Top 3 Things the U.S. Can Do Now To Better Fight COVID-19
One of the biggest challenges in understanding and combating COVID-19, especially with schools reopening soon, is the lack of consistent data and reporting. Public Democracy has been tracking COVID in the U.S. since January and working with partners to successfully engage marginalized and underrepresented communities to help mitigate the spread. That work — and the efforts by everyone striving to fight this disease — is being hampered by incomplete and inconsistent COVID data collection and reporting. These are huge problems, but they are also relatively easy to fix.
Public Democracy shifted most of our focus to COVID back in March, when our behavioral models began identifying COVID-19 outbreaks around the country before the first positive tests came back. These breakthroughs resulted in our joining the White House COVID-19 Data and Research Task force and the opportunity to work with Ochsner Health in New Orleans to help design and recruit what became the first large-scale COVID-19 trial to fully enroll a representative population sample. We’ll complete the second today in Baton Rouge.
We have learned a great deal through that process, especially about the vital information that is missing. Here are the 3 most important data efforts that can be done now to better understand how to defeat COVID:
- Standardize terms, reporting, and data usage
- Shift focus from daily test results and deaths to a more reliable and intentional sampling to establish actual infection rates (they are likely 10x higher than reported because of asymptomatic cases).
- Demand and expect full-inclusion of minorities in data collection and reporting (peer-review and current standards expect minorities to not be fully counted, which harms their communities and understanding in general).
Let’s review why each of these matter.
1. Standardize terms, reporting, and data usage
Lack of standardization has muddled national data reporting, compounding any inaccuracies already baked into state-level data. National reporting of Hispanic infection rates is one of the most eye-opening examples of how data consolidation does not work when localized data is already faulty and varied in what it measures.
For the first several months of the pandemic, patients who tested positive for the virus were not able to identify as Hispanic in many states. Instead, they were instructed to indicate “Other,” erasing key information for understanding who is being impacted by the virus. These massive holes in the data resulted in headlines highlighting high infection rates in the states that were counting and a very skewed understanding of Hispanic death rates.
“Without consistent reporting and agreement on terms and data, it is much harder to understand COVID and identify gaps where more data is needed.”
Even now that states are required to report Hispanic data, the confusion continues as there is still not a uniform method of recording race/ethnicity. Some states count ethnicity independently of race while others combine the two. Some report in real-time and others release tallies at month’s end. This inconsistent counting makes integrating state data into any national or historical database a sloppy, inaccurate process.
This issue is not limited to Hispanics. Across the board, there is lack of consistency in reporting on testing (overall and by group). Even COVID deaths and positive cases are not reported consistently to allow for effective comparisons between states or countries. China recently admitted to only counting “confirmed cases” for people who both test positive and actively exhibit symptoms, artificially lowering its infection rate compared to those of other countries. This matters as research publications and the John Hopkins COVID-19 map report China’s modified data with numbers in the same tables as the rest of the world, which count people who test positive as confirmed.
Without consistent reporting and agreement on terms and data, it is much harder to understand COVID spread and risk, as well as identifying gaps where more data is needed.
2. Intentional sampling
Holes in the data will only be filled with intentional sampling efforts designed to account for the many factors affecting an individual’s risk. Because of the United States’ diversity and how social determinants of health are biasing who is tested, the only way to collect all-inclusive data is with effective community-specific sampling. Prevalence studies can do this.
As the New Orleans prevalence study demonstrated, studies are able to account for a variety of community-specific social determinants — like New Orleans’s concentrated Vietnamese population and above-national-average rate of single mothers. Intentional sampling designed to address self-selection bias results in nuanced data that is able to zero in on infection rates by zip code, ethnicity, socioeconomic status, and even education level.
This also addresses another huge problem in current COVID data: most infectious people do not have symptoms, which makes them much less likely to have a doctor recommend they get tested or get tested on their own. Without good data on asymptomatic spreaders and infected Americans, we’ll reach the wrong conclusions.
Well-designed prevalence studies are an extremely accurate and insightful tool for data collection. They empower leaders with community-specific information to better respond to outbreaks in their specific jurisdictions. Currently, these studies indicate infection rates much higher and more unevenly distributed than self-reported testing is showing.
3. Demand and expect full-inclusion of minorities
Prevalence studies can also give an accurate picture of the pandemic’s impact on people of color, but they must be designed to enroll a diverse, representative population sample. Because prevalence studies can estimate COVID-19 infection rates independent of state data collection methods, groups like Hispanics that were excluded from many state-reported counts can be fully represented in these studies. But fully-inclusive samples are not the norm or expected (e.g., Black Americans are generally undercounted by ⅔ in clinical research). That must change — and it must change now in order to help the Americans most impacted by COVID-19.
In the New Orleans study, Public Democracy was able to recruit the sample for what became the first large clinical research study to achieve a fully-representative population sample. This complete, real-time picture helps reverse some of the damage done to minority communities during the first months of the pandemic, replacing inaccurate data and showing the extent to which they are being disproportionately affected by COVID-19.
Every community in America is experiencing COVID-19 differently. We must understand the details of how their experiences differ so that responses can be tailored to meet each locality’s unique needs as this pandemic continues to spread across the United States.
Over the next few weeks, we will release a series of posts providing a deeper look at what is possible when we fill holes in COVID data, who is most affected, and the risks to us all if we don’t fix things soon. Our next post will explore how better data on Hispanic infection rates provides key insights into how to best reopen schools and local economies.
Links to that piece and the others that follow will be posted below, so please check back here for further updates.
