In the United States, having health insurance is strongly associated with better self-reported health status [1–3], and gaining healthcare coverage has generally been found to improve healthcare access and outcomes [2, 4, 5]. Examples of specific improvements range from earlier diagnosis and control of certain chronic physical and mental health conditions [6–8] to lower mortality [9, 10], although results differ by research design and timing [11] and by insurance type [12–14].

Health insurance coverage among U.S. adults of working age differs by demographic characteristics. For 2023, the U.S. Census Bureau reported 10.9% uninsurance (having no health insurance) among all adults aged 19–64 and 10.0% among working adults in that age group [15]. The report found that Hispanic adults were most likely to be uninsured (23.6%), followed by Black adults (11.1%), non-Hispanic White adults (7.0%), and Asian adults (7.0%). Coverage also differed by marital status, with uninsurance most likely among adults who were separated (18.1%), followed by those who had never married (14.7%), and lowest among those who were married (7.9%). Other research has found that among uninsured adults aged 18–64, lack of affordability was the most common primary reason for not having health insurance, cited by 63.2% of respondents [16].

Among working adults, health insurance coverage differs not only by demographic characteristics but also by employment situation (employed vs. self-employed) and occupation. Analyses have found self-employed adults to be less likely to be insured than adults employed for wages [17–19]. By major occupational group, analyses of employed, self-employed, and recently unemployed (< 1 year) respondents to the 2022–2023 Behavioral Risk Factor Surveillance System (BRFSS) observed that more than 20% of workers were unemployed in four occupations: Farming, Fishing, and Forestry (28.8%); Building and Grounds Cleaning and Maintenance (27.2%); Construction and Extraction (24.7%); and Food Preparation and Serving (22.8%) [19]. Census data for working adults aged 19–64 in 2024 showed that uninsurance was most common in these same four groups [20], even though previous research found that farming, construction, and service were among occupations with the largest drops in uninsurance following implementation of the Patient Protection and Affordable Care Act (ACA) [21].

Among those with health insurance, healthcare access and utilization (often measured as unmet healthcare needs), as well as health status and outcomes, may vary by source of insurance. Insurance sources have often been divided into two broad categories: public and private. Private insurance historically encompassed employer-sponsored insurance (through a job or labor union, referred to as ESI) and non-group private insurance (individually purchased from an insurance broker), while public insurance included Medicare (for people aged 65 or older, and younger people with certain long-term disabilities) and Medicaid (for low-income individuals). However, both the health insurance landscape and categorization of insurance sources have become more complex since the ACA was introduced and expanded both public and private programs. These included, for example, Medicaid expansion and subsidized ACA Marketplace coverage (involving private purchase of coverage from an insurance broker or through federal or state insurance marketplace state-sponsored programs, and other programs with government involvement) [22].

According to 2023 US Census estimates, which capture coverage of an individual by multiple types of insurance, 54% of adults had ESI, with Medicaid the next most common at 15.9% [15]. As with insured status, the distribution of insurance sources differs by demographic characteristics. Among workers aged 19–64, 12.6% were classified by the US Census as having public insurance (Medicare, Medicaid, VA, or Champus [now TriCare]) and 80.1%, private insurance (employment-based, private, or through state or federal marketplace); this compared with 42.8% public and 46.7% private among adults in this age range who did not work at least 1 week [15].

Studies of the distribution of health insurance source by occupation have been limited. Limited data on access to ESI are available; one such study identified three broad occupational groups in which more than 40% of workers could not access ESI through their jobs: Farm, Fishing and Forestry; Construction; and Service [23]. Less attention has been given to the full distribution of workers’ health insurance sources by occupation, as well as healthcare access and outcomes. In the current research, we examined associations between demographic characteristics, occupation, health insurance status, primary source of health insurance, healthcare access, and self-reported health status.

BRFSS is conducted annually by all states and three U.S. territories to survey noninstitutionalized, U.S. civilian residents aged ≥ 18 years. The BRFSS core survey, administered to all respondents, includes questions about sociodemographic characteristics (including employment status), health behaviors, health conditions, and use of health-related services. BRFSS also includes topical modules that jurisdictions can opt to administer. The Industry and Occupation (I&O) optional module is for respondents who reported on the survey core that they were currently employed for wages (henceforth, “employed”), self-employed, or out of work for < 1 year. To ascertain occupation, respondents are asked, “What kind of work do you do, for example, registered nurse, janitor, cashier, auto mechanic?” Free-text responses are recorded and then coded by the National Institute for Occupational Safety and Health’s NIOSH Industry and Occupation Computerized Coding System (NIOCCS; https://csams.cdc.gov/nioccs/) into one of 22 two-digit standard occupational classification major groups (https://www.bls.gov/soc/2010/2010_major_groups.htm; specific occupational components of each major group are found at the same website). The I&O module was administered by 40 states, Guam, and the U.S. Virgin Islands for at least 1 year in 2022–2023. For 2022, median U.S. BRFSS response rates across jurisdictions were 46.3% for landline and 44.7% for cell phone users [24]. In 2023, median response rates were 54.3% for landline and 40.5% for cellphone users [25].

The BRFSS core survey includes a question about the primary source of insurance. In 2022, the question was “What is the current primary source of your health insurance?” and in 2023, respondents were asked “What is the current source of your primary insurance?” Response categories are employment (“A plan purchased through an employer or union [including plans purchased through another person’s employer]”); private (“A private nongovernmental plan that you or another family member buys on your own”); Medicare; Medigap; Medicaid; Children’s Health Insurance Program (CHIP); Military-related health care (“TRICARE (CHAMPUS)/VA health care/CHAMPVA”); Indian Health Service (IHS); state sponsored health plan; other government program; and no coverage of any type. For the “employment or union” category, we use the accepted term ESI, although we note that this category also includes insurance from labor unions. We refer to responses of “private” as “non-group private insurance.” For the purposes of these analyses, we combined Medigap with Medicare and CHIP with Medicaid. While enrolled members of Native American tribes are eligible for healthcare coverage under the IHS, the IHS is not a health insurance program, and participants with no other documented health insurance coverage are considered uninsured by the U.S. Census [15]; for comparability, we have followed this practice and grouped respondents reporting IHS as primary source with the uninsured. Because, with the exception of employment-based and non-group private health insurance, most health insurance sources have a governmental component, in this research, we describe results for the grouped BRFSS categories, rather than sorting into “private” and “public” categories.

BRFSS does not ascertain whether the respondent or another family member is the policyholder of this insurance and does not ask about non-primary health insurance sources. Thus, for example, some self-employed respondents may report having ESI although they are not “employees.” These respondents are likely a mix of business owners, those with ESI through family members, and respondents continuing insurance from previous employment for wages through temporary extensions via the Consolidated Omnibus Budget Reconciliation Act (COBRA) [26].

In addition to the question about primary insurance source, BRFSS has three questions about healthcare access. The first concerns whether the person has a primary provider: “Do you have one person (or a group of doctors) that you think of as your personal health care provider?” Answers of “no” were coded as positive for “no personal doctor.” Another question asked, “Was there a time in the past 12 months when you needed to see a doctor but could not because you could not afford it?” Positive answers were coded as positive for “could not afford needed care,” while noting that affording needed care has additional components. The third question was “About how long has it been since you last visited a doctor for a routine checkup?” with cutpoints of less than 1 year, 1–2 years, 2–5 years, and more than 5 years. Responses of 1 year or more were coded as positive for “no checkup past year.” We used each of these variables to measure the prevalence of lacking healthcare access (which we also refer to as unmet healthcare need).

BRFSS also has self-reported health status measures. General health is ascertained by “Would you say that in general your health is—Excellent, Very Good, Good, Fair, Poor?” We grouped answers of “fair” and “poor” into “fair or poor general health.” Two questions, one for physical health and one for mental health, ask the number of days of the last 30 during which the person’s health was not good. The physical health question asks “Now thinking about your physical health, which includes physical illness and injury, for how many days during the past 30 days was your physical health not good?” and the mental health question, “Now thinking about your mental health, which includes stress, depression, and problems with emotions, for how many days during the past 30 days was your mental health not good?” For each item, in line with the literature in this area [27], we dichotomized responses so that respondents reporting their health was not good for at least 14 days were classified as having “frequent physical distress” or “frequent mental distress.” All health status measures were self-reported, with no reference to diagnosis by a healthcare professional.

Study criteria include an age restriction of 18–64 years to focus on the sources of health insurance adults are likely to have access to during most years of their working lives (i.e., not Medicare, which is primarily age-based). We included respondents who indicated they were currently employed for wages (henceforth, “employed”) or self-employed; we excluded the unemployed to minimize recent health insurance changes due to employment status transitions. Because BRFSS is not representative of active-duty military personnel, respondents indicating this status were also excluded; however, some remaining respondents may have military-based insurance through a family member or as veterans currently working in civilian occupations.

Demographic characteristics of interest were sex, age group, race and Hispanic or Latino (henceforth, “Hispanic”) ethnicity, education level, marital status, and household income category. Sex was not missing for any respondents. Age and race/Hispanic ethnicity were imputed by the BRFSS program for 1.2% of respondents and 2.5% of respondents, respectively. While only 0.2% and 0.6% of respondents were missing educational attainment and marital status, respectively, almost 13% did not report household income. Because BRFSS measures household income rather than individual wages and had many missing observations, we also reported mean annual wage data from the U.S. Department of Labor for each occupational group. A total of 5821 respondents were missing information on primary insurance type; of these, 3985 reported that they did not know their primary insurance type, while 1836 declined to answer or were not asked the question. Respondents missing primary insurance source were included in calculations of the population mean but were not assigned a category for insurance coverage or source. Respondents missing information for other variables were retained but excluded from analyses involving those variables.

During 2022–2023, 283,760 BRFSS respondents indicated they were employed or self-employed. After exclusions for age, active-duty military, and lack of codable occupation, we retained 203,780 participants in the analytic dataset.

We calculated weighted prevalences and 95% confidence intervals (CIs) to evaluate several associations: (1) primary source of health insurance with healthcare access and health status; (2) demographic characteristics and employment situation with primary source of health insurance; (3) demographic characteristics with healthcare access and health status; (4) major occupation group with primary source of health insurance. We present weighted, unadjusted prevalences compared to the population mean prevalence to assess statistical significance, considering prevalence estimates with 95% CIs not overlapping those of the population mean to have statistically significant differences.

We also present a set of analyses more deeply exploring associations between major occupation groups and both healthcare access and health status. These include unadjusted prevalence, prevalence ratios (PRs), and adjusted prevalence ratios (aPRs) comparing healthcare access and health status measures in the occupation under study with those in all other occupations combined. We present both unadjusted PRs and three sets of adjusted PRs. The first adjustment accounts for sex (male, female); age (18–24, 25–34, 35–50, and 60–64); race/Hispanic ethnicity (non-Hispanic Black, non-Hispanic White, Hispanic, non-Hispanic other race or multiracial), and marital status (married, divorced, widowed, separated, never married, member of an unmarried couple). The second adjustment retains these and adds an adjustment for primary insurance source (ESI, military, non-group private, Medicare [including Medigap], state-sponsored, other governmental, Medicaid, and uninsured [including IHS]). The final adjustment retains everything included in the second adjustment and adds household income.

We used SAS version 9.4 (SAS Institute, Carey, N.C.) and SAS-callable SUDAAN version 11.0.1 (RTI International, Research Triangle Park) to account for the complex survey design. We used the SURVEYFREQ procedure to estimate population counts and weighted unadjusted prevalences and 95% CIs. We used the RLOGISTIC procedure (SAS-callable SUDAAN version 11.01.1, RTI International, Research Triangle Park) to perform logistic regression and estimate aPRs comparing the adjusted prevalence of each dependent variable for each major occupation to the adjusted prevalence for workers in all other occupations (except the one being compared) combined. Hence, the aPR compares the prevalence for workers in each group to the expectation based on the prevalence for all other workers combined, after adjustment. We considered aPRs with the lower 95% CI above 1.0 to indicate statistically significant prevalence elevations and aPRs with the upper 95% CI below 1.0 to be in a statistically significant deficit (i.e., below expectation). In the online supplemental materials (Supporting Information S1: Appendix A), we present this last set of results as prevalence figures with the same series of adjustments but compared to the population means.

This activity was reviewed by CDC, deemed not human subjects research, and was conducted consistent with applicable federal law (45 C.F.R. part 46.102(l)(2), 21 C.F.R. part 56; 42 U.S.C. Sect. 241(d); 5 U.S.C. Sect. 552a; 44 U.S.C. Sect. 3501 et seq.) and CDC policy.

Of the 203,780 respondents included in our analyses, 62.0% were covered by ESI (Table 1). The next most common sources of health insurance were non-group private (10.1%) and Medicaid (6.9%). State-sponsored, other governmental, and military-based insurance covered 3.3%, 2.9%, and 2.1% of workers, respectively. A total of 9.9% of respondents indicated that they were uninsured (including those enrolled in the IHS).

In our study of employed and self-employed 2022–2023 BRFSS respondents, being insured was associated with better healthcare access and health status. By insurance source, workers from occupations with high prevalences of ESI or non-group private insurance (generally high-wage occupations) reported the best access and health status. In contrast, workers from occupations with a high prevalence of uninsurance (generally low-wage occupations) had high prevalence of unmet healthcare need and adverse health status.

Despite working, more than 1 in 10 employed or self-employed adults in our study could not afford needed care, 1 in 5 had no personal doctor, and more than 1 in 4 reported no routine checkup in the past year. Among uninsured workers, more than one in three reported not being able to afford needed medical care, more than 60% had not had a routine checkup in the past year, and more than 60% did not have a personal doctor. Other research has shown that low-wage workers are less likely to have paid sick leave; the need to take unpaid time off to visit a healthcare provider is yet another economic barrier to care [28, 29].

Employment situation, wages, demographic characteristics, and occupation were associated with the prevalence of insurance and primary insurance source, as well as healthcare access and self-reported health status; these associations were intertwined. For example, healthcare access varied by occupation and was associated with the mean annual wages for an occupation. The prevalence of ESI was higher for occupations with higher mean annual wages, as was the ability to afford needed care. Other research has found that employee costs for both health insurance premiums and other out-of-pocket healthcare costs have been increasing [23]. Given our findings that the inability to afford needed care was highest among workers in occupations with the lowest mean household income, these cost increases could further decrease the ability of lower-wage workers to access healthcare, even if their employers offer ESI. The ability to afford needed care was uneven, even in occupations with more than half of workers reporting having ESI: in addition to occupations with high levels of uninsurance, several occupation groups with intermediate coverage levels and annual wages near $50,000 also had significantly elevated prevalences for being unable to afford needed care. These include Transportation and Material Moving, Sales and Related, Healthcare Support, and Office and Administrative Support. The decrease in these elevations in almost all occupations after adjustment for household income highlights the role of adequate wages, in addition to healthcare coverage, in healthcare access.

In our study, 13%–16% of workers with Medicare, Medicaid, state-sponsored, or other governmental insurance reported not being able to afford needed care in 2022–2023. The differing implementation of Medicaid expansions by state, with variations in income cutoffs for eligibility, coverage, and patient cost-sharing [30], may contribute to this finding. Given that these insurance sources are most commonly reported by low-wage workers and provide coverage for more than 50% of respondents with mean household incomes below $20,000, increases in healthcare costs, including higher premiums, would likely lead to a higher prevalence of being unable to afford needed care among these workers.

A number of occupations with high prevalence of uninsurance also had elevated PRs for fair or poor general health and, for several occupations, frequent physical distress. Adjustment for demographic covariates (to account for demographic sorting into occupations) led to aPRs for unmet health need and health status in these groups that were lower but remained significantly elevated. Several of the industries (Agriculture, Fishing, and Forestry; Transportation and Warehousing; and Healthcare and Social Assistance) employing the most workers in these occupations have some of the highest nonfatal occupational injury and illness incidence rates [31]. Access to comprehensive, affordable healthcare is important for ensuring these workers can obtain needed care in order to remain employed in these occupations, particularly in Agriculture, which is exempt from workers’ compensation coverage in some states [32]. The results of further adjustment for primary insurance source, which yielded lower but still statistically significantly elevated aPRs for unmet healthcare need, suggest that health insurance alone does not guarantee adequate financial means to utilize healthcare. The final adjustment for household income reduced the number of occupations with statistically significant elevations for unmet healthcare need, particularly for being unable to afford needed care. Other research has shown that being unable to afford care, particularly copays for prescriptions, can lead to strategies such as insulin rationing [33], highlighting the potentially serious effects of being uninsured or being unable to cover all healthcare costs, even if insured. The same adjustments reduced aPRs for fair or poor health in occupations with high levels of uninsurance.

In our study of health insurance, occupation, healthcare access, and health status, differences between uninsured workers and the insured, particularly those with ESI or non-group private insurance, show clear associations between health insurance coverage and the health status of U.S. workers. Longitudinal studies are needed to assess the causality of the associations observed in this study between occupation and both healthcare access and health status. Positive results would suggest that future disruptions to healthcare coverage could lead to overall declines in healthcare access and the health status of workers. This trajectory would contrast with results observed for Medicaid expansion, including associations with improved access to care [35], better awareness and control of chronic conditions [8], and increased early-stage diagnosis rates for cancer and lower rates for some causes of mortality [16].

Increases in the proportion of workers unemployed or working in occupations less likely to offer ESI or wages adequate for workers to pay premiums and other healthcare costs, along with changes in the health insurance landscape that would reduce workers’ access to Medicaid and other programs with a public element, could combine to leave more workers uninsured, with potentially deleterious effects on healthcare access and the health status of workers. The impact could be particularly pronounced on workers in two types of occupations: (1) those that already have a high percentage of workers uninsured and on Medicaid (Farming, Fishing, and Forestry; Building and Grounds Cleaning and Maintenance; Food Preparation and Serving); and (2) those with an intermediate percentage of uninsured workers but many workers reliant on Medicaid, state-sponsored programs, other governmental insurance, and Medicare. An example of the latter is Healthcare Support, one of the occupations expected to grow most quickly between 2023 and 2033 [36] as the U.S. population ages. Increasing premiums could lead to substantially lower levels of worker enrollment in ESI and Marketplace programs, and potentially higher levels of unmet healthcare needs and poor health status, than those seen in these analyses.

In this study of U.S. workers, we found large differences by occupation in the prevalence of healthcare coverage through ESI, uninsurance, healthcare access, and self-reported health status. Our results highlight the importance of both comprehensive health insurance and wages adequate to purchase and utilize healthcare coverage for the health and employability of all workers.

Additional supporting information can be found online in the Supporting Information section.

Supplemental Appendix A – Healthcare Access and Self-Reported Health Status by Broad Occupation (sorted by % uninsured), Behavioral Risk Factor Surveillance System 2022–2023*. Distribution of selected demographic characteristics by broad occupation groups, Behavioral Risk Factor Surveillance System 2022–2023*.