05/04/2021
Izabela SadejIzabela Sadej, MSW, is a policy analyst at Boston University School of Public Health. She tweets at @IzzySadej.
The COVID-19 pandemic has resulted in most communities taking precautionary, sometimes mandatory, measures to reduce the risk of spreading and contracting the virus. Physical distancing, also known as social distancing, has been one of the primary strategies adopted by various states and localities as a prevention tool. This typically includes the closure of schools and businesses and “stay-at-home” orders.
Since the onset of the pandemic in March 2020, existing systemic health disparities and social inequities have been magnified. Evidence indicates that residents of low-income neighborhoods were less likely to stay–at–home in response to COVID-19 compared to higher-income communities. These communities carry an unequal disease burden with higher confirmed caseloads and mortality rates, alongside financial constraints that impact low-income workers the most, given less of an ability to work-from-home. Many essential businesses are staffed by predominantly low-wage workers forced to choose between risking their income and exposure to COVID-19. This increased inequity, which remains unaddressed by public policy, has led to further research.
New Research
A recent study published in Nature Human Behaviour expanded the current evidence base by investigating the relationship between neighborhood income and physical distancing patterns during the COVID-19 pandemic in the United States. The authors hypothesized that: 1) the gap in physical distancing practices would be explained by work demands and not by visits to non-work locations, and; 2) state policies that ordered the closure of non-essential businesses and “stay-at-home” orders would contribute to the gap in physical distancing practices between low- and high-income communities.
(Academic affiliations of the authors for this study include Boston University School of Public Health Departments of Community Health Sciences; Global Health; Health Law, Policy and Management; and Epidemiology.)
Methods
Using longitudinal mobility data derived from smartphones and previous Census data on neighborhood income, the physical distancing practices of low-income neighborhoods were compared to higher-income neighborhoods. This was done by identifying mobility patterns for work-related activities and visits to non-work locations (liquor stores, carryout restaurants, convenience stores, hospitals, parks, places of worship, supermarkets) and comparing patterns two months before and after March 2020.
A series of analyses were conducted to distinguish the use of smartphone mobility data. This information was collected through SafeGraph, with an average sample size of 19 million smartphone devices used per day throughout the nation, aggregated by U.S. Census Block Groups (BGs). The three main mobility patterns were measured as follows:
- “Staying at home” was determined by a smartphone user’s overnight location for most nights during the previous six weeks, with the device being observed within the home location and nowhere else on a given night.
- “Working outside of the home” was observed through similar location tracking metrics as above and observing behavior consistent with full-time, part-time, or delivery work.
- “Visits to non-work activities” were observed through counting the number of visits to non-work locations listed above within BG income levels.
State-level social distancing policies were identified through a public database that tracked media coverage and verified government websites. To assess the impact of these policies on mobility, a differences-in-differences model was used to estimate how much physical distancing changed in each state after the implementation of a stay-at-home (SAH) order, calculated separately for each income quintile. This model allowed researchers to take advantage of the fact that states timed their SAH orders differently, if they implemented them at all.
Findings
Lower-income communities were found to increase physical distancing less than higher-income communities. There was no significant evidence that non-work activities contributed to these differences. Instead, differences were caused by lower-income community residents continuing to work outside of the home.
Physical distancing orders were associated with increased physical distancing activity, though the magnitude of policy effect was modest at every income level compared to overall trends. Residents of all-income levels began physical distancing before the implementation of state orders, with pre-trends steepest at high-income levels. State policies did little to level the disparities in social distancing among income levels.
The primary limitation of this study is the inability to generalize the findings. Smart phone ownership and usage varies across sociodemographic groups, with the possibility that teens from higher income neighborhoods may have been over-represented in the sample. Additionally, SafeGraph data could have systematically miscounted the number of smartphone users staying at home or going to work due to irregular data collection intervals. Lastly, local jurisdictions and mandates were not considered when measuring the impact of state policies.
Conclusions
The results of this study and the impacts of COVID-19 emphasize the importance of incorporating social and economic factors into public health responses. Differences in physical distancing patterns based on income level were noticeable yet state policies did not close this gap. It is crucial for policymakers to consider how existing health disparities within certain communities may be exacerbated by new policies. Many unintended consequences can be anticipated and, thus, mitigated.
Simultaneously employing other policies alongside physical distancing mandates, such as eviction moratoriums, mandating paid sick leave, and extended unemployment insurance would allow lower-income communities to better protect themselves. A more equitable COVID-19 response would include widespread adoption of these measures, just like the adoption of non-essential business closures and stay-at-home orders.
