Critical Issues for Parents with Mental Illness and their Families

Chapter I

The Scope of the Issue

CRITICAL ISSUE: There are no national data on the frequency with which adults with mental illness bear and care for children. Information on prevalence is drawn from existing data sets and is, therefore, limited by study rationale and methods. Large-scale descriptive data are not available on the characteristics of children whose parents have mental illness, such as where they are living or in whose custody or care.

Who are Parents with Mental Illness? The number of parents living with mental illness is likely to be in the millions. Every year, 22% of the American population is affected by psychiatric disorder. About 3% of the adult population (approximately 5 million individuals in 1990) is diagnosed with severe psychiatric disorder, as indicated by diagnosis and disability (Goodwin et al., 1993). According to U.S. Census Bureau data, the majority of American women and men are parents (Bachu, 1995, 1996). The likelihood is high that many of the 5 million Americans diagnosed with severe mental illness each year, those most challenged by disability, already are or will become parents.

We reviewed findings from existing data sets, including the National Institute on Disability and Rehabilitation Research-funded National Survey of Parents with Disabilities, and the National Comorbidity Survey, to begin to address this question more carefully. These studies were not designed to describe the circumstances of families or characteristics of children in which parents have mental illness, and the usefulness of findings must be qualified.

The National Survey of Parents with Disabilities
Within populations of parents known to have disabilities, many identify themselves as having psychiatric disabilities alone, or in combination with other physical or medical disabilities (Barker & Maralani, 1997). In the National Institute on Disability and Rehabilitation Research (NIDRR)-funded Research and Training Center (RTC) National Survey of Parents with Disabilities, data from a convenience sample of over 1,200 adults with disabilities was supplemented by a secondary analysis of the 1993 U.S. Census Bureau Survey of Income and Program Participation (SIPP; Barker & Maralani). In the NIDRR/RTC study, one in seven parents with disabilities identified a psychiatric disability as the primary disability. According to SIPP data, about 7 million parents with children under the age of 18 have disabilities. Applying the NIDRR/RTC rate of one in seven to the SIPP data, it is possible to estimate that approximately one million parents of children under the age of 18 have psychiatric disabilities.

This, however, is a gross underestimate of all individuals with mental illness who are parents for several reasons. First, the RTC Survey focused on parents with mental illness who identify themselves as disabled as a consequence. Therefore, parents with mental illness who do not define themselves as "ill" or who do not view themselves as disabled are not represented in the study. In addition, minority populations are underrepresented in the RTC sample. The RTC survey was only distributed in English; data were not obtained in other languages. The RTC sample also had a higher percentage of female, white, highly educated respondents than the U.S. Census Bureau SIPP data. Finally, the data do not include information about parents who do not have custody, those with children who are adults themselves, or those with children living in alternative family situations.

Mothers with mental illness, more frequently studied than fathers, have numbers of children consistent with the general population (Caton, Cournos & Dominguez, 1999; Mowbray et al., 1995b; White, Nicholson, Fisher & Geller, 1995). Therefore, extrapolating from the NIDRR/RTC data convenience sample estimate of about one million parents with psychiatric disabilities whose children are under the age of 18, at least two to three million children in American have parents with self-reported psychiatric disabilities. In the NIDRR/RTC data, nearly half of the children of parents with psychiatric disabilities have disabilities themselves, underscoring the importance of supports for children as well as adults, and the potential impact of improving parents' skills on the lives of millions of children. These data highlight the significance of considering families, rather than individual adults or children, as the client "unit."

The National Comorbidity Survey
Recently, data from the National Comorbidity Survey (NCS) of the United States, a survey of a national probability sample of over 8,000 respondents carried out between 1990 and 1992 by Dr. Ronald Kessler and colleagues, have become available for analysis regarding parenting issues and mental illness (Kessler, 1994). Demographic data from these respondents have been shown to be comparable to U.S. Census distributions (Kessler), suggesting that conclusions drawn from NCS data are likely to be generalizable to the larger U.S. population. In Phase II of the NCS, interviews were administered to 5,877 respondents, and included items on parenting and children, allowing preliminary findings on the rates at which individuals with psychiatric and/or substance abuse disorders, in various diagnostic categories, become parents.

Recent preliminary analyses of National Comorbidity Survey data regarding the prevalence of mental illness and parenting indicate that about 31% of American women and about 17% of men have a 12-month prevalence of at least one psychiatric disorder (not including substance abuse), and significant percents of these individuals are mothers (65%) and fathers (52%) (Nicholson et al., 2001). The NCS data suggest that women and men with psychiatric disorders or with co-occurring psychiatric disorders and substance abuse may be at least as likely if not more likely to be parents than those who do not meet criteria for psychiatric disorder or substance abuse.

Among women in the NCS sample who are mothers, 25% meet criteria for diagnoses in the affective disorder category, including major depression, mania, bipolar disorder and dysthymia. Approximately one-third of the NCS women who are mothers have a lifetime prevalence of anxiety disorders. Slightly over 11 percent meet criteria for post traumatic stress disorder, and less than one percent for a nonaffective psychosis disorder diagnosis.

Among men in the NCS sample who are fathers, almost one-sixth have a lifetime prevalence of affective disorders. About one-fifth have anxiety disorders. Approximately six percent meet criteria for post traumatic stress disorder, and fewer than one percent have a lifetime prevalence of nonaffective psychosis.

Unfortunately, it is impossible to describe the current family circumstances of adults who participated in the NCS in detail, other than to know they have had children. Information about the prevalence of custody loss among parents with mental illness, the frequency with which children require alternative caregiving when parents' functioning is compromised by illness, or the percent of children relinquished for adoption or raised by others is not available.

Public Sector Data
Available data from public sector management information systems sources, such as state mental health agency case management records, indicate a significant percent of adult clients receiving public sector mental health services are parents. Data from the New York State Office of Mental Health indicated that 45% of women under the age of 35 receiving intensive case management services have children, and of these, 20% are identified as the custodial parent (Blanch, Nicholson & Purcell, 1994). Analyses of Massachusetts Department of Mental Health (MA/DMH) case management rosters in the early 1990's indicated that approximately 9% of the total MA/DMH client population of adult women with serious and persistent mental illness were identified as having dependent children (White et al., 1995).

The knowledge gained from existing sources like these is routinely compromised by "sampling error," and deficits in the questions asked. For example, prevalence estimates culled from public sector mental health systems sources represent the sub-sample of adults diagnosed with mental illness who are in greatest need, who meet eligibility criteria regarding severity and duration of illness and functional impairment and, oftentimes, additional criteria such as multiple hospitalizations or homelessness. It is difficult to generalize from these data to the general population of individuals with mental illness that includes those receiving services in the private sector, or those not receiving services at all.

Often the right set of questions is not asked. For example, in completing the MA/DMH client tracking forms from the early 1990's, which have since been revised, the case manager checked a box if a client was known to "have a dependent child." This category was defined as "client is the primary caretaker for a minor child." First, a check in this box was based on the case manager's knowledge of the client's parenting status and history. Second, a check reflected knowledge only of those with children under the age of 18, with whom the adult was actively involved. Therefore, information was not recorded about children ever born who might be adults themselves, or those who might be living independently or in alternative family situations with relatives, foster parents or adoptive parents. In addition, information was not routinely obtained about pregnancy status or whether adults were considering becoming pregnant.

It is important to note that Massachusetts was not alone, by any means, in overlooking this information. In our original survey of SMHAs, conducted in 1990 and 1991, only 16 SMHAs routinely asked whether clients were parents (Nicholson et al., 1993). This state-level oversight is repeated at the clinical level in inpatient and outpatient psychiatric settings, where records do not contain this information (DeChillo, Koren & Schultze, 1994; Rudolph, Larson, Sweeny, Hough & Arorian, 1990). Information is not routinely obtained about whether adults with mental illness have even given birth to or fathered a child, the ages of these children, where they are living, or who is caring for them. And yet the likelihood that parents comprise a significant percent of inpatient or outpatient populations is high.

In Massachusetts, clinicians completed surveys at three sites-a Massachusetts Department of Mental Health case management site, the outpatient department of a community mental health center site, and a site providing both case management and outpatient treatment services (Nicholson, unpublished raw data). The percent of female outpatient clients who are mothers ranged from 33% to 50%. Twenty-one percent of male clients are known by clinicians to be fathers (Nicholson, Nason, Calabresi & Yando, 1999). The average number of children born to each parent is a little over two, suggesting these families are similar to the national average. Again, these prevalence figures must be considered underestimates as they are data obtained from clinicians, not the clients themselves, and these sites serve the small slice of the population diagnosed with severe and lengthy illnesses.

While individual states and agencies have modified forms and procedures to begin to collect this information, our recent survey of state mental health agencies, conducted in 1999, indicates that an even lower number of SMHAs (12) than the number in 1990 (16) are collecting this information on a routine basis. While SMHAs may justify this lack of data collection by pointing to the privatization of the case management function, or as the responsibility of managed care organizations and providers, this abdication of responsibility robs the SMHAs of opportunities to promote quality care through the development of relevant policies, regulations, procedures, and services. Adequate policy development, program planning or clinical care, to address the issues and meet the needs of adults with mental illness and their children, cannot occur without accurate prevalence data, and more detailed information about the circumstances, goals and needs of these families.

SUMMARY: There are no national data on the prevalence of adults with mental illness becoming parents and raising children, nor do data exist describing their children. Efforts to extrapolate these figures from existing data sets or community clinic sites are limited by the types of questions asked and populations served. Parents with mental illness are not routinely asked comprehensive questions about their family or household circumstances. Therefore policies and programs have not necessarily been developed with adequate information to meet the needs or goals of adults with mental illness as parents or the needs of their children. We know nothing about the prevalence of parents with mental illness in the private sector, nor about parents who are undiagnosed or not receiving treatment. Likewise, we know nothing about their children's circumstances or needs.

RECOMMENDATIONS: Adults with mental illness must be asked the "right" set of questions regarding their status as parents and their family circumstances. For example, women and men must be asked whether they have ever given birth to or fathered a child; the dates of birth and genders of their children; where their children are living; and who is caring for them. Adults with mental illness may be caring for adopted children, stepchildren or the children of others in informal arrangements as well. The issue of who has custody of children is a distinct question from the question of where children are living, who is caring for them, or the extent or nature of contact or involvement their parent with mental illness has with them. Policy makers and service providers must also ask about adults' wishes or goals for parenting, as they may differ from their current family situations. Adults' goals should be respected in service planning, as they affect adults' compliance with treatment recommendations, and outcomes for both parents and children. Children's circumstances and needs must be explored and documented.

Table of Contents | Previous | Next