Pandemic Strategies Should Factor Ethnicity and Social Contact Patterns, Study Finds

When the CDC rolled out its COVID-19 vaccination strategy in late 2020, the agency sorted people into priority groups by age, occupation, living situation, and medical risk. It was a logical framework. But it was also incomplete. A team of researchers at Iowa State University has now shown, through computational modeling of nearly three million different vaccination scenarios, that public health officials missed something crucial: the way disease spreads depends not just on who you are, but on who you interact with and where you live.

Claus Kadelka, an assistant professor of mathematics, led the work. His central argument is straightforward but carries weight: when you ignore ethnicity and social contact patterns, you get the epidemiology wrong. Most researchers in this field focus almost entirely on age—the intuitive question being whether to vaccinate the elderly first, since they face the highest risk of severe illness, or younger people first, since they have more contacts and spread disease more readily. But that framing leaves out entire dimensions of how disease actually moves through a population.

The data tells a stark story. In 2020, infection rates in predominantly Black counties were three times higher than in predominantly white counties. The Navajo Nation had more cases per capita than any state. These disparities were not random. People of color are concentrated in jobs that cannot be done remotely—transportation, grocery stores, meat processing—where physical distancing is impossible. They are also more likely to live in dense housing or multigenerational homes where quarantine is impractical. The virus found them where they worked and where they slept.

Kadelka's team built a model incorporating data from the CDC, the Census Bureau, and the Bureau of Labor Statistics. They layered in different contact rates and occupational hazards by both age and ethnicity, then ran their model against 2.9 million possible vaccination strategies using Iowa State's supercomputer. What emerged was a finding that challenges conventional wisdom: the strategy that minimized deaths was not the one that prioritized the oldest people overall. It was the one that prioritized the oldest people of color first—because they face immediate mortality risk—followed by working-age non-Hispanic whites and Asians in high-contact jobs, who could then stop transmitting the virus through their communities.

The researchers call this phenomenon "ethnic homophily," the tendency for people to interact more frequently with others from their own demographic group. When you account for it, the math changes. A vaccination strategy that includes ethnicity as a variable prevents more deaths than one that does not.

Kadelka is careful about claims of real-world applicability. Disease models are powerful tools, but they rest on assumptions about contact patterns and infection susceptibility that remain poorly understood. Still, the work is not purely theoretical. Montana and Vermont both opened vaccine eligibility to people of color ahead of the general public in spring 2021. Three dozen states set aside portions of their vaccine supply for residents facing disadvantages in income, education, and housing. The question now is whether future pandemic responses will be designed with this kind of granular thinking built in from the start, rather than added as an afterthought.