Employment is associated with independence and improved quality of life. In the United States, employment rates among people with disabilities have not changed substantially despite the Americans with Disabilities Act (ADA) and the Ticket to Work program [5,6]*. People with disabilities are twice as likely as people without disabilities to live below the government established poverty line [5,7–9]. Employment rates for people with substantial disabilities remain low, ranging from 15 to 44%, depending on the population studied [8–12].

Barriers to employment for people with disabilities include lack of transportation and environmental factors, such as the built environment, attitudes, and social practices [13,14]. Additional barriers include federal income support programs, such as Supplemental Security Income and Social Security Disability Insurance [8,15] and pervasive poverty [8]. Health-related factors may compromise the capacity to work, may force people with disabilities out of the workforce or may limit quality of life. These factors include adverse health behaviors [16,17]; chronic conditions, including secondary conditions [18–20]; lack of mental health care [21,22]; the lack of health promotion for people with disabilities [16,17,23,24]; significant out-of-pocket expenses, including lack of or inadequate health insurance coverage [12]; the lack of assistive technology [25] and repair [26]; and lack of access to care by knowledgeable providers [26].

Understanding the complex experience of people with disabilities is further complicated by definitional issues regarding what constitutes a disability [27]. Efforts to define disability recognize the importance of the person-environment interaction, suggesting that the built, policy, and attitudinal environment can serve as a barrier or facilitator enhancing or limiting the ability of people with disabilities to pursue social roles [28].The introduction of the International Classification of Functioning, Disability and Health (ICF) has codified the dimensional experience of disability [29]. Discussions of disability, however, offen treat the experience as a threshold concept – suggesting that people reach a tipping point where they become “disabled.” Dichotomous notions of disability fail to recognize variability of function. Severity of disability may be captured by concepts of work disability or eligibility for certain social support program, including Supplemental Security Income or Social Security Disability Income [15]. Much research compares people with and without disabilities in relation to health and social participation outcomes, including work, without considering severity of disability.

A recent examination of the Behavioral Risk Factor Surveillance System in the US identified differing health behaviors among three groups of people with disabilities (those reporting assistive device use and activity limitation, assistive device use only, and activity limitation only) compared with people with no disability [4].The investigators then modeled the interaction of work/nonwork to disability severity. The chief finding was that nonworkers with more severe disabilities reported a greater likelihood of physical inactivity. Our investigation advances this line of inquiry.

Rather than comparing people with and without disabilities, the aim of this study was to compare workers and nonworkers reporting any functional limitation by severity of limitation (mild, moderate or severe/complete) on several health-related and social characteristics to identify disparities in 19 different outcome indicators. These indicators include physical health; level of psychological distress; disability income and health insurance coverage; activities of daily living (ADLs), instrumental activities of daily living (IADLs), and use of special equipment; occurrence of comorbid chronic conditions; and health behavior practices. Since workers are generally healthier than nonworkers [30] and adults with lower levels of disability tend to be healthier than those with more severe disabilities [31], we attempted to pinpoint patterns of disparity in specific health-related and participation areas among workers and nonworkers with varying levels of functional limitations. Identification of these patterns of disparity might identify appropriate points of intervention for health promotion and workforce retention. To accomplish this goal, we asked the following research questions: In what areas of health and participation do workers and nonworkers with functional limitations differ most? How does level of limitation severity affect differences in health and participation among workers and nonworkers with functional limitations?

Our conceptual framework for this study was based on the ICF, a taxonomy developed by the World Health Organization (WHO) [30] as a companion tool to the International Classification of Diseases (ICD) [32]. The ICF promotes understanding of functioning, disability, and health across cultures by describing these concepts as health-related domains associated with body structures (s-codes), body functions (b-codes), activity limitations and participation (d-codes), and environmental factors (e-codes).

Our data source for this study was the Sample Adult and Person files from the 2000–2008 National Health Interview Survey (NHIS), a random sample survey [33]. Since 1997, the NHIS Core has included a Sample Adult Core module. An adult aged 18 years or older is selected from each surveyed family to participate as a Sample Adult. The Sample Adult Core contains specific information on commonly occurring chronic conditions and health behavior practices that is not found in other survey components.

We retained all records in the 9-year database of the combined survey files and excluded missing responses only in cases where respondents refused to answer the question or did not know the answer to the question. To ensure parity across data files over the 9-year period, we developed a crosswalk (grid) in which we entered all variables and their values. In some cases, recoding was necessary to achieve parity across files and make certain that variables and their values measured the same constructs across years. After our master database was in place and we identified survey questions of interest associated with targeted variables that were relevant to our research questions, we backcoded NHIS questions from the 2000–2008 surveys to the ICF taxonomy for all independent and dependent measures for which codes were available. This backcoding process helped us link our conceptual framework to our research questions and analytic procedures. We used SPSS 14.0 for our data management and SUDAAN 10.0 for our data analysis [34,35].

Our 9-year sample included 54,775 working-age adults with all levels of functional limitations. In this group, 22,908 respondents were aged between 18 and 44 years and 31,867 were between the ages of 45 and 64 years. Our sample included 20,619 males, 34,156 females, and included 18,581 respondents from minority respondents (African-Americans, Latinos/Hispanics, Asians, Native Americans, Pacific Islanders, and other races), and 36,194 non-minority whites (Caucasians). Our sample contained 5,501 military veterans. A total of 1,429 veterans had functional limitations severe enough (moderate, severe/complete) to be considered veterans with disabilities. A total of 14,150 individuals, regardless of military status, had functional limitations severe enough to be classified as disabilities, while 40,625 working-age respondents had mild functional limitations. Workers numbered 29,207, while 25,568 respondents with functional limitations did not currently work. Workers with moderate, severe/complete functional limitations numbered 3,051, while nonworkers with these functional limitations equaled 11,099.

Our population of workers and nonworkers with functional limitations is described in Table I.

Sixty-six percent of respondents (approximately 11.3 million people) with mild functional limitations but no disabilities were currently working, compared with 30.8% of people with moderate functional limitations and 9.5% of individuals with severe/complete functional limitations.Thus, an estimated 3,290,000 working-age adults with limitations severe enough to be classified as disabilities were actively engaged in the workforce, while an estimated 3.7 million adults with substantial disabilities were not working. An estimated 5.8 million working-age adults with mild functional limitations reported that they were not currently working.

Demographic characteristics of our working and nonworking adults are displayed in Table II.

Respondents in both work categories across functional limitation groups were more likely to be female. Minorities in all limitation groups were less likely to be working, especially those with moderate or severe/complete limitations. Workers with severe/complete limitations were more likely to have a high school education or less, as were their nonworking counterparts. Almost two-thirds (64.0%) of respondents having severe/complete limitations who did not work said they had no college training. About two-fifths (38.9%) of workers with severe/complete limitations reported their income was less than $20,000 annually, compared with half (50.1%) of their nonworking counterparts. Across limitation groups for both workers and nonworkers, respondents were more likely to be unmarried, especially workers with severe/complete limitations (61.2%). Nonworkers with moderate limitations were the most likely group to have served in the military. Respondents in all limitation categories across work categories were more likely to live in the South than in any other region. Nearly, one-half (46.2%) of nonworkers with severe/complete limitations resided in the South. Fewer workers and nonworkers with all levels of limitations lived in the Northeast. These results seem to be due to the over-all pattern of responses to the survey, but we could not determine whether pattern of response to the survey was the sole reason for this demographic pattern.

Outcomes for health and social participation factors are shown in Table III.

Having fair/poor health was strongly associated with limitation severity among workers and nonworkers, though more strongly so among nonworkers. Workers and nonworkers with severe/complete limitations were more likely to report having fair/poor health. About half of all workers in this limitation category rated their health as fair/poor (53.8%, adjusted odds ratio (AOR) = 6.04), compared with three-fourths of their nonworking counterparts (77.7%, AOR = 19.85).

Patterns of psychological distress in either workers or non-workers were not as straightforward as were patterns for fair/poor health. Psychological distress for all levels of limitation severity was more common among nonworkers than workers. Nonworkers with mild limitations were the most likely to report mild psychological distress (55.2%), while nonworkers with severe/complete limitations were more likely to indicate moderate/severe psychological distress (47.7%, AOR = 3.72). Among workers, individuals with severe/complete limitations were more likely to say they experienced mild psychological distress (57.9%, AOR = 1.83), and respondents with moderate limitations were more likely to say they experienced moderate/severe psychological distress (31.7%, AOR = 1.43).

We examined disability income as possible disincentive to work for people with functional limitations. A small number of workers with functional limitations reported receiving disability income, but most recipients were nonworkers, for whom receipt of disability income was strongly associated with limitations severity. Receipt of disability income ranged from 16.4% for people with mild limitations, 42.2% (AOR = 3.57) for people with moderate limitations, and 51.7% (AOR = 5.32) for nonworkers with severe/complete limitations. These findings should be viewed with caution, because at least one response category had less than 100 respondents per cell.

Findings on health insurance coverage were mixed. Workers with mild limitations were more likely to have health insurance coverage (84.0%, AOR = 1.30), while 79.0% (AOR = 0.97) of workers with moderate limitations had coverage, and 83.9% (AOR = 1.28) of workers with severe/complete limitations had health insurance. Among nonworkers, health insurance coverage was associated with disability severity. More than three-fourths (78.4%) of nonworkers with mild limitations had coverage, while 84% (AOR = 1.61) of moderately limited nonworkers and 86.3% (AOR = 1.93) of nonworkers with severe/complete limitations were covered. Notably, more than 16% of workers with severe/complete limitations and almost 14% of their nonworking counterparts reported having no health insurance coverage.

Very few workers with all levels of functional limitations had difficulties with ADLs, and almost one-fourth (24.1%, AOR = 5.95) of respondents with severe/complete limitations indicated difficulty with IADLs. Use of special equipment among workers was associated with severity of functional limitations, with workers who had severe/complete limitations having the highest usage (43.2%, AOR = 13.61).

Nonworker respondents in all functional limitation categories were more likely than worker respondents to report having difficulty with ADLs and IADLs and use of special equipment, and nonworkers with severe/complete functional limitations were more likely than nonworkers in other functional limitation categories to report difficulty in these areas. Almost one-quarter (24.5%, AOR = 18.87) of nonworkers with severe/complete limitations reported they experienced difficulties with ADLs, while more than two-fifths (44.5%, AOR = 14.08) of this group had difficulties with IADLs, and more than half (54.8%, AOR = 23.73) said they used some type of special equipment.

For both workers and nonworkers, hypertension was associated with level of functional limitation severity, with the strongest associations among nonworkers. This was also true for diabetes, heart problems, and breathing problems.

In regard to respondent-related conditions that were not diagnosed by a physician, workers were more likely than nonworkers to report having joint symptoms and low back pain, though respondents in both groups who had greater limitation severity were more likely to report having these conditions than were respondents in other functional limitation categories. Reports of hearing loss were mixed. Workers with mild limitations were more likely than their nonworking counterparts to report having hearing loss (19.1% v. 15.9%, AOR = 1.09), while nonworkers in the other limitation categories reported more hearing loss than their working counterparts. Nonworkers across all limitation groups were slightly more likely to report having vision loss than were workers.

Smoking was more likely among nonworkers with functional limitations, especially among individuals with moderate limitations (37.7%, AOR = 1.26). Workers were more likely to use alcohol weekly than nonworkers, but alcohol usage decreased as level of limitation severity increased for both workers and nonworkers with limitations. Almost three-fourths of workers with mild limitations (72.7%, AOR = 1.76) reported weekly alcohol use, compared with 59.6% of workers with moderate limitations and 48.1% of workers with severe/complete limitations. Patterns of overweight (25 ≤ BMI < 30 kg/m²), but not obesity, decreased with increasing level of disability severity, with workers having mild limitations being more likely than any other group to be overweight, but not obese (32.6%, AOR = 1.10). Workers were slightly more likely than nonworkers to be obese (BMI ≥ 30), with more than half of workers with severe/complete limitations reporting obesity (57.0%, AOR = 2.21). Conversely, nonworkers at all limitation levels were more likely to be physically inactive in their leisure time, with nearly four-fifths of nonworkers with severe/complete limitations reporting no regular leisure-time exercise (79.2%, AOR = 5.76).