COVID-19’s toll reflect social inequality? Early evidence from NYC
It is likely that people of color, immigrants, and the working class will bear a disproportionate burden of the COVID-19 pandemic in the US. Workers, and people who are black or indigenous (or from certain other racial/ethnic groups) already have high prevalence diabetes, heart disease, and other conditions that are associated with increased COVID-19 severity.
Add to this the “essential workers” who risk exposure while running what remains of our economy. These include delivery drivers, cashiers, garbage collectors, restaurant workers — occupations often filled by immigrants and people of color.
We do not yet have data on the race, ethnicity, nativity, or socioeconomic background of COVID-19 hospitalizations or deaths. Here, I explore early data from New York City to look at associations between characteristics of neighborhoods (using ZIP code tabulation areas) and indicators that may reflect COVID-19 prevalence.
Emergency Department Syndromic Surveillance
One data source we have is the number of emergency department (ED) visits for “influenza-like illness” (ILI) for each ZIP code in NYC — these are visits with symptoms such as fever and cough that could be the flu, but in 2020 are largely COVID-19. I calculated the rate of ED visits per neighborhood for the adult population in March 2020. I compared that to the average rate for March in 2016–2019 to calculate rate differences. There are caveats here — the health department warns that data from the last two weeks are not considered final, and even though I accounted for ED visit rates in prior years, it could be that people in some communities are more likely to seek care for COVID in the ED while others are using telemedicine or other sources.
Every ZIP code in NYC experienced an increase in ILI ED visits. The whitest neighborhoods saw the lowest rates of increase, however. Every neighborhood that experienced an increase of at least 5 visits per 1,000 population was majority people of color.
ED visit rates for ILI also increased in neighborhoods with larger Latinx and immigrant populations. The patterns weren’t as clear for % black or % Asian, which I don’t show here.
Looking at neighborhood poverty, we see the same pattern — higher-poverty neighborhoods have higher increases in ED visits for ILI.
The strongest apparent relationship was between crowding and ED visit increases (crowding is defined as the % of rental units where there is more than 1 occupant per room). We can’t infer causation here, but this is in line with observations showing within-household spread of SARS-CoV-2 is an important source of infection, and in a crowded household it would likely spread more easily.
Testing
Testing in NYC is being rationed and the city health department has advised providers to only test those who are hospitalized or have severe illness. As with the rest of the US, testing is not widely available enough to help with efforts to isolate people who are infected.
Even so, we see there is more testing in whiter and wealthier communities. These are not particularly strong relationships, however, and it’s likely other cities with less restrictive testing have stronger inequalities in testing rates.
The NYC health department makes several neighborhood testing measures available, including the percent of tests that come back positive. One concern here is that socially privileged people will have more access to testing even for those unlikely to have COVID-19. One could surmise that whiter and wealthier neighborhoods might have a low % of tests that come back positive solely because testing of less severe illness is more widespread. But this might not be as much of a concern in NYC. There is no clear linear relationship between a neighborhood’s testing rate and the percentage of tests that are positive for COVID. So % of tests that are positive might be a reasonable proxy when looking at the geographic distribution of COVID-19 in NYC.
Using % testing positive reveals much the same patterns as the ED visit data above. We see the same relationship, for example, of whiter neighborhoods having COVID-19 tests that are more likely to be negative.
