More than a month since the FDA issued an emergency authorization for the United States’ first COVID-19 vaccines, we have the beginnings of a useful federal dashboard for tracking the rollout, but many important data points are still not being published. We wrote about the government’s failure to release crucially important demographic data for The Atlantic this week, and also published a post about the lack of facility-level data for vaccinations in nursing homes and other long-term-care facilities. We suspect that federal agencies will eventually release the detailed data they have—but we need it now, while there’s still time to identify and remedy problems in distribution and administration.
Although we are not tracking vaccine data from states and territories—something other organizations, including Bloomberg and Johns Hopkins are doing—we have continued to study the metrics that jurisdictions are releasing. We are finding many of the same problems in vaccination data that we saw in other classes of COVID-19 data: Despite being required to report standard data to the CDC, states are making their own independent decisions about what data points to report to the public and how to report them. Our analysis of early COVID-19 immunization data reporting suggests there will be major obstacles to understanding who is getting vaccinated in the United States unless the federal government steps in.
Our detailed analysis of race demographics in vaccine data collects the category names used by the states for the race and ethnicity data they report, as well as the corresponding standard categories used by the Census and in other federal datasets. We are also tracking whether states and territories are providing race, age, and gender or sex data for doses administered, people vaccinated once or twice or ever—or if it’s unclear how their demographic breakdowns map back to vaccination metrics. In addition to data from the 50 states and territories reporting vaccine data so far, we include data from Philadelphia and New York City, because Pennsylvania and New York State exclude these cities from their figures.1
Demographic data is incomplete or entirely missing
In a recent topline analysis of the COVID-19 vaccine data reported by US states and territories, our researchers found only 17 states releasing data on race and ethnicity of vaccine recipients. That’s just a third of the 50 states and territories that are reporting immunization data. Only a few more states—19—report data on the gender or sex of vaccine recipients, and only 21 report age data.
Even when states do publish demographic data, most are leaving out critically important information on who is being vaccinated. Very few report data in ways that allow us to see recipients by age and race, for example. Many stumble over how to report demographic data across the two doses needed for the two vaccines the FDA has authorized.
Inconsistent categories make comparisons impossible
As with other demographic information about COVID-19, in the absence of federal guidelines for public reporting, states have chosen to define their demographic categories differently, preventing comparisons across states. Within states, comparisons are also difficult, because many use a different set of demographic categories for vaccine data than for all their other COVID-19 data. We’ve found that some states lump multiple race categories together, omit many categories entirely, or double-count recipients by reporting them in every one of the race/ethnicity categories they choose. We’ve also found that states are also using varying age brackets: For example, Alaska brackets age in increments of ten from zero, but Florida starts at 16, continues with 16-24, and then counts up in increments of ten years from there. These inconsistencies make accurate comparisons between states—or any national understanding of the vaccination campaign—impossible.
Ultimately, we believe that these problems cannot be resolved at the state level. We recommend that the federal government offer guidelines to states on the public reporting of all COVID-19 data, including vaccine data, to allow for greater consistency between state and federal data and across state datasets.
What the federal government should publish
When the CDC’s vaccination data tracker went up in late December, we outlined what we hoped to see in future releases. A little under a month later, the federal government has made some big improvements to its public data: It has added state-level data, started tracking individuals across their first and second doses, and added a topline count of individuals vaccinated in long-term care facilities through the federal Pharmacy Partnership program. As a next step, the federal government should post the following aggregate demographic data in absolute numbers (not percentages) including:
age data broken down by standard age brackets
race and ethnicity data broken down by standard categories
cross-tabulated data for as many categories as possible, including by age and race/ethnicity
Only the federal government has the resources and data access required to publish a maximally complete and consistent set of demographic data for COVID-19 vaccinations. Unless they take action to publish this data, we will not know if the country’s many vaccine campaigns are actually working: Important successes may never come to light, or grave inequities may escape public view until it is too late to correct course and protect our most vulnerable populations.
Charlotte Minsky contributed additional research and reporting for this article.
More “Demographic Data” posts
Publicly available federal race and ethnicity COVID-19 data is currently usable and improving, although it shares many of the problems we’ve found in state-reported data.
While our work to compile COVID-19 data has concluded, we will continue to share research, analysis, and documentation in the months ahead. We are enormously grateful to the hundreds of volunteers who made this work possible.
We know COVID-19 is affecting Black, Indigenous, Latinx, and other people of color the most. But we need more and more standardized data to truly understand the impact to these communities—and to mitigate those disparities.