Health coverage instability for mothers in working families.

Anderson, Steven G. ; Eamon, Mary Keegan

As women have entered the labor force in large numbers, they increasingly have gained access to the employer-based coverage that dominates the U.S. health care system (Mishel, Bernstein, & Schmitt, 2001). Yet, several studies have shown that many working women lack access to employer-based coverage (Currie & Yelowitz, 2000; Hoffman & Scholbohm, 2000; Van Loon, Borkin, & Steffen, 2002), and public health coverage options are limited by rigid income eligibility restrictions (Guyer, Broaddus, & Dude, 2002). The implementation of welfare reforms may have exacerbated this problem, as both work requirements and lifetime limits on welfare receipt have contributed to many women leaving welfare for low-paying jobs and then quickly exhausting transitional Medicaid benefits (Anderson & Gryzlak, 2002; Garrett & Holahan, 2000a). Globalization of the marketplace, increasing use of contingent and contractual labor, and rising health care premiums cast doubt on the future adequacy of employer-based coverage for working women (Davis, Aguilar, & Jackson, 1998; Keigher & Lowery, 1998; Kuttner, 1999).

In these dynamic economic and social policy environments, the extent to which mothers have access to either private or public health care coverage requires continuing scrutiny. We examined health insurance stability for a sample of mothers not receiving welfare when TANF programs were being implemented to determine prevalent health insurance coverage patterns for mothers as they experience varying employment and income scenarios. This was useful in ascertaining the health care situations of mothers in working families generally and in predicting coverage prospects for women less able to obtain Medicaid as welfare reform programs mature. We used a measure of health coverage stability, as opposed to coverage estimates at a single point in time, to provide a fuller picture of the health coverage disruptions experienced by women over time.

Background

In 2001 an estimated 41 million Americans lacked health insurance, and about three-fourths of the uninsured were in working families (Holahan & Kim, 2000; U.S. Census Bureau, 2002). Largely because they often obtain coverage through spouses, uninsurance rates for women were slightly lower than for men (U.S. Census Bureau, 2000). However, women are less likely to obtain coverage through their own employment, which leaves unmarried women particularly vulnerable to being uninsured (Currie & Yelowitz, 2000; Mishel et al., 2001).

The implications of these coverage problems are substantial. Research has demonstrated that health insurance coverage improves health care for low-income adults and their children (Almeida, Dubay, & Ko, 2001; Lillie-Blanton, 1999; Salganicoff & Wyn, 1999). In addition, the Medicaid program has been shown to have significantly narrowed socioeconomic differences in health care access and outcomes for nonelderly people (Andrulis, 1998; Berk & Schur, 1998; Currie & Gruber, 1996).

Being uninsured is much more prevalent among low-income groups, so poor single mothers who have been the focus of welfare reform are particularly likely to lack coverage in the absence of public health insurance programs. The Family Support Act of 1988 (P.L. 100-485) attempted to alleviate health coverage problems for those leaving welfare by establishing time-limited transitional Medicaid coverage (Garrett & Holahan, 2000b). This coverage continued with the devolution of welfare programs to the states under TANF, and some states have extended coverage beyond the one-year period allowed under federal regulations (Ellwood, 1999; Guyer et al., 2002).

Transitional Medicaid assistance is based on a philosophy of providing low-income workers a period of coverage while they gain the job experience needed to gain access to employer-based health care coverage. However, Garrett and Holahan (2000a) found that nearly half the people who leave TANF lack health care coverage one year later. These high uninsurance rates occur largely because welfare leavers typically obtain jobs without coverage or experience job instability (Anderson & Gryzlak, 2002; Anderson, Halter, Juenes, & Schuldt, 2000). Researchers also found that people leaving TANF underused Medicaid because of lack of knowledge about benefits, confusion about program rules, and other bureaucratic constraints (Ellwood, 1999; U.S. General Accounting Office, 1999).

Most women do not receive welfare benefits before entering the labor force, so they are not eligible for transitional Medicaid coverage. Welfare reform further constrains the possibility of women entering welfare programs, so a broader examination of whether and how women in working families obtain coverage is needed. The factors related to insurance coverage for these women and the pathways through which insurance is obtained also merit research attention.

We examined several of these issues using a sample of women without recent ties to welfare. This allowed us to analyze the health care environment of women in working families, which parallels the environment of former TANF recipients after they exhaust TANF eligibility and transitional Medicaid benefits. Unlike most studies, we measured health insurance coverage over a three-year period rather than at single points in time (Schoen & DesRoches, 2000). This method captures the dynamic movement in and out of coverage and provides a measure of the percentage of working women affected by episodes of uninsurance in a given period. Such a dynamic measure is especially important given women's frequent movement in and out of low-paying jobs (Anderson et al., 2000).

Data and Method

We used data from the National Longitudinal Survey of Youth (NLSY), a household survey of a representative sample of 12,686 noninstitutionalized civilian youths who were ages 14 to 22 years when first interviewed in 1979. The original NLSY contained oversamples of black, Hispanic, economically disadvantaged non-Hispanic white, and military youths. In 1998, 84 percent of eligible respondents were interviewed (U.S. Department of Labor, Bureau of Labor Statistics, 2000).

Data from female respondents for the two recent biannual survey rounds (1996 and 1998) were used. This allowed for the analysis of health insurance coverage for women in working families during the 1995-1997 period when TANF and predecessor Aid to Families with Dependent Children (AFDC) welfare reform programs were being developed and implemented. We limited the sample to women because they are the group most affected by TANF and the group on which welfare reform discussions generally have focused.

The respondents met four criteria to be included in the study. First, they received no income from AFDC or TANF during the study period. Second, the mother had at least one child in the home at the time of the 1996 and 1998 interviews. Third, the respondent and her spouse or partner jointly worked at least 960 hours in both 1995 and 1997 (approximately hall-time employment for one person). Fourth, the respondent had information on her health insurance coverage during the entire three-year study period.

The final study sample of 1,667 women in working families without welfare receipt included 390 black, 309 Hispanic, and 968 non-Hispanic white mothers between 32 and 39 years of age. The numbers reported are not weighted; they are the actual number of survey respondents. However, the percentages reported are weighted to produce estimates of a nationally representative sample of women who were 14 to 21 years of age on January 1, 1979.

Results

Only 3.1 percent of the 1,667 mothers in this study had no insurance coverage at any time during 1995-1997, whereas 14.9 percent lacked coverage in at least one month and 82.0 percent had coverage in all months. The construction of the insurance coverage variable only allowed us to determine whether a respondent had breaks in coverage during the period, as opposed to establishing the length of uncovered periods. We therefore combined those with coverage breaks with the small percentage of mothers who never had coverage during the period to form a "no or unstable coverage" category. Mothers in this category were compared with those who had stable coverage for the entire study period.

Because initial examination of the data confirmed the importance of income level in determining coverage instability, much of the analysis compares coverage instability among mothers at three income levels: less than 200 percent, 200 percent to 399 percent, and 400 percent or more of the poverty level. We refer to these groups as the low-, middle-, and high-income groups.

Characteristics of Mothers with Unstable Health Coverage

Several factors differentiated mothers who had stable health care coverage from those with unstable coverage (Table 1). The importance of income is demonstrated by sharp differences in coverage as income-to-needs ratios vary. Nearly one-half of mothers with incomes below 200 percent of the poverty level had unstable coverage, compared with 15.5 percent of those with incomes between 200 and 399 percent of poverty and only 4.3 percent of those with incomes of 400 percent or more of poverty.

Although all of these mothers were in working families, coverage varied considerably according to hours worked and job stability. For example, 55.3 percent of the mothers in households that averaged fewer than 1,920 hours of work (full-time for one person) had unstable coverage compared with only 20.9 percent for those that averaged 1,920 to 3,840 hours and 11.6 percent for those that averaged more than 3,840 hours. Coverage also decreased as study respondents changed jobs; respondents who changed jobs more than twice during 1995-1997 were nearly three times as likely to have unstable coverage as those who kept the same job (37.5 percent compared with 13.6 percent).

Coverage instability varied by demographic and educational characteristics as well. Nearly 42 percent of mothers who did not complete 12 years of school had unstable coverage, compared with 22.1 percent of those who completed 12 years of school and 12.1 percent who completed schooling beyond high school. About 74 percent of the respondents were married or living with a partner during the entire study period, and among these only 13.4 percent had unstable coverage. However, the risk of unstable coverage was much higher for single mothers (38.4 percent) and for those whose marital status changed during the study period (34.4 percent). Differences by race were less striking, with 22.8 percent of African American and Hispanic mothers having unstable coverage, compared with 17.2 percent of white mothers.

Source of Mothers' Insurance Coverage

Coverage sources varied by income for those who had coverage at both the 1996 and 1998 interviews (Table 2). The dominance of employer-based insurance is evident for all income groups, ranging from 86 percent of mothers with incomes below 200 percent of the poverty level to 93.6 percent for those with incomes of 200 percent to 399 percent of the poverty level. The lack of government health care support for women not on welfare is illustrated by the fact that only 5.2 percent of the low-income mothers who had health insurance in both years received government-based health coverage.

Two income-related differences in the types of employment-based coverage received by these women should be noted. First, mothers whose income was less than 200 percent of the poverty level were more likely to obtain health care insurance from their employer (49.0 percent) than were mothers with higher family incomes (37.9 percent of mothers in the middle-income group and 39.7 percent of mothers in the higher income group). Second, low-income mothers were less likely to obtain insurance coverage from their spouse or partners' employer (28.1 percent), compared with mothers in the two higher income categories (approximately 40 percent). This is not surprising, as lower income mothers are disproportionately likely to be single.

Social and Demographic Characteristics of Mothers with Unstable Coverage

Rates of coverage instability varied across and within income groups when selected social and demographic factors were examined (Table 3). The large differences in coverage instability between income groups for mothers with comparable social and demographic characteristics are the most notable findings. Mothers with incomes of less than 200 percent of the poverty level generally were at least two to three times as likely to have unstable coverage as mothers with comparable social and demographic characteristics who had income of 200 percent to 399 percent of the poverty level. The greatest discrepancy between income groups for these characteristics occurred in the northeast region of the United States, where 45.1 percent of low-income mothers but only 6.9 percent of middle-income mothers had unstable coverage.

Although coverage instability among low-income mothers was high regardless of social and demographic characteristics, the levels of instability were striking for those with selected characteristics. For example, 69 percent of low-income mothers with less than 12 years of schooling had unstable coverage. Low-income mothers living in the western region of the United States and in central city areas also had extremely high rates of unstable coverage (67.0 percent and 62.1 percent, respectively).

Although differences in coverage instability across social and demographic characteristics within income groups generally were smaller, several interesting variations may be observed. The most consistent difference pertains to education; mothers with less than 12 years of schooling were more likely to have unstable coverage than their more educated counterparts in each income group.

For other social and demographic characteristics, the coverage instability differences within income groups were more variable. For example, low-income mothers who remained single during the study period had higher levels of coverage instability than mothers who remained married or whose marital status changed. However, middle- and higher-income mothers who changed marital status had much higher instability rates than their consistently married or single counterparts.

Racial differences in coverage also varied within the three income groups. Among those with incomes of less than 200 percent of poverty, white mothers were most likely to experience coverage instability (51.0 percent compared with 42.1 percent for Hispanic mothers and 41.4 percent for African American mothers). However, Hispanic mothers were more likely to have unstable coverage than both white and African American mothers in the middle- and upper-income groups.

Job Characteristics of Mothers with Unstable Coverage

Five job characteristics related to health coverage instability were examined by income status: job stability, average yearly hours worked, union membership, type of employer, and type of employment (Table 4). Consistent with the pattern observed in Table 3, mothers in the lowest income group were more likely to have unstable insurance coverage relative to the higher income groups for each job characteristic examined. Nonetheless, job characteristics appeared to have important effects.

Job stability was particularly important to insurance coverage among low-income mothers; unstable insurance coverage ranged flora 39.9 percent for low-income mothers who did not change jobs between the 1996 and 1998 interviews to 75.0 percent for those with more than two job changes. Those with job changes also were much more likely to experience insurance instability in the 200 percent to 399 percent of poverty level group.

As would be expected in a health insurance system dominated by employer-based coverage, insurance instability generally declined as work hours increased. Approximately 74 percent of low-income mothers in households that averaged fewer than 1,920 hours of work in 1995 and 1997 had unstable insurance coverage. This level dropped to 47.8 percent and 43.9 percent, respectively, for those with 1,920 to 3,840 hours worked and those with more than 3,840 hours worked. Although coverage instability was much lower for each hours-worked category for those with incomes from 200 percent to 399 percent of the poverty level, the same pattern of lower levels of coverage instability as work hours increased was noted (Table 4). Mothers with higher incomes had low levels of unstable insurance coverage, regardless of hours worked.

Among low-income mothers and those with incomes of 200 percent to 399 percent of the poverty level, being associated with a labor union substantially decreased the likelihood of insurance instability. For example, only 14.1 percent of low-income mothers who were associated with a union in both 1995 and 1997 had unstable insurance coverage, compared with 54.4 percent of low-income mothers associated with a union in one or neither year. However, because only about 13 percent of these mothers were associated with a union in both years, the substantial health care benefits of union participation applied to only a small portion of the sample.

Coverage instability varied according to employer type and employment type, particularly for those with low incomes. Those who worked for the government in 1995 and in 1997 had the lowest percentage (28.7 percent) of unstable coverage among low-income mothers, and those who worked for a for-profit employer in both years had the highest percentage (57.6 percent). It is not surprising that mothers in each income group who were regular employees in both years were much more likely to have stable coverage than mothers with comparable incomes who worked in non-regular jobs, such as in temporary contract positions.

Discussion

The findings demonstrate the precariousness of insurance coverage for mothers in working families, especially for those with low incomes. Health coverage instability was uncommon for mothers with incomes over 400 percent of the poverty level. However, about one-sixth of mothers with incomes between 200 percent and 399 percent of the poverty level and nearly one-half of those with incomes below 200 percent of the poverty level had unstable coverage. Low-income mothers with poor education, single-parent marital status, low work hours, and frequent job changes were at even greater risk of coverage instability. There also were disparities in coverage instability by geographic region, with those living in central cities and in the western part of the country especially at risk.

These high levels of health coverage instability point to the importance of using measures of coverage that capture the dynamic circumstances facing families. For example, earlier research has found that about one-third of low-income people lack coverage at a point in time (Lillie-Blanton, 1999), and 31.2 percent and 21.4 percent, respectively, of the low-income respondents in this study at the 1996 and 1998 interviews were uninsured. In comparison, the finding that 49 percent of low-income mothers had unstable coverage indicates that one and one-half to more than two times as many low-income mothers experienced coverage disruptions in the three-year study period than would be suggested by point-in-time estimates.

The dynamic nature of health coverage perhaps is best illustrated by the findings for those who experienced marital disruptions or job changes. For example, whereas coverage instability among higher income mothers was negligible if they remained married of single for the three-year study period, more than one-fifth of higher income mothers whose marital status changed experienced health coverage disruptions. Similarly, women in the lower- and middle-income groups who changed jobs encountered much higher rates of coverage instability than those who kept the same job.

One unexpected finding concerns coverage differences for mothers of different ethnicities. Among the low-income mothers in the study, coverage instability was higher for white women than African American and Hispanic women. However, among higher income mothers, coverage instability was lowest for white women. Research that clarifies the reasons for racial differences in coverage and shows how such differences are affected by changing income levels would be useful.

The findings have troubling implications in terms of the health coverage prospects for mothers who leave TANF for work. Given that most TANF leavers can be expected to have even less employment experience than the low-income mothers in this study, coverage instability is likely to be common for TANF leavers once they exhaust transitional Medicaid. For example, the group of mothers in this sample that correspond most closely with working women who leave TANF are those who are not married and have incomes below 200 percent of the poverty level. More than 56 percent of sample members with these characteristics had unstable coverage over the three-year study period, which corresponded to the time that TANF and predecessor welfare reform programs were being implemented.

Two prominent policies advocated during the 2002 TANF reauthorization process were to either encourage marriage or penalize nonmarriage and to extend the number of hours that TANF recipients are required to work. Our findings suggest that such strategies would do little to improve health coverage stability in the absence of sustainable jobs and income gains. For example, whereas married low-income mothers in the sample fared better than those who were not married, 44 percent of low-income mothers married for the entire study period still experienced coverage instability. Similarly, those with more work hours were less likely to experience insurance instability; but even among low-income families averaging full-time work hours, more than two-fifths of mothers had unstable coverage.

The Challenge for Social Work

"Making work pay" was a slogan commonly used by welfare reform advocates and politicians, and this goal often was envisioned as including health care and child care for low-income working families more generally (Danziger, 1999). Yet, our findings suggest that consistent health coverage is likely to be unavailable to most TANF leavers and to about half of all low-income working mothers unless they are able to increase their earnings above low-income levels. Earlier research has shown that most women who leave TANF do not experience income gains quickly (Cancian, Haveman, Kaplan, Meyer, & Wolfe, 1999), and recent studies have found that employer-based coverage is eroding (Edwards, Doty, & Schoen, 2002; Holahan & Kim, 2000; Mishel et al., 2001). As a result, low-income working mothers will continue to be worse off in terms of health care coverage than those who remain on welfare, where Medicaid coverage generally is available to all. This asymmetry raises serious equity concerns, given that TANF work and training requirements and time limits force many women off welfare and into low-wage work environments. It indicates that one outcome of TANF will be to further undercut inadequate medical coverage for low-income women.

Working to rectify these health coverage inequities is consistent with social work's social justice mission. Continued advocacy on behalf of a national health insurance program is the most fundamental social policy direction in this respect. However, history suggests that short-term prospects for such a plan are dim. Social workers therefore must simultaneously focus on supporting incremental extensions of low-income health care coverage and on ensuring that eligible people gain full access to existing coverage.

One useful incremental policy involves raising the income eligibility limits for Medicaid and related public health programs. For example, the State Children's Health Insurance Program (SCHIP), together with previous Medicaid eligibility expansions targeted for children, is resulting in ma)or insurance coverage improvements for children living in families with incomes below 200 percent of the poverty level. Yet, working parents in these same families typically remain uninsured in the absence of federal or state policies to broaden coverage for low-income adults (Guyer et al., 2002). Some states have expanded coverage for low-income adults to correspond to that offered to children under SCHIP and Medicaid, using a variety of funding mechanisms (Jordan, Adamo, & Ehrmann, 2000). Continued advocacy at the federal and state levels could provide a more solid health insurance foundation for low-income working families.

More policy attention is needed to help low-income people with health coverage transitions when jobs are lost or when family disruptions occur. Although federal COBRA policy allows employees to continue their previous employer-based coverage for up to 18 months after leaving a job, high premiums often limit the usefulness of this policy for low-income people (Edwards et al., 2002). Programs that assist with COBRA payments or that make public coverage through Medicaid or other programs more readily available during transitional periods of job change and family disruption merit experimentation.

Social workers also have important roles to play in ensuring that low-income people get full access to existing medical coverage, because earlier research has demonstrated that Medicaid and other public health programs are underused (Ellwood, 1999). Outreach programs targeted at low-income neighborhoods is one approach that has been shown to be effective in improving Medicaid program participation (Jordan et al., 2000).

More fully developing case management roles for TANF caseworkers and for social workers in related human services agencies has promise for improving use of services (Hagen, 1999). Lack of knowledge about service availability has been identified as a primary reason for underutilization of services (Anderson, 2002). Because Medicaid eligibility rules tend to be complex and frequently changing, extensive and ongoing caseworker training on eligibility and application procedures is required if caseworkers are to be effective conveyors of health coverage information.

Caseworker efforts to improve health care access can proceed only if their service environments are accommodating. Thus, adequate funding and support of case manager positions, reasonable caseload sizes, and professional standards for workers are needed to improve the likelihood that caseworkers adequately fulfill information and referral roles. Administrators also can work to simplify eligibility rules and to calibrate procedures between related social services programs. Such efforts can lessen the complexity of benefit interpretations and correspondingly ease the information-processing tasks that clients encounter when attempting to obtain medical coverage and other services.

Improving the stability of health care coverage for low-income people is a complex human services problem for which social work efforts are required at the policy development, administrative, and direct service levels. Progress on this issue is a fundamental social justice concern and an important tool for empowering individuals to better control their economic and social well-being. Providing consistent coverage is critical in sending a social policy signal that the work efforts of low-income families will be positively reinforced.

Table 1

Stability of Health Insurance Coverage among Mothers, by Selected
Factors, 1995-1997 (N = 1,667)

 Stable No or Unstable
 Coverage Coverage
Factor N % %

Average family income-to-needs, 1995 and 1997
 Less than 200% of poverty 418 51.6 48.4
 200%-399% of poverty 747 84.5 15.5
 400% or more of poverty 502 95.7 4.3
Race/ethnicity
 African American 390 77.2 22.8
 Hispanic 309 77.2 22.8
 White 968 82.8 17.2
Educational level, 1996
 Less than 12 years 102 58.1 41.9
 12 years 722 77.9 22.1
 More than 12 years 843 87.9 12.1
Marital status, 1995-1997 (a)
 Single entire period 265 61.6 38.4
 Married entire period 1,240 86.6 13.4
 Marital status changed 162 65.6 34.4
Job stability, between 1996-1998
 Never changed jobs 1,003 86.4 13.6
 One job change 379 72.0 28.0
 Two job changes 107 71.3 28.7
 More than two job changes 38 62.5 37.5
 Never worked 140 86.1 13.9
Average yearly hours worked, 1995 and 1997 (b)
 Less than 1,920 hours 73 44.7 55.3
 1,920-3,840 hours 910 79.1 20.9
 More than 3,840 hours 684 88.4 11.6

(a) Single refers to women unmarried and unpartnered, and married
includes women living with a partner.

(b) Includes spouse or partner work hours, where applicable.

Table 2

Source of Health Insurance Coverage among Mothers Who Had Insurance
Coverage at the 1996 and 1998 Interviews, by Income Status (N = 1,424)

 Average Family
 Income-to-Needs
 1995 and 1997

 Less than
 200% of
Insurance Coverage poverty

Mothers' employee-based both years (%) 49.0
Spouse/partners' employee-based both years (%) 28.1
Combination mothers' and spouse/partners'
 employee-based both years (%) 8.9
 Total employee-based coverage (%) 86.0
Government-based either year (a) (%) 5.2
Other (b) (%) 8.8
N 264

 Average Family
 Income-to-Needs
 1995 and 1997

 200%-399%
 of
Insurance Coverage poverty

Mothers' employee-based both years (%) 37.9
Spouse/partners' employee-based both years (%) 40.7
Combination mothers' and spouse/partners'
 employee-based both years (%) 15.0
 Total employee-based coverage (%) 93.6
Government-based either year (a) (%) 1.0
Other (b) (%) 5.4
N 675

 Average Family
 Income-to-Needs
 1995 and 1997

 400% or
 more of
Insurance Coverage poverty

Mothers' employee-based both years (%) 39.7
Spouse/partners' employee-based both years (%) 40.1
Combination mothers' and spouse/partners'
 employee-based both years (%) 12.9
 Total employee-based coverage (%) 92.7
Government-based either year (a) (%) 0.0
Other (b) (%) 7.3
N 485

(a) Includes Medicaid, Medi-Cal, medical assistance, and welfare/
medical service.

(b) Includes coverage bought directly from a medical insurance company,
obtained from another source not includeD in the table, or obtained
from a combination of employee-based and other sources.

Table 3

Social and Demographic Factors Associated with Stability of Health
Insurance Coverage among Mothers, 1995-1997, by Income Status (N= 1,667)

Characteristic N (a)

Race/ethnicity
 African American 390
 Hispanic 309
 White 968
Education level, 1996
 Less than 12 years 102
 12 years 722
 More than 12 years 843
Marital status, 1995-1997
 Single entire period 265
 Married entire period 1,240
 Marital status changed 162
Region of residence, 1996 and 1998
 Northeast 234
 North central 406
 South 668
 West 316
 Changed during period 20
Urban residence, 1996 and 1998
 Not in SMSA 290
 SMSA, not in central city 867
 SMSA, in central city 120
 Changed during period 336

 No or Unstable
 Insurance
 Coverage

 Less than
 200% of
Characteristic poverty (b) (%)

Race/ethnicity
 African American 41.4
 Hispanic 42.1
 White 51.0
Education level, 1996
 Less than 12 years 69.0
 12 years 47.5
 More than 12 years 43.7
Marital status, 1995-1997
 Single entire period 56.1
 Married entire period 43.5
 Marital status changed 47.9
Region of residence, 1996 and 1998
 Northeast 45.1
 North central 39.4
 South 49.8
 West 67.0
 Changed during period 100.0
Urban residence, 1996 and 1998
 Not in SMSA 51.4
 SMSA, not in central city 49.3
 SMSA, in central city 62.1
 Changed during period 41.3

 No or Unstable
 Insurance
 Coverage

 200%-399%
 of
Characteristic poverty (b) (%)

Race/ethnicity
 African American 10.2
 Hispanic 17.7
 White 15.9
Education level, 1996
 Less than 12 years 26.8
 12 years 16.3
 More than 12 years 13.1
Marital status, 1995-1997
 Single entire period 17.7
 Married entire period 14.0
 Marital status changed 28.7
Region of residence, 1996 and 1998
 Northeast 6.9
 North central 8.5
 South 21.2
 West 23.2
 Changed during period 54.6
Urban residence, 1996 and 1998
 Not in SMSA 14.1
 SMSA, not in central city 16.6
 SMSA, in central city 11.6
 Changed during period 16.3

 No or Unstable
 Insurance
 Coverage

 400% or
 more of
Characteristic poverty (b) (%)

Race/ethnicity
 African American 8.5
 Hispanic 10.8
 White 3.8
Education level, 1996
 Less than 12 years 20.9
 12 years 7.5
 More than 12 years 2.5
Marital status, 1995-1997
 Single entire period 1.3
 Married entire period 3.5
 Marital status changed 20.8
Region of residence, 1996 and 1998
 Northeast 4.8
 North central 2.7
 South 4.2
 West 6.4
 Changed during period 0.0
Urban residence, 1996 and 1998
 Not in SMSA 7.4
 SMSA, not in central city 3.3
 SMSA, in central city 1.4
 Changed during period 6.5

NOTE: SMSA = standard metropolitan statistical area.

(a) N refers to number of respondents in each category who had
available data.

(b) Refers to average family income-to-needs, 1995 and 1997.

Table 4

Job Characteristics Associated with Stability of Health Care Insurance
among Mothers Who Worked in 1995 and 1997, by Income Status (N = 1,527)

Characteristic
 N (a)
Job stability, 1996 and 1998
 Never changed jobs 1,003
 One job change 379
 Two job changes 107
 More than two job changes 38
Average yearly hours worked, 1995 and 1997 (c)
 Less than 1,920 hours 60
 1,920-3,840 hours 784
 More than 3,840 hours 683
Union membership, 1995 and 1997 (d)
 Belonged to union in both years 180
 Belonged to union in neither or one year 1,208
Type of employer, 1995 and 1997
 Government both years 229
 For-profit both years 770
 Nonprofit both years 72
 Other and mixture of employer types (e) 317
Type of employment, 1995 and 1997
 Regular in both years 1,259
 Other type in one or both years (f) 128

 No or Unstable
 Insurance
 Coverage

 Less than
 200% of
Characteristic poverty (b) (%)

Job stability, 1996 and 1998
 Never changed jobs 39.9
 One job change 56.7
 Two job changes 66.0
 More than two job changes 75.0
Average yearly hours worked, 1995 and 1997 (c)
 Less than 1,920 hours 73.8
 1,920-3,840 hours 47.8
 More than 3,840 hours 43.9
Union membership, 1995 and 1997 (d)
 Belonged to union in both years 14.1
 Belonged to union in neither or one year 54.4
Type of employer, 1995 and 1997
 Government both years 28.7
 For-profit both years 57.6
 Nonprofit both years 38.7
 Other and mixture of employer types (e) 46.2
Type of employment, 1995 and 1997
 Regular in both years 48.9
 Other type in one or both years (f) 78.7

 No or Unstable
 Insurance
 Coverage

 200%-399%
 of
Characteristic poverty (b) (%)

Job stability, 1996 and 1998
 Never changed jobs 11.5
 One job change 27.6
 Two job changes 21.2
 More than two job changes 17.9
Average yearly hours worked, 1995 and 1997 (c)
 Less than 1,920 hours 40.7
 1,920-3,840 hours 16.5
 More than 3,840 hours 14.6
Union membership, 1995 and 1997 (d)
 Belonged to union in both years 2.7
 Belonged to union in neither or one year 17.5
Type of employer, 1995 and 1997
 Government both years 10.0
 For-profit both years 16.2
 Nonprofit both years 6.4
 Other and mixture of employer types (e) 21.3
Type of employment, 1995 and 1997
 Regular in both years 13.9
 Other type in one or both years (f) 31.4

 No or Unstable
 Insurance
 Coverage

 400% or
 more of
Characteristic poverty (b) (%)

Job stability, 1996 and 1998
 Never changed jobs 4.6
 One job change 4.3
 Two job changes 3.3
 More than two job changes 0.0
Average yearly hours worked, 1995 and 1997 (c)
 Less than 1,920 hours 0.0
 1,920-3,840 hours 7.6
 More than 3,840 hours 2.4
Union membership, 1995 and 1997 (d)
 Belonged to union in both years 0.0
 Belonged to union in neither or one year 4.5
Type of employer, 1995 and 1997
 Government both years 3.6
 For-profit both years 4.3
 Nonprofit both years 0.0
 Other and mixture of employer types (e) 3.9
Type of employment, 1995 and 1997
 Regular in both years 3.4
 Other type in one or both years (f) 8.4

(a) N refers to number of respondents in each category who had
available data.

(b) Refers to average family income-to-needs, 1995 and 1997.

(c) Includes spouse/partner work hours, where applicable.

(d) Includes membership in a labor union or association or current job
covered by a union contract.

(e) Other includes self-employment and working in a family business.

(f) Other includes temporary worker, contractor, consultant, or
employee of a contractor.

References

Almeida, R., Dubay, L., & Ko, G. (2001). Access to care and use of health services by low-income women. Health Care Financing Review, 22(2), 27-47.

Anderson, S. G. (2002). Ensuring the stability of welfare-to-work exits: The importance of recipient knowledge about work incentives. Social Work, 47, 162-170.

Anderson, S. G., & Gryzlak, B. M. (2002). Social work advocacy in the post-TANF environment: Lessons from early TANF research studies. Social Work, 47, 301-314.

Anderson, S. G., Halter, A., Julnes, G., & Schuldt, R. (2000). Job stability and wage progression patterns among early TANF leavers. Journal of Sociology and Social Welfare, 27(4), 39-59.

Andrulis, D. P. (1998). Access to care is the centerpiece in the elimination of socioeconomic disparities in health. Annals of Internal Medicine, 129, 412-416.

Berk, M. L., & Schur, C. L. (1998). Access to care: How much difference does Medicaid make? Health Affairs, 17(3), 169-180.

Cancian, M., Haveman, R., Kaplan, T., Meyer, D., & Wolfe, B. (1999). Work, earnings, and well-being after welfare. In S. H. Danziger (Ed.), Economic conditions and welfare reform (pp. 161-186). Kalamazoo, MI: W. E. Upjohn Institute for Employment Research.

Currie, J., & Gruber, I. (1996). Health insurance eligibility, utilization of medical care, and child health. Quarterly Journal of Economics, 3, 431-466.

Currie, J., & Yelowitz, A. (2000). Health insurance and less skilled workers. In D. Card & R. M. Blank (Eds.), Finding jobs: Work and welfare reform (pp. 233-261). New York: Russell Sage Foundation.

Danziger, S. H. (1999). Introduction. In S. H. Danziger (Ed.), Economic conditions and welfare reform (pp. 1-14). Kalamazoo, MI: W. E. Upjohn Institute for Employment Research.

Davis, K. E., Aguilar, M. A., & Jackson, V. H. (1998). Save low-income women and their children first [Guest Editorial]. Health & Social Work, 23, 83-85.

Edwards, J., Doty, M., & Schoen, C. (2002, August). The erosion of employer-based health coverage and the threat to workers' health care: Findings from the Commonwealth Fund 2002 Workplace Health Insurance Survey [Issue Brief]. New York: Commonwealth Fund.

Ellwood, M. (1999). The Medicaid eligibility maze: Coverage expands, but enrollment problems persist (Assessing the New Federalism, Occasional Paper No. 30). Washington, DC: Urban Institute.

Family Support Act of 1988, P.L. 100-485, 102 Stat. 2343.

Garrett, B., & Holahan, J. (2000a). Health insurance coverage after welfare. Health Affairs, 19(1), 175-184.

Garrett, B., & Holahan, J. (2000b). Welfare leavers, Medicaid coverage, and private health insurance. In New Federalism: National Survey of America's Families (Series B, No. B-13). Washington, DC: Urban Institute.

Guyer, J., Broaddus, M., & Dude, A. (2002). Millions of mothers lack health insurance coverage in the United States. International Journal of Health Services, 32, 89-106.

Hagen, J. L. (1999). Public welfare and human services: New directions under TANF? Families in Society, 80, 78-89.

Hoffman, C., & Scholbohm, A. (2000). Uninsured in America: A chart book. Washington, DC: Kaiser Commission on Medicaid and the Uninsured.

Holahan, J., & Kim, J. (2000). Why does the number of uninsured Americans continue to grow? Health Affairs, 19(4), 188-196.

Jordan, J., Adamo, A., & Ehrmann, T. (2000). Innovations in Section 1115 demonstrations. Health Care Financing Review, 22(2), 49-59.

Keigher, S. M., & Lowery, C. T. (1998). The sickening implications of globalization [National Health Line]. Health & Social Work, 23, 153-158.

Kuttner, R. (1999). The American health care system: Health insurance coverage. New England Journal of Medicine, 340, 163-167.

Lillie-Blanton, M. (1999). A review of the nation's progress and challenges in assuring access to health care for low-income Americans. In M. Lillie-Blanton, R. M. Martinez, B. Lyons, & D. Rowland (Eds.), Access to health care: Promises and prospects for low-income Americans (pp. 19-37). Washington, DC: Kaiser Commission on Medicaid and the Uninsured.

Mishel, L., Bernstein, J., & Schmitt, J. (2001). The state of working America: 2000/2001. Ithaca, NY: Cornell University Press.

Salganicoff, A., & Wyn, R. (1999). Access to care for low-income women: The impact of Medicaid. Journal of Health Care for the Poor and Uninsured, 10, 453-467.

Schoen, C., & DesRoches, C. (2000). Uninsured and unstably insured: The importance of continuous insurance coverage. Health Services Research, 35(1, part II), 187-206.

U.S. Census Bureau. (2000). Health insurance coverage. Current Population Reports (Series P60-211). Washington, DC: U. S. Government Printing Office.

U.S. Census Bureau. (2002). Health insurance coverage: 2001. Washington, DC: U. S. Government Printing Office.

U.S. Department of Labor, Bureau of Labor Statistics. (2000). NLS handbook, 2000. Washington, DC: Author.

U.S. General Accounting Office. (1999). Medicaid enrollment: Amid declines, state efforts to ensure coverage after welfare vary (GAO/HEHS-99-163). Washington, DC: Health, Education, and Human Services Division.

Van Loon, R. A., Borkin, J. R., & Steffen, J. J. (2002). Health care experiences and preferences of uninsured workers. Health & Social Work, 27, 17-26.

Steven G. Anderson, PhD, is assistant professor, and Mary Keegan Eamon, PhD, ACSW, is assistant professor, School of Social Work, University of Illinois at Urbana-Champaign, 1207 West Oregon, Urbana, IL 61801; e-mail: sandersn@uiuc.edu. The authors would like to thank Meenakshi Venkataraman, Joo-Hui Ryo, and Karen Sodowsky for their research assistance.

Original manuscript received October 3, 2002 Accepted February 3, 2003