Monday, April 30, 2012

Child welfare in Minnesota: Arbitrary, capricious and cruel

            Every state has thousands of pages of rules, regulations and laws concerning the investigation of alleged child abuse and the removal of children from their homes.  But despite all that exists on paper, in real life child welfare is, in the words of Diane Redleaf, executive director of the Family Defense Center “a lawless system.”

Child welfare systems are arbitrary, capricious, and cruel.  A child’s fate depends on where that child lives, which caseworker shows up at the door and what mood she’s in.
            This can be seen in the fact that when one compares states, or even counties within a state, there is wild variation in decision making: The rate at which a call is accepted for investigation (“screened in”) the rate at which investigated cases are “substantiated” – a misleading term which usually means that a caseworker has decided, on her own that she thinks there is slightly more evidence than not of maltreatment – and the rate at which children are torn from everyone they know and love and consigned to foster care all vary enormously.

            So, for example, even after factoring in rates of child poverty, a child in Nebraska is more than five times more likely to be taken from his or her parents than a child in Illinois.  In California, where child welfare systems are run by counties, a child is nearly three times as likely to be taken away in Butte County than in Tulare County. 

            But some of the strongest evidence of the fundamental lawlessness of child welfare has come from studies in which the people who decide whether to “screen in” a phone call alleging abuse or neglect are given hypothetical cases.  Studies in New York in 1986 and Florida in 1991 found dramatic variation from county to county or even hotline operator to hotline operator.

            The most recent, and most comprehensive such study that I know of comes from the Office of the Legislative Auditor in Minnesota – a state which, year after year, takes away children at rates far above the national average.  It was released in February.  It is a strikingly thorough assessment of how more than 80 different county run child welfare systems make their screening decisions.  The study used several methods, including giving the screeners ten hypothetical cases, based on actual, typical calls to child protection “hotlines,” and asking if they would be accepted for investigation or not.

            In only three cases did even 80 to 82 percent of the screeners agree (and in at least one of those cases, as is discussed below, they were dreadfully wrong).  In three more cases, 64 to 71 percent of screeners agreed.  And in the remaining four they split almost 50/50.


            The case that produced the most agreement that it should be screened in, actually was one of those where the evidence was strongest that the best thing the county child welfare agency could have done was leave the family alone.  This is the hypothetical case:

Dr. Jones calls to report that Emily Blackdeer tested positive for marijuana after giving birth to a baby boy yesterday. He says the child’s meconium was not tested due to a mix up. Jones reports that Emily also tested positive for marijuana during her pregnancy. Jones said Emily told him she smoked marijuana during her pregnancy to help her with her appetite.

            So in 80 percent of Minnesota counties, the stress any mother faces right after giving birth could be compounded by a child protective services worker coming to the door, poking and prying into the home asking mom about the most intimate aspects of her life, questioning friends, neighbors and relatives. (For reasons discussed below it might not play out that way in Minnesota, but the risk remains considerable.) All because, like a great many Americans, mom smoked pot.  All this could happen to the mother, and the child could be placed at risk of foster care, in spite of the fact that Mom clearly was getting prenatal care - something she probably won’t do next time, since her doctor turned her in.

            The enormous potential for harm in just these circumstances was well documented in stories in The New York Times and Philadelphia Citypaper.

            In part this may be because Minnesota’s idiotically, broad, vague definition of “neglect” (as one county screener admitted, they could, if they chose, screen in almost any call) is specific on one point: “prenatally exposing a child to controlled substances” automatically is defined as “neglect.”  But in this hypothetical, we don’t know if the infant was born with marijuana in his system.

            Another possible reason: The mother’s fictional name is Blackdeer, suggesting she might be Native American.  The hypotheticals used a wide variety of names, some of which suggested the family’s race.  Unfortunately, the study did not go further and try giving identical hypotheticals with different racial descriptions to the screeners to test for the common problem of racial bias in child welfare decision-making.

            The study notes that “Staff at one agency explained that marijuana use is so common in the area that when a positive marijuana test is the only concern, they screen out the referral and offer child welfare services instead.”  But another county “screens in marijuana use because of the risk of mis-socializing a child.”

            Right.  Because nothing “socializes” a child better than a few foster homes.

            In contrast, only 47 percent of counties would screen in a case in which a mother allegedly “is drunk every day to the point of throwing up and has withdrawal tremors from alcohol.”  The father, who does not have custody, says the mother’s friends brought their four-year-old son to him because mom “was too drunk to take care of” the boy.

            Of course in this case, the controlled substance is legal, and there are no ethnically-identifiable surnames.


            And it gets weirder still.  In 54 percent of counties, this case would be “screened-in:”

Police fax the following report: I responded to a report of five-year-old Davie Michaelson wandering in town. I met with Ann Johnson, a passerby who had found this child. While I was speaking with Ann, a man approached who said he knew the child. He directed me to a house at the end of the block. The yard was fenced, but the gate and front door were open. I entered the house and found Tammy Michaelson (Davie’s mother) sleeping on the couch. I awakened her and she explained that she had worked the third shift at the gas station last night and had left the boy to watch cartoons while she napped. The TV was on with a children’s DVD playing. Tammy said she had locked the door, but Davie must have unlocked it and left.
A [records] search shows that Tammy, age 21, received children’s mental health services as a child.

            I’ve written before about the double standards that abound in these types of cases.  In an example of the kinds of unwritten rules that substitute for the rule of law in child welfare, a screener in one Minnesota county said that “caregivers generally get one free pass on wandering child allegations because such incidents often are accidental.

            Other appalling decisions include:

● Twenty-nine percent of counties saying they would screen in a case where the only allegation is that a father who was evicted from his apartment now lives with his ten and 12-year-old sons in a trailer with no plumbing or electricity.  In other words, he’s poor.

● More than one-third of counties would investigate a mother whose only crime was to have been punched and choked by the father while the children were upstairs playing video games and may or may not have heard what was going on.  At least one county said they recently had screened in a case that involved parents yelling at each other, with no allegation of physical maltreatment.

This in spite of the fact that the danger to the child can be removed by arresting the abuser (an approach pioneered in Minneapolis) and in spite of the enormous body of research on the extra measure of harm done to children if those children are taken from their mothers under these circumstances.


To some extent screening in cases like the pregnant mother who smoked marijuana or the father who simply was poor might be explained by the fact that Minnesota pioneered “differential response,” in which some cases that normally would be screened out, and some which otherwise would get a full-scale investigation instead receive an “assessment” in which the agency offers help without the threat of removal.

But differential response exists statewide in Minnesota, yet there remains enormous variation in deciding whether to screen in cases.  And differential response was never intended to be used in cases where the only issue is poverty or a mother who smokes pot while pregnant.

A more likely explanation can be found in this question, asked in a survey of screeners in Minnesota counties.  They were asked if they agreed with the statement that “At times, child protection interventions can be more harmful than helpful to families.”  Now, keep in mind, virtually everyone in child welfare knows that you are supposed to at least pretend to agree with that statement, given the gigantic body of evidence showing that it is true.

Yet 22 percent of screeners disagreed or strongly disagreed with that statement.  In other words, nearly one time in four, a call to a child welfare hotline in Minnesota is likely to be taken by someone who has no clue that the investigation ever can do any harm to anyone.

That should cause a lot of people in Minnesota to be afraid – especially if they live in poor neighborhoods or on Indian reservations.

Making things worse: A bizarre Minnesota practice called a “child welfare check” and an insidious financial incentive to screen in cases.  That story in a future post.