Our findings reinforce the value of workplace-based vaccination strategies: the overrepresentation of undervaccinated demographic groups in occupations at higher risk of SARS-CoV-2 exposure can be leveraged to synergistically protect vulnerable workers and communities. Industry and occupation data should be collected at vaccination, to help identify workers at disproportionate risk of disease.
Keywords: SARS-CoV-2, job exposure matrix, vaccine promotion, non–health care workers, occupational health
Abstract
Background
Occupation, a risk factor for SARS-CoV-2 exposure, is excluded from immunization records. Identifying undervaccinated workers could optimize interventions to protect vulnerable populations.
Methods
We analyzed health department case interviews (June 2021 to May 2022) to describe 3763 non–health care workers with COVID-19 in Chicago. Job exposure matrices categorized SARS-CoV-2 exposure risk through frequency of indoor work and proximity to the public and to coworkers. Logistic regression quantified associations between occupational risk and vaccination status.
Results
Pre-Omicron vaccination rates were lower among high-risk versus lower-risk workers (41% vs. 49%). Post-Omicron rates did not differ (75%). Undervaccinated groups had disproportionately high occupational risk: 66% of Black–non-Latinx, 57% of 18- to 29-year-olds. Demographically adjusted models found no associations between occupational risk and vaccination.
Conclusions
Given demographic patterns of employment, workplace-based COVID-19 vaccine promotion can help reach undervaccinated communities. To inform interventions, immunization records must include occupation.
CME Learning Objectives
After completing this enduring educational activity, the learner will be better able to:
Summarize methods for defining occupation-based disease risk: utilizing characteristics of pathogen transmission, job-exposure matrices, and O*NET data.
Discuss differences in SARS-CoV-2 exposure risk by occupation, and pre- and 33 post-Omicron vaccination among workers with COVID-19 in Chicago, including by occupation and demographic subgroups.
Describe the value of (1) capturing work-related data at immunization, and (2) workplace-based strategies to improve COVID-19 vaccine uptake.
In early phases of the COVID-19 pandemic, large outbreaks were identified among workplaces with poor ventilation and prolonged close proximity between workers.1–3 Although they did not necessarily represent all jurisdictions, these experiences informed national vaccination prioritization by industry.4,5 In the United States, initial vaccine allocations included health care personnel in Phase 1a, followed by frontline (non–health care) essential workers (first responders, corrections officers, US Postal Service, and workers in food and agriculture, manufacturing, grocery, public transit, education and childcare) in Phase 1b. In Phase 1c, other essential workers (transportation and logistics, water and wastewater, food service, shelter and housing, finance, information technology and communications, energy, legal, media, public safety and public health) were prioritized over the broader working-age population (Phase 2).4,5 In general, work-related data are not collected at the time of vaccination, precluding studies of vaccination rates or identification of coverage disparities among workers by industry or occupation.
To identify occupations at highest risk of SARS-CoV-2 exposure, the Council of State and Territorial Epidemiologists (CSTE) devised the SARS-CoV-2 Occupational Exposure Matrix (SOEM).6 This framework estimates three categories of risk (low, medium, and high) of workplace-acquired SARS-CoV-2 exposure for non–health care occupations described in the US Department of Labor's Occupational Information Network (O*NET) survey.7 The risk index is based on frequency of three factors: (1) working indoors, (2) close proximity to coworkers (e.g., on assembly lines), and (3) public facing work. These factors have since been assessed in studies of COVID-19 incidence by occupation and used to compare risk among workers by demographic groups. For example, analyses of Wisconsin's COVID-19 cases from September 2020 through May 2021 found that occupational groups categorized as high risk in the SOEM (e.g., personal care and service, food and beverage sales, personal appearance, law enforcement) had the highest COVID-19 incidence of all non–health care groups.8 A 2019 national survey from the Bureau of Labor and Statistics found that workers identifying as Black, non-Latinx, or Latinx9 were overrepresented in occupations that involve working in close proximity to others. This is important because COVID-19 vaccination disparities among these groups in the United States have been well described.10–12 Data showing that occupational factors are associated with both COVID-19 incidence and race-ethnicity suggest that outreach to high-risk workplaces could help reduce disease burden and vaccine hesitancy among working-age Americans and their networks.
Vaccination records for COVID-19 in Illinois have not routinely included industry or occupation as data elements. However, because COVID-19 case interviews conducted by the Chicago Department of Public Health (CDPH) queried occupation, matching interview data to vaccination records enabled description of Chicagoans who contracted COVID-19 by both occupational risk and vaccination status. This analysis aimed to (1) describe working-age Chicagoans who contracted COVID-19 by level of risk of work-related transmission and (2) explore associations between occupational risk and vaccination status in both Pre-Omicron and Post-Omicron periods. Results could inform workplace-based vaccine promotion (e.g., messaging, incentivization, on-site vaccination, employer requirement) to address low vaccination rates.
METHODS
Data Sources and Inclusion Criteria
Analyses were conducted using SAS v9.4. They included all Chicagoans aged 18 to 64 years who completed routine case investigation with CDPH after laboratory-confirmed SARS-CoV-2 infection between June 1, 2021, and May 31, 2022 (n = 25,047). Case and vaccination data originated from Illinois' National Electronic Disease Surveillance System (I-NEDSS) and the State immunization record (Illinois Comprehensive Automated Immunization Registry Exchange or I-CARE), respectively. Case interviews were conducted using a standardized questionnaire, in which occupation was collected as a free-text field following best practices of the National Institute for Occupational Safety and Health.13 Industry was not collected. Occupations were standardized to 6-digit 2010 Standard Occupational Classification (SOC) codes using NIOCCS (version) 4 (NIOSH Industry and Occupation Computerized Coding System)14 and a minimum match probability of 0.90. Employer name was collected and retained in place of industry only to supplement occupation coding in NIOCCS. This enabled inclusion of some records with missing or incomplete occupation, for which an occupation group could be coded with high confidence based on the employer name. Employer name was not coded or analyzed separately as a proxy for industry.
CSTE initially constructed the SOEM to estimate risk among non–health care workers, who were not afforded the same infection control precautions (i.e., personal protective equipment and vaccination) as health and congregate care workers. This analysis thus excluded 929 health care workers and further excluded those in education (n = 579) and corrections (n = 10), given surveillance and vaccination initiatives dedicated to workers in health care and congregate settings during Chicago's COVID-19 response. Other exclusions due to missing or indeterminate occupation or SOEM classification data are summarized in Figure 1 and in supplemental content (Supplemental Digital Content, Supplementary Table 1, http://links.lww.com/JOM/B801). A total of 3763 Chicagoans working in non–health care, noncongregate occupations were included in the final sample. Of these, 74% (n = 2782) had both occupation and employer name available for coding, 20% had occupation only, and 6% were classified by employer name alone.
FIGURE 1.
Cases included in analyses of COVID-19 vaccination status by occupational risk among Chicagoans (n = 3763). *Cases aged 18 to 64 years at time of laboratory-confirmed COVID-19 were included; ages 16 to 17 years were excluded for comparability with existing data sources describing case burden and vaccination coverage. **Among 3763 records included in the final sample, 74% (n = 2782) had both employer name and occupation, 20% (n = 763) had occupation alone, and 6% (n = 218) had employer name only. ***Excluded due to missing exposure classification: “All Other” occupations and military (n = 217), unpaid (n = 34). Abbreviations: CDPH, Chicago Department of Public Health; NIOSH; National Institute of Occupational Safety and Health; NIOCCS, NIOSH Industry and Occupation Computerized Coding System; SOC, standard occupational classification; SOEM, SARS-CoV-2 Occupational Exposure Matrix; NHNCW, non–health care, noncongregate workers.
The CDPH Institutional Review Board determined that this retrospective study was exempt from review: all data were collected previously for public health surveillance purposes and in accordance with 45 CFR Part 46.102(k), 46.102(l)(2) (“Public Health Authority and Surveillance Activities”).
Exposure Definition: SARS-CoV-2 Occupational Exposure
Case data were classified by SOC and level of risk of workplace-acquired infection (low, medium, high) using the first version of the CSTE SOEM risk index and framework (April 2021). This composite risk index incorporates frequency of work that is (1) public facing, (2) indoors, and (3) done in proximity to others. The framework defines two measures of proximity: one based on O*NET (worker-reported) data alone and one including review of O*NET data by occupational health experts. We used the latter, in which employees' self-reported frequencies of proximity to others were reviewed manually, with some revised to more closely represent potential exposure to an infectious agent. Occupations could meet SOEM criteria as “high risk” for SARS-CoV-2 exposure by scoring “high” on at least two of the three environmental parameters. Occupations scored as neither close nor public facing were classified as “low risk,” and all other occupations were defined as “medium risk”. In this analysis, risk was dichotomized for modeling (high vs. low or medium risk), to facilitate implementation of any recommendations that stemmed from findings: high-risk occupations and environments may be more easily distinguished from those of medium or low risk.
Outcome Definition: Vaccination Status
In primary analyses, “vaccination status at infection” referred to the CDC's definitions of full vaccination and a breakthrough case at the time of study15: a second dose of mRNA vaccine (Pfizer or Moderna) or first dose of Johnson & Johnson/Janssen vaccine at least 14 days prior to positive COVID-19 test date. Case patients with no corresponding vaccine information in the patient-level dataset were considered unvaccinated at time of infection for this analysis.
Demographic Variables
Demographic variables (age group, sex, and a combination variable for race-ethnicity) were defined categorically for consistency with CDPH's COVID-19 surveillance data. The “Other, non-Latinx” race-ethnicity category included all who identified as not being of Latinx ethnicity while identifying as more than one race or a race other than White, Black, or Asian alone. These data were derived primarily from I-NEDSS and supplemented with race and ethnicity data from case interviews to decrease missingness. Case interview data improved completeness for 8% of the final sample (307/3763 cases): three-quarters (75%, 307/407) of patients missing race/ethnicity data in I-NEDSS reported some race/ethnicity data during interviews. Age was defined categorically (18–29, 30–49, 50–64 years). These groups were collapsed from those used in citywide coverage estimates,16 both to avoid issues of small cell size and to align with age-based vaccine and treatment recommendations. Zip codes were used to classify cases into the six Healthy Chicago Equity Zones (HCEZ) defined in 2021 to assist in localized public health initiatives.17
Time Periods
To explore differences in association between occupational risk and vaccination status before and after Omicron variant predominance, cases were classified by test date as either Pre-Omicron (through December 14, 2021) or Post-Omicron (December 15, 2021, to May 31, 2022).18
Analytic and Statistical Methods
Demographics, major occupational group, occupational risk, and vaccination status at infection were described overall and by period. Correlations between demographics and occupational risk were evaluated using Cramér's V tests. Distributions of characteristics by vaccination status and demographics by occupational risk level were compared using Pearson's χ2 tests. Logistic regression analyses were conducted separately for the Pre-Omicron and Post-Omicron periods. Associations between occupational risk and vaccination status by age, sex, race-ethnicity, and city region were described in single-factor stratified analyses. All demographic variables were a priori defined confounders in multivariable modeling, and those with differing stratum-specific odds ratios (ORs) were evaluated as effect modifiers. To avoid small cell size, models with single interaction terms were compared to each other and the fully adjusted model of only main effects. The final model was selected based on lowest Akaike Information Criterion (AIC).
Two sensitivity analyses were conducted. The first evaluated selection bias related to exclusion of cases with <0.9 probability occupational match in NIOCCS: demographic and occupational risk distributions of cases coded with 0.80 to 0.89 probability were compared to the final sample (≥0.9 probability). Second, associations between occupational risk and being up-to-date with COVID-19 vaccination (i.e., including receipt of booster doses, if eligible) were also explored among cases in the Post-Omicron period (boosters were recommended for non-immunocompromised working-age Chicagoans effective November 29, 2021).18 Cases were classified as up-to-date according to booster recommendations on the date of positive test and by manufacturer of initial vaccine series and described as (1) unvaccinated, (2) vaccinated and booster-eligible (but not yet boosted), (3) vaccinated and not eligible for boosting, or (4) boosted at time of SARS-CoV-2 infection. These categories were dichotomized (the first two as “not up-to-date,” the others as “up-to-date”) in logistic regression models.
RESULTS
Characteristics of Cases by Period
Characteristics of the final sample are shown in Table 1. From Pre- to Post-Omicron periods, the proportion of cases who were fully vaccinated at infection increased substantially (44% to 75%). Cases with low occupational risk comprised a greater proportion of infections Post-Omicron than Pre-Omicron (32% vs. 27%). Whereas most occupational groups comprised a similar proportion of cases across periods, the proportion in transportation and material moving decreased from Pre- to Post-Omicron (13% to 7%). At the same time, proportions increased slightly among business and financial operations (8% to 11%), legal occupations (2% to 5%), and life, physical, and social sciences (1% to 4%).
TABLE 1.
Demographic and Occupational Characteristics of Non–Health Care, Noncongregate Worker COVID-19 Cases in Chicago, Illinois, June 2021–May 2022 (N = 3763), by Period
| All N = 3763 | Pre-Omicron n = 2455 | Post-Omicron n = 1308 | |
|---|---|---|---|
| Age group (years), n (%) | |||
| 18–29 | 1063 (28.2) | 763 (31.1) | 300 (22.9) |
| 30–49 | 2001 (53.2) | 1277 (52.0) | 724 (55.3) |
| 50–64 | 699 (18.6) | 415 (16.9) | 284 (21.7) |
| Sex, n (%) | |||
| Male | 1910 (50.7) | 1292 (52.6) | 618 (47.2) |
| Female | 1844 (49.0) | 1159 (47.2) | 685 (52.3) |
| Unknown | 9 (0.2) | 4 (0.1) | 5 (0.3) |
| Race-ethnicity group, n (%)* | |||
| Latinx | 848 (22.5) | 546 (22.2) | 302 (23.0) |
| Black, non-Latinx | 950 (25.2) | 726 (29.5) | 224 (17.1) |
| White, non-Latinx | 1493 (39.6) | 911 (37.1) | 582 (44.4) |
| Asian, non-Latinx | 162 (4.3) | 78 (3.1) | 84 (6.4) |
| Other, non-Latinx | 200 (5.3) | 126 (5.1) | 74 (5.6) |
| Unknown | 110 (2.9) | 68 (2.7) | 42 (3.2) |
| City region, n (%) | |||
| North Central | 1054 (28.0) | 596 (24.2) | 458 (35.0) |
| Northwest | 849 (22.5) | 559 (22.7) | 290 (22.1) |
| West | 607 (16.1) | 405 (16.4) | 202 (15.4) |
| Southwest | 506 (13.4) | 361 (14.7) | 145 (11.0) |
| Far South | 399 (10.6) | 299 (12.1) | 100 (7.6) |
| South | 335 (8.9) | 228 (9.2) | 107 (8.1) |
| Unknown | 13 (0.3) | 7 (0.2) | 6 (0.4) |
| Hospitalized, n (%) | |||
| Yes | 108 (2.8) | 89 (3.6) | 19 (1.4) |
| No | 526 (13.9) | 277 (11.2) | 249 (19.0) |
| Unknown | 3129 (83.1) | 2089 (85.0) | (79.5) |
| Deceased, n (%) | 2 (−) | 1 (−) | 1 (−) |
| Fully vaccinated, n (%) | 2089 (55.5) | 1104 (44.9) | 985 (75.3) |
| Occupational risk, n (%) | |||
| High risk | 1991 (52.9) | 1318 (53.6) | 673 (51.4) |
| Medium risk | 684 (18.1) | 466 (18.9) | 218 (16.6) |
| Low risk | 1088 (28.9) | 671 (27.3) | 417 (31.8) |
| Major occupation group, n (%) | |||
| Office and administrative support | 480 (12.7) | 307 (12.5) | 173 (13.2) |
| Management | 446 (11.8) | 278 (11.3) | 168 (12.8) |
| Transportation and material moving | 418 (11.1) | 327 (13.3) | 91 (6.9) |
| Business and financial operations | 348 (9.2) | 192 (7.8) | 156 (11.9) |
| Sales and related | 295 (7.8) | 211 (8.5) | 84 (6.4) |
| Food preparation and serving | 288 (7.6) | 198 (8.0) | 90 (6.8) |
| Protective service | 257 (6.8) | 174 (7.0) | 83 (6.3) |
| Personal care and service | 167 (4.4) | 118 (4.8) | 49 (3.7) |
| Building and grounds cleaning and maintenance | 166 (4.4) | 115 (4.6) | 51 (3.8) |
| Computer and mathematical | 150 (3.9) | 84 (3.4) | 66 (5.0) |
| Construction and extraction | 146 (3.8) | 105 (4.2) | 41 (3.1) |
| Arts, design, entertainment, sports, and media | 131 (3.4) | 87 (3.5) | 44 (3.3) |
| Legal occupations | 120 (3.1) | 59 (2.4) | 61 (4.6) |
| Production | 111 (2.9) | 76 (3.0) | 35 (2.6) |
| Life, physical, and social science | 82 (2.1) | 29 (1.1) | 53 (4.0) |
| Architecture and engineering | 76 (2.0) | 42 (1.7) | 34 (2.5) |
| Installation, maintenance, and repair | 70 (1.8) | 47 (1.9) | 23 (1.7) |
| Educational instruction and library | 7 (0.1) | 2 (−) | 5 (0.3) |
| Farming, fishing, and forestry | 5 (0.1) | 4 (0.1) | 1 (−) |
*For 407 cases with no race/ethnicity in the state surveillance system, 307 (75%) reported some race-ethnicity during interview with the health department; race-ethnicity data derived from interview only thus account for approximately 8% of all cases in the analysis (307/3763). Supplementing the case records with details from case interviews thus improved the completeness of race-ethnicity data and decreased the proportion with unknown race-ethnicity from 407 records (11%) to 110 records (3%).
Vaccination by Occupational Risk and Demographics
Comparisons of vaccination status by occupational risk and demographic groups are shown in Figure 2. Being unvaccinated at infection was more common among high- versus lower-risk workers in the pre-Omicron period only (P < 0.0001). By major occupational group (Supplemental Digital Content, Supplementary Table 2, http://links.lww.com/JOM/B802), pre-Omicron vaccine coverage rates were lowest among workers in protective service (28%); installation, maintenance, and repair (30%); transportation (31%); and construction (31%). Post-Omicron coverage rates were lowest among construction (49%), transportation (57%), and production (66%) occupations. In both periods, vaccination rates were lower among 18- to 29-year-olds and among Latinx, Black, non-Latinx, or Other, non-Latinx race-ethnicity groups than among older, Asian, non-Latinx, and White, non-Latinx groups. Coverage varied widely across regions, from 24% in the Far South to 62% in North Central region Pre-Omicron, and 63% in the Near South to 81% in the North Central region Post-Omicron.
FIGURE 2.
Vaccination rates among non–health care, noncongregate workers with COVID-19 in Chicago, Illinois, June 2021 to May 2022 (N = 3763) by period. Proportions represent rates of being fully vaccinated at time of SARS-CoV-2 infection, according to CDC guidance at time of study. *P < 0.001 (Pre-Omicron only); **P < 0.001 (both periods).
Bivariate Analyses: Occupational Risk by Demographics
Just over half (53%) of all NHNCW cases were in occupations classified as high risk. Demographic distributions of cases reporting high-risk occupations are shown in Figure 3 and differed significantly by age, sex, race-ethnicity, and city region. The youngest and oldest age groups were more likely to work in high-risk occupations: 57% of 18- to 29-year-olds and 57% of 50- to 64-year-olds. Black, non-Latinx workers were most likely to work in high-risk occupations (66%), as were residents of Near South (65%), Far South (62%), and Southwest regions (58%). Major occupational groups represented over both periods are shown in supplemental content (Supplemental Digital Content, Supplementary Fig. 1, http://links.lww.com/JOM/B803), by proportion classified as high risk. Risk among occupations included in Phase 1b/1c vaccine eligibility groups (i.e., frontline and other essential workers) varied dramatically. For example, food preparation and serving and protective service groups consisted entirely of high-risk occupations, whereas <1% of construction and extraction occupations were high risk.
FIGURE 3.
Frequency of high-risk occupations among non–health care, noncongregate workers with COVID-19 in Chicago, Illinois, June 2021 to May 2022 (N = 3763). *P < 0.0001.
Single-Factor Analyses of Vaccination Status by Occupational Risk
Unadjusted models found an association between high occupational risk and being unvaccinated among Pre-Omicron cases only (odds ratio [OR], 1.4; 95% CI, 1.19–1.64) (Supplemental Digital Content, Supplementary Table 3, http://links.lww.com/JOM/B804). Single-factor adjusted Pre-Omicron models suggested confounding by region and race-ethnicity, and effect modification by age, sex, and region. Specifically, workers in high-risk occupations had greater odds of being unvaccinated at time of SARS-CoV-2 infection, compared to those in lower-risk occupations, among 18- to 29-year-olds (adjusted odds ratio [aOR], 2.2; 95% CI, 1.58–2.92), females (1.9 [1.46–2.37]), and those living in the Southwest region (1.8 [1.15–2.78]).
Multivariable Models of Occupational Risk and Vaccination Status
The final multivariable logistic regression model for the Pre-Omicron period included all a priori confounders, with interaction by age (Table 2). Among 18- to 29-year-olds, those in higher-risk environments had statistically significantly greater odds of being unvaccinated at infection (aOR, 1.5; 95% CI, 1.10–2.14). No associations between SOEM occupational risk category and being unvaccinated were found among Post-Omicron cases.
TABLE 2.
Multivariable-Adjusted Models by Period: Odds of Being Unvaccinated at SARS-CoV-2 Infection by Occupational Risk Among Non–Health Care, Noncongregate Workers in Chicago, Illinois, June 2021 to May 2022
| Pre-Omicron | Post-Omicron | ||||||
|---|---|---|---|---|---|---|---|
| OR | (95% CI) | P | OR | (95% CI) | P | ||
| Age group (years) | 0.001 | <0.0001 | |||||
| 18–29 | 3.2 | (2.40–4.17) | 2.5 | (1.69–3.75) | |||
| 30–49 | 1.8 | (1.40–2.30) | 1.3 | (0.91–1.87) | |||
| 50–64 | (ref) | (ref) | |||||
| Sex at birth | 0.01 | 0.39 | |||||
| Female | 0.79 | (0.65–0.94) | 0.9 | (0.68–1.17) | |||
| Male | (ref) | (ref) | |||||
| Race-ethnicity group | <0.0001 | <0.001 | |||||
| Latinx | 2.1 | (1.66–2.70) | 1.7 | (1.16–2.39) | |||
| Black, non-Latinx | 4.7 | (3.59–6.18) | 1.7 | (1.11–2.71) | |||
| Asian, non-Latinx | 0.9 | (0.57–1.55) | 0.9 | (0.46–1.58) | |||
| Other, non-Latinx | 3.6 | (2.37–5.43) | 2.7 | (1.59–4.60) | |||
| White, non-Latinx | (ref) | (ref) | |||||
| City region | <0.0001 | 0.10 | |||||
| Far South | 2.5 | (1.76–3.65) | 2.2 | (1.28–3.84) | |||
| Northwest | 1.4 | (1.07–1.79) | 1.3 | (0.87–1.86) | |||
| South | 1.9 | (1.29–2.87) | 1.2 | (0.65–2.13) | |||
| Southwest | 2.0 | (1.48–2.80) | 1.5 | (0.90–2.35) | |||
| West | 1.4 | (1.07–1.92) | 1.3 | (0.84–1.96) | |||
| North Central | (ref) | (ref) | |||||
| Occupational risk | |||||||
| High* | 0.9 | (0.70–1.21) | 0.55 | ||||
| High, by age group† | 0.029 | ||||||
| 18–29 | 1.5 | (1.10–2.14) | |||||
| 30–49 | 1.1 | (0.84–1.39) | |||||
| 50–64 | 0.7 | (0.48–1.14) | |||||
OR, odds ratio; ref, reference group.
*Pre-Omicron model adjusted for demographics listed (main effect is reported by age group due to interaction observed).
†Post-Omicron model adjusted for all demographic variables listed, main effect is reported as a single measure (no interaction observed).
Sensitivity Analyses
Broadening the inclusion criteria to a 0.8 probability SOC match from NIOCCS (compared to ≥0.9) increased the final sample size by 6% (3763 to 3996) with no difference in proportions excluded by occupational group (Supplemental Digital Content, Supplementary Fig. 2, http://links.lww.com/JOM/B805). Cases with a 0.8 to 0.89 match were comparable to those in the final sample by demographics, occupational risk, and vaccination status (Supplemental Digital Content, Supplementary Table 4, http://links.lww.com/JOM/B806). When cases were reclassified following booster recommendations (Supplemental Digital Content, Supplementary Table 5, http://links.lww.com/JOM/B807; Supplementary Fig. 3, http://links.lww.com/JOM/B808), low-risk workers were most likely to be up-to-date. No associations were found between occupational risk and not being up-to-date, including when adjusting for demographics. Model selection yielded results like those in primary Post-Omicron analyses (aOR, 1.1 [95% CI, 0.86–1.40] in fully adjusted model with no interaction).
DISCUSSION
This analysis included novel descriptions of (1) COVID-19 cases by occupational risk and (2) associations between demographics and occupational risk, to help assess the utility of workplace-based vaccine outreach. Our findings indicate that coverage disparities among high-risk occupational groups can be explained, in part, by vaccine hesitancy among demographic groups overrepresented in these jobs. The observed relationships between race-ethnicity, age, and occupation can be leveraged during immunization campaign planning: workplace-based initiatives focused on higher-risk workers can supplement community-based initiatives serving the same groups (e.g., younger and Black, non-Latinx workers). We observed decreased coverage disparities from Pre- to Post-Omicron periods. These are consistent with trends in citywide coverage rates by race-ethnicity19 and may be partially attributable to hyperlocal outreach efforts conducted in lowest-coverage HCEZ from December 2021 to July 2022.20
The major occupation groups with lowest coverage in our study were also among those most likely to be vaccine hesitant in a large Pre-Omicron survey of working-age Americans (April 20, 2021, through May 19, 2021; N = 338,226).21 King et al.21 found that workers in construction and extraction (35%); installation, maintenance, and repair (29%); protective service (26%); and transportation and material moving (22%) were most likely to say they would “definitely not” be vaccinated against COVID-19. Vaccine skepticism and fear of side effects were the most cited reasons across these groups, followed by perceptions of low COVID-19 risk (e.g., due to working outside). This speaks to the difference between describing occupations as high-risk for SARS-COV-2 exposure versus critical infrastructure, as echoed in our findings: not all groups defined as critical infrastructure during the COVID-19 pandemic22 were classified as high risk by the SOEM.
Overrepresentation of racial and ethnic minority groups in precarious work environments is well documented in occupational health studies,23 and here we bolster the evidence specific to risk of respiratory disease transmission. We also suggest that associations between workplace characteristics and individual-related health attributes can be exploited to protect vulnerable workers and, by extension, their communities.23 In their analyses of the nationally representative American Communities Survey and O*NET data,24 Goldman et al. compared demographics of high-risk versus lower-risk occupations. Occupational standing (i.e., “proportion of workers with at least some college education”) was used as a proxy for educational attainment and use of preventive measures. Racial and ethnic minority groups were more likely to have a higher risk for workplace-acquired COVID-19 and also more likely to have lower occupational standing. These findings emphasize the value of workplace-based initiatives in educating and protecting workers who may otherwise lack knowledge or resources to practice preventive health.
Limitations
A lack of industry and occupation data in COVID-19 laboratory and vaccination datasets was a major limitation to these analyses. Descriptions of industry (among workers with COVID-19) would have enabled analyses of occupational risk and vaccination coverage stratified by industry-based vaccine eligibility phases. Although case investigations provide a rich data source describing Chicagoans who have contracted COVID-19, they underrepresent workers with natural or vaccine-induced immunity, who are less likely to be included in a case-based sample. This was especially true Pre-Omicron, when vaccine was more protective against infection. Moreover, restriction to those who practiced health-seeking behaviors (laboratory-based testing, case interview) likely overrepresented residents of more affluent city regions.
As shown in Figure 1, only a small fraction of all cases complete interviews with the health department; reliance on interview data for occupation details further limits the generalizability of our sample. Other analyses of COVID-19 case investigation data in Chicago indicate that case interviews became less representative of all working-age cases over our study period. Possible explanations are multifactorial and include changing case investigation practices, health department capacity, and patient compliance.25 Including only cases with complete interview and occupation data likely had similar effects: underrepresenting younger workers and those of race-ethnicity groups other than White, non-Latinx. This was especially true Post-Omicron (Supplemental Digital Content, Supplementary Table 6, http://links.lww.com/JOM/B809). Although our study includes only laboratory and provider-reported cases, the overrepresentation of business, financial, and legal workers in the Post-Omicron period was consistent with an increased proportion of such workplaces directly reporting clusters and outbreaks to CDPH in the same period.25
Because I-CARE excludes data from federal vaccine registries and out-of-state vaccinations, workers vaccinated through federal efforts (e.g., protective service) or those who elected to travel for vaccination may have been misclassified as unvaccinated at time of infection. This could have overestimated or underestimated coverage disparities by occupation or other demographics. The number of such workers is unknown. Because Illinois providers may have updated I-CARE on their behalf, it is possible that they are classified correctly in our analyses. Sensitivity analyses could not account for earlier booster eligibility based on individual comorbidities and thus may have overestimated proportions of workers who were not up-to-date.
Occupational risk may have been overestimated in this analysis, as we could not identify remote workers or those in modified work environments (e.g., with masking, increased distancing) during the pandemic. At the time of this publication, however, many workplaces have reduced pandemic-era controls (such as remote work and distancing), and isolation guidance for COVID-19 has been relaxed,26 suggesting that occupational risk frameworks are increasingly relevant to workplace infection control.
Future studies should leverage the most up-to-date SOEM for inclusion of more occupations; more recent versions27 include 45 non–health care occupations that were missing risk measures in the framework we utilized. These comprised a small proportion of our sample (n = 217 or 2% of all cases with available occupation data), so we expect that our overall findings would have been similar with the new framework. Also, our sensitivity analyses suggest that a threshold probability criterion of 0.9 could be lowered, to increase sample size without introducing selection bias. This should be replicated in other analyses. As utilized in other studies,8 a lower threshold match, coupled with manual review (e.g., 0.5 and manual review of cases with <0.5 probability) could have permitted a larger sample with more precise classification. However, this approach must be weighed against the resource intensity of manual review.
Although coverage rates in this report cannot be extrapolated to present-day seasonal vaccines, current data indicate very poor uptake of seasonal COVID-19 vaccines, both locally and nationally.28 In Chicago, the proportion of Chicagoans who are up-to-date on COVID-19 vaccines as of July 2024 is estimated to be low across all working-age groups (from 9% among 18- to 29-year-olds to 20% among 50- to 64-year-olds). Disparities between racial and ethnic groups persist, with rates among Black, non-Latinx, and Latinx Chicagoans (both 11%) less than half those of White, non-Latinx Chicagoans (23%).29 Therefore, the “two-for-one” benefit of workplace-based vaccine outreach is still much needed and valuable for protecting both vulnerable worker groups and their networks.
Although not without limitations, our approach enabled exploration of coverage disparities among occupational groups in Chicago. It also illustrated a need for public health data modernization: until occupation and industry are routine elements of immunization records, analyses of vaccination disparities among workers will continue to be a challenge.
CONCLUSIONS
Because low-coverage demographic groups are overrepresented in higher-risk occupations, workplace-based COVID-19 vaccine promotion can help protect undervaccinated communities. After adjusting for race, ethnicity, and age, we found no differences between odds of being unvaccinated among Chicagoans in high-risk versus lower-risk occupations. To enable (1) identification of workers with low coverage rates and (2) evaluation of job-related vaccination efforts, immunization records must include industry and occupation data.
ACKNOWLEDGMENTS
Jennifer Cornell, Li Liu, Stacey Marovich, Peter Ruestow, Bethlehem Solomon, Mary Turyk, and Case Investigation teams at the Chicago Department of Public Health.
Footnotes
Ethical considerations and disclosures: None.
Funding sources: None to disclose.
Lendacki, Forst, Mehta, and Kerins have no relationships/conditions/circumstances that present potential conflict of interest.
The JOEM editorial board and planners have no financial interest related to this research.
Author contributions: F.L. designed the study, conducted analyses, and wrote the manuscript with support from L.F., S.D.M., J.K., S.D.M., and J.K. supervised the project.
Data availability: Datasets generated during this study are not publicly available, as these contain information that can potentially be used to identify patients. Deidentified data are available from the corresponding author on reasonable request.
No AI was utilized at any stage of research development, design, data collection, analysis, or manuscript preparation.
Supplemental digital contents are available for this article. Direct URL citation appears in the printed text and is provided in the HTML and PDF versions of this article on the journal’s Web site (www.joem.org).
Contributor Information
Linda Forst, Email: lforst@uic.edu.
Supriya D. Mehta, Email: supriyad@uic.edu.
Janna L. Kerins, Email: Janna.Kerins@cityofchicago.org.
REFERENCES
- 1.Dyal JW. COVID-19 among workers in meat and poultry processing facilities ― 19 states, April 2020. MMWR Morb Mortal Wkly Rep 2020;69. doi: 10.15585/mmwr.mm6918e3. [DOI] [PubMed] [Google Scholar]
- 2.Waltenburg MA. Update: COVID-19 among workers in meat and poultry processing facilities ― United States, April–May 2020. MMWR Morb Mortal Wkly Rep 2020;69. doi: 10.15585/mmwr.mm6927e2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Waltenburg MA Rose CE Victoroff T, et al. Coronavirus disease among workers in food processing, food manufacturing, and agriculture workplaces. Emerg Infect Dis 2021;27:7–249. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.McClung N Chamberland M Kinlaw K, et al. The advisory committee on immunization practices' ethical principles for allocating initial supplies of COVID-19 vaccine — United States, 2020. MMWR Morb Mortal Wkly Rep 2020;69:1782–1786. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Dooling K Marin M Wallace M, et al. The advisory committee on immunization practices' updated interim recommendation for allocation of COVID-19 vaccine — United States, December 2020. 2021;69:4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Council of State and Territorial Epidemiologists, Occupational Health Subcommittee . Characterizing the risk of exposure to SARS-CoV-2 Among non–health care occupations based on three workplace risk factors: public facing work, working indoors, and working in close physical proximity to others. August 18, 2021. Available at: https://resources.cste.org/soem-documents/Combined%20SOEM%20Report%20-%208.18.21%20Final.pdf. Accessed March 4, 2022.
- 7.US Department of Labor . About O*NET at O*NET Resource Center. Available at: https://www.onetcenter.org/overview.html. Accessed November 5, 2021.
- 8.Pray IW Grajewski B Morris C, et al. Measuring work-related risk of coronavirus disease 2019 (COVID-19): comparison of COVID-19 incidence by occupation and industry—Wisconsin, September 2020 to May 2021. Clin Infect Dis 2023;76:e163–e171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Hawkins D. Differential occupational risk for COVID-19 and other infection exposure according to race and ethnicity. Am J Ind Med 2020;63:817–820. doi: 10.1002/ajim.23145. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Barry V Dasgupta S Weller DL, et al. Patterns in COVID-19 vaccination coverage, by social vulnerability and urbanicity — United States, December 14, 2020–May 1, 2021. MMWR Morb Mortal Wkly Rep 2021;70:818–824. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Kriss JL. COVID-19 vaccination coverage, by race and ethnicity — National Immunization Survey Adult COVID module, United States, December 2020–November 2021. MMWR Morb Mortal Wkly Rep 2022;71:757–763. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Centers for Disease Control and Prevention . COVID-19 Data Tracker. March 28, 2020. Available at: https://covid.cdc.gov/covid-data-tracker. Accessed November 12, 2021.
- 13.National Institute of Occupational Safety and Health . Collecting and Using Industry and Occupation Data. September 8, 2023. Available at: https://www.cdc.gov/niosh/topics/coding/collect.html. Accessed October 3, 2024.
- 14.National Institute of Occupational Safety and Health . NIOSH Industry and Occupation Computerized Coding System (NIOCCS) — NIOSH. Available at: https://csams.cdc.gov/nioccs/Default.aspx. Accessed November 5, 2021. [DOI] [PMC free article] [PubMed]
- 15.Centers for Disease Control and Prevention . COVID-19 Vaccination. January 9, 2023. Available at: https://www.cdc.gov/coronavirus/2019-ncov/vaccines/stay-up-to-date.html. Accessed January 17, 2023. [PubMed]
- 16.City of Chicago . Latest Data. Available at: https://www.chicago.gov/content/city/en/sites/covid-19/home/covid-dashboard.html. Accessed October 4, 2021.
- 17.City of Chicago . Healthy Chicago Equity Zones. Available at: https://www.chicago.gov/content/city/en/sites/covid-19/home/healthy-chicago-equity-zones.html. Accessed February 16, 2022.
- 18.City of Chicago . COVID-19 Vaccine Boosters. COVID-19 Vaccine Boosters. Available at: https://www.chicago.gov/content/city/en/sites/covid19-vaccine/home/boosters.html. Accessed January 17, 2023.
- 19.City of Chicago . COVID-19 Vaccine Coverage Trends. COVID-19 Vaccine Coverage Trends. Available at: https://www.chicago.gov/content/city/en/sites/covid-19/home/covid-dashboard/covid19-vaccine-coverage-trends.html. Accessed October 3, 2024.
- 20.City of Chicago . Healthy Chicago Equity Zones (HCEZ) One-Pager 04.25.24. April 25, 2024. Available at: https://www.chicago.gov/content/dam/city/depts/cdph/community-health/hcez/downloads/HCEZ-One-Pager_04.25.24.pdf. Accessed October 3, 2024.
- 21.King WC, Rubinstein M, Reinhart A, Mejia R. COVID-19 vaccine hesitancy January–May 2021 among 18–64 year old US adults by employment and occupation. Prev Med Rep 2021;24:101569. doi: 10.1016/j.pmedr.2021.101569. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Cybersecurity & Infrastructure Security Agency . Identifying Critical Infrastructure During COVID-19. Available at: https://www.cisa.gov/identifying-critical-infrastructure-during-covid-19. Accessed November 5, 2021.
- 23.Murray LR. Sick and tired of being sick and tired: scientific evidence, methods, and research implications for racial and ethnic disparities in occupational health. Am J Public Health 2003;93:221–226. doi: 10.2105/AJPH.93.2.221. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Goldman N, Pebley AR, Lee K, Andrasfay T, Pratt B. Racial and ethnic differentials in COVID-19-related job exposures by occupational standing in the US. PloS One 2021;16:e0256085. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Lendacki FR, Forst L, Weber E, Mehta SD, Kerins JL. COVID-19 clusters and outbreaks among non–health care, noncongregate workers in Chicago, Illinois: surveillance through the first omicron wave. J Occup Environ Med 2023;65:e211. doi: 10.1097/JOM.0000000000002782. [DOI] [PubMed] [Google Scholar]
- 26.Centers for Disease Control and Prevention . Respiratory virus guidance. Respiratory Virus Guidance March 1, 2024. Available at: https://www.cdc.gov/respiratory-viruses/guidance/index.html. Accessed July 3, 2024.
- 27.Council of State and Territorial Epidemiologists, Occupational Health Subcommittee . OH Subcommittee Publication: SARS-CoV-2 Occupational Exposure Matrix-Industry Extension (SOEM-IE): an approach to using occupational information to identify industries at high risk of exposure to SARS-CoV-2. March 4, 2022. Available at: https://www.cste.org/page/OHPublications. Accessed October 3, 2024.
- 28.Centers for Disease Control and Prevention . COVID-19 Vaccination Coverage and Vaccine Confidence Among Adults. Available at: https://www.cdc.gov/vaccines/imz-managers/coverage/covidvaxview/interactive/adults.html. Accessed July 3, 2024.
- 29.City of Chicago . COVID-19 Vaccine Coverage. COVID-19 Vaccine Coverage. Available at: https://www.chicago.gov/city/en/sites/covid-19/home/covid-19-vaccine-coverage.html. Accessed July 3, 2024.