Abstract
Objectives. This study investigated whether health problems among poor mothers of chronically ill children affect their ability to obtain and maintain employment.
Methods. Mothers of children with chronic illnesses were surveyed at clinical and welfare agency sites in San Antonio, Tex.
Results. There were distinct health differences according to mothers’ TANF and employment status. Mothers without TANF experience reported better physical and mental health and less domestic violence and substance use than did those who had TANF experience. Those not currently working had higher rates of physical and mental health problems.
Conclusions. Poor maternal health is associated with need for cash assistance and health insurance. Policymakers must recognize that social policies promoting employment will fail if they do not address the health needs of poor women and children.
The landmark welfare reform legislation passed by Congress in 1996—the Personal Responsibility and Work Opportunity Reconciliation Act—revealed politicians’ beliefs that the primary solutions to poverty reside in stringent work and personal behavior requirements on individuals.1 The law posited that requiring poor women with children to work and restricting their sexual and reproductive behaviors would lift poor families out of poverty within the federally mandated 5-year time limit and make them self-sufficient.
Many have debated the ethical and causal assumptions of this approach. We highlight another component not considered in the legislation: the health status of poor women and children. The Personal Responsibility and Work Opportunity Reconciliation Act ended the entitlement to a basic level of subsistence that President Roosevelt’s New Deal established with the Social Security Act of 1935; it also abdicated the long-standing guarantee to protect the “health and well-being” of the poor.2 This was unfortunate, given the indisputable evidence that the poor bear a disproportionate share of the disease burden in this country.3
Beyond basic data on caseload statistics, numbers of sanctions, and employment trends, the Personal Responsibility and Work Opportunity Reconciliation Act did not require states to report a great deal of information to the federal government. Thus, few national data are available with which to assess any association between the new welfare policies and the health of the target population.4
One research project that examined health issues among women who had received public assistance in 4 urban areas (Cleveland, Los Angeles, Miami, and Philadelphia) revealed that these women (and their children) had higher rates of physical and mental health problems in comparison with US women overall.5 Furthermore, working women who had left welfare (“leavers”) continued to experience these health problems and often lacked health insurance. Leavers who were not working had the worst health situations: increased health problems, no health insurance, and unmet health care needs.
Another city-level study (conducted in Boston, Chicago, and San Antonio) that collected health-related data showed that a quarter of women remaining on welfare reported health conditions that prevented them from working (as compared with 11% of former recipients and 8% of nonrecipients); also, these women experienced higher levels of depression and domestic violence compared with nonrecipients.6 Other reports of health-related problems among individuals who have had contact with the welfare system support these findings, although the primary foci of these research efforts have often been economic and, therefore, limited in their health content.7–9
In the present study, we examined whether health problems among mothers of chronically ill children affect their ability to comply with the increased emphasis on employment. Data were derived from a study conducted in San Antonio, Tex, a city with a population of just over 1 million people of primarily Hispanic (59%), non-Hispanic White (32%), and African American (7%) ethnicity.10 Nationally, Texas ranks poorly on many social and health indicators, including overall population living in poverty (10th), school-aged children living in poverty (13th), recipients of Temporary Assistance for Needy Families (TANF) (36th) and food stamps (31st) per 100 people in poverty, number of uninsured children (2nd), and per capita spending on public health (44th).11 In 1999, Texas had the highest proportion of low-income families in the country.12 Given these data, it is important to determine whether poor families that have children with chronic illnesses are experiencing other problems as well.
METHODS
Study Design
The research findings reported here were derived from the baseline round of a longitudinal study of 504 low-income mothers in San Antonio, Tex, who had children with specific chronic illnesses. A closed-ended survey instrument was administered in person by trained bilingual (English and Spanish) interviewers who recruited women at 1 of 8 clinical sites or 1 of 2 welfare agency sites. The survey required approximately 40 minutes to administer.
We undertook special efforts to recruit participants from diverse sites (e.g., walk-in clinics, inpatient wards, a private pediatric office, a public hospital, and TANF job centers) and thus minimize bias in the sample. Eligible participants were the parents (usually the mothers) or primary caretakers of children aged 2 to 12 years with 1 of 7 diagnoses: asthma, diabetes, hemophilia, sickle-cell anemia, cystic fibrosis, seizure disorder, or cerebral palsy (or other serious neurological impairment).
Survey Instrument
The primary purpose of the survey instrument was to elicit information on the health status of the child and the child’s mother (or primary caretaker), health insurance, and TANF status. The survey also collected information on employment, child care, mental health, domestic violence, substance use, and demographics. Although we were able to include previously validated items in some survey sections (e.g., mental health [SF-36 Health Survey13,14], domestic violence [Partner Violence Screen15], and substance abuse [CAGE16,17]), our unique focus required us to develop many measures specifically for this study (e.g., different categories of welfare status, association between health and compliance with welfare and work requirements). We hoped that including questions about depression, domestic violence, and substance use would help us augment findings from previous work on the prevalence of these problems among the poor.18
The principal researchers, clinicians, and San Antonio health department personnel extensively reviewed and revised the survey. It was translated into Spanish as part of a multistep process involving native speakers from the area. The survey was pretested in a clinical setting and further revised to clarify areas of misunderstanding and to incorporate potential responses to field-coded questions not previously considered by the researchers.
Data Collection
Informed consent was obtained from study participants, who were interviewed while waiting for their children’s medical appointments (at the clinical sites) or their own caseworker appointments (at the TANF sites). All of the recruitment sites provided a special location in which to conduct the interviews (e.g., an empty examination room or office) so as to ensure privacy. Participants received a $10 gift card to the local grocery store as thanks for their time and participation. The study coordinator conducted standard validation analyses of 10% of the completed surveys.
Measures and Data Analysis
The findings presented here focus on prevalence of maternal health problems according to welfare status. Welfare status was defined as follows: no contact with the welfare system (had never applied for or received benefits), denied (had applied for but been denied benefits), pending (had pending applications), former (had received benefits in the past), or current (receiving benefits at the time of enrollment in the study). We present bivariate associations between maternal health and welfare status and employment and examine important employment issues such as availability of child care and ability to keep medical appointments. We conducted multivariate analyses of predictors of welfare status and employment barriers related to health using binary and multinomial logistic regression.
RESULTS
Sample Description
Most (91%) of the survey respondents were mothers of children with chronic illnesses, and the remainder consisted of other female relatives, legal guardians, or fathers. (Because the overwhelming majority of respondents were either mothers or other female caretakers, the analyses described subsequently collectively refer to mothers, maternal health, and so forth.) The mean age was 31 years. This was a low-income sample, with more than half of the respondents reporting a monthly income of less than $1000 for a mean household of 4.7 people.
Fewer than half of the respondents had health insurance, and about one third (35%) did not have high school diplomas. The majority of the sample reported Hispanic ethnicity, being born in the United States, and English as their primary language. About half of the respondents were married or cohabiting, and half were single or separated. Although fewer than half were employed at the time of enrollment in the study, two thirds of those not working had been employed in the previous 3 years.
As can be seen in Table 1 ▶, distinct profiles were revealed among those who had never been welfare recipients, those currently receiving TANF, those who had left TANF (or were denied benefits), and those applying for TANF. Respondents with any TANF experience (i.e., applicants, current or former recipients) were more likely than nonrecipients to have monthly incomes below $1000. Similarly, those with any TANF experience had significantly fewer years of schooling than did nonrecipients. Compared with TANF nonrecipients, more of the women with any TANF experience were single or separated. The proportions of Spanish speakers among TANF nonrecipients and recipients did not differ, although former recipients were less likely to be Spanish speakers.
TABLE 1.
—Sociodemographic Characteristics of Mothers of Chronically Ill Children, by Welfare Status (n = 504)
| Total | Nonrecipients | Current | Former | Denied | Pending | |
| Age, y, mean | 31.0 | 30.6 | 31.8 | 30.3 | 32.3 | 32.0 |
| Income less than $1000/mo, %*** | 53.6 | 36.2 | 79.0 | 64.7 | 52.8 | 80.0 |
| Highest grade, y, mean** | 11.3 | 11.6 | 10.1 | 11.5 | 10.9 | 10.5 |
| No high school diploma, % | 35.2 | 32.9 | 54.0 | 26.7 | 37.7 | 40.5 |
| Marital status, %*** | ||||||
| Single/separated | 49.1 | 30.7 | 68.3 | 64.2 | 54.7 | 69.0 |
| Cohabitating/married | 35.6 | 54.7 | 17.5 | 20.0 | 28.3 | 35.6 |
| Cohabitating/not married | 15.3 | 14.7 | 14.3 | 15.8 | 17.0 | 15.3 |
| Ethnicity, %*** | ||||||
| Mexican American/Hispanic | 62.2 | 69.2 | 62.3 | 45.0 | 78.8 | 53.7 |
| Black/African American | 22.0 | 12.2 | 23.0 | 39.2 | 17.3 | 29.3 |
| White/Anglo | 10.5 | 14.9 | 4.9 | 7.5 | 3.8 | 12.2 |
| Other | 5.3 | 3.6 | 9.8 | 8.3 | 0 | 29.3 |
| Born in US, % | 87.9 | 84.1 | 87.3 | 93.3 | 86.8 | 95.2 |
| Primary language Spanish, %* | 11.5 | 15.1 | 15.9 | 3.3 | 13.2 | 7.1 |
| Household size, mean No. | 4.7 | 4.7 | 4.8 | 4.5 | 5.1 | 5.0 |
| Employment,% | ||||||
| Currently employed*** | 42.3 | 48.2 | 17.5 | 47.5 | 52.8 | 19.0 |
| Worked in past 3 years** | 63.2 | 53.8 | 51.9 | 81.0 | 68.0 | 76.5 |
| Tried or wanted to work in past 3 years | 40.2 | 31.5 | 52.0 | 58.3 | 50.0 | 25.0 |
| Has health insurance, %*** | 45.7 | 40.3 | 85.7 | 34.5 | 47.2 | 45.2 |
| Hardship in past 6 months | ||||||
| Hardship scale, mean score** | 1.2 | 1.0 | 1.3 | 1.4 | 1.3 | 1.7 |
| Any hardship, %*** | 60.8 | 47.1 | 69.4 | 71.7 | 71.7 | 76.2 |
| Housing, % | 32.1 | 28.8 | 27.0 | 36.7 | 32.1 | 45.2 |
| Food, %*** | 40.3 | 29.6 | 44.4 | 50.8 | 45.3 | 54.8 |
| Telephone service cut off, %* | 25.6 | 19.9 | 27.0 | 27.5 | 34.0 | 38.1 |
| Utility service cut off, % | 14.7 | 12.4 | 17.5 | 15.8 | 18.9 | 14.3 |
*P < .05; **P < .01; ***P < .001 (across-group trends).
About half of nonrecipients, former recipients, and denied applicants were currently employed, as compared with fewer than one fifth of current recipients and pending applicants. Notably, those with any experience with the welfare system reported higher levels of previous employment than did nonrecipients. Current TANF recipients had approximately double the rate of health insurance coverage of the total sample and of most of the other comparison groups; only about one third of former TANF recipients had health insurance. There were significantly higher rates of food insecurity and telephone disconnection among those with any TANF experience than among nonrecipients.
Maternal Health According to Welfare and Employment Status
Maternal health measures included specific health problems, limitations in activity due to health, depression, domestic violence, substance use, and emergency department visits (although also considered a measure of health care use, emergency department visits were used here as a marker of poor health). Mothers without TANF experience reported significantly better physical and mental health and lower rates of domestic violence and substance use than did those with any TANF experience (Table 2 ▶). Although current and former TANF recipients had substantially higher rates of these health problems, the highest rates were seen among those who had applied for TANF and been denied benefits or for whom decisions were pending. The pending group had the lowest mental health scores (i.e., highest levels of depression) and the highest rates of domestic violence and emergency department visits.
TABLE 2.
—Health-Related Characteristics of Mothers of Chronically Ill Children, by Welfare and Employment Status (n = 504)
| TANF Status | Currently Working | |||||||
| Variable | Total | Nonrecipients | Current | Former | Denied | Pending | No | Yes |
| General health | ||||||||
| Routinely suffer from any of 9 chronic health conditions, % | 70.5 | 60.8** | 82.0 | 76.1 | 78.4 | 77.5 | 71.9 | 68.4 |
| Health problems, mean No. | 1.7 | 1.2*** | 2.4 | 1.9 | 2.3 | 2.1 | 1.9** | 1.5 |
| Health problems make activities of daily living difficult, % | 63.6 | 56.3† | 72.0 | 63.7 | 63.4 | 81.3 | 68.3* | 56.7 |
| One or more visits to emergency department in past 6 months, % | 33.7 | 26.1** | 42.9 | 35.0 | 39.6 | 50.0 | 37.5* | 28.6 |
| Depression | ||||||||
| Depression, mean score | 61.6 | 67.6*** | 57.0 | 58.7 | 56.5 | 51.3 | 59.4* | 64.8 |
| Routinely suffer from depression, % | 26.5 | 16.4*** | 42.9 | 28.6 | 32.1 | 42.9 | 33.0*** | 17.5 |
| Experienced domestic violence, % | 23.8 | 16.4* | 31.7 | 29.5 | 28.6 | 40.0 | 27.4†† | 18.7 |
| Any negative mention of alcohol/drug use, % | 14.7 | 9.4* | 17.5 | 20.0 | 20.8 | 16.7 | 16.2 | 12.8 |
*P < .05; **P < .01; ***P < .001; †P < .06; ††P = .052 (across-group trends).
Similarly, there were significant differences on most health measures between respondents who were and were not employed at the time of enrollment in the study. Those who were not working reported more health problems, limitations due to health, emergency department visits, depression, and domestic violence than did those who were employed (Table 2 ▶).
Respondents were asked if they were currently employed, if they had been employed within the past 3 years , or if they wanted or tried to work during this time. When factors related to finding employment, work absenteeism, and job loss were assessed, significantly higher rates of health-related (own, child’s, and other family member) and child-care–related barriers were reported for most measures among those not currently working. Specifically, compared with employed women, significantly more women not currently employed reported that they experienced difficulty finding work owing to their own health (36% vs. 64%), their child’s health (44% vs. 56%), and lack of child care (38% vs. 62%). Similarly, greater job loss among currently unemployed women was associated with their own health (34% vs. 70%), their child’s health (43% vs. 57%), the health of another family member (38% vs. 62%), and lack of child care (41% vs. 59%). Moreover, examination of the effect of these barriers to finding work showed that rates of barriers regarding child health (72% vs 34%), mother’s own health (42% vs 20%), the health of other family members (30% vs 15%), and child care (58% vs 40%) were twice as high among those who had wanted or tried to work in the previous 3 years but had not been able to do so as among those who were currently employed or who had been employed within the past 3 years (data not shown).
Child Care and Ability to Keep Medical Appointments According to Welfare Status
We examined the often conflicting relationships of employment responsibilities with child care responsibilities and medical appointment attendance for children with chronic illnesses (data not shown). Those with current (40%), former (33%), denied (49%), and pending (40%) TANF status reported significantly more difficulty obtaining child care because of their child’s health problems in comparison with nonrecipients (23%; P < .01). Similarly, nonrecipients were significantly less likely to miss children’s medical appointments because of work or school (19%) than were those in the current (25%), former (35%), denied (55%), and pending (23%) groups (P < .01).
Predictors of Health-Related Employment Barriers
We conducted multivariate analyses assessing maternal health barriers to employment. Three binary logistic regression models examined whether mothers’ health status (1) made it difficult to find a job, (2) caused missed work days, or (3) caused loss of a job. Variables included measures of maternal health and child health, maternal and child health insurance status, and relevant demographic factors (Table 3 ▶). In the first model, worse maternal mental health (i.e., depression), more other maternal health problems, and more maternal visits to the emergency department were significantly associated with difficulty finding a job. Specifically, increased maternal health problems and emergency department visits were associated with a 50% increase in emergency department visits and with a fivefold increase, respectively, in difficulty finding work. Mothers who had less education, were older, and were members of households with fewer people also had more difficulty finding work. Maternal health problems and visits to the emergency department were 2 and 4 times more likely, respectively, to be associated with job absenteeism (model 2). Depression, maternal health problems, and lack of maternal health insurance all were associated with a greater likelihood of job loss (model 3).
TABLE 3.
—Binary Logistic Regression Models Predicting Maternal Health Barriers to Employment Among Mothers of Chronically Ill Children
| Health Made It Hard to Find a Job, OR (95% CI) | Health Caused Missed Work Days, OR (95% CI) | Health Caused Loss of Job, OR (95% CI) | |
| Maternal mental health | 0.97** (0.95, 0.99) | 0.99 (0.97, 1.0) | 0.97* (0.95, 0.99) |
| Domestic violence experienced | 0.83 (0.3, 2.1) | 0.77 (0.3, 1.7) | 1.91 (0.8, 4.6) |
| Substance use reported | 0.67 (0.2, 2.1) | 1.35 (0.6, 3.1) | 0.70 (0.2, 2.1) |
| Maternal health problems | 1.47** (1.1, 1.9) | 2.09*** (1.6, 2.7) | 1.44** (1.1, 1.9) |
| Mother visited emergency department (past 6 mo) | 4.99*** (2.1, 11.7) | 4.31*** (2.2, 8.4) | 1.61 (0.7, 3.8) |
| Child activities limited | 1.35 (0.5, 3.4) | 0.74 (0.4, 1.5) | 0.98 (0.4, 2.4) |
| Child visited emergency department (past 6 mo) | 1.12 (0.4, 2.9) | 1.01 (0.5, 2.1) | 1.60 (0.6, 4.3) |
| Child visited hospital (past 6 mo) | 0.91 (0.3, 2.7) | 0.67 (0.3, 1.6) | 1.75 (0.7, 4.6) |
| Child has health insurance | 0.99 (0.3, 3.0) | 0.85 (0.3, 2.1) | 1.63 (0.5, 5.0) |
| Mother has health insurance | 0.70 (0.3, 1.7) | 0.78 (0.4, 1.6) | 0.39* (0.15, 0.98) |
| Harder to find child care | 1.08 (0.4, 2.6) | 1.37 (0.7, 2.8) | 1.15 (0.5, 2.8) |
| Years of education | 0.76* (0.6, 0.9) | 1.17 (1.0, 1.4) | 0.94 (0.8, 1.2) |
| Married (vs single) | 0.93 (0.5, 1.6) | 0.77 (0.5, 1.1) | 1.19 (0.7, 2.0) |
| Born in US | 0.77 (0.1, 4.7) | 1.18 (0.3, 5.2) | 2.87 (0.4, 21.6) |
| Age (of mother) | 1.08* (1.01, 1.15) | 1.03 (1.0, 1.1) | 1.05 (1.0, 1.1) |
| Speak English/both most comfortably | 2.01 (0.3, 13.3) | 3.75 (0.8, 18.4) | 0.42 (0.1, 2.5) |
| No. of household members | 0.76* (0.6, 0.99) | 1.05 (0.9, 1.3) | 0.84 (0.7, 1.1) |
| Income (>$1000 mo) | 0.86 (0.3, 2.3) | 0.91 (0.4, 1.9) | 0.57 (0.2, 1.5) |
| Model χ2 (df) | 79.86*** (18) | 102.99*** (18) | 46.82*** (18) |
Note. OR = odds ratio; CI = confidence interval.
*P < .05; ** P < .01; ***P < .001.
Predictors of Welfare Status
Multinomial logistic regression was used to test correlates of TANF status, categorized as current/former recipients vs nonrecipients (model 1) and denied/pending applicants vs nonrecipients (model 2) (Table 4 ▶). In model 1, having maternal health insurance, being single or separated, and having a monthly income of less than $1000 were associated with current or former TANF status. In model 2, maternal health problems, child health–related limitations in daily activities, lack of child health insurance, and being single or separated were associated with having applied for TANF. Specifically, women with more health problems were 25% more likely to apply for TANF, women without health insurance for their children were more than 2.5 times more likely to apply, and women with children whose activities were limited were 60% more likely to apply.
TABLE 4.
—Multinomial Logistic Regression Model Predicting Temporary Assistance for Needy Families Status Among Mothers of Chronically Ill Children
| Current/Former vs Nonrecipients, OR (95% CI) | Denied/Pending vs Nonrecipients, OR (95% CI) | |
| Maternal mental health | ||
| Less depressed | 1.59 (0.7, 3.3) | 1.50 (0.6, 3.7) |
| More depressed (reference) | 0.00 | 0.00 |
| Domestic violence | ||
| No experience | 0.53 (0.2, 1.2) | 0.42 (0.2, 1.1) |
| Experience (reference) | 0.00 | 0.00 |
| Substance use | ||
| No report | 0.50 (0.2, 1.1) | 0.49 (0.2, 1.3) |
| Experience with substance use (reference) | 0.00 | 0.00 |
| Maternal health problem scale score | 1.16 (1.0, 1.4) | 1.27* (1.02, 1.59) |
| Maternal visits to emergency department (past 6 mo) | ||
| None | 1.21 (0.7, 2.2) | 0.83 (0.4, 1.8) |
| One or more (reference) | 0.00 | 0.00 |
| Child health affects activities | ||
| No | 0.75 (0.4, 1.4) | 0.38* (0.2, 0.9) |
| Yes (reference) | 0.00 | 0.00 |
| Child visits to emergency department (past 6 mo) | ||
| None | 0.86 (0.5, 1.6) | 0.90 (0.4, 2.0) |
| One or more (reference) | 0.00 | 0.00 |
| Child hospitalizations (past 6 mo) | ||
| None | 0.90 (0.5, 1.8) | 2.30 (0.9, 5.9) |
| One or more (reference) | 0.00 | 0.00 |
| Child health insurance | ||
| No | 1.39 (0.6, 3.1) | 2.62* (1.02, 6.72) |
| Yes (reference) | 0.00 | 0.00 |
| Maternal health insurance | ||
| No | 0.44** (0.2, 0.8) | 0.59 (0.3, 1.3) |
| Yes (reference) | 0.00 | 0.00 |
| Impact of child’s health on child care | ||
| None | 0.65 (0.3, 1.3) | 0.51 (0.2, 1.1) |
| Made it harder to find (reference) | 0.00 | 0.00 |
| Years of education | 0.97 (0.9, 1.1) | 0.91 (0.8, 1.1) |
| Marital status | ||
| Single/separated | 4.39*** (2.2, 8.8) | 3.00* (1.2, 7.2) |
| Married (reference) | 0.00 | 0.00 |
| Born in US | ||
| No | 1.03 (0.3, 3.2) | 0.77 (0.1, 3.9) |
| Yes (reference) | 0.00 | 0.00 |
| Age (mother) | 1.02 (1.0, 1.1) | 1.05 (1.0, 1.1) |
| Language spoken most comfortably | ||
| Spanish | 0.64 (0.2, 2.1) | 0.53 (0.1, 2.9) |
| English or both (reference) | 0.00 | 0.00 |
| No. of household members | 1.09 (0.9, 1.3) | 1.18 (1.0, 1.4) |
| Income | ||
| < $1000/mo | 2.66** (1.5, 4.9) | 1.74 (0.8, 3.8) |
| ≥ $1000/mo (reference) | 0.00 | 0.00 |
| Model χ2 (df) | 124.49*** (40) | |
Note. OR = odds ratio; CI = confidence interval.
*P < .05; **P < .01; ***P < .001.
DISCUSSION
To examine the potential association between health and welfare in the context of recent welfare policy changes, we studied women who had children with chronic illnesses. We found that women who had had any contact with the TANF system (i.e., the current, former, denied, and pending groups) were significantly more likely to report health problems and recent visits to the emergency department in comparison with nonrecipients. In addition, 2 separate mental health measures revealed worse depression scores and increased prevalence of depression in this group overall (and, most saliently, among the pending group). The same pattern was seen in reports of domestic violence, with 40% of the pending group affected. This group also reported the greatest material hardship, particularly insufficient amounts of food and loss of telephone service.
Although it is not surprising that health problems and barriers to work associated with health and child care were reported more frequently among those not currently employed, the higher prevalence of jobs lost because of health problems and lack of child care should be cause for alarm. Conversely, the increased difficulty in keeping children’s medical appointments, particularly among the former recipient and denied applicant groups, may indicate that health needs are not being adequately addressed in these families. Their health appears to hamper their ability to obtain and maintain employment, and their employment appears to hamper their ability to adequately address their health needs. This predicament was most apparent among those who had not worked in the past 3 years but had wanted or tried to work. Their reports of having twice as many health and child care barriers to employment as do those currently or previously employed indicate that health problems are serious impediments to employment.
Maternal health problems (including depression) were clearly associated with difficulty finding work after adjustment for child health status, child health insurance, child care, and demographic variables. Maternal health variables were also associated with missed work and job loss. The association of lack of maternal health insurance with lost employment is dramatic, especially because findings from state welfare “leaver studies” have indicated that need for health insurance is a top reason for TANF recidivism.19
The persistence of maternal health insurance status as a predictor of current or former TANF status provides additional evidence that need for health insurance is an important factor in TANF application. That low income is associated with TANF status is logical; similarly, the association between being single or separated and TANF status probably represents the long-standing prohibition against providing benefits to married women. Health is a factor in applying for assistance, because maternal health problems, children’s health-related limitations in daily activities, and lack of child health insurance are all associated with having applied for TANF. In the multivariate analysis, low income was not a significant factor for applicants, which suggests that there may be a group that does not meet the income eligibility requirements, yet still has special health needs and may therefore apply for TANF because of the need for health insurance.
Our findings have limited generalizability in that they are based on a sample of mothers of chronically ill children from 1 urban center. However, the finding of increased rates of health problems among families receiving welfare is consistent with other reports.20,21 The cross-sectional nature of these data does not permit conclusions concerning causality; however, the follow-up round of data collection is under way. Information on maternal and child health was based on the mother’s self-report; thus, validity may be an issue owing to imperfect recall (e.g., estimates of medical visits, missed school days). Respondents were recruited at both clinical and TANF centers, and thus the study participants differed in that half were seeking care for their child at the time of the interview and half were not (although the latter were required to have a child with 1 of the same 7 diagnoses). However, this sampling approach minimized sample bias by allowing for the inclusion of a broad group of individuals who might have contact with the welfare system.
CONCLUSIONS
When the Personal Responsibility and Work Opportunity Reconciliation Act was passed in 1996, the major goal of policymakers was to move the poor “from welfare to work.” Almost immediately, the TANF caseload began to drop; it is currently less than half of what it was in 1996.22 This decline was accompanied by a drop in Medicaid caseloads, which has left many without health insurance.23 Although many did leave for employment,19 others left as a result of sanctions,24 confusion about the new policies, time limits, or diversion policies.25 Proponents have interpreted the absolute drop in welfare caseloads as a de facto success,26,27 but others have asked the following obvious questions: Are the leavers really working? Have their incomes improved? Are they still living in poverty? Are they able to care for their children? Are their children faring better economically and socially? Do women and children have health insurance? Do women have health or social problems that may interfere with their ability to obtain employment and become self-sufficient?28,29
Despite the “creaming” that occurred shortly after the law passed in 1996—with the most employable leaving TANF for work first—various studies have reported adverse outcomes among leavers, such as lack of employment, incomes at or below the poverty level, lack of work-related benefits, reductions in child supervision, poor educational outcomes among adolescents,30 and return to TANF.31 Our report clearly documents (1) the increased prevalence of health problems among the poor, (2) the association between poor maternal health and need for cash assistance and health insurance, and (3) maternal health barriers to employment and job retention (for details on the impact of poor child health on maternal employment, see the Smith et al.32 and Wood et al.33 articles elsewhere in this issue).
Few require convincing of the correlation between health and work. Unfortunately, in their zeal to promote the spirit of the Personal Responsibility and Work Opportunity Reconciliation Act and get the poor to work, many states have overlooked 2 related barriers to employment: health problems and lack of health insurance. Our data from San Antonio suggest that women who have sick children (but who have relatively higher incomes) may apply for TANF partly because they need health insurance. Our findings clearly indicate that being poor is associated with health problems and that health problems, in turn, are associated with difficulty in finding employment. Policymakers should recognize that social policies that narrowly promote employment are destined to fail if they do not take into account the close association between health and work.
Acknowledgments
This research was supported in part by funding from the Ford Foundation, the General Service Foundation, the Maternal and Child Health Bureau, the Moriah Fund, the Office of Population Affairs and the Open Society Institute.
We gratefully acknowledge Fernando Guerra, MD, MPH, William Parry, MD, Steven Enders and the administrators at the San Antonio Texas Works offices, without whose assistance we would not have been able to conduct this research. We would also like to express our appreciation to Monica Trevino, the San Antonio site coordinator; Tammy Draut, Lauren Oshman, Naomi Lince, and Julia Choe for overall study coordination; the various clinical and agency site personnel; our dedicated interviewers; and the families and children who so generously agreed to participate in this important project. Thanks to the many reviewers for their thoughtful suggestions, to Rebecca Shoai for her assistance with background research, and to Ann Gavaghan for coordinating the review and submission of this cluster of articles.
Human Participant Protection This research was approved by the institutional review boards of CHRISTUS Santa Rosa Hospital, the University of Texas Health Science Center, and the University Health Center (Downtown), San Antonio; the Columbia-Presbyterian Medical Center; and the Boston University Medical Center. Signed informed consent was obtained from all participants.
Peer Reviewed
D. Romero, W. Chavkin, P. H. Wise, L. A. Smith, and P. R. Wood contributed to the development of the hypotheses, to the planning of the study, and to the preparation of the article. D. Romero analyzed the data and wrote the article.
References
- 1.Personal Responsibility and Work Opportunity Reconciliation Act, Pub L No. 104-193 (1996).
- 2.US Social Security Administration. Preamble to the Social Security Act of 1935. Available at: http://www.ssa.gov/history/35actpre.html. Accessed December 17, 2001.
- 3.Raphael D. Health inequalities in the United States: prospects and solutions. J Public Health Policy. 2000;21:394–427. [PubMed] [Google Scholar]
- 4.Klerman LV. Child health and well-being. In: Besharov DJ, ed. Family Well-Being After Welfare Reform. College Park, Md: University of Maryland; 2002:9-11–9-22.
- 5.Polit DF, London AS, Martinez JM. The Health of Poor Urban Women: Findings From the Project on Devolution and Urban Change. New York, NY: Manpower Demonstration Research Corp; 2001.
- 6.Moffitt R, Cherlin A. Disadvantage Among Families Remaining on Welfare. Chicago, Ill: Joint Center for Poverty Research, Northwestern University/University of Chicago; 2000.
- 7.Ahluwalia SK, McGroder SM, Zaslow MJ, Hair EC. Symptoms of Depression Among Welfare Recipients: A Concern for Two Generations. Washington, DC: Child Trends; 2001.
- 8.Danziger S, Corcoran M, Danziger S, et al. Barriers to the Employment of Welfare Recipients. Ann Arbor, Mich: University of Michigan, Poverty Research and Training Center; 2000.
- 9.The CALWORKS Project: Overcoming Mental Health, Alcohol and Other Drugs and Domestic Violence Barriers to Employment. Six County Case Study: Moving Beyond Implementation to Identification and Service. Sacramento, Calif: California Institute for Mental Health; 2001. Project report 2.
- 10.Selected Health Demographics and Health Indicators. Washington, DC: US Bureau of the Census; 2000.
- 11.Center for Public Policy Priorities. Recent changes in Texas rankings. Available at: http://www.cppp.org. Accessed April 10, 2002.
- 12.The Snapshot of Texas. Washington, DC: Urban Institute; 1999.
- 13.Hemingway H, Nicholson A, Stafford M, Roberts R, Marmot M. The impact of socioeconomic status on health functioning as assessed by the SF-36 questionnaire: the Whitehall II study. Am J Public Health. 1997;87:1484–1490. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Ayuso-Mateos JL, Lasa L, Vazquez-Barquero JL, Oviedo A, Diez-Manrique JF. Measuring health status in psychiatric community surveys: internal and external validity of the Spanish version of the SF-36. Acta Psychiatr Scand. 1999;99:26–32. [DOI] [PubMed] [Google Scholar]
- 15.Feldhaus KM, Koziol-McLain J, Amsbury HL, Norton IM, Lowenstein SR, Abbott JT. Accuracy of 3 brief screening questions for detecting partner violence in the emergency department. JAMA. 1997;277:1357–1361. [PubMed] [Google Scholar]
- 16.Samet JH, O’Connor PG. Alcohol abusers in primary care: readiness to change behavior. Am J Med. 1998;105:302–306. [DOI] [PubMed] [Google Scholar]
- 17.Fiellin DA, Reid MC, O’Connor P. Screening for alcohol problems in primary care: a systematic review. Arch Intern Med. 2000;160:1977–1989. [DOI] [PubMed] [Google Scholar]
- 18.Raphael J. Saving Bernice: Battered Women, Welfare and Poverty. Boston, Mass: Northeastern University Press; 2000.
- 19.Welfare Reform: Information on Former Recipients’ Status. Washington, DC: US General Accounting Office, Health, Education, and Human Services Division; 1999. GAO/HEHS publication 99-48.
- 20.Smith LA, Hatcher JL, Wertheimer R. The association of childhood asthma with parental employment and welfare receipt. J Am Med Womens Assoc. 2002;57:11–15. [PubMed] [Google Scholar]
- 21.Lennon MC, Blome J, English K. Depression among women on welfare: a review of the literature. J Am Med Womens Assoc. 2002;57:27–31. [PubMed] [Google Scholar]
- 22.US Dept of Health and Human Services, Administration for Children and Families. TANF recipients. Available at: http://www.acf.dhhs.gov/news/stats/recipients.htm. Accessed April 12, 2002.
- 23.Chavkin W, Romero D, Wise PH. State welfare reform policies and declines in health insurance. Am J Public Health. 2000;90:900–908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Welfare Reform: State Sanction Policies and Number of Families Affected. Washington, DC: US General Accounting Office, Health, Education, and Human Services Division; 2000. GAO/HEHS publication 00-44.
- 25.Maloy KA, Pavetti LA, Shin P, Darnell J, Scarpulla-Nolan L. A Description and Assessment of State Approaches to Diversion Programs and Activities Under Welfare Reform. Washington, DC: George Washington University Medical Center, Center for Health Policy Research; 1998.
- 26.Implementation of Welfare Reform Work Requirements and Time Limits: Hearings Before the Subcommittee on Human Resources of the House Committee on Ways and Means, 107th Cong, March 5, 2002 (opening statement of Wally Herger).
- 27.Toner R, Pear R. Bush urges work and marriage programs in welfare plan. New York Times. February 27, 2002:A18.
- 28.TANF Reauthorization Act of 2001, HR 3113, 107th Cong (2001) (legislation proposed by Rep. Mink).
- 29.Next Step in Reforming Welfare Act, HR 3625, 107th Cong (2002) (legislation proposed by Rep. Cardin).
- 30.Brooks JL, Hair EC, Zaslow MJ. Welfare Reform’s Impact on Adolescents: Early Warning Signs. Washington, DC: Child Trends; 2001.
- 31.Greenberg M, Richer E, Sankarapandian V. Welfare caseloads are up in most states. Available at: http://www.clasp.org/pubs/TANF/Fyo1%20Caseload%20Data.htm. Accessed January 4, 2002.
- 32.Smith LA, Romero D, Wood PR, Wampler NS, Chavkin W, Wise PH. Employment barriers among welfare recipients and applicants with chronically ill children. Am J Public Health. 2002;92:1453–1457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Wood PR, Smith LA, Romero D, Bradshaw P, Wise PH, Chavkin W. Relationships between welfare status, health insurance status, and health and medical care among children with asthma. Am J Public Health. 2002;92:1446–1452. [DOI] [PMC free article] [PubMed] [Google Scholar]