Low-income adults with mental disorders may receive income support to help cover their basic needs through two federally supported programs, Temporary Assistance for Needy Families (TANF) and Supplemental Security Income (SSI). Although individuals must have a medically determined disability and be unable to engage in substantial gainful activity to qualify for SSI, they have to demonstrate only financial need and the presence of children to qualify for TANF benefits. However, recent studies indicate that mental disorders are prevalent among TANF recipients. One-fourth to one-third of current recipients have symptoms associated with a mental condition (1,2). Estimates differ depending on how disorders are defined and measured and on the data collection strategy and research methodology used. One study that examined specific types of mental disorders found major depression to be the most common mental disorder (26.7%), followed by posttraumatic stress disorder (14.6%) and generalized anxiety (7.3%) (3).
In recent years there has been increasing interest in exploring whether adult TANF beneficiaries with a mental disorder may be better served by SSI or by a new program for individuals with a temporary disability (4). This interest arises out of several concerns. First, TANF was not designed to serve individuals with mental disorders. Few TANF staff members have the skills to identify individuals with a mental disorder, and many are not familiar with the mental health programs available (5,6). Second, TANF recipients with mental disorders have low rates of employment (7,8,9). States are therefore at risk of having financial penalties imposed upon them for not meeting required work participation rates embedded within the TANF program. Finally, the SSI program provides a more stable income source for individuals with chronic conditions. TANF recipients are subject to a federal five-year lifetime limit on benefits—a period that may not be sufficient for individuals with mental disorders (5).
A question that often arises but for which data have not been readily available is, how many TANF recipients with mental disorders have work limitations that are severe enough to qualify them for SSI? The answer to this and related questions can help inform discussions on how to best address the needs of TANF recipients with mental disorders. In this study, we used a unique data set of Medicaid claims to address the following questions: How many TANF recipients have diagnosed mental disorders and are using mental health services? How do the characteristics and types of mental disorders of TANF recipients compare with those of SSI recipients? And how many TANF recipients have levels of service use comparable with those of the SSI population?
The tables in this article are based on data from the 2003 Medicaid Analytic eXtract (MAX) files (the most recent files available at the start of the study) in four states: Illinois, Maine, Texas, and Vermont. The MAX data contain information on beneficiaries, service use, prescription drugs, and expenditures in state Medicaid programs. All claims, except those for prescription drugs, indicate a primary diagnosis based on ICD-9-CM codes, the amount of the combined state and federal Medicaid reimbursement, and the type of service that generated the claim. The service codes are used to identify treatment settings (such as emergency rooms) and to distinguish psychiatric from general hospitals.
Data from beneficiaries in fee-for-service care for at least one month of the year were included in this analysis. Because these analyses are based solely on administrative data, institutional review board approval was not obtained. We used fee-for-service Medicaid claims because they yield more complete information on service use than data from Medicaid managed care organizations. For beneficiaries enrolled for part of the year in Medicaid managed care and part of the year in a fee-for-service system, we excluded data for the months of managed care enrollment.
The study population includes beneficiaries with a primary diagnosis of a mental disorder on at least one fee-for-service Medicaid claim during the year and those treated in a psychiatric facility during the year. The diagnoses used to identify beneficiaries with mental disorders, identified primarily by the first three digits of ICD-9-CM codes, are those considered by most payers to be mental disorders (10) (listed in Table 1). The tables in this article are restricted to female adults aged 19 to 64 because that group represents the majority of TANF recipients subject to work requirements; all beneficiaries who met our selection criteria were included in the analysis.
SSI and TANF recipients were grouped by MAX uniform eligibility codes. Beneficiaries were considered to be SSI recipients if their eligibility code indicated that they were receiving cash assistance and had met criteria for being aged, blind, or disabled for at least one month. Although this should have captured almost everyone receiving federal SSI benefits, it may also have captured a small number of non-SSI beneficiaries in state grant supplement programs. Beneficiaries were identified as TANF recipients if they were coded in the MAX data for at least one month as categorically needy as specified in Section 1931 of the Social Security Act. Although nearly all TANF recipients receive Medicaid through Section 1931, many non-TANF recipients may be eligible for Medicaid through Section 1931. To ensure that our sample contained mostly TANF recipients, we excluded from the analysis states in which the number of Section 1931 Medicaid beneficiaries was more than 5 percentage points higher than the TANF population in the state. We also excluded states in which more than 33% of the Medicaid beneficiaries were enrolled in Medicaid managed care plans and in which the MAX data did not meet quality standards related to mental health and substance abuse services (10). The four states in our analysis are those that remained after these exclusions.
We identified beneficiaries with a costly physical health condition by using the Chronic Illness and Disability Payment System (CDPS). The CDPS categorizes diagnoses into one of 13 body systems; within these categories, diagnoses are sorted into low-, medium-, high-, or very-high-cost diagnoses. Beneficiaries with costly physical health conditions were defined as beneficiaries with diagnoses (excluding the psychiatric and substance use categories) defined by the CDPS as medium to very high cost, such as congestive heart failure and breast cancer (11). Cases where a beneficiary had at least one claim with a primary mental disorder diagnosis and at least one claim with a substance use disorder diagnosis were classified as having co-occurring mental and substance use disorders.
Across the four states, 13% (N=22,691) of female TANF and 32% (N=65,303) of female SSI beneficiaries were identified as having mental disorders. Compared with SSI beneficiaries, TANF beneficiaries with these disorders were younger and more often white or Hispanic (Table 1). Compared with all TANF recipients, those with mental disorders were slightly younger and substantially more often white (results not shown): 59% of TANF recipients with a mental disorder were white, compared with 36% of all TANF recipients. We observed a similar pattern for SSI recipients, although differences in the racial distribution were much less striking than those among TANF recipients (55% of SSI recipients with a mental disorder were white, compared with 47% of all SSI recipients).
Among TANF beneficiaries, the most prevalent diagnoses included neurotic and other depressive disorders (42%) and major depression and affective psychoses (30%). Few TANF beneficiaries had schizophrenia or other psychoses (1%); in contrast, 23% of SSI beneficiaries had these disorders. About one-quarter (24%) of SSI recipients were diagnosed as having neurotic and other depressive disorders, and 41% had major depression and affective psychoses.
Beneficiaries with mental disorders had other challenges as well. Almost one in ten TANF beneficiaries (8%) had a physical comorbid condition that was considered costly. The percentage of those with diagnosed physical comorbid conditions among SSI beneficiaries was substantially higher, at 31%. About 3% of TANF and 4% of SSI beneficiaries had a substance use disorder. The percentage of TANF recipients with a mental disorder who were caring for a newborn (11%) was higher than the percentage of SSI recipients with a mental disorder and caring for a newborn (1%).
The extent to which TANF and SSI beneficiaries with a diagnosed mental disorder utilized mental health treatment services differed considerably. For TANF beneficiaries, treatment was mostly short term (Table 2). Almost three-quarters (73%) received treatment for three or fewer months. Compared with TANF beneficiaries, on average, SSI beneficiaries participated in treatment for twice as long, and less than half (41%) received treatment for three or fewer months. (Because we observed only one year of data, the maximum length of treatment in the analysis was 12 months.) Five percent of TANF recipients and 10% of SSI recipients used inpatient treatment to address their disorders; 11% of TANF recipients and 14% of SSI recipients had one or more emergency room visits for mental health treatment.
TANF and SSI recipients differed somewhat in the type and number of psychotropic drugs they used. Among TANF beneficiaries, 66% used antidepressants and 31% used antianxiety agents; only 13% used antipsychotic drugs. Among SSI beneficiaries, 69% used antidepressants, 47% used antipsychotics, and 39% used antianxiety agents. TANF beneficiaries were also less likely than SSI beneficiaries to rely on more than one class of psychotropic drug (32% compared with 54%).
The use of the emergency room or an inpatient facility for mental health treatment, use of one or more classes of psychotropic drugs, and the co-occurrence of a substance use disorder or physical comorbidity can be used to gauge the extent to which recipients' mental disorders may limit their ability to fully support themselves through paid work. In Table 3, these indicators classify TANF and SSI beneficiaries into three groups, representing low, moderate, and high levels of mental health service use. TANF recipients were concentrated in the low category, whereas SSI recipients were distributed more evenly across all three categories. Among TANF recipients, 60% were in the low-use category, 29% were in the moderate category, and 11% were in the high category. Among SSI recipients, the distributions were 40%, 31%, and 29% for low-, moderate-, and high-use categories, respectively. The median expenditures for the high-use TANF group were comparable with the median expenditures for all SSI recipients, suggesting that the high-use TANF group may be most comparable to SSI recipients with respect to their service use and more limited employment potential.
This study found that 13% of TANF beneficiaries were receiving treatment for diagnosed mental disorders. Common disorders included neurotic and other depressive disorders and major depression and affective psychoses. Schizophrenia and other psychoses that are relatively common in the SSI population with mental disorders were rare for the TANF population. Although this study confirmed existing research documenting the types of mental disorders among TANF recipients, the proportion of TANF recipients with diagnosed mental disorders who received mental health treatment was substantially lower than estimates of the prevalence of mental disorders among TANF recipients. This finding is consistent with the perceptions of TANF staff, who have noted that many TANF recipients with mental disorders have never been diagnosed or treated (6).
For TANF beneficiaries with a diagnosed mental disorder, mental health treatment was primarily short term. From these data, we cannot determine whether this finding is reflective of the nature of their disorder or because TANF recipients had difficulty participating in treatment for longer periods. Other studies have documented that many TANF recipients have personal and family challenges that interfere with work (such as lack of transportation or child care, cognitive impairments, low levels of education, domestic violence issues, or unstable housing) (12,13,14,15). These conditions may create obstacles to participation in regularly scheduled treatment or appropriate medication management.
Looking at levels of service use data, we found that 11% of TANF beneficiaries receiving treatment for a mental disorder may be the most likely to meet the eligibility criteria for SSI. According to their levels of service use, most TANF beneficiaries appeared to face fewer obstacles than faced by most SSI recipients. Over half of TANF recipients had low levels of service use, suggesting that with additional support to help them manage their mental disorder, they may be able to meet their TANF work requirements and eventually find paid employment. Recipients with moderate use of mental health services may need more intensive support and may benefit from modified work requirements until their conditions are stabilized. This latter group, accounting for just under a third of TANF recipients receiving treatment for a mental disorder, may pose the greatest challenge for TANF agencies because they do not appear to have limitations serious enough to meet SSI eligibility criteria but may be unable to work the 20–30 hours per week that TANF requires.
A critical challenge facing mental health providers and TANF agencies is how to implement work programs that recognize the special needs that TANF recipients with mental disorders may face. A key characteristic that distinguishes TANF from SSI recipients is that TANF recipients are raising children, often as single parents, and face all the challenges that come with parenting in addition to the challenges that arise from their mental disorder. A mental disorder in combination with other adversities may result in greater treatment and employment challenges than programs for childless or higher-income individuals typically address.
In recent years, some have questioned whether the TANF system is capable of addressing the needs of families with mental disorders, and a new work-based service system has been proposed that is designed to meet the needs of parents who experience personal and family challenges, including mental disorders, that limit their ability to work but do not create limitations serious enough to qualify them for SSI (4). These proposals acknowledge that some parents need more assistance and more time than others to make the transition to paid employment, and, for some, steady full-time employment may not be a realistic goal.
In some states, TANF agencies have cultivated partnerships with mental health systems where colocated mental health providers train TANF staff to identify the signs of potential disorders or provide clinical assessments and referrals. Such systems have not been rigorously studied but may offer opportunities for implementing evidence-based practices for TANF recipients with mental disorders. As another alternative, a new work-based service system might promote individualized employment planning based on the strengths and functional limitations of each participant. Such a system could provide necessary work-based and personal supports to TANF-eligible families and divert some from the SSI caseload.
Our study has some limitations. First, because we included Medicaid claims data from only four states, our results may not be generalizable to other states or to the country as a whole. And, owing to considerable state policy and program variation, this problem may be more acute for TANF than for SSI beneficiaries.
Second, we may have underestimated the number of TANF and SSI recipients who have a diagnosed mental disorder. Some beneficiaries with these disorders may not have received Medicaid-covered services during 2003 and therefore would have generated no claims. Some beneficiaries may have used only prescription drugs for their treatment; because these claims do not include diagnostic codes, we were not able to identify individuals with qualifying diagnoses. Other beneficiaries with mental disorders may have used services, but the mental disorder was not included on the claim as the primary diagnosis. We also do not have a completely accurate representation of TANF and SSI recipients generally. By using the MAX uniform eligibility codes, we may have coded some non-TANF and non-SSI beneficiaries as recipients (and vice versa).
Third, the analysis excluded beneficiaries enrolled in Medicaid managed care plans. Few beneficiaries were excluded from our analysis as a result of being in managed care for the entire year (only 5% in Illinois, 19% in Texas, and none in Vermont or Maine). Texas' exclusion rate was somewhat higher than in other states because it is the only state in our analysis that carves out mental health services to managed care organizations for some beneficiaries (it does so in only parts of the state and does it in the same way across all bases of eligibility). We also excluded months in which fee-for-service beneficiaries were enrolled in managed care. The number of months that we excluded varied slightly by the basis of eligibility—11% of months among disabled adults and 16% among nondisabled adults—but the difference was much lower than nationally (26% of months among disabled adults and 41% among nondisabled adults). Thus, although the managed care exclusions affected the overall number and characteristics of beneficiaries identified as having used mental health services, it should not have substantially affected our comparisons between TANF and SSI beneficiaries.
This study provides evidence to suggest that a nontrivial portion of the TANF caseload has a diagnosed mental disorder for which they are receiving treatment. The findings from the study are not indicative of the prevalence of mental disorders, because many individuals go undiagnosed and untreated. Indeed, the number of TANF recipients identified as having mental disorders in this study is lower than in previous studies that gathered data on symptoms associated with mental disorders. Among those with mental disorders in this study, some had patterns of service use indicating potential eligibility for SSI, but most appeared to have more modest limitations that could be addressed through a combination of work accommodations and more intensive work support. A key challenge facing mental health providers and TANF agencies is how to implement work programs that recognize the special needs that these TANF recipients may face.
This study was supported by contract 280-2003-00015 to Mathematica Policy Research, Inc., from the Substance Abuse and Mental Health Services Administration. The authors thank Jeffrey A. Buck, Ph.D., and Judith L. Teich, M.S.W., for their substantial assistance;Ellen Singer, M.A., and Mei-Ling Mason, M.S., for their programming support; and Henry Ireys, Ph.D., and Marilyn Ellwood, M.S.W., for their substantive and technical support.
The authors report no competing interests.