Skip to main content
PMC home page
PLOS Medicine logoLink to PLOS Medicine
. 2023 Jan 31;20(1):e1004167. doi: 10.1371/journal.pmed.1004167

Inequities in COVID-19 vaccine and booster coverage across Massachusetts ZIP codes after the emergence of Omicron: A population-based cross-sectional study

Jacob Bor 1,2,*, Sabrina A Assoumou 3,4, Kevin Lane 5, Yareliz Diaz 6, Bisola O Ojikutu 7, Julia Raifman 6, Jonathan I Levy 5
PMCID: PMC9888673  PMID: 36719864

Abstract

Background

Inequities in Coronavirus Disease 2019 (COVID-19) vaccine and booster coverage may contribute to future disparities in morbidity and mortality within and between Massachusetts (MA) communities.

Methods and findings

We conducted a population-based cross-sectional study of primary series vaccination and booster coverage 18 months into the general population vaccine rollout. We obtained public-use data on residents vaccinated and boosted by ZIP code (and by age group: 5 to 19, 20 to 39, 40 to 64, 65+) from MA Department of Public Health, as of October 10, 2022. We constructed population denominators for postal ZIP codes by aggregating census tract population estimates from the 2015–2019 American Community Survey. We excluded nonresidential ZIP codes and the smallest ZIP codes containing 1% of the state’s population. We mapped variation in ZIP code-level primary series vaccine and booster coverage and used regression models to evaluate the association of these measures with ZIP code-level socioeconomic and demographic characteristics. Because age is strongly associated with COVID-19 severity and vaccine access/uptake, we assessed whether observed socioeconomic and racial/ethnic inequities persisted after adjusting for age composition and plotted age-specific vaccine and booster coverage by deciles of ZIP code characteristics.

We analyzed data on 418 ZIP codes. We observed wide geographic variation in primary series vaccination and booster rates, with marked inequities by ZIP code-level education, median household income, essential worker share, and racial/ethnic composition. In age-stratified analyses, primary series vaccine coverage was very high among the elderly. However, we found large inequities in vaccination rates among younger adults and children, and very large inequities in booster rates for all age groups. In multivariable regression models, each 10 percentage point increase in “percent college educated” was associated with a 5.1 (95% confidence interval (CI) 3.9 to 6.3, p < 0.001) percentage point increase in primary series vaccine coverage and a 5.4 (95% CI 4.5 to 6.4, p < 0.001) percentage point increase in booster coverage. Although ZIP codes with higher “percent Black/Latino/Indigenous” and higher “percent essential workers” had lower vaccine coverage (−0.8, 95% CI −1.3 to −0.3, p < 0.01; −5.5, 95% CI −7.3 to −3.8, p < 0.001), these associations became strongly positive after adjusting for age and education (1.9, 95% CI 1.0 to 2.8, p < 0.001; 4.8, 95% CI 2.6 to 7.1, p < 0.001), consistent with high demand for vaccines among Black/Latino/Indigenous and essential worker populations within age and education groups. Strong positive associations between “median household income” and vaccination were attenuated after adjusting for age. Limitations of the study include imprecision of the estimated population denominators, lack of individual-level sociodemographic data, and potential for residential ZIP code misreporting in vaccination data.

Conclusions

Eighteen months into MA’s general population vaccine rollout, there remained large inequities in COVID-19 primary series vaccine and booster coverage across MA ZIP codes, particularly among younger age groups. Disparities in vaccination coverage by racial/ethnic composition were statistically explained by differences in age and education levels, which may mediate the effects of structural racism on vaccine uptake. Efforts to increase booster coverage are needed to limit future socioeconomic and racial/ethnic disparities in COVID-19 morbidity and mortality.

In a population-based cross-sectional study, Dr. Jacob Bor and colleagues, explore inequities in COVID-19 vaccine and booster coverage across Massachusetts ZIP codes after the emergence of Omicron.

Author summary

Why was this study done?

  • Vaccination, including the booster shot, is a critical line of defense against severe disease due to Coronavirus Disease 2019 (COVID-19) infection.

  • We sought to understand the geography of vaccine and booster coverage across Massachusetts (MA) ZIP codes and to assess coverage inequities 18 months into the MA’s general population vaccine rollout.

What did the researchers do and find?

  • Data on numbers vaccinated and boosted from MA Department of Public Health were combined with ZIP code denominators constructed de novo from US Census data.

  • As of October 2022, there were large differences in primary series vaccine and booster coverage across MA ZIP codes. The share of children vaccinated ranged from under 40% to over 90% across ZIP codes. The share of elderly adults boosted ranged from under 60% to 100%.

  • Primary series vaccine and booster coverage increased with ZIP code-level income and education and fell with percent Black/Latino/Indigenous and percent essential workers. Education levels were the strongest predictor of vaccine and booster uptake.

  • After adjusting for age and education levels, vaccine and booster uptake was higher in ZIP codes with many Black/Latino/Indigenous residents or essential workers. Access, not “hesitancy,” may drive persistent vaccination gaps in these communities.

What do these findings mean?

  • Inequities in vaccine and booster coverage may lead to future inequities in morbidity, mortality, and economic losses due to COVID-19.

Introduction

Vaccination against Coronavirus Disease 2019 (COVID-19)—including recommended booster shots—is a critical line of defense against severe illness, hospitalization, and death. In Fall 2022, the United States had ended most public health measures amidst a continued high number of daily deaths and the potential for future surges due to holiday travel, waning immunity, or emergence of new variants. Communities with low vaccination rates are particularly vulnerable to future COVID-19 infection [1,2]. Vaccine uptake in Massachusetts (MA) is high relative to the national average. While the Winter 2021–2022 wave resulted in few excess deaths in MA, there was a rise in hospitalizations [3]. Still, as of October 2022, there were an appreciable number of eligible individuals who remained unvaccinated and a greater number who had not received boosters.

Nationally, vaccination rates are lower for less educated and lower-income people, and vaccine uptake has been slower in Black and Latino populations [4,5]. These populations, as well as the overlapping population of essential workers, have been at elevated risk for COVID exposure, infection, morbidity, and mortality throughout the pandemic [6–12] Inequities in vaccine coverage could further exacerbate disparities in health outcomes.

“Vaccine hesitancy” has dominated media coverage on racial disparities, often without mention of the historical roots of medical mistrust [13]. However, other factors associated with structural racism and classism have also undermined vaccine uptake in people of color and lower-income people. When the vaccine was first rolled out, eligibility for the vaccine was limited to healthcare workers and persons ages 65 and older, excluding many nonelderly, low-income, essential workers, and people of color who were at high risk [14]. Lack of access to a regular health provider and accurate health information contributed to low vaccine uptake among the uninsured, even though the vaccine was free for all who could access it. Additionally, whereas initial vaccine rollout focused on centralized mass distributions sites, community-based delivery has since been shown to be more effective in reaching lower-income people and people of color [15,16]. Lack of paid time off to get vaccinated and recover, limited outreach to non-English speakers, and low trust in the health system have also been barriers [17,18]. Structural racism also drives socioeconomic differences by race/ethnicity that affect access to transportation and time to get vaccinated. Federal and state leaders began to message that the pandemic was ending in May 2021, removing masking requirements when less than 40% of the US population was vaccinated and did not reinvigorate vaccine communications and delivery when the Delta and Omicron variants began to spread.

MA allocated funds to improve equity in COVID-19 vaccine coverage. In February 2021, the MA Vaccine Equity Initiative (VEI) aimed to improve vaccine administration rates in the 20 most disproportionately impacted communities [19]. These communities were identified based on COVID-19 case rates, CDC-defined Social Vulnerability Index, and the percentage of people of color [20]. The VEI ensured a vaccine pipeline to these communities during the initial rollout and has awarded $58 million to 198 community organizations to support vaccine education and outreach [21]. In March 2022, the state legislated a “COVID-19 vaccination equity plan… to eliminate disparities in the rates of vaccination” including boosters [21].

In this study, we assess variation in primary vaccine series and booster coverage across MA ZIP codes, analyzing data 18 months into MA’s general population vaccine rollout. Boosters are a critical line of defense in mitigating the impacts of future waves [22,23]. Although the first year of the vaccine rollout involved extensive community outreach [16], the booster campaign (including the bivalent booster that became available in September 2022) has received less public attention despite its importance for maintaining immunity and protecting against Omicron variants. Closing vaccination gaps requires an understanding of barriers to uptake and identifying geographic areas with low coverage. ZIP codes are small enough to capture sociodemographic heterogeneity that is obscured in city, town, or county-level estimates. We additionally stratified our analysis by age, the leading risk factor for severe illness due to COVID-19 [24] and a strong correlate of vaccination rates. While racial disparities have often been attributed to vaccine hesitancy or medical mistrust, structural racism also drives large racial differences in income, education, and occupation that shape access to information, health services, and time off from work and could affect vaccination rates. We therefore assessed the extent to which socioeconomic factors mediated disparities in vaccination by ZIP code racial/ethnic composition. Our analysis complements aggregate reporting by the MA Department of Public Health (DPH) on the 20 cities and towns that have been the focus of the VEI, as well as prior reports of vaccination patterns at the city/town level and across Boston ZIP codes [25].

Methods

Data sources

Vaccines and booster shots delivered

We extracted data on counts of vaccinations by ZIP code and sex and by ZIP code and age group, published by MA DPH as “Weekly COVID-19 Municipality Vaccination Data” [26] Data are reported to the state by health facilities and vaccination sites. Residential postal ZIP codes and patient age are extracted from paperwork filed at the vaccination site. We analyzed data on vaccines administered in MA from the start of the vaccine rollout through October 10, 2022. MA rolled out the vaccine in phases, first to clinicians and nursing homes in December 2020, then to members of the general population in February 2021, starting with adults over 75 years. This study is reported as per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (S1 STROBE Checklist). The study did not have a prespecified analysis plan.

We extracted data on two key constructs reported in the MA DPH vaccine database:

  • Primary vaccine series: one shot if Ad26.COV2.S (e.g., Janssen/Johnson and Johnson), two shots if mRNA-1273 (e.g., Moderna) or BNT162b2 (e.g., Pfizer-BioNTech) vaccine; this definition corresponds with the “fully vaccinated” data reported by MA DPH.

  • Booster shot (any booster shot after completing initial vaccine schedule).

As of October 2022, all MA residents ages 5 years and over were eligible to be vaccinated and to receive the booster shot if their primary series vaccine was completed at least 2 months prior.

MA DPH reports data on vaccines administered by ZIP code, stratified by age, sex, and race/ethnicity. ZIP code-level totals, which are not reported, can be constructed through aggregation. To protect confidentiality, MA DPH suppresses exact numbers in cells with fewer than 30 people vaccinated/boosted. To minimize the influence of missing data in the ZIP code-by-age data, we instead used ZIP code-by-sex data to estimate aggregate ZIP code totals for people ages 5 and older. The ZIP code-by-sex data were nearly complete for self-identified “males” and “females.” Data for the category “neither male nor female” were frequently suppressed due to small numbers (75% of ZIP codes were missing data on this group for vaccines, and 96% were missing data on boosters). We imputed data for “neither male nor female” as follows. We computed the ratios of people vaccinated/boosted among persons “neither male nor female” relative to the numbers of people vaccinated/boosted among persons either “male” or “female” for those ZIP codes reporting all three categories. These ratios were 0.4% for vaccinated and 0.3% for boosted. We then used these ratios to estimate the number of people “neither male nor female” vaccinated/boosted based on the number either “male” and “female” for those ZIP codes that did not report numbers for “neither male nor female.” Where the estimated number “neither male nor female” exceeded the data suppression threshold of 30, we imputed a value of 30. We then aggregated across all sexes to obtain counts of ZIP code residents who had received the primary vaccine series and/or booster shot.

In addition to ZIP code totals, we assessed ZIP code-by-age primary series vaccine and booster coverage. MA DPH reports the following age categories: 5 to 11, 12 to 15, 16 to 19, 20 to 29, 30 to 39, 40 to 49, 50 to 59, 60 to 64, 65 to 69, 70 to 74, 75+ years. Some ZIP-by-age cells were missing data due to suppression rules (0.04% of cells for vaccinated, 3.7% of cells for boosted). To impute values, we estimated a mixed effects Poisson model with fixed effects for each age group and interactions of each age group with median income, ZIP code random effects, and ACS population denominators (described below) as the offset. The model predictions were strongly correlated (0.98) with the observed values. We obtained predicted counts for the missing cells and imputed either that value or 30, whichever was smaller. We then aggregated to broader age groups for the analysis: 5 to 19, 20 to 39, 40 to 64, and 65+ years.

Approximately 3% of MA residents receiving primary series vaccines and 2% of MA residents receiving boosters did not report their ZIP code and were excluded from the analysis.

Population denominators

We constructed ZIP code-by-age population counts using publicly available data from the American Community Survey (ACS) 5-year combined estimates (2015 to 2019) [27]. The ACS is a random sample survey of approximately 3 million people in the US population each year conducted by the US Census Bureau. Because postal ZIP codes (reported on vaccine forms) differ from Census-defined ZIP code tabulation areas (ZCTAs), we constructed population denominators de novo, aggregating up from census tracts. Age-specific population data were extracted from ACS at the census tract level. We used the Housing and Urban Development tract to ZIP code 2019 fourth quarter crosswalk to assign these census tract populations to postal ZIP codes [28]. Most census tracts are fully contained within single ZIP codes; however, some are split across ZIP codes. For these, we allocated age-specific population counts to ZIP codes proportionately based on the populations of the underlying census blocks and their ZIP code membership. We note that sampling variability in the population denominators implies that the estimated population size may sometimes be smaller than the number of residents vaccinated.

ZIP code characteristics

Census tract-level ACS 5-year estimates for 2019 were obtained for race/ethnicity (B03002), median household income (B19001), educational attainment (B15003), and occupation (S2404). Essential workers were defined based on definitions developed by the American Civil Liberties Union, which identified those job types considered “essential” during the pandemic such as work in healthcare, transportation, and food preparation [29] and has been used to assess COVID-19 inequities in MA [10–12]. We aggregated to postal ZIP codes using the tract-to-ZIP crosswalk, weighting by tract population. Using the Census ACS estimates, we derived ZIP code-level “percent college graduates over the age of 25 years,” “percent Black, Latino, or Indigenous,” “percent essential workers,” and “median household income.” Race/ethnicity in the ACS was determined by self-report. People were identified as Black, Latino, or Indigenous if they reported Black or American Indian/Alaska Native among their “races” or if they reported Hispanic ethnicity. We aggregated Black, Latino, and Indigenous MA residents into a single measure due to the high concentration of each of these groups in a relatively small number of ZIP codes. Aggregation of income data yielded population-weighted averages of median household income across census tracts within each ZIP code. For convenience, we refer to this ZIP code-level measure as “median household income.” We additionally constructed a series of variables to capture differences in age composition across ZIP codes: percentage of ZIP code residents aged 0 to 4 years, 5 to 19 years, 20 to 39 years, 40 to 64 years, and 65+ years. Finally, we created an indicator for whether the ZIP code was within one of the state’s 20 VEI communities: Boston, Brockton, Chelsea, Everett, Fall River, Fitchburg, Framingham, Haverhill, Holyoke, Lawrence, Leominster, Lowell, Lynn, Malden, Methuen, New Bedford, Randolph, Revere, Springfield, and Worcester.

Exclusions

We excluded ZIP codes that did not correspond to residential areas, ZIP codes that corresponded to specific universities or businesses, and ZIP codes assigned to post office boxes and not to residential addresses. This reduced the total number of ZIP codes from 648 to 481. To avoid unstable estimates due to very small denominators, we excluded the smallest ZIP codes containing 1% of the total population (n = 63). After these exclusions, we had estimates of vaccination and booster coverage for all remaining ZIP codes (n = 418).

Analysis

We constructed estimates of “percent vaccinated” (primary series) and “percent boosted” for MA ZIP codes and for ZIP code-by-age-group cells. Different age categories were reported in the state vaccination data and the ACS. To harmonize age groups and reduce the number of small cells, we collapsed age to the following categories: 5 to 19 years, 20 to 39 years, 40 to 64 years, and 65+ years.

Uncertainty exists when estimating vaccine coverage for specific ZIP codes due to sampling error in the population denominators. The ACS population estimates are published with standard errors; however, these are reported at the census tract (not ZIP code) level. We captured this uncertainty when reporting estimates for specific ZIP codes (and ZIP-by-age cells) by constructing 90% confidence intervals (CIs) using a resampling approach. For each census tract by age group cell, we simulated 101 population estimates from a normal distribution defined by the ACS point estimates and standard error for that observation. We then aggregated to ZIP codes and computed the constructs of interest holding the numerator constant. To capture uncertainty in the estimates, we then ranked these simulated estimates and used the fifth and 95th ranked estimate as the lower- and upper-bound for our 90% CI.

ZIP code-level analysis

We assessed spatial heterogeneity in vaccine coverage by mapping the percent of residents ages 5 years and older who have received the primary series vaccine or booster shot. We created higher-resolution maps for densely populated areas. We then investigated the association of ZIP code vaccination and booster coverage with ZIP code characteristics: median household income, percent college graduates, percent Black/Latino/Indigenous, percent essential worker, and percent in each age group: 5 to 19, 20 to 39, 40 to 64, and 65+ years. We present scatter plots and estimated bivariate linear regression models with robust standard errors.

We also estimated multivariable regression models to determine whether the bivariate associations are statistically explained by other factors. For example, higher-income MA residents are on average older, which could lead to a positive association between ZIP code income and vaccination rates. Of particular interest, we assessed whether percent Black, Latino, and Indigenous was independently associated with vaccine coverage, after adjusting for age shares and after adjusting for income, education, and essential worker shares—factors that may influence access to vaccination. If vaccine hesitancy related to racial/ethnic identity constrained uptake in communities of color, then we would expect to see a persistent negative association between percent Black, Latino, and Indigenous and vaccine uptake in these adjusted models.

ZIP code-by-age analysis

To illustrate heterogeneity in age-specific primary series vaccination and booster rates, we plotted ranked coverage rates, ordered by ZIP code. In these distribution plots, we censored the top and bottom 5% of observations to facilitate visualization of the rest of the distribution. To assess the association between vaccination rates and ZIP code population characteristics, we constructed population-weighted deciles for each characteristic. We then computed binned averages and 95% CIs within each decile to estimate the percent vaccinated and percent boosted. We also assessed the association of these ZIP code characteristics with percent vaccinated and boosted in age-stratified multivariable regression models, allowing for different relationships within each age group. To test for effect heterogeneity, we fit regression models on all age groups, interacted the covariate of interest with age, and tested the null hypothesis that the coefficients on the interaction terms were jointly equal to zero.

Analysis plan

As the study did not have a prespecified analysis plan, we describe the timing of our analytic choices here. The socioeconomic and demographic variables used in the assessment of vaccine uptake disparities were selected prior to the analysis and were chosen based on the prior literature on disparities in COVID-19 impact and early vaccine uptake. The approach to constructing population denominators followed prior work by the research team. The age ranges used in the analysis were determined a priori based on harmonization of ages reported in the state vaccination data and ACS denominator data. The decision to exclude ZIP codes with small populations was made after observing the high rates of data suppression and instability (wide CIs) of estimates in those ZIP codes. The methods for the decile-based descriptive analysis and multiply-adjusted regression analyses were planned a priori. At the suggestion of a reviewer, we tested the null hypothesis that the VEI had the same association with vaccine coverage as with booster coverage, given that community outreach was strongest during the initial vaccine rollout. We also assessed associations stratified by male/female sex.

Ethics

The Boston University Medical Campus IRB does not require ethical review for secondary analyses of publicly available, deidentified data.

Results

We analyzed data on 418 ZIP codes containing 97% of the MA population (Table 1). Of these ZIP codes, 184 (44%) were in cities or towns containing multiple ZIP codes. The mean population of the included ZIP codes was 15,967 (range 2,014 to 61,099) individuals. On average, the study population resided in ZIP codes where 41% of residents were college educated, 19% were Black, Latino, or Indigenous, 32% were essential workers, and where median household income was $41,100.

Table 1. Characteristics of the study population.

ZIP codes (N) 418
Total population (N) 6,674,243
ZIP code population (mean) 15,967
ZIP code population (range) 2,014 to 61,099
Percentage with a primary series vaccine Mean (SD)
 Aged 5+ years 83% (10%)
 5–19 years 65% (17%)
 20–39 years 79% (14%)
 40–64 years 85% (10%)
 65+ years 100% (10%)
Percentage with a booster shot Mean (SD)
 Aged 5+ years 50% (11%)
 5–19 years 24% (12%)
 20–39 years 39% (12%)
 40–64 years 55% (11%)
 65+ years 81% (10%)
ZIP code characteristics Mean (SD)
 Median household income $41,100 ($12,500)
 % college graduates 41% (17%)
 % Black, Latino, or Indigenous 19% (20%)
 % essential workers 32% (6%)
Open in a new tab

Note: Mean and standard deviation (SD) of vaccination coverage and ZIP code characteristics are weighted by ZIP code population ages 5 years and older. Percent with a primary series vaccine and booster shot are as of October 10, 2022.

On average, an estimated 83% (5,239,410/6,321,016) of residents ages 5 years and older had received the primary series vaccine, and 50% (3,163,944/6,321,016) had received a booster shot. We estimated that 100% (1,079,095/1,074,418) of residents ages 65 years and older had received the primary series vaccine. (We have not constrained the numerator to be less than the denominator in these estimates.) Coverage for the primary series vaccine was lower in younger age groups, with 85% (1,888,624/2,223,264) of persons 40 to 64 years, 79% (1,444,242/1,837,095) of persons 20 to 39 years, and 65% (765,259/1,186,227) of persons 5 to 19 years vaccinated. Booster coverage was 81% (871,749/1,074,418) among residents 65 years and older, 55% (1,230,146/2,223,264) among persons 40 to 64 years, 39% (725,306/1,837,095) among persons 20 to 39 years, and 24% (286,712/1,186,227) among persons 5 to 19 years.

The maps in Fig 1 show ZIP code-level variation in vaccination (top) and booster (bottom) coverage for residents ages 5 years and older. The panels show similar geographic patterns of vaccine and booster uptake, with the highest coverage among ZIP codes in Boston’s Western suburbs, the North and South Shore, Cape Cod and the Islands, and in the college towns of the Pioneer Valley. Lower vaccine and booster coverage was observed in lower-income ZIP codes of the urban centers—greater Boston, Worcester, Springfield, New Bedford, Fall River, Lawrence, and Lowell (S1 Fig shows higher-resolution maps of these areas). Additionally, lower coverage was observed in nonurban ZIP codes in Central, Western, and Southeast MA, and in MA ZIP codes along the borders with New Hampshire, Rhode Island, and Connecticut.

Fig 1. Percentage of residents with a COVID-19 (a) vaccine or (b) booster by ZIP code as of October 10, 2022.

Fig 1

Open in a new tab

Note: We analyzed data on 418 ZIP codes containing 97% of the MA population. ZIP codes labeled “missing/suppressed” were excluded because they corresponded to specific businesses or universities, to post office boxes rather than residential addresses, or because were the smallest 1% of ZIP codes, which we excluded due to instability of estimates. To facilitate comparisons across ZIP codes within each panel (a) and (b), the scales in the panels differ. Shape files were obtained from the US Census Bureau, accessed October 20, 2022: https://www.census.gov/geographies/mapping-files/time-series/geo/tiger-geodatabase-file.html. Copyright protection is not available for any work of the United States Government (Title 17 U.S.C., Section 105).

Fig 2 shows scatter plots of vaccine and booster coverage against ZIP code characteristics. We observed strong correlations between primary series vaccine and booster coverage and the four sociodemographic indicators. A $10,000 increase in a ZIP code’s median household income was associated with a 4.5 percentage point (95% CI 3.8 to 5.2, p < 0.001) increase in primary series vaccine coverage and a 6.8 percentage point (95% CI 6.2 to 7.4, p < 0.001) increase in booster coverage. From the lowest to the highest income levels, booster coverage increased from about 30% to over 70%. A 10 percentage point increase in percent college graduates was associated with a 2.9 percentage point (95% CI 2.2 to 3.6, p < 0.001) increase in vaccinations and a 5.1 percentage point (95% CI 4.5 to 5.6, p < 0.001) increase in boosters. A 10 percentage point increase in percent essential workers was associated with a −5.5 percentage point (95% CI −7.3 to −3.8, p < 0.001) reduction in primary series vaccination coverage and a −11.0 percentage point (95% CI −12.4 to −9.6, p < 0.001) reduction in booster coverage. We observed a weaker relationship between percent Black, Latino, and Indigenous and vaccination coverage, although a 10 percentage point increase in this population share was associated with a −2.6 percentage point (95% CI −3.0 to −2.2, p < 0.001) decline in booster coverage.

Fig 2. Percent vaccinated and boosted according to individual ZIP code characteristics.

Fig 2

Open in a new tab

Note: Data on vaccine and booster coverage are as of October 10, 2022. Fitted lines are from bivariate regressions. Coefficients and confidence intervals for bivariate and multivariable models are presented in Table 2. To preserve the scale across the plots, three outliers with estimated vaccine coverage over 115% were suppressed in the scatter plots but contribute to the fitted lines.

Table 2 shows results from both bivariate and multivariable models regressing percent vaccinated/boosted on variables describing the age composition of the ZIP code (defined above) as well as the four sociodemographic characteristics. In these multivariable regression models, percent college graduates emerged as the strongest predictor of vaccine and booster coverage. Each 10 percentage point increase in percent college graduates was associated with a 5.1 percentage point (95% CI 3.9 to 6.3, p < 0.001) increase in vaccine coverage and a 5.4 percentage point (95% CI 4.5 to 6.4, p < 0.001) increase in booster coverage. Although median household income was strongly positively associated with vaccination and booster rates in the multivariable analyses, these associations were substantially attenuated and lost statistical significance after adjusting for age composition, suggesting that age distribution confounds the relationship between ZIP code median income and vaccination and booster rates when controlling for other factors (S1 Table).

Table 2. Association of ZIP code characteristics with the percent of residents aged 5 years and older that are vaccinated for COVID-19 and/or have had a COVID-19 booster shot.

a) Percent vaccinated
Bivariate Multivariable
OLS regression models (1) (2) (3) (4)
beta (95% CI) p-value beta (95% CI) p-value beta (95% CI) p-value beta (95% CI) p-value
Median household income ($10k increase) 4.5*** (3.8,5.2) <0.001 4.8*** (3.3,6.3) <0.001 1.5 (−0.1,3.1) 0.060 1.5 (−0.1,3.1) 0.059
Percent college graduates (10% point increase) 2.9*** (2.2,3.6) <0.001 2.3*** (1.1,3.6) <0.001 5.1*** (3.9,6.3) <0.001 5.1*** (3.9,6.3) <0.001
Percent Black, Latino, Indigenous (10% point increase) −0.8** (−1.3,−0.3) 0.002 1.4*** (0.8,2.0) <0.001 1.8*** (1.1,2.6) <0.001 1.9*** (1.0,2.8) <0.001
Percent essential worker (10% point increase) −5.5*** (−7.3,−3.8) <0.001 4.5*** (2.1,6.8) <0.001 4.8*** (2.6,7.1) <0.001 4.8*** (2.6,7.1) <0.001
VEI community (0,1) −4.0** (−6.4,−1.6) 0.001 −0.3 (−3.6,3.0) 0.850
Adjusted for age distribution No No Yes Yes
R2 NA 0.38 0.47 0.47
b) Percent with a booster shot (dependent variable)
Bivariate Multivariable
OLS regression models (1) (2) (3) (4)
beta (95% CI) p-value beta (95% CI) p-value beta (95% CI) p-value beta (95% CI) p-value
Median household income ($10k increase) 6.8*** (6.2,7.4) <0.001 3.7*** (2.2,5.2) <0.001 0.6 (−0.8,2.0) 0.392 0.6 (−0.7,1.9) 0.351
Percent college graduates (10% point increase) 5.1*** (4.5,5.6) <0.001 2.8*** (1.6,4.1) <0.001 5.5*** (4.5,6.4) <0.001 5.4*** (4.5,6.4) <0.001
Percent Black, Latino, Indigenous (10% point increase) −2.6*** (−3.0,−2.2) <0.001 −0.1 (−0.4,0.3) 0.687 0.8*** (0.4,1.2) <0.001 1.0*** (0.5,1.5) <0.001
Percent essential worker (10% point increase) −11.0*** (−12.4,−9.6) <0.001 0.2 (−1.7,2.2) 0.801 0.3 (−1.3,1.9) 0.714 0.4 (−1.2,1.9) 0.630
VEI community (0,1) −11.1*** (−13.4,−8.8) <0.001 −2.1 (−4.4,0.2) 0.079
Adjusted for age distribution No No Yes Yes
R2 NA 0.67 0.76 0.76
Open in a new tab

Note: N = 418 ZIP codes. In column (1), each cell shows the result of a separate bivariate OLS regression model. In columns (2)–(4), each column shows the results of multivariable (adjusted) OLS regression models. Coefficients are scaled so that they can be interpreted as the percentage point change in primary series (or booster) coverage associated with a 10% (or $10K) increase in the continuous covariates. 95% CIs based on heteroskedasticity-robust standard errors are shown in parentheses. We adjusted for age by including variables denoting the percentage of ZIP code residents in the following age groups: 5–19, 20–39, 40–64, and 65+ years.

*p < 0.05,

**p < 0.01,

***p < 0.001.

CI, confidence interval; OLS, ordinary least squares; VEI, Vaccine Equity Initiative.

We observed notable changes in the coefficients for “percent Black, Latino, or Indigenous” and “percent essential workers” after adjusting for age composition and education. In bivariate models, these ZIP code characteristics were negatively associated with percent vaccinated; however, in multivariable models, the associations turned positive. ZIP codes with a larger share of Black, Latino, and Indigenous residents and with more essential workers had higher vaccination rates than would be otherwise expected based on the education and age composition in those ZIP codes. Adjusting for age and education, each 10 percentage point increase in “percent Black, Latino, or Indigenous” was associated with a 1.9 percentage point (95% CI 1.0 to 2.8, p < 0.001) increase in vaccine coverage and each 10 percentage point increase in “percent essential workers” was associated with a 4.8 percentage point (95% CI 2.6 to 7.1, p < 0.001) increase in vaccine coverage. For boosters, while the associations from the bivariate models were similarly attenuated after adjusting for age and education, there was a less pronounced association in the multivariable models with percent essential workers (null) and percent Black, Latino, or Indigenous (modestly positive).

Table 2 includes an indicator for whether the ZIP code was a part of one of the 20 VEI communities targeted by the state for enhanced vaccine outreach. VEI communities had lower vaccine coverage (−4.0 percentage points, 95% CI −6.4 to −1.6, p < 0.01) and booster coverage (−11.1 percentage points, 95% CI −13.4 to −8.8, p < 0.001) than non-VEI communities. After adjusting for sociodemographic characteristics and age composition, the coefficient on VEI status was close to zero: −0.3 percentage points (95% CI −3.6 to 3.0, p = 0.85) for vaccine coverage and −2.1 percentage points (95% CI −4.4 to 0.2, p = 0.079) for booster coverage. The VEI coefficients in the adjusted vaccine and booster models were significantly different (p = 0.049). (We note that other factors not included in our model may explain the lower vaccination or booster rates in VEI communities and that vaccination and booster rates might have been even lower in the absence of the VEI.) In sex-stratified models, results were similar for men and women (S2 Table).

We also evaluated whether ZIP code inequities in vaccination coverage varied by age group. Fig 3 (left-sided graphs) shows the distribution of “percent vaccinated” in each ZIP code, stratified by age group. The point estimates contain estimation uncertainty in the denominator, which explains why some percentages exceed 100%. The figure shows very high vaccination rates among the elderly (65+ years), with relatively little variation across ZIP codes. Variation in primary series vaccine coverage was more pronounced for younger age groups, in particular, for children (5 to 19 years), where vaccination rates vary from as low as 40% to as high as 90%. Fig 3 (right-sided graphs) shows “percent boosted” in each ZIP code, stratified by age group. The share boosted varied widely in all age groups, even for the elderly. For people over 65 years, the share boosted varied from approximately 60% in the lowest coverage ZIP codes to nearly 100% in the highest-coverage ZIP codes.

Fig 3. Percentage of residents who have received a COVID-19 vaccine by ZIP code and age.

Fig 3

Open in a new tab

Note: Data are ranked by percent vaccinated with the primary series or booster shot. Vertical lines are 90% confidence intervals, which capture uncertainty due to sampling error in the population denominators. Due to the presence of sampling error in the denominators, neither the point estimates nor confidence bounds are constrained to be within the [0%,100%] interval.

Fig 4 shows the share of MA adults by age group who are vaccinated and boosted by median household income and by the percent of ZIP code residents that are college graduates, Black, Latino, or Indigenous, and essential workers. MA has achieved very high vaccination coverage among 65+-year-olds across ZIP codes with different population characteristics (bottom graphs of the figure). Vaccination coverage among MA residents 65+ years was above 95% in all deciles of all four ZIP code characteristics, and close to 100% for most. Vaccination coverage among adults ages 40 to 64 years was also relatively high and equitably distributed (third set of graphs of the figure). At least 75% of residents were vaccinated in all ZIP code covariate deciles, and vaccination rates were highest in ZIP codes with a greater share of Black, Latino, or Indigenous residents. Larger gaps (and inequities) in vaccine coverage were apparent for MA adults ages 20 to 39 (second set of graphs) and children ages 5 to 19 (top graphs of the figure).

Fig 4. Percent vaccinated or boosted in each age group, by decile of ZIP-code characteristics.

Fig 4

Open in a new tab

Note: Plot displays % vaccinated (blue squares) and % boosted (yellow circles) for MA residents ages 5–19, 20–39, 40–64, and 65+ years, displayed by decile of four ZIP code characteristics. The first decile is the bottom 10% of ZIP codes, ranked by each covariate; the 10th decile is the top 10% of ZIP codes. Vertical spikes are 95% confidence intervals. Due to sampling error in the population denominators, neither point estimates nor CIs are constrained to be below 100%.

By comparison, there was wide variation in coverage of the booster shot across all age groups. The share of MA residents over 65 years who had received a booster was under 80% in ZIP codes with the lowest income, lowest share of college graduates, and highest shares of Black, Latino, or Indigenous residents and essential workers. Patterns were similar for middle-aged adults (40 to 64 years), with pronounced disparities and rates under 50% for ZIP codes in the lowest deciles of education and income and the highest deciles of Black, Latino, or Indigenous and essential worker shares. Booster coverage was nearly 70% in higher-income ZIP codes. Large disparities in booster coverage were also apparent for younger adults and children, with gaps in coverage of over 30 percentage points between the top and bottom deciles of ZIP code income and gaps over 20 percentage points by deciles of education and essential worker share.

Table 3 presents age-stratified bivariate and multivariable regression models, replicating Table 2 for each age group—5 to 19, 20 to 39, 40 to 64, and 65+ years. Inequities in vaccine and booster uptake were most pronounced in younger age groups. In the bivariate models (panels (a) and (c)), income and education had positive associations with vaccine and booster coverage across all age groups, although the gradient was steeper at younger ages. Percent Black, Latino, or Indigenous was negatively associated with vaccine coverage among children and younger adults and with booster coverage at all ages, but positively associated with vaccine coverage among adults age 40 to 64. Percent essential workers was negatively associated with vaccine and booster coverage for all age groups. For all ZIP code characteristics, we rejected the null hypothesis of effect homogeneity across age groups (F-test, p < 0.0001).

Table 3. Association of ZIP code characteristics with percent vaccinated and boosted, by age.

a) Percent vaccinated: Age-stratified bivariate OLS regression models
Age group 5–19 years 20–39 years 40–64 years 65+ years
Parameters beta (95% CI) p-value beta (95% CI) p-value beta (95% CI) p-value beta (95% CI) p-value
Median household income ($10k increase) 9.2*** (8.1,10.2) <0.001 5.3*** (4.1,6.5) <0.001 0.8* (0.0,1.6) 0.040 1.7*** (0.9,2.6) <0.001
% college graduates (10% point increase) 5.8*** (4.7,6.9) <0.001 2.9*** (1.9,3.9) <0.001 1.1** (0.5,1.8) 0.001 1.3*** (0.7,1.9) <0.001
% Black, Latino, Indigenous (10% point increase) −2.1*** (−2.9,−1.3) <0.001 −1.0** (−1.7,−0.3) 0.006 1.4*** (0.8,1.9) <0.001 −0.8* (−1.4,−0.2) 0.010
% essential worker (10% point increase) −12.6*** (−15.5,−9.7) <0.001 −4.6** (−7.6,−1.6) 0.003 −2.8*** (−4.3,−1.2) 0.001 −1.5* (−2.9,−0.1) 0.042
VEI community (0,1) −11.5*** (−15.5,−7.5) <0.001 −3.0 (−6.3,0.3) 0.077 3.9*** (1.6,6.3) 0.001 −3.4** (−5.9,−0.9) 0.008
b) Percent vaccinated: Age-stratified multivariable OLS regression models
Age group 5–19 years 20–39 years 40–64 years 65+ years
Parameters beta (95% CI) p-value beta (95% CI) p-value beta (95% CI) p-value beta (95% CI) p-value
Median household income ($10k increase) 10.1*** (7.9,12.3) <0.001 6.1*** (3.9,8.3) <0.001 −0.2 (−1.7,1.3) 0.759 1.2 (−0.7,3.0) 0.221
% college graduates (10% point increase) 2.5** (0.7,4.3) 0.008 2.7*** (1.2,4.2) <0.001 3.6*** (2.2,5.1) <0.001 2.1* (0.4,3.8) 0.015
% Black, Latino, Indigenous (10% point increase) 2.8*** (1.8,3.8) <0.001 1.5** (0.5,2.5) 0.004 2.8*** (1.8,3.8) <0.001 0.4 (−0.7,1.4) 0.481
% essential worker (10% point increase) 4.1** (1.0,7.3) 0.010 7.1*** (3.4,10.8) <0.001 2.0 (−0.8,4.7) 0.162 4.7** (1.5,7.9) 0.004
VEI community (0,1) −2.9 (−7.7,1.9) 0.232 0.1 (−4.7,5.0) 0.962 −0.7 (−4.4,3.0) 0.707 −1.9 (−6.0,2.2) 0.353
R2 0.57 0.27 0.25 0.07
c) Percent boosted: Age-stratified bivariate OLS regression models
Age group 5–19 years 20–39 years 40–64 years 65+ years
Parameters beta (95% CI) p-value beta (95% CI) p-value beta (95% CI) p-value beta (95% CI) p-value
Median household income ($10k increase) 7.6*** (7.0,8.2) <0.001 7.7*** (6.8,8.7) <0.001 5.3*** (4.6,5.9) <0.001 3.8*** (2.9,4.6) <0.001
% college graduates (10% point increase) 5.5*** (4.9,6.1) <0.001 5.5*** (4.8,6.2) <0.001 4.8*** (4.3,5.2) <0.001 3.1*** (2.6,3.6) <0.001
% Black, Latino, Indigenous (10% point increase) −2.4*** (−2.9,−2.0) <0.001 −2.6*** (−3.1,−2.1) <0.001 −1.0*** (−1.4,−0.7) <0.001 −1.8*** (−2.3,−1.3) <0.001
% essential worker (10% point increase) −13.3*** (−15.0,−11.6) <0.001 −11.9*** (−14.4,−9.5) <0.001 −11.1*** (−12.3,−9.9) <0.001 −5.5*** (−6.7,−4.2) <0.001
VEI community (0,1) −11.9*** (−14.3,−9.6) <0.001 −9.2*** (−12.2,−6.2) <0.001 −5.4*** (−7.7,−3.1) <0.001 −7.7*** (−10.1,−5.4) <0.001
d) Percent boosted: Age-stratified multivariable OLS regression models
Age group 5–19 years 20–39 years 40–64 years 65+ years
Parameters beta (95% CI) p-value beta (95% CI) p-value beta (95% CI) p-value beta (95% CI) p-value
Median household income ($10k increase) 5.5*** (4.2,6.7) <0.001 4.4*** (3.1,5.8) <0.001 −0.2 (−1.5,1.1) 0.783 0.7 (−1.1,2.5) 0.433
% college graduates (10% point increase) 1.5** (0.5,2.5) 0.005 3.9*** (3.1,4.7) <0.001 5.7*** (4.6,6.7) <0.001 3.8*** (2.2,5.4) <0.001
% Black, Latino, Indigenous (10% point increase) 1.1*** (0.6,1.5) <0.001 0.8* (0.2,1.4) 0.011 1.9*** (1.3,2.5) <0.001 0.3 (−0.4,1.1) 0.407
% essential worker (10% point increase) −3.1*** (−4.9,−1.3) <0.001 1.6 (−0.8,4.1) 0.194 −0.6 (−2.4,1.2) 0.519 4.1** (1.4,6.9) 0.004
VEI community (0,1) −2.3 (−4.9,0.2) 0.074 −2.1 (−5.5,1.2) 0.213 −2.3 (−5.1,0.6) 0.119 −3.7* (−7.3,−0.1) 0.044
R2 0.76 0.67 0.62 0.27
Open in a new tab

Note: N = 418. In panels (a) and (c), each cell shows the coefficient estimate from a separate bivariate OLS regression model. In panels (b) and (d), each column shows adjusted coefficients from multivariable regression models. Heteroskedasticity-robust standard errors are shown in parentheses.

*p < 0.05,

**p < 0.01,

***p < 0.001.

OLS, ordinary least squares; VEI, Vaccine Equity Initiative.

In multivariable models (panels (b) and (d)), ZIP code income was strongly associated with vaccine and booster coverage for children (5 to 19 years) and young adults (20 to 39 years), but not for older ages. Education levels maintained strong positive associations across all age groups. After adjusting for income and education, ZIP codes with larger Black, Latino, or Indigenous populations had similar or higher vaccination coverage across all age groups, in contrast to the bivariate models. In the adjusted models, residing in a VEI community was not associated with vaccine or booster coverage in any age group except for persons 65+ years who were −3.7 percentage points (95% CI −7.3 to −0.1, p < 0.05) less likely to be boosted if they lived in a VEI community.

Discussion

We assessed geographic and sociodemographic equity in the state of Massachusetts (MA)’s COVID-19 vaccination program as of October 10, 2022, through analysis of newly released data on primary series vaccinations and boosters by age and ZIP code. Our findings indicate that coverage of the primary series vaccine among elderly MA residents (65 years and older) was high and equitable, reaching over 95% in the lowest-income decile of ZIP codes and in those with the greatest share Black, Latino, and Indigenous. This finding likely reflects the state’s consistent emphasis on vaccinating older adults, who are at high risk for severe complications. Primary series vaccine coverage for younger MA residents was lower and exhibited large disparities by ZIP code-level education, income, percent essential workers, and racial composition. In addition, very large inequities were observed for booster coverage, with gaps of more than 30 percentage points between the lowest- and highest-income ZIP codes. These observed gaps in vaccine-induced protection could have significant adverse health consequences during future waves of COVID-19 infection. Protection against the Omicron variant is diminished in the absence of a booster [30,31].

Racial disparities in vaccination coverage were statistically explained by differences in age and education levels. Despite lower vaccination coverage in communities with large shares of essential workers and Black, Latino, or Indigenous residents, we found that these characteristics were not independently associated with lower vaccine or booster uptake. Far from “vaccine hesitancy” [32], these findings suggest greater demand for vaccination among populations that have been most affected by COVID, after adjusting for age and socioeconomic factors. A causal interpretation of our model would imply that ZIP codes with many essential workers or Black/Latino/Indigenous residents had lower vaccination or booster rates because their populations were on average younger, had lower levels of educational attainment and lower incomes, and because the vaccination campaign did not adequately address these barriers to access.

Medical mistrust was widely noted early in the rollout, and efforts at community outreach may have been important in increasing vaccine uptake among Black, Latino, and Indigenous populations. The high density of community organizations in MA may have facilitated diffusion of this information. Additionally, MA’s communities of color were hit hardest by the pandemic in 2020, and fear of COVID-19 was higher in these communities, e.g., resulting in higher support for remote schooling among Black MA residents. These factors might explain why demand for COVID-19 vaccination has been high in MA’s communities of color—after adjusting for age, income, and education levels. The more substantial disparities in boosters may relate to reduced outreach efforts subsequent to the initial vaccine rollout. In addition, demand for boosters could have been lower among communities and subpopulations with higher rates of infection, given the perception that acquired immunity (with or without vaccination) is sufficiently protective.

Epidemiological analyses of race/ethnicity demand attention to underlying causal pathways [33]. Differences in education, income, and occupation between racial groups are a product of structural racism. (In MA, for example, reliance on local property taxes for education funding perpetuates schooling inequities and residential segregation.) Therefore, it is appropriate to think of these variables as mediators rather than as confounders of the relationship between racial composition and vaccine coverage. Disparities by racial composition should not be ignored even if they can be statistically explained by socioeconomic factors.

Additionally, Black, Latino, and Indigenous residents and essential workers have experienced disproportionate infection rates, morbidity, and mortality from COVID-19 [10–12] and would therefore be expected to benefit more from vaccination than other groups that are less likely to be exposed to COVID and/or less likely to experience severe illness.

Along with racial/ethnic patterns, substantial disparities related to income and education attainment persist, and local health departments continue to call for more long-term investments to close these gaps. Interventions to address vaccine access barriers related to education and poverty—including offering more convenient clinic times, paid sick leave for potential side effects, providing information through trusted community-based sources in multiple languages, and conducting outreach outside of the formal medical sector—will be essential to improving equity across multiple dimensions, including by race and occupation [16].

Our study has some limitations. First, our numerator and denominator data come from different sources. People may misreport their ZIP code at the vaccination site or may move to another ZIP code since receiving the vaccine. The population denominator data are estimated for ZIP codes based on aggregation of census tract-level ACS estimates and under the assumption that the population for 2020 to 2021 was similar to the population for 2015 to 2019. Second, sampling error in the denominators results in uncertainty in coverage estimates for specific ZIP codes. Third, a small percentage of MA residents did not provide ZIP codes when they received their primary series or booster vaccine. Fourth, we excluded 167 ZIP codes because they were nonresidential or institutional ZIP codes, including PO boxes, universities, and businesses with dedicated ZIP codes. Exclusion of these ZIP codes yielded a data set with greater comparability across units covering 97% of the MA population; however, some MA residents were not represented in our analysis. Due to these exclusions—particularly of university-specific ZIP codes—our totals for vaccination and booster coverage differ somewhat from published statewide estimates. Fifth, we further excluded the smallest ZIP codes (n = 63) representing 1% of the population as the estimates were unstable due to very small denominators. Sixth, we lacked individual-level data that would enable inferences on the experiences of essential workers, individuals with different income and education levels, and of different race/ethnicities. Our inferences are therefore restricted to population-level statements about people in ZIP codes with different characteristics. Seventh, we do not distinguish between essential workers who were in healthcare and typically required to be vaccinated and in other sectors. Eighth, in assessing ZIP code racial/ethnic composition, we used a single metric for percent Black, Latino, or Indigenous, as these populations are concentrated in a relatively small number of MA ZIP codes. Our analysis misses important differences in the experiences of Black, Latino, and Indigenous MA residents. Ninth, it is possible that vaccine or booster coverage could be underestimated due to failure to link vaccine records longitudinally for individual residents; although all vaccinations are mandated to be reported into the Massachusetts Immunization Information System (MIIS), the foundation for the data used in our analysis, some data may be missing or incomplete, especially for those who may have received a shot outside of the state. Tenth, MA is one of the wealthiest states in the country, and patterns observed in MA may not hold elsewhere in the US.

These limitations should be considered alongside the study’s strengths, namely, the use of official, state-reported data on primary series vaccine and booster shots that was aggregated from data on place of residence collected at vaccination facilities; de novo construction of population denominators at the postal ZIP code level enabling analysis of small-area variation in vaccine coverage not previously reported statewide; assessment of inequities across a range of ZIP code characteristics relevant to the epidemiology of COVID-19 and vaccine uptake; and assessment of inequities stratified by age. Finally, our analysis includes all persons vaccinated through October 10, 2022, about 18 months into the general population vaccine rollout.

Closing vaccine coverage gaps in MA and in other states will require ongoing concerted effort. The initial MA vaccination campaign in Spring 2021 was accompanied by daily media coverage and an explicit focus on equity. The late 2021 Delta and Omicron waves refocused public attention on the importance of completing the primary vaccine series and getting a booster shot. Once the 2021/2022 Omicron wave ended, vaccinations slowed. From March to September 2022, coverage of the primary series vaccine increased from 80% to 83% and booster coverage increased from 43% to 50%. A large share of MA residents have yet to receive the additional protection afforded by booster shots. Bivalent boosters were approved in September 2022, providing an opportunity to implement strategies in the near term to address inequitable access and coverage.

Our findings show where MA should focus its efforts: ensuring high, equitable booster coverage including among middle-aged and older adults, and vaccinating children and younger adults. Recent data indicate that vaccines for children are effective in reducing hospitalizations [34]. Strategies used successfully during the initial vaccine rollout, including community outreach efforts and establishment of convenient venues for vaccination, need to be continued during the rollout of boosters to close coverage gaps in lower-educated communities regardless of racial composition.

In conclusion, our analyses indicate large geographic and sociodemographic inequities in vaccine and booster coverage, which should allow for targeted outreach efforts that leverage local infrastructure, including school-based immunization, workplace vaccination drives, community-based campaigns in multiple languages, and routine clinical care. In particular, the relatively lower uptake of booster shots (which is seen across the US) coupled with large disparities in who has received boosters indicates that increasing bivalent booster coverage should be a high-priority effort to protect against severe outcomes. Low booster coverage will lead to higher hospitalization and mortality rates in the future, with associated need for enhanced public health measures. Ensuring access and communicating the ongoing importance of vaccination and boosters will be essential.

Supporting information

S1 STROBE Checklist. STROBE Checklist.

(PDF)

Click here for additional data file. (155.2KB, pdf)
S1 Fig. Percent (A) vaccinated and (B) boosted in urban areas.

Note: S1 Fig displays higher-magnification maps of urban areas from the maps shown in Fig 1. To facilitate comparisons across ZIP codes within each panel (a) and (b), the scales in the panels differ. Shape files were obtained from the US Census Bureau, accessed October 20, 2022: https://www.census.gov/geographies/mapping-files/time-series/geo/tiger-geodatabase-file.html. Copyright protection is not available for any work of the United States Government (Title 17 U.S.C., Section 105).

(PDF)

Click here for additional data file. (6.4MB, pdf)
S1 Table. Age-adjusted bivariate associations of ZIP code characteristics with percent vaccinated and boosted.

Note: Each cell shows results of a separate OLS (ordinary least squares) regression model. The crude models are bivariate. The age-adjusted include controls for ZIP code age shares in the following groups: 5–19, 20–39, 40–64, and 65+ years. Heteroskedasticity-robust 95% confidence intervals are shown in parentheses. *p < 0.05, **p < 0.01, ***p < 0.001.

(PDF)

Click here for additional data file. (67.3KB, pdf)
S2 Table. Sex-stratified models: Associations of ZIP code characteristics with percentage vaccinated and boosted.

Note: Each column shows results of a separate OLS (ordinary least squares) regression model. The models are stratified by sex. Each includes controls for ZIP code age shares in the following groups: 5–19, 20–39, 40–64, and 65+ years. Heteroskedasticity-robust 95% confidence intervals are shown in parentheses. *p < 0.05, **p < 0.01, ***p < 0.001.

(PDF)

Click here for additional data file. (103.4KB, pdf)

Abbreviations

ACS

American Community Survey

CI

confidence interval

COVID-19

Coronavirus Disease 2019

DPH

Department of Public Health

MA

Massachusetts

MIIS

Massachusetts Immunization Information System

VEI

Vaccine Equity Initiative

ZCTA

ZIP code tabulation area

Data Availability

Data are publicly available for download from the Massachusetts Department of Public Health (https://www.mass.gov/info-details/massachusetts-covid-19-vaccination-data-and-updates) and US Census Bureau (https://www.census.gov/programs-surveys/acs/data/data-tables.html).

Funding Statement

The author(s) received no specific funding for this work.

References

Decision Letter 0

Richard Turner

18 May 2022

Dear Dr Bor,

Thank you for submitting your manuscript entitled "Inequities in COVID-19 vaccine and booster coverage across Massachusetts ZIP codes: a population-based study following the Winter 2022 Omicron wave" for consideration by PLOS Medicine.

Your manuscript has now been evaluated by the PLOS Medicine editorial staff and I am writing to let you know that we would like to send your submission out for external assessment.

However, we first need you to complete your submission by providing the metadata that are required for full assessment. To this end, please login to Editorial Manager where you will find the paper in the 'Submissions Needing Revisions' folder on your homepage. Please click 'Revise Submission' from the Action Links and complete all additional questions in the submission questionnaire.

Please re-submit your manuscript within two working days, i.e. by May 20 2022 11:59PM.

Login to Editorial Manager here: https://www.editorialmanager.com/pmedicine

Once your full submission is complete, your paper will undergo a series of checks in preparation for assessment.

Feel free to email us at plosmedicine@plos.org if you have any queries relating to your submission.

Kind regards,

Richard Turner, PhD

Senior Editor, PLOS Medicine

rturner@plos.org

Decision Letter 1

Beryne Odeny

20 Jul 2022

Dear Dr. Bor,

Thank you very much for submitting your manuscript "Inequities in COVID-19 vaccine and booster coverage across Massachusetts ZIP codes: a population-based study following the Winter 2022 Omicron wave" (PMEDICINE-D-22-01626R1) for consideration at PLOS Medicine.

Your paper was evaluated by a senior editor and discussed among all the editors here. It was also sent to independent reviewers, including a statistical reviewer. The reviews are appended at the bottom of this email and any accompanying reviewer attachments can be seen via the link below:

[LINK]

In light of these reviews, I am afraid that we will not be able to accept the manuscript for publication in the journal in its current form, but we would like to consider a revised version that addresses the reviewers' and editors' comments. Obviously we cannot make any decision about publication until we have seen the revised manuscript and your response, and we plan to seek re-review by one or more of the reviewers.

In revising the manuscript for further consideration, your revisions should address the specific points made by each reviewer and the editors. Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments, the changes you have made in the manuscript, and include either an excerpt of the revised text or the location (eg: page and line number) where each change can be found. Please submit a clean version of the paper as the main article file; a version with changes marked should be uploaded as a marked up manuscript.

In addition, we request that you upload any figures associated with your paper as individual TIF or EPS files with 300dpi resolution at resubmission; please read our figure guidelines for more information on our requirements: http://journals.plos.org/plosmedicine/s/figures. While revising your submission, please upload your figure files to the PACE digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at PLOSMedicine@plos.org.

We expect to receive your revised manuscript by Aug 10 2022 11:59PM. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

We ask every co-author listed on the manuscript to fill in a contributing author statement, making sure to declare all competing interests. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. If new competing interests are declared later in the revision process, this may also hold up the submission. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT. You can see our competing interests policy here: http://journals.plos.org/plosmedicine/s/competing-interests.

Please use the following link to submit the revised manuscript:

https://www.editorialmanager.com/pmedicine/

Your article can be found in the "Submissions Needing Revision" folder.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

We look forward to receiving your revised manuscript.

Sincerely,

Beryne Odeny,

PLOS Medicine

plosmedicine.org

-----------------------------------------------------------

Requests from the editors:

1) Please revise your title according to PLOS Medicine's style. Your title must be nondeclarative and not a question. It should begin with main concept if possible. Please place the study design in the subtitle (i.e., after a colon), e.g., a retrospective cohort study. Please place the statement, “following the winter 2022 Omicron wave” before the colon.

2) Please include line numbers in your next draft.

3) Abstract:

a) Please ensure that all numbers presented in the abstract are present and identical to numbers presented in the main manuscript text.

b) Please mention the study design at the start of the Methods and Findings section

c) Please quantify the main results (please present both 95% CIs and p values).

d) In the last sentence of the Abstract Methods and Findings section, please describe the main limitation(s) of the study's methodology.

e) Conclusions:

i) Please interpret the study based on the results presented in the abstract, emphasizing what is new without overstating your conclusions.

ii) Please mention specific implications substantiated by the results.

4) Author Summary – under the subheading, “What did the researchers do and find?” please trim the text to no more than 4 bullet points

5) Please conclude the “Introduction” with a clear description of the study question or hypothesis.

6) Did your study have a prospective protocol or analysis plan? Please state this (either way) early in the Methods section.

a) If a prospective analysis plan (from your funding proposal, IRB or other ethics committee submission, study protocol, or other planning document written before analyzing the data) was used in designing the study, please include the relevant prospectively written document with your revised manuscript as a Supporting Information file to be published alongside your study, and cite it in the Methods section. A legend for this file should be included at the end of your manuscript.

b) If no such document exists, please make sure that the Methods section transparently describes when analyses were planned, and when/why any data-driven changes to analyses took place.

c) In either case, changes in the analysis-- including those made in response to peer review comments-- should be identified as such in the Methods section of the paper, with rationale.

7) Please provide the name(s) of the institutional review board(s) that provided ethical exemption.

8) Please ensure that the study is reported according to the STROBE and include the completed STROBE checklist as Supporting Information. Please add the following statement, or similar, to the Methods: "This study is reported as per the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guideline (S1 Checklist)."

a) When completing the checklist, please use section and paragraph numbers, rather than page numbers.

9) How was race/ethnicity defined and by whom?

10) Please provide p values (where appropriate) in addition to 95% CIs in the main text and tables

11) Figures and tables:

a) Please define all abbreviations in Tables and Figures e.g., SD, HH, Blk, Hisp, Indig, OLS

b) Please confirm that the appropriate usage rights apply to the use of this map. Please see our guidelines for map images: https://journals.plos.org/plosmedicine/s/figures#loc-maps

c) Please consider avoiding the use of red and green in order to make your figure more accessible to those with color blindness

12) References:

a) Please select the PLOS Medicine reference style in your citation manager. In-text reference call outs should be presented as follows noting the absence of spaces within the square brackets, “…population [1,2].”

b) References should have six names before et al.

c) Please provide a weblink for Reference #22

d) Please ensure all weblinks are accessible

e) Please ensure that journal name abbreviations consistently match those found in the National Center for Biotechnology Information (NCBI) databases. https://journals.plos.org/plosmedicine/s/submission-guidelines#loc-references.

Comments from the reviewers:

Reviewer #1: I confine my remarks to statistical aspects of this paper. These were well done and I recommend publication.

Peter Flom

Reviewer #2: Overall comments:

This is a well-written manuscript examining inequity in vaccination and boosters in Massachusetts using aggregated state-level data. The authors use sophisticated methods to grapple with some of the discrepancies between the state-level vaccination data and population census data. The authors' present interesting and new data on associations between zip code level characteristics and vaccination and booster rates; it is an important finding to see how these disparities differ for initial vaccination and boosters. Most of the comments focus on areas that could be strengthened or clarified a bit. Mostly, I think the framing of the fact that associations reverse with adjustment could be sharper…I think having this crystal clear is important so readers do not take away the "wrong" message from it. I think additional exploration of the VEI program would be quite interesting…any insights into what potential impact it may have had I think would be very important to know (acknowledging there may be many limitations to being to do such an analysis).

Specific comments:

Intro

* The introduction is well-written. I think the systemic nature of root causes of the disparities could be spelled out a bit more. Many of the mechanisms of the systemic racism in healthcare are there, but they could be tied together. Also could bring in the fact that there are different phenotypes of vaccine hesitancy/confidence, which is also rooted in disparities.

Methods:

* I wasn't fully clear on the imputation approach for the neither male or female. Did you use the ratio for the zip codes where this information was available (there were a few zip codes with this information) to impute all the rest? Where there instances where the imputation estimated greater than 30 people for a zip code (in which case you would expect that data should not have been missing…perhaps needs a cap too). Is neither male nor female available in census data to understand the ratio in the total population.

* Also, do estimates change at all if you used age (I imagine this data is also fairly complete?). I think another approach that could be considered is multiple imputation that could potentially incorporate all the reported vaccine data and zip code level characteristics. Hopefully all of these would generate fairly similar results, but some sensitivity analyses to confirm might be nice.

* For the zip by age analyses, it is possible to do a similar imputation process to not exclude those with missing cells. This might require a more complicated imputation process since age distributions will likely vary across zip code level characteristics too. Just unclear what, if any, impact this missingness has on results. Might also be helpful to present this missingness information so that extent of it is transparent.

* The methods used to estimate percent vaccinated or boosted that accounts for the uncertainty in population estimates is sophisticated. However, I wasn't clear where the 90% confidence intervals for percent vaccinated came into play for any of the analyses. Figures 3 to 5 seem to be using something else, but I may be mistaken.

* I might separate out the different analyses under zip code-level and zip code by age (each its own paragraph). Particulalry for the regression and interpretation of adjustment, there is enough to warrant separating it out. I had a bit of a hard time matching them up with figures an results as I went through ti.

Results

* I wonder if by the geospatial plots it would be helpful to make the scale for vaccinated and boosted plot the same. Currently they seem to be color-coded by decile, but the most prominent finding then becomes that zip codes in the top/bottom deciles for the initial vaccine series remain there for boosters as well. Perhaps that is the point trying to be conveyed though. Either way, a figure legend that has a bit of walkthrough to highlight the main findings would be helpful.

* Figure 2 is quite informative. It shows clear relationships between zip code characteristics and vaccination (and stronger relationship for boosters). Reporting out the confidence intervals around the estimates would nice and also the p-value or correlation coeffiecient (I see that it is in Table 2 and so could also just make clear the estimates are there).

* The paragraph starting with "Bivariate association…" seems like it should be included in the methods section to explain the rationale for the different regression models.

* I found Table 2 kind of hard to interpret. The reversal of the slope with adjustments seems a bit unexpected. Is there concern for overfitting (not sure there would be with 418 zip code)? The authors do explain this in the results and discussion, though it is still hard to align with other data in the literature (which often show independent associations of all these factors). It may be that the in Massachusetts, the effects of systemic racism are almost fully mediated through income and education (as opposed to many other areas of the countries where this is not the case).

* I think something else worth considering is that it may also be reflective of success of the VEI program in that it was able to attenuate many of these associations (or reverse them). In the final model, being a VEI program was associated with only 0.5% difference in vaccination…which seems very low. Though not able to assess here, I would expect that difference to likely be higher.

* Ultimately, I might try to frame some of the discussion on causal interpretation thinking about mediators of racism and disparities. I always find Camara Jones work quite informative in interpreting these types of analyses/findings (https://academic.oup.com/aje/article/154/4/299/61900). There are complex relationships between income, education, race, employment, and zip code and it is not as straightforward to be able to just adjust them away. Careful interpretation is required (which I think that authors do) but perhaps could go a bit deeper.

* The impact of age structure is actually quite interesting and something I think that hasn't always been taken into account when considering raw estimates. It might also be interesting to see how this impacts the bivariate estimates too.

* I think a bivariate estimate for VEI could also be interesting. To accomodate all of the different estimates, could consider presenting all bivariate estimates in a single column.

* Consider providing details in table 2 about coefficient interpretation (i.e., X% change per 1% increase…could also scale similarly to Figure 2).

* Is it possible to include case and death rates as zip code level covariates too? Not sure how readily available that data is, but is quite relevant to the targeting that happened by the VEI program it would seem.

* For Figure 3, I do know that it common enough that vaccination rates seem to exceed population estimates, though it is still odd to see. Most of the approaches to exclusion or missing data would seem to underreport vaccination rates. I am not sure if there is any strategy to address it, but could consider truncating at 100%. Would address the approaches to this in the methods.

* For Figure 4, consider putting vaccinated and boosted on same panel (and color code differently) and aligning the a-d on one page. Would also add figure legend that explains lowest to highest decile for quicker orientation for the reader. Also consider putting children on the same figure (as opposed to figure 5).

* The authors assess age strata by zip code characteristic interactions. Is it also possible to assess race by zip code characteristic interactions also. This data seems available, though missingness may be a problem. It doesn't seem possible to essentially analyze at individual level using frequency weights, but this could be included as a limitation.

* Are the authors' aware of changes to the VEI efforts during the initial campaign vs. for boosters. I imagine there were. If the VEI initiative was more prominent during the initial vaccination campaign, could explore the interaction of its effect during these two periods. This is already alluded to in results for vaccination vs. booster but could be more prominent and formally assessed. Should also be included in the discussion.

*

Discussion

* I think something that ends up somewhat de-emphasized in the discussion how associations changed after adjustments is that being a minority in the US and experiencing systemic racism is can't really be disaggregated from income, education, where you live, and healthcare access. The ultimate conclusions the authors reach are fine, but I think the meaning of why some of these associations attenuated after adjustment needs to be sharpened. Again Camara Jones' work is illustrative here. What comes across a bit—and I don't think the authors intended this—is that race, employment aren't the root causes and it is just income and education.

* I think this finding probably needs to be contextualized in the setting of Massachusetts (I would expect the findings would differ in many other parts of the US) and perhaps also in the context of the VEI program. I am not sure it would be too far fetched to hypothesize that VEI program may be part of the reason zip codes with more minorities and essential workers outperformed expectations after adjustment specifically they were targeted for additional resources. The authors can consider whether this is reasonable or not.

* I think additional limitations to consider are that 1) this analysis is primarily ecological and could not use individual-level data, 2) missing data, and 3) also the discrepancy in denominators for vaccination and the census data (this is alluded to already but could be more specific).

* I think in general more contextualization is needed in terms of what the findings mean beyond the different analytic strategies (e.g., this analysis was adjusted this one was not). What is larger of importance of what the results mean because of these analytic differences.

* For example, I think more contextualization is needed in terms of this finding "In fact, the higher risk of severe COVID-19 in low-income households seems to be strongly driven by personal occupational status and a foreign background." Why would this be the case? What are the mechanisms

Aaloke Mody

Reviewer #3: This is an excellent, well-written article on disparities in COVID-19 and vaccine booster uptake in Massachusetts. The authors take a careful look at racial/ethnic and other group differences in vaccine/booster uptake, adjusting for age and other factors. Among their key and important findings are that many racial/ethnic differences that were assumed to be due to differences in vaccine hesitancy are likely instead attributable to differences in age and SES distributions. The authors' framing of the issues is excellent. The methodological approaches are thoughtful and rigorously applied. The data are very timely (through March 2022). Both primary vaccine and booster uptake are examined.

I have only a few minor suggestions that the authors may want to consider:

1. There have been some concerns about the accuracy of vaccine coverage data from state and local health department vaccine registries. The authors mention many of the reasons why there are some limitations to the data. However, one that they don't mention is that accurately estimating vaccine coverage requires linkage of longitudinal information across individuals (i.e., dose 1, dose 2, and booster doses). If there is not good matching, this can result in underestimating or overestimating coverage. When fully vaccinated person A's dose 1 and dose 2 are not linked, it could be counted as two individuals with single doses of vaccine only. When person B and person C's first doses are incorrectly linked, then one person will be counted as fully vaccinated and not two as under vaccinated. etc. I am not sure how much of an issue this is for Massachusetts, but the authors may have some information worth sharing in the article about the data quality of the numerator. In other jurisdictions and nationally, there have been some implausible estimates of coverage by race/ethnicity, particularly on the higher end. But they may want to say more about the data generating mechanisms and data sources for the numerator data.

2. Could the authors clarify the source of ZIP level occupation data?

3. The authors mention that a key aim of the regression analysis is to determine whether bivariate associations are explained by other factors. Would it be helpful to include DAGs in the appendix, to differentiate potential confounding vs potential mediating? Or at least to discuss some of the possible reasons why bivariate associations may be fully or partially or not at all explained by other factors?

4. While not possible to estimate at the ZIP code level without joint distributions, it may be important for the authors to recognize the potential role of prior history of SARS-CoV-2 infection as a possible factor weighing on people's decisions to pursue a vaccine or booster. i.e., those who had covid in the past, including recent past, may be less likely to pursue a booster. And also that in considering at least the short term risks for severe outcomes during surges, it would be helpful to know the extent to which populations (say ZIP codes or by SES strata) the proportion of the population that has protection due to vaccination/boosting only or vaccination/boosting PLUS a recent history of COVID. This may also explain some of the differences between groups, including essential workers (and other groups who have endured a disproportionate share of the pandemic's burden) vs. others.

Reviewer #4: Thank you for the opportunity to review this research.

This is a well written paper.

1. For the multivariable regression（Table 2a and Table 2b), how was age distribution adjusted? Do it mean age, as a continuous variable, was included in the models?

2. Figure 1: what's the difference between "missing" and "suppressed". How many zip codes were excluded because of missing ness, and how many were excluded because they corresponded to specific universities or businesses, they did not correspond to residential areas, or they were assigned to post office boxes and not to residential addresses, o

3. Figure 3: why the point estimate and upper bound of the confidence exceed 1? Is there any method that could make the simulated estimate less than 1?

4. Through stratified analysis, it's interesting that larger inequities in vaccination rates among younger adults and children were found. The author visually presented the the difference in the inequities in vaccination rates across different age groups (Figure 4), but whether the differences are statistically significant have not been tested (Table 3).

5. In addition to age, could sex be a potential effect modifier as well?

Any attachments provided with reviews can be seen via the following link:

[LINK]

Decision Letter 2

Philippa Dodd

12 Dec 2022

Dear Dr. Bor,

Thank you very much for re-submitting your manuscript "Inequities in COVID-19 vaccine and booster coverage across Massachusetts ZIP codes in Fall 2022: a population-based cross-sectional study" (PMEDICINE-D-22-01626R2) for review by PLOS Medicine.

I have discussed the paper with my colleagues and the academic editor and it was also seen again by 3 reviewers. I am pleased to say that provided the remaining editorial and production issues are dealt with we are planning to accept the paper for publication in the journal.

The remaining issues that need to be addressed are listed at the end of this email. Any accompanying reviewer attachments can be seen via the link below. Please take these into account before resubmitting your manuscript:

[LINK]

***Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.***

In revising the manuscript for further consideration here, please ensure you address the specific points made by each reviewer and the editors. In your rebuttal letter you should indicate your response to the reviewers' and editors' comments and the changes you have made in the manuscript. Please submit a clean version of the paper as the main article file. A version with changes marked must also be uploaded as a marked up manuscript file.

Please also check the guidelines for revised papers at http://journals.plos.org/plosmedicine/s/revising-your-manuscript for any that apply to your paper. If you haven't already, we ask that you provide a short, non-technical Author Summary of your research to make findings accessible to a wide audience that includes both scientists and non-scientists. The Author Summary should immediately follow the Abstract in your revised manuscript. This text is subject to editorial change and should be distinct from the scientific abstract.

We expect to receive your revised manuscript within 1 week. Please email us (plosmedicine@plos.org) if you have any questions or concerns.

We ask every co-author listed on the manuscript to fill in a contributing author statement. If any of the co-authors have not filled in the statement, we will remind them to do so when the paper is revised. If all statements are not completed in a timely fashion this could hold up the re-review process. Should there be a problem getting one of your co-authors to fill in a statement we will be in contact. YOU MUST NOT ADD OR REMOVE AUTHORS UNLESS YOU HAVE ALERTED THE EDITOR HANDLING THE MANUSCRIPT TO THE CHANGE AND THEY SPECIFICALLY HAVE AGREED TO IT.

Please ensure that the paper adheres to the PLOS Data Availability Policy (see http://journals.plos.org/plosmedicine/s/data-availability), which requires that all data underlying the study's findings be provided in a repository or as Supporting Information. For data residing with a third party, authors are required to provide instructions with contact information for obtaining the data. PLOS journals do not allow statements supported by "data not shown" or "unpublished results." For such statements, authors must provide supporting data or cite public sources that include it.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript.

Please note, when your manuscript is accepted, an uncorrected proof of your manuscript will be published online ahead of the final version, unless you've already opted out via the online submission form. If, for any reason, you do not want an earlier version of your manuscript published online or are unsure if you have already indicated as such, please let the journal staff know immediately at plosmedicine@plos.org.

If you have any questions in the meantime, please contact me or the journal staff on plosmedicine@plos.org.  

We look forward to receiving the revised manuscript by Dec 19 2022 11:59PM.   

Sincerely,

Philippa Dodd MBBS MRCP PhD

PLOS Medicine

plosmedicine.org

------------------------------------------------------------

Requests from Editors:

GENERAL

Please respond to all editor and reviewer comments detailed below, in full

Line numbers should start at 1 on the title page and continue in sequence to the end of the manuscript. In your response to revisions please indicate where in the manuscript the revision can be located (by section and line number)

DATA AVAILABILITY STATEMENT

Thank you for making your data available. Your data availability statement requires revision. Please also provide the URL for the Massachusetts Department of Public Health and US Census Bureau, in the manuscript submission form.

TITLE

Thank you for revising your title. Previously you detailed “following the Winter 2022 Omicron wave” now you detail “across Massachusetts ZIP codes in Fall 2022”. In your abstract you state “vaccination and booster coverage 18 months into the general population vaccine rollout…as of October 2022”. The different time descriptions used in the two versions of your title describe different time periods (suspect that in the former you mean Winter 2021 Omicron wave?).

Does the time frame need to be defined by season? Also, the term “fall” is not universally adopted to describe the Autumn season. Suggest instead: “Inequities in COVID-19 vaccine and booster coverage across Massachusetts ZIP codes after the emergence of Omicron: a population-based cross-sectional study” or : “Inequities in COVID-19 vaccine and booster coverage across Massachusetts ZIP codes in October 2022: a population-based cross-sectional study” or a similar variation of either.

ABSTRACT

Line 19-20: would it be helpful to also report the data that you refer to in this statement?

AUTHOR SUMMARY

Please remove the box and shading surrounding the author summary – it should be included as text rather than in figure form (and should be included in the line numbering also)

“Percent college graduates was the strongest predictor of vaccine and booster uptake.” Would it be helpful to simply state “education level” or something similar?

“These impacts are preventable.” Suggest tempering of language here, especially as you state “inequities…may lead to…”. Suggest “Inequities in vaccine and booster coverage may lead to preventable inequities in morbidity, mortality, and economic losses due to COVID-19.” Or something similar.

INTRODUCTION

Page 4, line 1: Please change the heading “Background” to “Introduction”

Page 5, line 2: Please define MA at first use here as follows: “Massachusetts (MA)”

METHODS and RESULTS

Page 7 lines 29-30: suggest the following “….across ZIP codes: percentage of ZIP code residents aged 0-4 years, 5-19 years, 20-39 years…” Note the removal of the repeated % symbol. This also makes your reporting consistent with that on page 6, line 34. Please check and amend throughout all sections of the manuscript

Page 10 paragraph 1 – it would be helpful to enumerate the percentages reported here

Page 12, line 40: “40 CI -7.3 to -0.1, p<.05” please report p as <0.05 – please check and amend throughout including tables and figures as necessary

Figure 3: when describing figure 3 you refer to “columns” but they are not columns they are graphs stacked upon each other. Please refer to plots or graphs “(left sided graphs)” for example

*** Regarding the reviewer’s comment re: sex as a potential modifier. Any changes to analyses that are peer review driven should and can be included in the manuscript as part of PLOS Medicine’s transparency of data reporting policy. Changes in the analysis should be identified as such in the Methods section of the paper, with rationale. Outcome of the analysis should be reported and/or discussed io the relevant section of the manuscript. The data table can be included in the supporting files. Please include a caption which describes its content. As for all tables, please report p-values (not only asterisks), 95% CIs, define all abbreviations, report whether any other variables are adjusted for, report unadjusted analyses for comparison. ***

FIGURES and TABLES

All figures and tables should have an appropriate figure caption such that the contents of the figure/table are accessible without the manuscript text. Please include as required including the supplementary files.

Figure 2 Title: suggest “Percent vaccinated and boosted according to individual ZIP code characteristics” or something similar

Figure 4: Please use full term “indigenous” the abbreviation could be considered offensive. Without the manuscript text to refer to, it is not easy to understand what the figure shows. What does "ZIP code decile" mean? Please revise in-line with above comments.

TABLES

Table 1: Considering the study shows disparity between ZIP codes, is it necessary to show combined data in a table here? If so, suggest revising the title to “Mean Vaccination Coverage (percentage) Across the State of Massachusetts According to Age” or something similar.

We ask that all authors provide a table showing the baseline characteristics of the study population and specific to your study, ZIP code demographics should also be included. Perhaps the currently presented data referred to above, could be incorporated into this table.

Table 2: Thank you for reporting p-values. We request that these are reported in full in a separate column, rather than defined by asterisks. Please revise accordingly, this may require you to split your tables. For main outcome measures, where analyses are adjusted, we request of all authors that the unadjusted analyses also be reported. Please indicate which factors are adjusted for in the table caption.

We note reviewer comments re: presentation of Table 2 and agree that accessibility could be improved, the empty cells in the tables are rather distracting and in light of the aforementioned required revisions, we request that the table is revised accordingly.

Table 3: As above. Please report the p-value result in a separate column in place of asterisks. Please provide unadjusted analyses as necessary.in the table caption please indicate which factors are adjusted for.

Please define VEI in the table caption

DISCUSSION

Please ensure that the Discussion is organized and presented as follows: a short, clear summary of the article's findings; what the study adds to existing research and where and why the results may differ from previous research; strengths and limitations of the study; implications and next steps for research, clinical practice, and/or public policy; one-paragraph conclusion.

REFERENCES

Page 4, line 15-16: In-text reference callouts should be placed in square parentheses and preceding punctuation as follows: “…please check [1,2,3,4].” Please note the space after the text before the opening parenthesis. Please check and revise throughout the manuscript.

In your bibliography please ensure you list up to but no more than 6 authors (followed by et al where more than 6 authors contribute to the cited study

Journal name abbreviations should be those found in the National Center for Biotechnology Information (NCBI) databases.

Please see out guidelines for further information here: https://journals.plos.org/plosmedicine/s/submission-guidelines#loc-references

SOCIAL MEDIA

To help us extend the reach of your research, please provide any Twitter handle(s) that would be appropriate to tag, including your own, your coauthors’, your institution, funder, or lab. Please respond to this email with any handles you wish to be included when we tweet this paper.

COMMENTS FROM THE ACADEMIC EDITOR

Identifying potential areas of inequity for COVID-19 vaccines is certainly an important area. I read through the author responses and the article. It has done a nice job using available data to provide some insights for MA. Would agree with the reviewers and accept at this time.

Comments from Reviewers:

Reviewer #2: The authors were very responsive to reviewer comments. I have no further suggestions.

Reviewer #3: The authors have been very responsive to comments from reviewers. It is an excellent paper!

Reviewer #4: My questions have been sufficiently answered. I do not have any further comments. Thnak you.

Any attachments provided with reviews can be seen via the following link:

[LINK]

Decision Letter 3

Philippa Dodd

28 Dec 2022

Dear Dr Bor, 

On behalf of my colleagues and the Academic Editor, Dr. Amitabh Suthar, I am pleased to inform you that we have agreed to publish your manuscript "Inequities in COVID-19 vaccine and booster coverage across Massachusetts ZIP codes after the emergence of Omicron: a population-based cross-sectional study" (PMEDICINE-D-22-01626R3) in PLOS Medicine.

Prior to publication please address the following final revision:

* Page 10, line 376 onwards: regarding descriptive information and percentages, please also include numerators (and denominators) used to derive these – i.e. what number of people aged 65+ did the 100% coverage equate to? Apologies for my previous lack of clarity regarding this point.

Before your manuscript can be formally accepted you will need to complete some formatting changes, which you will receive in a follow up email. Please be aware that it may take several days for you to receive this email; during this time no action is required by you. Once you have received these formatting requests, please note that your manuscript will not be scheduled for publication until you have made the required changes.

In the meantime, please log into Editorial Manager at http://www.editorialmanager.com/pmedicine/, click the "Update My Information" link at the top of the page, and update your user information to ensure an efficient production process. 

PRESS

We frequently collaborate with press offices. If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximise its impact. If the press office is planning to promote your findings, we would be grateful if they could coordinate with medicinepress@plos.org. If you have not yet opted out of the early version process, we ask that you notify us immediately of any press plans so that we may do so on your behalf.

We also ask that you take this opportunity to read our Embargo Policy regarding the discussion, promotion and media coverage of work that is yet to be published by PLOS. As your manuscript is not yet published, it is bound by the conditions of our Embargo Policy. Please be aware that this policy is in place both to ensure that any press coverage of your article is fully substantiated and to provide a direct link between such coverage and the published work. For full details of our Embargo Policy, please visit http://www.plos.org/about/media-inquiries/embargo-policy/.

To enhance the reproducibility of your results, we recommend that you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. Additionally, PLOS ONE offers an option to publish peer-reviewed clinical study protocols. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols

Thank you again for submitting to PLOS Medicine. We look forward to publishing your paper. 

Best wishes, 

Philippa Dodd, MBBS MRCP PhD 

PLOS Medicine

Associated Data

    This section collects any data citations, data availability statements, or supplementary materials included in this article.

    Supplementary Materials

    S1 STROBE Checklist. STROBE Checklist.

    (PDF)

    Click here for additional data file. (155.2KB, pdf)
    S1 Fig. Percent (A) vaccinated and (B) boosted in urban areas.

    Note: S1 Fig displays higher-magnification maps of urban areas from the maps shown in Fig 1. To facilitate comparisons across ZIP codes within each panel (a) and (b), the scales in the panels differ. Shape files were obtained from the US Census Bureau, accessed October 20, 2022: https://www.census.gov/geographies/mapping-files/time-series/geo/tiger-geodatabase-file.html. Copyright protection is not available for any work of the United States Government (Title 17 U.S.C., Section 105).

    (PDF)

    Click here for additional data file. (6.4MB, pdf)
    S1 Table. Age-adjusted bivariate associations of ZIP code characteristics with percent vaccinated and boosted.

    Note: Each cell shows results of a separate OLS (ordinary least squares) regression model. The crude models are bivariate. The age-adjusted include controls for ZIP code age shares in the following groups: 5–19, 20–39, 40–64, and 65+ years. Heteroskedasticity-robust 95% confidence intervals are shown in parentheses. *p < 0.05, **p < 0.01, ***p < 0.001.

    (PDF)

    Click here for additional data file. (67.3KB, pdf)
    S2 Table. Sex-stratified models: Associations of ZIP code characteristics with percentage vaccinated and boosted.

    Note: Each column shows results of a separate OLS (ordinary least squares) regression model. The models are stratified by sex. Each includes controls for ZIP code age shares in the following groups: 5–19, 20–39, 40–64, and 65+ years. Heteroskedasticity-robust 95% confidence intervals are shown in parentheses. *p < 0.05, **p < 0.01, ***p < 0.001.

    (PDF)

    Click here for additional data file. (103.4KB, pdf)
    Attachment

    Submitted filename: Bor vaccines, RR1 Response to Reviewers v2.docx

    Click here for additional data file. (479.9KB, docx)

    Data Availability Statement

    Data are publicly available for download from the Massachusetts Department of Public Health (https://www.mass.gov/info-details/massachusetts-covid-19-vaccination-data-and-updates) and US Census Bureau (https://www.census.gov/programs-surveys/acs/data/data-tables.html).

    Articles from PLOS Medicine are provided here courtesy of PLOS

    RESOURCES