U.S. counties with lower per capita income and greater Republican orientation were associated with significantly reduced physical distancing during the COVID-19 pandemic, according to a new study that shows how government policies failed to close disparities between communities throughout almost the entirety of the pandemic.

As the COVID-19 pandemic developed, researchers saw how it was hurting communities and how difficult it was for many people in their community to protect themselves from infection, according to Nolan Kavanagh, a co-author of the study and medical student at the University of Pennsylvania and lecturer at the University of Michigan.

"We wanted to understand the structural factors, whether social, economic or political, that were putting communities at risk of infection," he said. "And so, our goal was to produce a guide for policymakers and other concerned individuals to rapidly identify communities that were at risk of infection, and ideally, develop policies and communications to direct resources to protect them."

This research, published March 24 in the American Journal of Preventive Medicine, follows a similar study, published in March, that showed that states with GOP governors had worse COVID-19 outcomes. 

Using the cellphone records of 15-17 million people in 3,037 counties obtained from Unacast, researchers were able to calculate county-level averages of distance traveled per person. The researchers used the distance that cellphones traveled as a proxy for social distancing from the beginning of the pandemic, set as March 9, 2020, through Jan. 17. 

The researchers also collected data from the American Community Survey to determine county-level socioeconomic status, measured as income per capita. Political orientation, in turn, was measured as the 2016 vote share for Donald Trump, obtained from the MIT Election Data and Science Lab. 

Since the data the researchers collected was not individual-level data, they were not able to tie demographics, socioeconomic status or political affiliations to the individuals whose devices were followed. Rather, this research speaks to a group-level analysis, capturing community-level determinants of social distancing.

"What we found is that nearly a year into the pandemic," Kavanagh said, "There continue to be disparities and physical distancing between communities for higher and lower socioeconomic status, and between communities that lean Democratic and that lean Republican."

The researchers were not surprised to find that low-income communities and Republican-leaning communities were less likely to physically distance when they started the study in April, according to Kavanagh, since their findings were consistent with news reports and the little information there was on the virus and policy responses to it. 

However, as the pandemic continued to unfold, the researchers were surprised to find that these disparities between communities did not narrow as time went on, suggesting "that to some degree, our policy and communication responses to the pandemic haven't been able to level the playing field between communities," Kavanagh said. 

Specifically, the researchers found that greater county-level income per capita was highly correlated with more physical distancing, or less movement, relative to the other counties studied. 

Using the interquartile range — comparing the 25th percentile to the 75th percentile — as a rough measure, the researchers were able to illustrate how wide the distances were between counties of lower socioeconomic status and higher socioeconomic status. 

For the month of May, the researchers found that increasing per capita income by the interquartile range — from $22,700 to $29,918 — resulted in a 3.8 to 4.5 percentage point decrease in average movement, depending on the week. 

These findings reflect the reality that lower-income households may be less able to work from home, and they may not have sufficient money on hand to buy items in bulk, necessitating more trips to the store, according to the study.  

Additionally, the researchers found that counties with a greater share of votes for Trump in 2016 were less likely to physically distance and had increased movement relative to the other counties studied.

Similar to income per capita, for the month of May, the researchers found that an interquartile range increase in support for Trump from 54.3% to 74.6% resulted in a 4.8 to 6.7 percentage point increase in average movement. 

This finding may reflect to some degree the beliefs of individuals, "but for the most part, it's a reflection of the messaging that comes from our political leaders," Kavanagh said. "And it starts at the top. There has consistently been polarization among political leaders over the pandemic that has shaped how individuals respond."

April 2020 was one of the months with the highest engagement in physical distancing, according to Kavanagh, who reported that the average county had a 31% drop in movements compared to before the pandemic. 

However, once November arrived, the average county had about 19% less movement than before the pandemic, according to Kavanagh. 

"For any county, that might correspond to thousands and thousands of miles traveled, depending on the size of the county and the number of people in it," he told The Academic Times.

Researchers also found that during the early months of the study period, counties with greater shares of Black and Hispanic residents were less likely to engage in physical distancing, which was explained by structural barriers facing Black and Hispanic residents.

For example, Kavanagh said, Black and Hispanic people are more likely to work low-wage jobs, be essential workers and have less wealth available to weather a crisis like COVID-19.

"On top of that, racial and ethnic minorities experience racism and discrimination in government and health care institutions," he added. "All of these structural barriers put them at higher risk of getting COVID-19 and suffering its severest complications. These barriers, rather than cultural differences, likely explain the racial and ethnic disparities observed at some points in the study."

Kavanagh said these disparities suggest that policymakers need to do a better job designing policies and communications that support communities struggling the most to socially distance.

"That could take the form of economic policies that support families that are otherwise unable to engage in social distancing," he said, "but it also requires leadership from our political leaders and public health leaders to ensure that the communities that have yet to be reached by our messaging are in fact reached."

The study "County-level socioeconomic and political predictors of distancing for COVID-19," published March 24 in the American Journal of Preventive Medicine, was co-authored by Nolan M. Kavanagh, Atheendar S. Venkataramani, and Rishi R. Goel, University of Pennsylvania.