Objective. Although one third of young adults in the United States are uninsured, lack of insurance in this age group has been the subject of few published studies. Because opportunities to obtain public and private insurance are likely to differ for men and women, the objective of this study was to describe the gender-specific relationship of sociodemographic variables and lack of insurance among young adults.
Methods. We examined data for 6884 young adults (aged 19–24 years) who completed the Sample Adult Questionnaire of the National Health Interview Survey for 1998, 1999, and 2000. Gender-stratified multiple logistic regression was used to estimate the odds of being uninsured associated with race/ethnicity, household income, major activity in the previous week, marital status, and pregnancy (women).
Results. Overall, 32% of male participants and 27% of female participants reported being uninsured at the time of the survey. Uninsured men outnumbered insured men in several sociodemographic categories, including Hispanic men (58% uninsured), men not attending high school (85%), and men employed in a workplace that did not offer health insurance (51%). High rates of uninsurance were reported by women not attending high school (65%), Hispanic women (46%), those who were keeping house (41%), and women with a household income between $10 000 and $20 000 (41%). In multiple logistic regression models, many of the sociodemographic variables studied were similarly correlated with health insurance for both men and women. Employment in a workplace where the young adult was not offered health insurance coverage, low household income, low educational attainment, and Hispanic ethnicity were associated with increased odds of being uninsured for both genders. Having attended college, higher household income, and being a student or employed in a workplace that offers health insurance coverage were associated with lower odds of being uninsured for both genders.
Conclusion. This study suggests that additional opportunities for health insurance coverage are needed for young adults—particularly men, Hispanics, and those in low- and middle-income households. Increasing the availability of employment-based health insurance, discouraging attrition from primary and secondary education, and the creation of insurance opportunities for minorities and near-poor and middle-income households are potentially important target areas for programs that seek to reduce the number of uninsured young adults.
One third of young adults aged 19 to 24 are uninsured, making this age group nearly twice as likely to be uninsured as children or older adults under 65 years of age.1,2 Exclusion from a parent’s family health insurance plan or government plan because of age, employment in part-time or entry-level positions without health benefits, inability to afford health insurance independently, and/or lower priority placed on obtaining insurance have been hypothesized by others as contributors to the high rates of uninsurance among young adults.3,4 Medicaid and the State Children’s Health Insurance Program have provided public health insurance coverage for an increasing number of children and adolescents.5 Beyond age 18, Medicaid and State Children’s Health Insurance Program make provision for adults who are severely impaired and for those who are pregnant or parents and meet income guidelines. Other than these special circumstances, there is little availability of health insurance for young adults not covered privately. Surprising is that uninsurance among young adults has been the focus of only a few studies in the medical literature.4,6,7
Lack of health insurance is likely to present a significant barrier to health care access for young adults. Studies of adolescents and adults have implicated lack of health insurance coverage as a reason for missing needed care and forgoing routine care.8–12 Although perceived as healthy, young adults are at high risk for acute health problems (eg, sexually transmitted infections, intentional and unintentional injuries).13–15 Furthermore, because of improved diagnosis and management, an increasing number of children with previously fatal chronic medical conditions are now surviving to adulthood.16 Thus, the problem of uninsured young adults is of special relevance to pediatric providers—particularly those who continue to provide care for young adult patients or attempt to help them make the transition to adult-oriented health care.
It is consistently reported that young male adults are more likely to be uninsured than young female adults. Still, previous nationally representative studies have examined uninsurance among young adults as a single group. The role of gender is a potentially important consideration because many of the sociodemographic factors thought to influence insurance status differ for men and women. For example, the US Census Bureau reported that for the year 2000, young male adults (aged 20–24 years) were more likely than young female adults to be employed (83% vs 73%) and less likely ever to have been married (16% vs 27%).17 In addition, insurance opportunities may differ by gender: a nondisabled young female adult with limited income may be insured through Medicaid because of pregnancy or motherhood, whereas a young male adult of similar economic status may not qualify. To describe better the gender-specific impact of sociodemographic variables on being uninsured among young adults, we analyzed data from the National Health Interview Survey (NHIS), a nationally representative health survey.
We examined data from young adults who were aged 19 to 24 years and completed the NHIS for 1998, 1999, and 2000. The NHIS is an ongoing cross-sectional national household survey conducted by the US Census Bureau and the Centers for Disease Control and Prevention to collect information on the health status and use of health services by the US civilian noninstitutionalized population.18 The sampling plan uses a multistage area probability design that permits the representative sampling of households. Weights are provided to adjust the sample data to reflect the age, gender, and race distribution of the US population. Using formulas provided by NHIS, we calculated the response rate for all adults selected for the Sample Adult questionnaire during the years 1998, 1999, and 2000 to be 71.8%.18
The NHIS consists of a group of surveys that collect data on households and individuals. Data are released for public use in several data files. Data for this study were drawn from the person-level file (containing sociodemographic and basic health information for each member of a household collected directly from individuals or from other adult members of the household if the individual is not at home or is unable to answer) and from the sample adult file (containing additional health, health care utilization, and behavior data obtained from 1 randomly selected adult in each household).
Main Outcome Measure
The main outcome measure was health insurance coverage at the time of the interview. Consistent with the definition used by the National Center for Health Statistics and the US Census Bureau, the uninsured were people who reported having no coverage under private insurance, Medicare, Medicaid, Children’s Health Insurance Program, a state-sponsored health plan, other government programs, or military health plan. Respondents with only Indian Health Service coverage were also considered to be uninsured.2,19
The NHIS provided recoded responses to questionnaire items regarding educational attainment, marital status, and household income. The education and household income recodes were derived by collapsing responses to the questionnaire items. The marital status recode was derived by combining responses to >1 questionnaire item. After examining frequencies and bivariate comparisons of insured and uninsured young adults in each of the recoded variable categories, we further collapsed the response categories to preserve adequate cell sizes. We collapsed categories to reflect whether the young adult’s highest level of educational attainment was ≤8th grade, 9th through 12th grade (no diploma), high school graduate, general educational development (GED) recipient, or some college (including those with and without a degree). Marital status was recoded to single (including those who were widowed, divorced, or never married), married (including those who were separated), and unmarried but living with a partner. Information provided about household income from all sources was collapsed into the following categories: <$10 000, $10 000 to $19 999, $20 000 to $34 999, $35 000 to $54 999, $55 000 to $74 999, and $75 000 or over. Respondents were asked whether they were working at a job or business in the previous week. Those who were not working were asked why they were not working. From this information, young adults were classified according to the major activity in which they were involved in the week before the interview: working at a job or business, going to school, keeping house, or something else. Those who were working at a job or business were further categorized according to whether they reported that health insurance was offered to them through the workplace. Male participants who reported keeping house were included in the “something else” category.
Comparisons were made between insured and uninsured participants using χ2 analysis. Multiple logistic regression models were analyzed to estimate the adjusted odds of being uninsured associated with gender, race/ethnicity, household income, level of education, major activity in the previous week, and marital status. Gender-stratified multiple logistic regression models were created to estimate the adjusted odds of being uninsured associated with race/ethnicity, household income, level of education, major activity in the previous week, marital status, and, for women, current pregnancy. Final analyses were conducted using SUDAAN20 to account for the complex sampling design and clustering of data. The study protocol was approved by the Institutional Review Board of Vanderbilt University (Nashville, TN).
The total number of 19- to 24-year-old participants was 8751. Half of the weighted sample was female, and approximately two thirds were white non-Hispanic (Table 1). Of the total sample, 6884 participants (3154 men, 3730 women) provided complete information (79%) and were included in regression analyses. Almost all excluded cases were attributable to missing or incomplete information regarding household income. Compared with participants for whom income data were available, those who did not provide income data were more likely to be uninsured, to be a member of a racial or ethnic minority, and to have lower educational attainment. In the weighted sample, 32% of male participants and 27% of female participants reported being uninsured at the time of the survey. In multivariate analyses, men were found to have significantly higher adjusted odds of being uninsured (adjusted odds ratio [AOR]: 1.46; 95% confidence interval [CI]: 1.25–1.70) than women (reference group). On the basis of these findings, all subsequent analyses were stratified by gender.
In bivariate analyses, race/ethnicity, educational attainment, major activity in the last week, marital status, and household income were each significantly associated with insurance status for both genders, and pregnancy was associated with insurance status for women (Table 2). Uninsured men outnumbered insured men in several sociodemographic categories, including Hispanic men (58% uninsured) and men who had not received a high school diploma (85% of those with <9th grade education, 55% of men with a 9th- through 12th-grade education, and 54% of men with a GED). More than half of young male adults who were working at a job that did not offer health insurance to them (51%) or men who reported doing “something else” in the previous week (62%) were likewise uninsured. Least likely to report being uninsured were men who were going to school (12%) or who were employed in a workplace that offered health insurance coverage to them (15%). Men in the highest income brackets and those who had attended some college also had lower rates of uninsurance.
Although women generally were less likely to report being uninsured than men, high rates of uninsurance were reported by young female adults with ≤8th-grade education (65%), Hispanic women (46%), women with a 9th- through 12th-grade education but no diploma (41%), those who were keeping house in the previous week (41%), and those with a household income between $10 000 and $20 000 (41%). The lowest rates of uninsurance were reported by women in the highest household income category (12%), those who were working at a job that offered health insurance (13%), pregnant women (16%), women who had attended college (18%), and those who were going to school in the previous week (18%).
Multiple Logistic Regression
Among young male adults, major activity in the previous week, highest level of educational attainment, and household income were the strongest predictors of being uninsured (Table 3). Compared with men who were going to school in the previous week, men who reported working at a job that did not offer health insurance (AOR: 7.01; 95% CI: 4.55–10.80) and men who reported doing “something else” had 7 times the adjusted odds of being uninsured. In contrast, young male adults who were working at a job that offered health insurance coverage had similar odds of being uninsured as men who were going to school.
An inverse relationship was observed between the highest level of educational attainment and the odds of being uninsured among young male adults. Men with ≤8th-grade education had nearly 4 times the adjusted odds of being uninsured (AOR: 3.95; 95% CI: 2.07–7.55) compared with high school graduates (reference group). Men who had attended college had the lowest adjusted odds of being uninsured (AOR: 0.45; 95% CI: 0.35–0.58). Although considered educationally equivalent, men with a GED had significantly higher odds of being uninsured than those who graduated from high school in the traditional sense. Men who were living in households with incomes between $10 000 and 34 999 had more than twice the adjusted odds of being uninsured as those in the highest income. Unlike the findings with education, men in the lowest income category did not have significantly different adjusted odds of being uninsured as those in the highest income category.
Compared with white non-Hispanic young male adults, men in each of the other race/ethnicity categories had higher adjusted odds of being uninsured. Hispanic men (AOR: 2.05; 95% CI: 1.60–2.64) and men in the “other” category (AOR: 2.19; 95% CI: 1.29–3.71) had the highest odds of being uninsured.
In the multiple logistic regression model for women, significantly higher adjusted odds of being uninsured were seen for women with household income between $10 000 and $19 999 (AOR: 3.90; 95% CI: 2.36–6.45) and $20 000 and $34 999 (AOR: 2.62; 95% CI: 1.53–4.47) compared with women in the highest income category. Women who were employed in a workplace that did not offer health insurance to them had more than twice the odds of being uninsured (AOR: 2.37; 95% CI: 1.60–3.50) as women who were going to school. In contrast, young female adults who were employed in a workplace that offers health insurance had significantly lower odds of being uninsured (AOR: 0.59; 95% CI: 0.39–0.89) than current students. Similar to men, Hispanic women had twice the odds of being uninsured as white non-Hispanic women. Women who were currently pregnant had significantly lower odds of being uninsured (AOR: 0.28; 95% CI: 0.18–0.42) than nonpregnant women.
Marriage was not a significant predictor of insurance status for either gender model. However, young male adults who were unmarried but living with a partner had higher adjusted odds of being uninsured (AOR: 1.47; 95% CI: 1.02–2.12) than single men.
This article makes a new contribution to the literature by considering the association of sociodemographic and insurance status among young male and female adults separately. In addition to finding increased odds of uninsurance among young male adults, we identify groups that are at higher risk of being uninsured for both genders. Documenting these associations may help programs that seek to reduce the rates of uninsured young adults to target more effectively services to discrete segments of the population. This study also adds to our understanding of changes in risk of uninsurance that occur as adolescents make the transition to adulthood. In a study that used 1995 NHIS data, Newacheck et al9 reported that 14% of adolescents aged 10 to 18 were uninsured and that the highest odds of uninsurance were among older adolescents, Hispanics, and adolescents below 200% of the poverty level. Newacheck found no significant gender difference in the odds of being uninsured among adolescents. In our study, the rate of uninsurance among young adults was more than twice that of the adolescents in Newacheck’s study, with a significant difference in uninsurance rates between men and women. Similar to Newacheck’s study of adolescents, we found that Hispanic young adults and those in low to middle income levels had higher adjusted odds of being uninsured.
Young adults of both genders were less likely to be uninsured when they were students. The lower rate of uninsurance among students serves as an example of the positive impact that special provisions and incentives may have on the problem of uninsurance. Full-time students are frequently required by their institution or state to carry health insurance coverage. To facilitate this requirement, private insurers frequently allow full-time students to remain covered through a parent’s family plan into their early 20s. In addition, many educational institutions offer school-sponsored health insurance to those who are not covered privately or through a parent’s plan.
Similar to students, low rates of uninsurance were also reported by young adult workers who were offered health insurance through their place of employment. After adjustment, young adults of both genders had similar odds of uninsurance as young adult students. In contrast, men in workplaces that do not offer health insurance to them had 7 times the odds of being uninsured as male students, and women had more than twice the odds of being uninsured. Among young adults who reported working at a job in the previous week, almost half of men and more than half of women reported that they were not offered health insurance coverage through the workplace. A 1997 National Center for Health Statistics report indicated that >80% of all adult private workers in the United States worked in establishments that offered health insurance to their employees, that 68% of the employees were eligible for health benefits, and that 58% participated in the employer-sponsored health plan.21 Although lacking information on employment-based insurance availability, eligibility, or acceptance for young adults, our findings suggest that young adults are less likely than older adults to have access to health insurance through their employers. Although this inequality is not surprising considering that young adults are likely to work in service industries, occupy entry-level employment positions, or work part time,4 it may be a substantial contributor to the high rates of uninsurance in the young adult population. Young adults in workplaces that do not offer health insurance represent a potentially important target for programs and policies that seek to reduce the number of uninsured young adults. Similar to policies directed at young adult students, requirements or incentives for young adult workers to have health insurance coverage accompanied by provisions from private insurers and employers could potentially lead to lower rates of uninsurance in this population.
As expected, male and female young adults in the highest household income category were the least likely to report being uninsured. That there was no significant difference in the adjusted odds of being uninsured for men in the lowest income category compared with the highest income category (and that women in the lowest income category had only slightly greater odds of being uninsured than those in the highest income category) highlights a difficulty in using income data alone to determine risk for uninsurance in this age group. These findings are more readily understood when considering that a larger proportion of young adults in the lowest income category were students than in any other income category. Among those who provided household income information, 39% of male students and 29% of female students reported living in households with incomes <$10 000. For both genders, the highest odds of being uninsured were seen among young adults in households with an income between $10 000 and $34 999. According to the US Census Bureau, the weighted average poverty threshold in 2000 was $8794 per year for a 1-person household and $17 603 per year for a 4-person household.22 In addition to agreeing with other reports describing inequities in coverage of the poor and near-poor of all ages,2,23–25 this study suggests that young adults with household incomes that may approach and even exceed 200% of the poverty threshold are also at significant risk for being uninsured. This group is a potentially large segment of the young adult population. In this sample, 43% of men and 46% of women for whom income information was available lived in households with an income of <$35 000—an estimate that is likely to be conservative given the demographics and high uninsurance rate reported among the 20% of participants for whom income data were not available.
Black non-Hispanic and Hispanic young adults of both genders were more likely to report being uninsured than white non-Hispanic young adults. After adjustment, the highest odds of being uninsured were seen among Hispanic young adults. These findings are consistent with other studies of adults and children showing that Hispanic Americans are more likely than other racial or ethnic groups to be uninsured.4,26–29 These studies have implicated differences in the availability of employment-based health insurance, citizenship concerns, and other cultural factors as contributing factors to the high rates of uninsurance among Hispanic Americans. More study is needed to characterize the barriers to health insurance for this heterogeneous population.
There are several potential limitations to consider. The NHIS is based on self-report or the report of other members of the household and may be subject to recall error. We calculated that 28% of those selected as sample adults did not complete the survey. This may introduce bias, although the weighting of the data partially takes into account this nonresponse. These analyses did not validate insurance status through another source, and measures of health insurance adequacy or inquiries about gaps in insurance coverage among the currently insured were not included. Because major activity in the previous week was used as a proxy for student and employment status, misclassification may have occurred. We anticipate that this misclassification would lead to an underestimation of the associations between student and employment status and health insurance status.
Approximately one quarter of female young adults and one third of male young adults who participated in the Sample Adult interview of the NHIS from 1998 to 2000 reported being uninsured at the time of the survey. In general, sociodemographic variables were similarly correlated with health insurance for both men and women. Employment in a workplace that did not offer health insurance coverage, low household income, low educational attainment, and Hispanic ethnicity were independently associated with increased odds of being uninsured for both genders. Having attended college, higher household income, and being a student or an employee in a workplace that offers health insurance coverage were associated with decreased odds of being uninsured for both genders. For many sociodemographic variables studied, the magnitude of association with uninsurance was less pronounced for women than for men. Although this attenuation may, in part, reflect the additional safety net of Medicaid coverage available to young female adults, additional study is needed.
Pediatric providers are in a unique position to have a positive impact on the problem of uninsurance among young adults at an individual and policy level, although substantial improvement in the problem will require efforts from government, insurers, employers, and the larger medical community. An awareness of potential risk factors for being uninsured and knowledge of state and national programs may improve the provider’s ability to match some uninsured young adult patients (or those at risk for uninsurance) to existing resources. Pediatric providers should use their individual and collective influence to advocate for local and national programs and policy change that increase opportunities for young adults to obtain and retain health insurance coverage.
Our findings suggest that additional opportunities for health insurance coverage are needed for young adults—particularly men, Hispanic young adults, and those in low- and middle-income households. Increasing the availability of employment-based health insurance coverage, discouraging attrition from primary and secondary education, and the creation of insurance opportunities for minorities and young adults in households with incomes that exceed the poverty level are potentially important target areas for programs that seek to reduce the number of uninsured young adults. Additional research should evaluate the impact of being uninsured on health and health care access for young adults in general and for special populations of young adults such as those with chronic medical conditions.
CONFERENCE DEPLORES CORPORATE INFLUENCE ON ACADEMIC SCIENCE
“A backlash against academic science collaborating with industry has been building for years, inspired by instances of suppression of research unfavorable to corporate sponsors, hidden conflicts of interest, distorted clinical trials, and retaliation against uncooperative scientists. … These and other misdeeds were publicly chronicled and deplored on July 11 in Washington, DC, USA, at an all-day conference. The meeting is a good candidate for landmark status in the mounting concerns about the erosion of scientific independence under stretched academic budgets, the lure of corporate funds, and government pressures for closer relations between the two sectors as a source of economic growth. … The proclaimed hero of the day was Herbert Needleman, of the University of Pittsburgh, who was pilloried by the lead industry for his research to establish lead’s toxic effects on children. In recognition of his struggles, which included accusations of scientific misconduct, Needleman received the CSPI’s [Center for Science in the Public Interest] inaugural Award for Integrity in Science in Honor of Rachel Carson.”
Greenberg DS. Lancet. July 26, 2003
Noted by JFL, MD
We thank Lynn S. Walker, PhD, for thoughtful review of previous versions of this manuscript.
- Received February 18, 2003.
- Accepted May 28, 2003.
- Reprint requests to (S.T.C.) Division of Adolescent Medicine, Vanderbilt University Medical Center, 436 Medical Center South, Nashville, TN 37212-3100. E-mail:
- ↵Rhoads JA. The uninsured in America 2001. Medical Expenditure Panel Survey. Statistical Brief #4. Baltimore, MD: Agency for Healthcare Research and Quality; 2001. Available at: www.meps.ahrq.gov/papers/st4/stat04.pdf
- ↵Mills RJ. Health Insurance Coverage: 2001. Washington, DC: US Census Bureau; 2002 (Current Population Reports No. P60-220)
- ↵Institute of Medicine. Health Insurance Is a Family Matter. Washington, DC: National Academy Press; 2002:51–60
- ↵Quinn K, Schoen C, Buatti L. On Their Own: Young Adults Living Without Health Insurance. New York, NY: The Commonwealth Fund; 2000. Available at: www.cmwf.org
- ↵Ryan JM. SCHIP turns five: taking stock, moving ahead. NHPF Issue Brief.2002;(781) :1– 12
- ↵McManus MA, Greaney AM, Newacheck PW. Health insurance status of young adults in the United States. Pediatrics.1989;84 :709– 716
- ↵Newacheck P, Brindis CD, Uhler-Cart C, Marchi K, Irwin CE. Adolescent health insurance coverage: recent changes and access to care. Pediatrics.1999;104 :195– 202
- ↵Division of STD Prevention. Tracking the Hidden Epidemics: Trends in STD’s in the United States 2000. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention; 2000. Available at: www.cdc.gov/nchstp/dstd/Stats_Trends/Trends2000.pdf
- ↵American Academy of Pediatrics. A consensus statement on health care transitions for young adults with special health care needs. Pediatrics.2002;110(suppl 2) :1304– 1306
- ↵US Census Bureau. Statistical Abstract of the United States 2001: The National Data Book. Washington, DC: US Census Bureau; 2001:363–377
- ↵National Center for Health Statistics. National Health Interview Survey survey description. US Department of Health and Human Services, Centers for Disease Control and Prevention; 2002. Available at: www.cdc.gov/nchs/nhis.htm
- ↵National Center for Health Statistics. Health, United States, 2001 With Urban and Rural Health Chartbook. Washington, DC: US Government Printing Office; 2001:396
- ↵SUDAAN User’s Manual, Release 8.0. Research Triangle Park, NC: Research Triangle Institute; 2002
- ↵National Center for Health Statistics. Employer Sponsored Health Insurance: State and National Estimates. Washington, DC: US Department of Health and Human Services, Centers for Disease Control and Prevention; 1997:6–10 (Report No. [PHS] 98-1005)
- ↵Dalaker J. Poverty in the United States: 2000. Washington, DC: US Census Bureau; 2001 (Current Population Reports No. P60-214)
- ↵Kuttner R. The American health care system: health insurance coverage. N Engl J Med.1990;340 :163– 168
- ↵Institute of Medicine. Coverage Matters: Insurance and Health Care. Washington, DC: National Academy Press; 2001:70–100
- Monheit AC, Bistnes JP. Race/ethnicity and health insurance status: 1987 and 1996. Med Care Res Rev.2000;57(suppl 1) :11– 35
- ↵Rhoades JA, Vistnes JP, Cohen JW. The uninsured in America: 1996–2000. Rockville, MD: Agency for Healthcare Research and Quality; 2002;17–24 (MEPS Chartbook No. 9. AHRQ Pub. No. 02-0027)
- Copyright © 2004 by the American Academy of Pediatrics