In the early 1990s in Illinois, a number of mothers with schizophrenia and other forms of mental illness killed their children. In at least one of these highly publicized cases the child welfare system had returned a child from foster care to live with his mother. The deaths prompted the state to form a task force to look into how the child welfare system assesses the ability of parents with severe mental illness to care for their children. The result is one of the nation's most extensive examinations of what does and does not work in evaluating parents with mental illness. D r. Laura Miller, a professor and director of the Women's Mental Health Program at the University of Illinois' Department of Psychiatry, was part of the task force. She says the researchers found something she had long suspected but that had never been directly studied: A mental health diagnosis is not an effective way to assess whether someone is capable of being a good parent.

She explains that for mothers with mental illness, there are two key factors that correlate well with good parenting. The first is a mother's insight into her own mental illness--including whether she is able to see the warning signs of an episode coming on. The second is a parent's insight into their child's development. A mother who has unrealistic ideas about what can be expected from children at certain ages is more likely to mistreat a child than a mother who holds realistic expectations. A mother who is willing to acknowledge that she has problems and is willing to accept help is less likely to mistreat a child than one who denies she needs help. Social supports are also key: mothers who have supportive partners, neighbors or family members are less likely to abuse their children than mothers who are socially isolated.

The task force helped develop guidelines for measuring parenting competence. It recommended that parent evaluations be performed by a team that includes a psychiatrist, a psychologist, and a social worker, rather than by just one clinician. The task force recommended that evaluators observe parents interacting with their children, preferably in a home setting. It said evaluators should use multiple sources of information (medical records, criminal reports, drug screens, school records and child welfare files) and minimize the risk of bias by not having the evaluation paid for by any part of an adversarial court proceeding. Finally, the task force recommended using a standard of minimal--rather than optimal--parental competence to measure fitness.

In many ways, this is a significant departure from common practice. In New York City, for example, most evaluations are performed by a single clinician; many are performed without ever meeting the children, much less observing a parent's interaction with them; and evaluations are frequently performed and paid for as part of a court proceeding.

The task force used its findings to develop a model "assessment team" to evaluate parents with severe mental illness in Chicago. The team prioritized two types of cases: those in which a parent had been hospitalized with mental illness within the past eight years, and those in which previous evaluations had been contradictory or confusing to such a degree that child welfare workers and the courts were unable to make a recommendation on whether to return a child home or terminate parental rights.

Under the model, child welfare workers or the court could refer a family to the team with specific questions, such as: Will this mother's mental illness affect her parenting capacity? If so, what kind of interventions would help? These questions focused the examinations, which included a psychiatric evaluation as well as an assessment of how much insight a parent has into her mental illness, and how much practical knowledge she has of child development.

The assessments included direct observations of the mother with her children; a developmental screening of each child; interviews with each child; videotaped observations of the parent interacting with the child both in a clinic and at home; parents' responses to questionnaires; and a review of previous records and psychiatric evaluations. Each individual on the team wrote a report summarizing his or her area of the evaluation, and a team coordinator compiled the reports, wrote a summary of the findings and made recommendations for services and supports likely to help the parent.

One drawback to these evaluations is their cost--about $1,700 per family. But the task force suggested that a comprehensive evaluation might cost no more than a series of less detailed evaluations. Before this comprehensive method was developed, parents would often undergo multiple evaluations at a cost of about $500 each--either because the parent's lawyer challenged an evaluation as biased, or because a court believed an evaluation was incomplete or confusing. Hasty, incomplete evaluations can lead to longer stays in foster care because courts are afraid of making incorrect decisions, notes Miller.

"At present, many individuals with severe and chronic mental disorders are evaluated several times over a long period using unsound methodologies that do not directly assess parenting competency or potential for violence," explains the report, titled "Focus on Women: Mentally Ill Mothers Who Have Killed," which Miller co-authored. "Use of comprehensive, sound assessments when a case first enters the system could help considerably in shortening the length of time that children of mentally ill parents remain in foster care."

In Chicago, says Miller, the team quickly gained a reputation for helping "to clear cases that had been conundrums for years held up in the child welfare system."

However, there were unintended consequences of the evaluation reform. The original mission of the team was to conduct evaluations when a mother first came into contact with the foster care system, in order to determine what services and supports she needed to bring her children home. Instead, when Chicago attorneys saw how seriously the judges took these evaluations, they began to use them to speed the termination of parental rights for mothers whose children had been in the system for years.

Rather than using the evaluation team to help mothers who were new to the foster care system, says Miller, "they would find it was much easier to take a case that was languishing for five or six years and ask us to evaluate the case and subpoena us, and they could get the case off their hands much faster because the judges always listened to us because our evaluations were very sound. And it wasn't that these cases shouldn't have been terminated. It was just a poor use of our resources that ran counter to our mission. We weren't in the business of wanting to spend our time terminating rights for mothers."

As a result, the original parenting assessment team closed shop. However, two other groups in Chicago are using a similar method to evaluate parents involved in some of the most severe cases of abuse and neglect whose children have recently entered foster care.