Using child welfare data to learn from the past: why is it so unpopular?

Miracle Jackson, a seven-month-old in Detroit, died with a sock stuffed down her throat and her face covered in duct tape at the hands of her father in 2000. During the same week in the same city, a five-month-old named Jamar was severely beaten. It turned out that Miracle’s mother and Jamar’s parents had abused or neglected their previous children seriously enough that their rights to parent those children were terminated. Yet, when Miracle and Jamar were born, nobody checked on them to make sure they were safe. But that was about to change in Michigan, which became the first state to match birth and child welfare data to identify new children born to parents who had severely abused or neglected previous children – a practice that has become known as “birth match.”

The logic behind birth match is simple. Research suggests that in parenting as in other areas, past behavior is often the best predictor of future actions. Current technology makes it possible to match existing databases maintained by the child welfare and health agencies in order to identify infants born to parents who have had their parental rights terminated, been convicted of a crime against a child or have other history identifying them as a safety risk to a newborn. So it is not surprising that the Committee to Eliminate Child Abuse and Neglect Fatalities (CECANF) in its 2016 report recommended birth match as one strategy to identify children at high risk of maltreatment so that action can be taken to keep them safe. Yet, only four other states have adopted birth match, and only one (Missouri) has adopted it since the CECANF recommendation.

In a report called Learning from the Past: Using Child Welfare Data to Protect Infants Through Birth Match Policies, published by the American Enterprise Institute, I discussed what we know about birth match in the five states that use it. As the report illustrates, birth match policies and procedures varied widely from state to state.

All of the states that use birth match identify infants born to parents who had their rights terminated because of abuse or neglect, with some specific differences. It is not surprising that they all identify parents with a termination of parental rights (TPR), because a TPR usually means that there has been severe abuse or neglect and and the parent has been given multiple chances to ameliorate the behaviors or conditions that caused the child’s removal.

Each state has chosen to include certain other parents in addition to those who had a TPR. Maryland has the most limited policy, including (in addition to those who had their rights terminated) only parents who have been convicted of the murder, attempted murder, or manslaughter of a child. Minnesota includes the broadest group of parents–all those who were determined to have committed “serious maltreatment,” the highest of four categories of severity that are assigned to all substantiated instances of maltreatment. States also differ in how far back they look in time for evidence of dangerous parental behavior: Texas looks back only two years, Maryland and Missouri look back ten years, and Michigan and Minnesota match all available records, regardless of when the maltreatment or termination occurred.

States also differ in whether they treat birth match referrals as allegations of abuse and neglect, requiring a regular CPS investigation. The first two states to adopt birth match, Michigan and Minnesota, already had a category of child maltreatment called “threatened harm” or “threatened injury.” Birth matches in those cases receive a CPS investigation of an allegation of threatened harm or injury. In Texas, matched infants and their families also receive a regular investigation, but the type of allegation depends on the content of the report.¹ In general, investigations result in a finding on the truth of the allegation; if it is “substantiated,” or found to be true, it may result in the removal of a child or children into foster care, the provision of in-home services and monitoring to ensure their safety, or a possibly a placement with a relative or family friend with the consent of the parent.

In contrast to the other three states, Maryland and Missouri treat birth match referrals differently from allegations of child abuse and neglect. In Missouri, birth match referrals are treated as “Non-Child Abuse/Neglect Referrals” and receive a “Newborn Crisis Assessment,” a special type of investigation that was designed to respond to calls from hospital personnel who are hesitant to release newborns from the hospital because of safety concerns. If no safety concerns are identified, parents can decline any services that are offered; if safety concerns are identified, social workers have the same choices as in a regular investigation: they may go to court to request immediate custody, allow the child to stay at home under a safety plan supervised by the department, or negotiate a voluntary placement with a relative.

In Maryland matched families receive an “assessment,” which is less comprehensive than a regular investigation. Families can refuse to participate, unless there is “reason to believe a child has been abused or neglected or is at substantial risk of abuse or neglect,” in which case the local department of social services is directed to make a report to CPS. Similarly, the department is directed to call CPS if there is such a concern at any time during the birth match assessment process.

The lack of data makes it difficult to assess the impact of existing birth match processes. Other than Missouri, where birth match has been in use for less than a year, none of the states publishes data on the results of these programs as part of their regular reporting, and it appears that administrators do not review this data internally. In response to the request for data for the report, child welfare officials had to generate new tables from their databases. But the data raised many questions and without knowing exactly how it is obtained, one cannot judge its accuracy. There were some anomalies that state administrators were unable to explain, like the fact that the total number of matches in Michigan dropped from 1186 in FY 2019 to to 873 in FY2020 and then down to 515 in FY2021–a drop of 50 percent in two years! It appeared that state administrators were unaware this anomaly before being asked about it, and they were unable or unwilling to provide an explanation. 

If the data provided by the states is approximately accurate, birth match is identifying significant numbers of children. The number of matched infants identified in FY2019 (before the pandemic) was 1,188 in Michigan, 1,138 in Texas, 420 in Minnesota, and 243 in Maryland. Between half and two-thirds of these children already had an open investigation or case. It is encouraging that so many of these infants were known to CPS without birth matching, but it also shows that a sizable number and proportion of infants at risk due to their parents’ earlier behavior would be unidentified in the absence of this tool.

But the effectiveness of birth match depends on the quality of the investigations or assessments that are conducted and whether they result in actions to ensure child safety. The limited evidence is not encouraging. The number and percentage of matched children and families reported to be actually receiving services was surprisingly low. In Texas, of the 302 families investigated due to birth match in FY2019, only 70 received in-home services and 28 had a child or children removed. In Michigan, of the 484 investigations due to birth match, only 49 cases opened for services and 24 had a removal of a child. In Maryland, only four of the 89 families investigated due to birth match were documented to have received services. Minnesota provided no data beyond the number of matches. Without better data and case reviews, it is impossible to know why so few families received services.

The fact that the data requested had to be specially generated suggests that child welfare administrators in birth match states have little interest in the implementation and effects of of birth match. That was not always the case, at least in Michigan. One former CPS director in Michigan, who had served as a CPS worker and supervisor earlier in his career, had a strong belief in the potential of the process to protect children if correctly implemented. He conducted an internal review of birth match cases and found that 75 percent of the investigations resulted in no finding of threatened harm to the child, and only 6.5 percent of the cases eventually went to court for removal or court-ordered services. He concluded that investigative workers were not following agency policy and that supervisors were nevertheless approving the findings of the flawed investigations. He was working on ways to improve implementation through oversight of supervisory decisions. But with a change of personnel, those efforts never came to fruition. Now, birth match is under review in Michigan as part of a “front end redesign” of the child protection system.

Many former birth match advocates appear to have lost interest as well. In Texas, birth match was adopted in response to a recommendation by the State Child Fatality Review Team (SCFRT). But after requesting updates on implementation in FY2013 (which were never provided) and recommending expanding the program to look back five years in FY2018 (a recommendation which DFPS rejected), the SCFRT stopped making recommendations about the program. In Maryland, advocates pushed to strengthen the program by increasing the “lookback” period from five to ten years. But after such legislation was passed in 2018, it does not appear that advocates asked about its implementation nor about the effects of the expansion. Moreover, in passing the 2018 legislation, the legislature included a provision that appears to be aimed at finding less controversial alternatives to birth match.

The changing ideological climate might be the reason for the loss of interest in birth match among officials and advocates in the first four states to adopt it. In today’s atmosphere, identifying parents based on their past involvement in child welfare or criminal justice is likely to be criticized because these systems involve Black people at a rate that is disproportionate given their share of the population, though proportionate to their rate of abuse and neglect compared to other populations. There is no escaping the conclusion that birth match is simply at odds with the current zeitgeist in child welfare. Missouri was the only state to institute birth match since it was recommended by CECANF in 2016.

The report makes three recommendations. Due to its support in research and common sense, birth match should be added to every state’s set of tools to prevent child abuse and neglect and Congress should consider mandating birth match as a requirement to receive funds under the Child Abuse Prevention and Treatment Act (CAPTA). Birth match provisions should include all parents who committed severe abuse or neglect whether or not they had a TPR or criminal conviction. And finally, states with birth match programs should track and publish data on the children matched and should conduct case reviews to assess the implementation of their programs. But it is not likely that any of these recommendations will be widely adopted until the pendulum swings toward the needs of children living in unsafe homes.

When a new baby is born to parents who had their rights terminated to a previous child due to severe abuse or neglect, or who killed or severely harmed another child, the child welfare agency should be notified, and a professional should make contact with the family to ensure the child is safe and offer the parents any assistance needed. It is such a commonsense idea that it’s hard to imagine anyone would oppose it. Nevertheless, only five states have adopted such a program, and and the four states with programs that have been in effect for more than one year have displayed what appears to be little interest in assessing or improving their implementation; on the contrary, there seems to be some interest in eliminating the programs among administrators and legislators in some states. The current ideological climate in child welfare may be be responsible for our failure to use a simple tool to protect children.

Notes

1. How the allegation type is determined and by whom, and how maltreatment can be found before it has occurred are unclear. Birth match is not mentioned in the department’s policy manual and DFPS’ Media Relations Director was not able or willing to answer these questions.