health insurance, typically provided by an employer, is ''the dominant mechanism for paying for health services'' in the United. States.1(p79) According to the ...
RESEARCH AND PRACTICE
Health Care Coverage and the Health Care Industry Chiu-Fang Chou, DrPH, Pamela Jo Johnson, PhD, MPH, Andrew Ward, PhD, MPH, and Lynn A. Blewett, PhD
For complex socioeconomic reasons, private health insurance, typically provided by an employer, is ‘‘the dominant mechanism for paying for health services’’ in the United States.1(p79) According to the Kaiser Commission on Medicaid and the Uninsured and the Urban Institute, analyses of data from the Current Population Survey (CPS) show that, in 2006, 54% of the US civilian, noninstitutionalized population had employer-sponsored health insurance; 5% had private, nongroup health insurance; and 26% had public health insurance coverage. Approximately 46 million US residents (16% of the population) are currently uninsured.2 Numerous studies have shown that, relative to people with health insurance, uninsured people receive less preventive care, are diagnosed at more advanced disease stages, and, once diagnosed, tend to receive less therapeutic care and have higher mortality rates.3–8 Although national uninsurance trends are well-documented, the rate of uninsurance within the health care workforce has received scant attention. Given that health care employment rates are increasing at a more rapid pace than overall employment rates, this lack of attention is especially worrisome. According to the Bureau of Labor Statistics, nearly half of the 30 occupations in which employment opportunities are growing fastest are health care occupations. For example, whereas the Bureau of Labor Statistics projects that overall employment will increase about 10% from 2006 to 2016, employment opportunities for personal and home care aides are projected to increase nearly 51%, and opportunities for physical therapist assistants are expected to increase by a third. The Bureau of Labor Statistics also projects that, by 2016, new job opportunities for registered nurses will increase by approximately 24% (approximately 587 000 new jobs).9 Although the overall employment outlook for health care workers is promising, what is less clear is to what degree employment in
Objectives. We examined rates of uninsurance among workers in the US health care workforce by health care industry subtype and workforce category. Methods. We used 2004 to 2006 National Health Interview Survey data to assess health insurance coverage rates. Multivariate logistic regression analyses were conducted to estimate the odds of uninsurance among health care workers by industry subtype. Results. Overall, 11% of the US health care workforce is uninsured. Ambulatory care workers were 3.1 times as likely as hospital workers (95% confidence interval [CI] = 2.3, 4.3) to be uninsured, and residential care workers were 4.3 times as likely to be uninsured (95% CI = 3.0, 6.1). Health service workers had 50% greater odds of being uninsured relative to workers in health diagnosing and treating occupations (odds ratio [OR] = 1.5; 95% CI = 1.0, 2.4). Conclusions. Because uninsurance leads to delays in seeking care, fewer prevention visits, and poorer health status, the fact that nearly 1 in 8 health care workers lacks insurance coverage is cause for concern. (Am J Public Health. 2009;99:2282–2288. doi:10.2105/AJPH.2008.152413)
health care is associated with health insurance coverage. A 2001 General Accounting Office report suggested that one fourth of nursing home aides and one third of home health care aides were uninsured.10 The Kaiser Family Foundation reported that the uninsured rate among workers in the health and social services industry was 23% in 2007.11 On the basis of a review of the literature in the health and human services occupations, Ebenstein concluded that the health insurance plans offered to direct care workers in the developmental disabilities field are ‘‘inferior . . . with less coverage and more out-of-pocket expenses’’ and that fewer direct care workers ‘‘are able to afford health coverage even if they are eligible.’’12(p132) Taking a more comprehensive look at the US health care workforce, Himmelstein and Woolhandler13 used 1991 CPS data to estimate uninsurance rates among physicians and other health care personnel. They reported that, overall, 9% of health care workers were uninsured, along with more than 20% of nursing home workers. Examining CPS data from 1988 to 1998, Case et al. found that uninsurance rates among all health care workers rose from 8% to 12%, that rates increased more for health care workers than for workers in other industries, and that rates
2282 | Research and Practice | Peer Reviewed | Chou et al.
differed according to occupation and place of employment.14 For example, occupation-specific uninsurance rates were 23.8% among health aides, 14.5% among licensed practical nurses, and 5% among registered nurses, whereas placespecific rates were 20% among nursing home workers, 8.7% among medical office workers, and 8.2% among hospital workers.15 In their studies, Himmelstein and Woolhandler13 and Case et al.14 used national-level data to estimate uninsurance trends among health care workers. However, these trends were not adjusted for health care workers’ social, demographic, or economic characteristics, which would have helped explain variation across categories or over time. Moreover, with the growth of the health care workforce, estimates from these older studies probably do not reflect the current situation. As a result, the picture of uninsurance as it pertains to the health care workforce lacks the precision and currentness necessary for sound policy decisions. In an effort to expand knowledge in this area, produce more up-todate estimates, and provide support for possible policy decisions, we used data from the National Health Interview Survey (NHIS) to examine uninsurance among workers in the health care industry.
American Journal of Public Health | December 2009, Vol 99, No. 12
RESEARCH AND PRACTICE
METHODS We assessed the uninsurance status of US health care workers by health care industry subtype and workforce category using NHIS data from 2004 to 2006. With the exception of industry and occupation variables, all of the data used in our study were retrieved from the Integrated Health Interview Series, a Webbased data resource containing harmonized NHIS data from 1969 to the present.15,16 The industry and occupation variables were retrieved from the original NHIS public use files and linked with the Integrated Health Interview Series variables. Our sample consisted of adults who were aged 20 to 64 years and who reported being employed in the health care industry during the week preceding their interview. We excluded respondents who did not report their occupation or industry and those with missing data with respect to full-time employment status. Data from 2004 to 2006 were pooled to obtain a sufficient sample size of health care workers. Sampling weights were adjusted to account for the pooling of 3 years of survey data. The final, unweighted sample comprised 5192 adults employed in the health care industry.
Measures Data on all measures were obtained through respondents’ self-reports during the in-person NHIS survey. The outcome of interest was health insurance coverage status, dichotomized as ‘‘insured’’ (covered) or ‘‘uninsured’’ (not covered). Respondents were considered insured if they reported being covered by private health insurance, military health insurance, Medicare, Medicaid, the State Children’s Health Insurance Program, or another state-sponsored health plan. To identify health care workers, we used data on the respondents’ self-reported main occupation during the week prior to the interview. We classified health care workforce occupations using the 3 categories identified in the NHIS data: health diagnosing and treating occupations (e.g., physicians, nurses), health technicians (e.g., clinical laboratory technologists, licensed practical nurses), and health service workers (e.g., home health aides, medical assistants). We
included a residual category for all other workers (e.g., clerks, secretaries). We used the 3 industry subtypes identified in the NHIS data—ambulatory health care services, hospitals, and nursing and residential care facilities—to subdivide the health care industry. We included a number of additional covariates in our analyses. Self-reported race and ethnicity were combined to create categories representing non-Hispanic Asian, non-Hispanic Black, Hispanic, non-Hispanic White, and other non-Hispanic (which included American Indian/Alaska Native, multiracial, and any other designation not included in the preceding categories). Age was categorized into 3 groups: 20 to 34 years, 35 to 49 years, and 50 to 64 years. Educational attainment (college degree versus no college degree), marital status (married versus not married), and full-time employment status (either working or not working 35 hours or more a week) were dichotomized. Total income in the preceding year was categorized into 3 groups: less than $20 000, $20 000 to $44 999, and greater than $45 000.
Data Analysis We examined the extent to which employees in the 3 health care industry subtypes differed in terms of sociodemographic background characteristics potentially associated with insurance coverage; in these analyses, we used cross tabulations and design-based F tests to account for the complex sample design. We used multivariate logistic regression to estimate the odds of uninsurance for health care industry subtypes and workforce categories adjusted for age, race/ ethnicity, gender, marital status, education, employment status, and total income. Finally, in the case of each of the 3 health care industry subtypes, we estimated the odds of uninsurance for health care workforce categories with separate logistic regression models adjusted for all of the covariates just mentioned. We used Stata statistical software, which produces unbiased estimates from data collected through complex sampling designs such as the one used in the NHIS, in conducting all of our analyses.17,18 We
December 2009, Vol 99, No. 12 | American Journal of Public Health
produced variance estimates using Taylor series linearization.
RESULTS Table 1 describes the characteristics of the study population by health care industry subtype. The majority of health care workers were women (78%), White (70%), married (62%), and college educated (74%), and a majority had incomes between $20 000 and $44 999 in the year preceding the survey (42%). Of those employed in the health care industry, 43% worked in ambulatory care, 40% worked in hospitals, and 17% worked in residential care. Approximately 29% of health care workers were in health diagnosing and treating occupations, 12% were health technicians, and 19% were health service workers; nearly 41% were employed in other health care occupations. The majority (78%) of health care workers were employed full time. As can be seen in Table 1, the distribution of workforce categories differed according to health care industry subtype. Health service workers were primarily employed in residential care (34%) and ambulatory care (20%) settings, whereas workers in the health diagnosing and treating occupations were primarily employed in hospitals (41%). Table 1 also shows that the background characteristics of health care workers differed significantly across health care industry subtypes. Figure 1 shows uninsurance rates by employment status (full or part time) for the 3 health care industry subtypes. In hospital settings, the difference in uninsurance rates among full-time and part-time workers was nominal. However, there were significant differences between part-time and full-time workers in ambulatory and residential care settings, with uninsurance rates being 6 percentage points higher for part-time ambulatory workers and 7 percentage points higher for part-time residential care workers than for fulltime workers. Figure 2 presents uninsurance rates by health care workforce category and health care industry subtype between 2004 and 2006. The rate of uninsurance among all workers aged 20 to 64 years in the health care industry
Chou et al. | Peer Reviewed | Research and Practice | 2283
RESEARCH AND PRACTICE
TABLE 1—Characteristics of Workers in the US Health Care Industry, by Health Care Industry Subtype: National Health Interview Survey, 2004–2006 Total (n = 5192), % (SE)
Ambulatory Care (n = 2209), % (SE)
Hospital (n = 2038), % (SE)
Residential Care (n = 945), % (SE)
Insurance status Insured
89.0 (0.52)
87.5 (0.8)
95.8 (0.49)
76.9 (1.61)
Uninsured
11.0 (0.52)
12.5 (0.8)
4.2 (0.49)
23.1 (1.61)
Ambulatory care
43.1 (0.86)
...
...
...
Hospital
39.8 (0.82)
...
...
...
Residential care
17.1 (0.61)
...
...
...
Health diagnosing and treating Health technician
29.2 (0.73) 11.5 (0.51)
25.1 (1.13) 11.1 (0.77)
40.9 (1.32) 13.6 (0.9)
12.3 (1.45) 8.0 (1.14)
Health service worker
18.5 (0.63)
19.8 (1.0)
10.6 (0.83)
33.8 (1.83)
Other
40.7 (0.77)
44.0 (1.19)
34.9 (1.18)
45.9 (1.91)
Part time
23.4 (0.72)
25.5 (1.03)
20.6 (1.06)
24.7 (1.75)
Full time
76.6 (0.72)
74.5 (1.03)
79.4 (1.06)
75.3 (1.75)
20–34 35–49
31.9 (0.86) 41.8 (0.81)
33.0 (1.26) 42.5 (1.25)
30.5 (1.2) 42.3 (1.22)
32.1 (1.89) 38.8 (1.91)
50–64
26.4 (0.7)
24.5 (1.06)
27.2 (1.09)
29.1 (1.81)
Women
78.1 (0.7)
76.4 (1.08)
77.0 (1.12)
84.7 (1.48)
Men
21.9 (0.7)
23.6 (1.08)
23.0 (1.12)
15.3 (1.48)
37.6 (0.76)
35.0 (1.2)
36.8 (1.18)
46.0 (2.01)
62.4 (0.76)
65.0 (1.2)
63.2 (1.18)
54.0 (2.01)
< College
25.6 (0.69)
22.7 (1.01)
19.0 (1.07)
48.5 (2.0)
College
74.4 (0.69)
77.3 (1.01)
81.0 (1.07)
51.5 (2.0)
Asian, non-Hispanic
5.0 (0.38)
4.5 (0.63)
5.9 (0.58)
4.2 (0.89)
Black, non-Hispanic
15.8 (0.61)
11.7 (0.77)
16.0 (0.91)
25.7 (1.71)
8.5 (0.41)
10.4 (0.7)
6.8 (0.56)
7.7 (0.98)
0.5 (0.06) 70.2 (0.78)
0.5 (0.07) 72.8 (1.12)
0.6 (0.12) 70.7 (1.15)
0.2 (0.02) 62.1 (2.0)
< 20 000
28.0 (0.79)
28.2 (1.21)
20.1 (1.07)
45.9 (1.98)
20 000–44 999
42.4 (0.84)
41.2 (1.3)
43.3 (1.26)
43.4 (1.95)
> 45 000
29.6 (0.79)
30.6 (1.26)
36.5 (1.29)
10.7 (1.36)
P
