COVID-19 across United States congressional districts

As of July 14, 2020, confirmed coronavirus disease 2019 (COVID-19) cases and deaths in the United States (US) were close to 3.4 million, and 135,000, respectively (Fig. 1). One characteristic of the discourse related to the pandemic was the omnipresence of COVID-19 statistics at the state- and county-levels, spurring important analysis that are of public and policy significance. However, a critical geographic unit for which COVID-19 mortality and case statistics have not been presented are the US congressional districts (CDs). The CDs represent “a territorial division of a state from which a member of the US House of Representatives is elected.”1 The lack of reporting of COVID-19 statistics for CDs is true of other data on population health.2

Fig. 1
Daily COVID-19 cases and deaths for the United States from January 22, 2020 to July 14, 2020.
COVID-19 = coronavirus disease 2019.

• Click for larger image
• Download as PowerPoint slide

One of the reasons statistics on population health and well-being indicators are not routinely available for CDs is because CDs are not straight-forward aggregations of existing smaller Census-based administrative geographies, such as counties; a unit where a wide array of population health data is routinely available.2 Yet, having CD-level statistics can substantially enhance the ability of elected officials and the populations they represent to monitor, evaluate and design programs specific to the needs and vision of the constituents. Other smaller administrative units, such as counties, do not have similar single elected representatives, and thus lack the same level of visibility, and therefore, accountability, that the CD has.

In this study, we provide estimates of COVID-19 cumulative and recent cases, deaths (per 1,000 people) and the case fatality rate (CFR), defined as the ratio of cumulative COVID-19 deaths to cases, for the 436 CDs in the US. The cumulative cases and deaths provide information about the overall burden of COVID-19 in a CD, while COVID-19-related statistics for last month are presented so that elected representatives and their constituents can deliberate their respective reopening strategies.

All but 15 CDs are not straight-forward aggregation of US counties. When only portions of counties lie in a district, it is not a simple aggregation exercise of county rates to a CD, as these portions are irregular in size and shape with different population fractions in them. We used a previously published population-weighted methodology, to derive CD-level COVID-19 statistics from county-level data.2 The approach we have taken falls under the category of dasymetric mapping techniques that accounts for the spatial distribution of the underlying population for performing such an interpolation.3, 4

We used county-level COVID-19 cases and deaths data from January 22, 2020 to July 13, 2020 in the US from USAFacts (https://usafacts.org/visualizations/coronavirus-covid-19-spread-map/, last accessed July 14, 2020) that aggregates official COVID-19 statistics from the Centers for Disease Control and Prevention, state and local-level public health agencies.5 We focused on COVID-19 statistics of cases, deaths (per 1,000 people) and the case fatality rate (CFR) for the period January 22, 2020–July 13, 2020 (“cumulative”), and COVID-19 cases and deaths (per 1,000 people) for June 13 to July 13, 2020 (“recent”).

We retrieved the population for each of the CDs in effect for the 116^th US Congress (January 2019 to January 2021) from the National Historical Geographic Information System, derived from the 2014–2018 American Community Survey (https://www.nhgis.org/, Last accessed July 14, 2020). We retrieved census block and county population data from the US Census Bureau's 2010 decennial census (latest available at this geographic scale; https://www.nhgis.org/, Last accessed July 14, 2020). Census blocks nest inside both CDs and counties. We assigned county COVID-19 cases and deaths per capita to census blocks in an unweighted manner. We then summed the collection of block-level rates in a CD using the census block/CD population weights to obtain CD-level estimates of the different in every CD. Finally, we categorized cumulative CD-level cases and deaths per 1,000 and CFR, and the number of recent cases and deaths per 1000 residents in CD-level cases and deaths per 1,000 in each CD by decile. We provide a detailed table of the estimates for each CD in Supplementary Table 1.

Although state and local governments are at the front line of the COVID-19 crisis, the US Congress plays an important role in designing federal legislation to help the country counter COVID-19. The CDs form a critical link between the local, state, and federal instruments of government, and offer real opportunities for convergence between the administrative and political aspects of policy governance.2

More than 800 legislative bills designed to combat COVID-19 are pending in Congress.6 These bills range from designing the means to improve emergency unemployment relief for governmental entities, establishing loan programs for affected businesses, instituting emergency housing protections and so on. Such federal legislation remains key to the safe reopening of the US economy. Data at the CD-level can equip congressional representatives to further customize legislation and vote for bills most suited for their constituent's needs.

Access to COVID-19 statistics can also help representatives decide which bills to push for, how to effectively communicate with their constituents, make decisions about the required increasing capacity of healthcare centers in their districts and the kind of socially distancing measures the need to enforce. Data is integral to monitoring and evaluating the plans to reopen their respective districts safely. Such data should also facilitate learning from other CDs, regardless of their “success” or “failures”. Equally importantly, access to data at the CD-level can inform and empower the constituents to engage with their representatives and foster accountability and improve governance mechanisms.

The substantial variation in CD-level COVID-19 statistics within states suggests that access to only state-level information is insufficient to capture the nature of the disease burden in a CD.

The COVID-19 data presented here can be layered on with other data related to clinical and non-clinical risk factors of COVID-19, as well as health care capacities or socioeconomic indicators of well-being for better targeting of strategies to contain and mitigate the impact of COVID-19. As COVID-19 outcomes are dynamic and change every day, a dynamic national-level monitoring system needs to be set up to monitor changes in CDs for representatives to better evaluate their plans.

Indeed, in all likelihood, congressional representatives or their constituents are perhaps aware of the burden in a general sense. However, data can sharpen and bring precision to the intensity of the COVID-19 related burden, and in turn, triggering targeted action. For example, Senator Andrew Jones, when presented with CD-level opioid prescription rates in his district said, “I always knew it was a serious problem here. But to learn that we were ground zero for prescribing rates was eye-opening.”7

In conclusion, our aim is to fill an important data void by providing an evidence-base for US congressional representatives and their constituents, which we hope can be leveraged to decide on the containment and mitigation strategies for the COVID-19 infection, as well as to ensure a safe opening up of our economy and society.

Supplementary Table 1

Estimates of the CD-level COVID-19 outcomes: cumulative cases per 1,000 residents as of July 13, cumulative deaths per 1,000 residents as of July 13, CFR as of July 13, 2020, cases per 1000 between June 13 and July 13, cases per 1000 between June 13 and July 13, the ranking of reach CD for each outcome and the decile in which each CD falls in for each outcome