Inequities among vulnerable communities during the COVID-19 vaccine rollout

Importance

Federal and state governments sought to prioritize vulnerable communities in the vaccine rollout through various methods of prioritization, and it is necessary to understand whether inequities exist.

Objective

To assess whether vulnerable counties have achieved similar rates of coverage to non-vulnerable areas, and how vaccine acceptance varies by vulnerability.

Design, Setting, and Participants

We use population-weighted univariate linear regressions to associate the COVID-19 Community Vulnerability Index (CCVI) and its 7 constituent themes with a county-level time series of vaccine coverage and vaccine acceptance. We fit a multilevel model to understand how vulnerability within and across states associates with coverage as of May 8, 2021.

Main Outcome(s) and Measure(s)

The COVID-19 Community Vulnerability Index was used as a metric for county-level vulnerability. County-level daily COVID-19 vaccination data on both first doses administered and people fully vaccinated from April 3, 2021 through May 8, 2021 were extracted from the Covid Act Now API. County-level daily COVID-19 vaccine acceptance survey data from January 6, 2021 through May 4, 2021 were obtained via the Carnegie Mellon University Delphi Group’s COVIDcast API.

Results

Vulnerable counties have consistently lagged less vulnerable counties. As of May 8, the top third of vulnerable counties in the US had fully vaccinated 11.3% fewer people than the bottom third (30.7% vs 34.6% of adult population; linear regression, p= 2.2e-16), and 12.1% fewer initiated vaccinations (40.1% vs 45.6%; linear regression, p= 2.2e-16)). Six out of seven dimensions of vulnerability, including Healthcare System Factors and Socioeconomic Status, predicted lower coverage whereas the Population Density theme associated with higher coverage. Vulnerable counties have also consistently had a slightly lower level of vaccine acceptance, though as of May 4, 2021 this difference was observed to be only 0.7% between low- and high-vulnerability counties (high: 86.1%, low: 85.5%, p=0.027).

Conclusions and Relevance

The vaccination gap between vulnerable and non-vulnerable counties is substantial and not readily explained by a difference in acceptance. Vulnerable populations continue to need additional support, and targeted interventions are necessary to achieve similar coverage in vulnerable counties compared to those less vulnerable to COVID-19.

Key Points