What is the cause of racial disparity in child welfare?

There is no doubt that Black children and families are reported to child abuse hotlines, investigated, and removed from their homes more than White children. But many leading voices in child welfare today have made the dubious assumption that racial bias in reporting and child protective services is the underlying reason for these disparities. Unfortunately, based on this assumption, they propose policy solutions that risk destroying existing protections for Black children or even for all abused and neglected children. A star-studded group of researchers has collaborated on a paper that ought to put this presumption to bed for good. I hope that this brilliant paper is able to change the minds of some who have unquestionably adopted the fashionable theory that is being promoted by the child welfare establishment.

There is no dispute that Black children are reported to authorities, investigated for abuse or neglect, and placed in foster care at a higher rate than White children. The federal publication, Child Maltreatment 2021, reports that Black children are nearly twice as likely as White children to be the subject of a screened-in report and almost twice as likely to be substantiated as a victim of child abuse or neglect. In 2020, Black children were 14 percent of the child population but 20 percent of the children entering foster care. Kim et al estimated that 53 percent of Black children will experience a CPS investigation by the age of 18, compared with 28 percent for Whites. But are these disparities greater than what would be expected given the higher rates of poverty and other social problems among Black children? That’s the question that a group of 13 researchers addressed in a recent article on racial and ethnic differences in child protective services reporting, substantiation and placement, published in the leading child welfare journal, Child Maltreatment. The authors include most of the top researchers in the field, such as lead author Brett Drake and his co-authors Richard Barth, Sarah Font, Emily Putnam-Hornstein, Jill Duerr-Berrick, and Melissa Jonson-Reid–an accumulation of starpower rarely seen collaborating on a single article.

Previous studies cited in the paper have already concluded that when adjusting for income and family context, Black children were actually reported to CPS at similar or slightly lower rates than White children and that Black children who are the subject of investigations were no more likely to be substantiated or placed in foster care than White children. Despite these results, the belief that racial disparities are due to anti-Black bias in reporting and child protective services (CPS) decisionmaking has been asserted as established fact in publications by the federal government, numerous child welfare groups, the American Bar Association, the American Civil Liberties Union and Human Rights Watch, and many media outlets. A report by a leading legal advocacy group and the Columbia Law School Human Rights Institute urged the UN to investigate the American child welfare system for racial discrimination. After conducting its own review, a UN Committee recommended that the United States “take all appropriate measures to eliminate racial discrimination in the child welfare system, including by amending or repealing laws, policies and practices that have a disparate impact on families of racial and ethnic minorities.” Rather than advocating for reform of child welfare systems, some individuals and organizations, such as upEND, press for the extreme step of abolishing the entire child welfare system.

The new paper provides a needed antidote to the certainty that racial bias is the principal source of racial disproportionalities in reporting, substantiations, and foster care placements, and provide compelling evidence against it. The authors use universal national data to ask two questions:

  1. Are Black-White and Hispanic-White disparities in CPS reporting lower than, similar to, or higher than disparities in non-CPS measures of social risk and child harm?
  2. Once referred and accepted for investigation, do Black or Hispanic children experience substantiation and removal into foster care at rates lower, similar or higher than White children?

The authors focused on Black, White and Hispanic populations. Native American populations, which are also disproportionately involved in child welfare, are difficult to study because many are served by tribal child welfare systems and may not be reflected in the national data that the authors use. The authors used CPS data from the National Child Abuse and Neglect Data System (NCANDS), which gathers information from all 50 states, the District of Columbia, and Puerto Rico about reports of child abuse and neglect and their handling by child welfare agencies. Data from certain states and years had to be eliminated because of data quality problems and missing data. The elimination of all data from New York and Pennsylvania is unfortunate, but it is unlikely that these omissions changed the overall trends. Data for income and other indicators of risk and harm came from the Census Bureau, the Kids Count Data Center, National Vital Statistics records, and the Centers for Disease Control.

Question One: Reporting Disparities

The authors posit that the “expected rate” of Child Protective Services (CPS) involvement for a particular group of children should be “the rate at which children in that population experience child abuse, neglect, or imminent risk thereof.” But the authors explain that we cannot actually observe the incidents of abuse or neglect, as they are not always reported to authorities. And when reports are made, the system may not always make the correct decision when it decides whether or not to “substantiate” or confirm the allegations made by the reporter. To estimate the “expected rate” of being reported to CPS, Drake and colleagues used several categories of risk and harm that are known to be highly correlated with the risk of child abuse and neglect. Indicators of “social risk” included the numbers of children in poverty, children in single parent families, teen birth rate, and adults without a high school degree. To indicate harm to children, the authors used “very low birthweight,” “very preterm births,” infant mortality, homicide injury, and “unintentional death.”

Drake and his colleagues calculated “disparity ratios (DR’s),” by dividing the incidence of social risk or harm for Black or Hispanic children by the rates for White children by year. They found that the DR’s for all the measures of risk, and all of the measures of harm except accidental deaths, were greater than the DR’s for CPS reports. In other words, there was a greater disparity in risk and harm to Black children than there was in CPS reporting. Thus, given their likelihood of being abused or neglected, Black children appear to be reported to CPS less than are White children.

The tables below illustrate the incidence of risk, harm and CPS reports for Black children compared to White children. While Black children were reported to CPS at a rate close to twice the rate of White children throughout the period studied, their poverty rate was three times that of White children in 2019, the proportion of Black children in single-parent households was 2.5 times as as that of Whites, and the disparity in the rate of single-parent households and adults without a high school degree was almost as great. In terms of harm, Black children were four times as likely to be a homicide victim in 2019, nearly three times as likely to have a very low birth weight, and more than twice as likely to die of maltreatment, in 2019.

Disparities in Substantiation and Removal

To address disparities in substantiation and removal following investigation, Drake and coauthors compared the raw data and also ran regressions to adjust for demographic factors that might affect placement, such as poverty. They found that in both adjusted and unadusted estimates, Black children, once investigated, have been less likely to be substantiated and placed in foster care in more recent years. Before 2011, Black children were slightly more likely to be substantiated and placed in foster care than White children before the trend reversed. The unadjusted estimates are shown below.

When they compared Hispanic children to White children, the authors found a very different pattern. While Hispanic children face much more exposure to social risks like poverty than White children, they experience harm and CPS reporting at about the same rate as White children. This pattern is consistent with what is known as the “Hispanic paradox.” This term describes a well-documented phenomenon in the child welfare and medical literatures wherein Hispanic children and families have indicators of well-being similar to their White non-Hispanic counterparts, despite having much higher indicators on risk factors like poverty. For Hispanic children, there were slightly greater unadjusted rates of substantiation and placement than for White children, but these differences disappeared when statistical controls were added.

Conclusions and Implications

The authors draw two primary conclusions from their research. First, “Black-White CPS reporting disparities were consistently lower than Black-White disparities in external indicators of social risk and child harm.” Black children were exposed to more risk and experienced harm at greater rates than White children, and these disparities were consistently greater than the disparities in reporting. If either group is overreported in relationship to their risk it is White children. It is still possible, the authors point out, that all children are overreported to CPS in relation to external indicators of risk and harm. But “if there is systemic overreporting, it is not specific to Black children and thus, unlikely to be driven by racial animus.” They also found no evidence that once investigated, Black children were disproportionately substantiated or placed in foster care.

Second, the authors found continued evidence for the “Hispanic paradox” in CPS reporting compared to observed risk exposure. Although Hispanic children face substantially greater social risks than White children, they experience harm and CPS reporting at about the same rate as White children. This supports the well-documented pattern whereby more recently immigrated Hispanic families, despite having higher risk factors, tend to have indicators of well-being similar to Whites.

In the authors’ own words:

It is indisputable that despite progress in certain areas, the United States has not overcome the legacy of slavery, segregation and Jim Crow. This legacy lingers most clearly in the patterns of segregation that emerge in many of our metro areas…To assert that these patterns, and the poverty and chronic stress they perpetuate, would have no impact on behavioral and psychosocial functioning among the individuals and families in those neighorhoods is to reject decades of scientific consensus on human development. Indeed, this history and its unresolved legacy is essential to understanding why Hispanic families face similar individual socioeconomic disadvatage but appear to have sigificantly lower rates of CPS involvment than Black children.

If I have one quibble with the authors of this brilliant and essential article, it is their lack of attention to the possile psychological impacts of intergenerational trauma from the history of slavery, Jim Crow, and racial hatred and violence. As the child of Holocaust survivors, I can attest that the six years of trauma that my parents suffered after the Nazis invaded Poland has affected me and even my daughter. For families in which nearly every generation going back almost 400 years suffered the trauma imposed by living within slavery, Jim Crow, or a culture of virulent and violent racism that continues in some form today, it would be surprising if there was no current mental health impact on the generation that is parenting children today. Such a impact might include elevated levels of mental illness as well as self-medication through drugs and alcohol, both of which are associated with child maltreatment.

In the section on Implications, the authors assert the need to address the factors that underlie the differing rates of risk and harm to Black children, outside the CPS system itself–factors such as poverty and racial segregation. The belief that abolition of child protections would in and of itself help Black children, the authors point out, relies not only the assumption that CPS is racially discriminatory, which this paper has debunked. It also relies on the assumption that CPS provides no protection to children. Certainly there is room for improvement in our child protection systems, particularly in the quality of care they provide to children removed from their homes. Yet, foster youth testimonies such as “being placed in a foster home saved me,” or “Using my voice is the reason I am no longer in a household that is broken,” as well as the silent testimony of the more than two thousand children who die of abuse and neglect every year,1 are a testament to the untruth of this statement.

The authors suggest three courses of action for the future. First, we should acknowledge and address the true drivers of racial inequity among families, such as multigenerational poverty, underresourced schools, and lack of access to quality substance abuse and mental health treatment programs. Second, despite their results, we must acknowledge that racial bias may exist in certain localities and be prepared to address it. And third, “there is clearly room to consider restructuring child and family policy generally to include a focus on providing preventive services, including material assistance, to families. (See my discussion of universal yet targeted programs to prevent child maltreatment.)

The authors go on to state that “It is possible that a narrow focus on reducing Black children’s CPS involvement without addressing the pronounced inequities documented by the external indicators will result in systematic and disproportionate unresponsiveness to abuse and neglect experienced by Black children.” And indeed, there are already reports that professionals are already more reluctant to report Black children and CPS employees are more reluctant to substantiate or remove them.2 Or to put it more bluntly, the standards for parenting Black children will be lowered, and the level of maltreatment that Black children are expected to endure before getting help will be raised. Ironically, this calls to mind some manifestations of racism that have been cited by scholars and advocates, such as treating Black children as if they are older than their actual age, and thinking that Blacks have a higher pain threshhold than Whites. Of course if the child welfare abolitionists have their way, the entire system will be abolished, destroying protections for all children. That is unlikely to happen, but what is more likely is a weakening or repeal of critical laws like the Child Abuse Prevention and Treatment Act or the Adoption and Safe Families Act, which are both currently under attack, to eliminate or weaken provisions like mandatory reporting.

Sadly, few leaders on either side of our increasingly polarized political scene will be open-minded enough to read, understand and accept the conclusions of this important paper. While the progressive mainstream (and even many others in the child welfare establishment) has blindly accepted the notion that racial bias is the primary driver of child welfare disparities, conservatives remain obsessed with reducing the size of government and cutting taxes, refusing to recognize the need for massive spending, even a domestic Marshall Plan, to rectify the result of centuries of slavery and anti-Black racism in America.


  1. States reported 1,820 child maltreatment fatalities to NCANDS in 2021. But experts cied by the National Commission to Eliminate Child Abuse and Neglect Fatalities (p. 9) estimate that the actual number is at least twice as many as that reported to NCANDS.
  2. See, for example, Safe Passage for Children of Minnesota, Minnesota Child Fatalities from Maltreatment, 2014-2022. The report authors found evidence that raised the question of whether Minnesota child welfare agencies may have tended to leave Black children in more high-risk situations for longer periods of time than children of other races and ethnicities. See also Stacey Patton, The Neglect Of 4 Texas Brothers Proves That The Village It Takes To Raise A Black Child Is The Same Village That Stands By And Watches Them Die, Madamenoire, November 2, 2021. She states that “To reduce the number of Black children entering into foster care as a result of abuse, child welfare professionals are increasingly “screening out” calls for suspected child abuse.  There haven’t been any state or national level studies to show whether disproportionately higher numbers of calls of Black child abuse are being screened out to avoid claims of racial discrimination.  However, in my work as a child advocate, I keep hearing stories of non-Black child welfare professionals who don’t report abuse because they either don’t want to be accused of racism, or they just accept that beating kids is an intrinsic part of Back culture.”

No room for child advocates: Why I was kicked off DC’s Child Fatality Review Committee

Until recently, I was one of three “community representatives” on the District of Columbia’s Child Fatality Review Committee. Community representatives are the only members who are not paid to sit on this panel; the rest are agency representatives who sit on it as part of their jobs. My service on the panel was an important aspect of my advocacy for abused and neglected children in the District. But this work ended abruptly for me in March of this year when I was told that my service was over. As described below, I have some ideas about why the panel decided to dismiss perhaps its most engaged, passionate and productive member.

On March 2, 2023 I got a call from the Director of the Mayor’s Office of Talent and Appointments (MOTA). He said he was calling about my position on the District’s Child Fatality Review Committee (CFRC). I told him I had already received a call several weeks earlier from a MOTA staffer telling me that she was working on my reappointment, which should have happened earlier but was backlogged due to the pandemic. She asked me to submit an updated resume and told me she would be back in touch shortly to help me prepare for my DC Council confirmation hearing. But on March 2, the Director told me there had been a mistake. I was not being reappointed to the committee, and since my term had already expired, I was now off the panel.

When I asked why I was not being reappointed, I was told that it was time to give other people a chance to serve. This explanation made no sense. I was one of only three “community members” on the panel, out of eight authorized by DC Code. In her 2017 report, the DC Auditor noted the many vacant seats for community members and the importance of these community representatives, who are not tied to a specific agency. In her remarks preceding the 2017 report, CFRC Co-Chair Cynthia Wright wrote that “the addition of new community members [of which I was one] who provide a fresh perspective to our work …. has increased the vitality of the CFRC.” I doubt that there are five people lined up waiting to be appointed, or even one person ready to replace me. It’s not surprising that there is no long line of community members who want to volunteer two to four hours of their time each month in meetings about children who die, not to mention reading the sad case histories before the meetings. It was clear that my expulsion was not intended “to make room for somebody else.”

My de facto expulsion certainly did not stem from a lack of commitment or shoddy performance. I attended all 13 meetings of the Child Fatality Review Team in FY 2022 and the first quarter of FY 2023. According to the government’s responses to the oversight questions posed by the Committee on the Judiciary and Public Safety, the two other community members attended nine and seven out of 13 meetings respectively. I read every case study in advance of the meeting, and came prepared with questions and comments. Based on the questions asked at the meetings, it was clear that most members never read the case histories (sometimes as long as 20 single-spaced pages) and instead relied on the quick presentations given by Committee staff. In addition, I was a main source of new ideas on the panel; indeed, the two most recent presentations before the committee before my exit stemmed from my suggestions.1 So there must be another reason I was not reappointed. And I think I know what it is, but let me first say something about the Committee and why I joined it.

As described on the website of the Office of the Chief Medical Examiner (OCME), under which the CFRC is located, the goal of the CFRC is to “reduce the number of preventable child fatalities in the District of Columbia through identifying, evaluating, and improving programs and systems, which are responsible for protecting and serving children, and their families.” Based on the information it reviews about the histories of children who died, the CFRC makes findings and recommendations to prevent such deaths in the future. The CFRC is comprised of two teams, the Infant Mortality Review Team (IMRT), which reviews deaths of District infants from birth through twelve months, and the Child Fatality Review Team (CFRT), which reviews the deaths of children aged one to 18 years old as well as youths aged 18 through 21 who were known to the child welfare system within four years of their deaths or to the juvenile justice system within two years of their deaths.2 Their are child fatality review teams in all 50 states and some tribal nations as well.

I joined the CFRC because of my concern about children who are abused or neglected and my belief that CFRC had the potential to have a broader impact beyond preventing fatalities because the conditions that lead to child deaths also cause harm to many more children who do not die. The DC Auditor reported hearing this from several individuals who likened the fatality cases that are examined to a “canary in a coal mine.” I had a particular interest in monitoring the work of the Child and Family Services Agency (CFSA), which is charged with protecting maltreated children in the District. When I joined CFRC, I had recently left my job as a social worker at a private agency that provided foster care as part of the child welfare system led by CFSA. In that capacity, I had heard children’s lawyers express their fear that due to the recent sharp drop in removals of children from theirt homes into foster care, many were being left in dangerous situations that might eventually result in deaths or irreversible emotional or physical damage.

And indeed, upon joining CFRC, I found a number of reasons for concern about CFSA’s effectiveness in protecting children. It was astounding to learn how many children died after having some contact with CFSA. According to CFRC’s annual reports, 69 percent of families of decedents reviewed by CFRT in 2019 had prior CFSA involvement; that figure could not be calculated for IMRT reviews. Of the cases reviewed by the CFRC3 in 2020, 15 out of 18 (or 83 percent) of the decedents’ families had prior CFSA involvement. Reading the CFSA histories of these families often revealed as many as 20 reports to the hotline over the years. Many of these reports were not even accepted for investigation. Those that were investigated were often not “substantiated” or verified by the investigators, which is required for opening a case, despite what seemed like ample evidence of abuse or neglect cited in the case summaries. Even when a report was substantiated and a case was opened for in-home services, more calls often came in about the same families and investigators continued to find dangerous conditions and parenting practices. Even after the cases closed, the reports would continue to arrive, suggesting that nothing had changed as a result of CFSA’s intervention. And even when children were removed to foster care, they were often returned home with no evidence of improved parenting or conditions, and the reports continued to come in.

But when I expressed my concerns about CFSA’s response to frequently reported families and suggest that a finding or recommendation might be in order, I was repeatedly accused of “picking on” CFSA. It is as though CFSA was a child needing protection from bullying rather than an agency responsible for protecting children. Perhaps I shouldn’t be surprised. In its July 2017 report on CFRC, the DC Auditor reported this exact concern — that several panel members believed “defensive or territorial behavior remains an impediment to productive deliberations.”

It was perhaps during my second term at the CFRC, starting in 2020, that another set of issues arose that also put me outside the mainstream of CFRC members. The District was already at the forefront of a national movement to drastically reform what was described as a racist child welfare system by reducing foster care placements and government intervention in the lives of families. The murder of George Floyd and calls to abolish the police intensified this movement, with some even calling for the abolition of child welfare agencies, which were labeled as a “family policing system.” An effect of this type of thinking was an unwillingness to suggest that parents were unfit, no matter how abusive or neglectful they may have been, or to suggest that CFSA should have intervened more aggressively to protect children who later died. While my concern was for the safety of children, other members of the committee were more interested in demonstrating their opposition to governmental interference in the lives of families, regardless of the cost to children’s lives or safety.

When I joined the panel in 2017, there was more tolerance for diverse viewpoints and more concern for the needs of vulnerable children, regardless of race. There were frequent discussions about how to work with the parents who were repeatedly reported to CFSA but did not ever seem to change. Such families are well-known in the child welfare literature as “chronically neglectful,” “chronically maltreating” or “frequently reported” families. Many of these parents had problems with substance abuse, mental illness, domestic violence, or some combination of these three factors that impaired their ability to parent. They had been offered numerous services to help address these issues, which they either declined, dropped out of or completeded without any apparent benefit. Discussions of these families often led to suggestions that the agency make more use of a tool called “community papering,” which means filing a petition for court intervention to compel parental participation in services when a child is not being removed from the home. This resulted in a recommendation in the 2017 report that CFSA should use this tool more consistently for families that need some pressure to participate in services. In the same report, the panel also recommended that CFSA strengthen its policy and practice to “ensure families with multiple referrals to Child Protective Services receive an intensive historical review.” After 2017, there were no more recommendations for strengthening CFSA interventions with frequently reported families.

The changing ideological climate manifested itself in other ways. Serving on the CFRC, I soon realized that a striking number of child fatalities happen in extremely large families, with six, seven or as many as 12 children. Perhaps it is not so surprising. It’s hard to imagine safely caring for that number of children, especially if they are closely spaced. There was a time when this topic could be discussed, especially on the IMRT, whose members were concerned with protecting vulnerable infants. In the 2016 report, two paragraphs described discussion by the IMRT of “the concept of developing a public service media and marketing campaign focused on the health and economic benefits of family planning for all age ranges.” Clearly there was not enough support for this idea to result in a recommendation, but the discussion was robust enough to warrant inclusion in the report. Even in my earlier years on the Committee, this issue was occasionally raised by public health professionals. But it was no longer apparently an acceptable topic for discussion by the time my service ended in 2022.

The changing ideological climate also seemed to affect the CFRC’s willingness to address substance abuse. Parental use of alcohol, marijuana or illegal substances is a common factor cited in the cases reviewed by the panel. That includes the case of Trinity Jabore, who was born with marijuana in her system and later found dead at only seven weeks old, having suffered starvation, thirteen fractured ribs, and severe diaper rash. As the prosecutor of her parents put it, “They deliberately chose not to feed or take care of their infant and to instead smoke marijuana, PCP, get high and take selfies all day.” In 2018, the IMRT formed a subcommittee to look at the impact of marijuana usage on families in the District, in light of concerns raised by the legalization of cannabis use. In the 2019 report, the IMRT expressed concern about the role of marijuana and illicit substances in inducing a deep sleep from which parents did not rouse even as their dying babies fought for breath. But in the 2020 Annual Report, parental substance use was mentioned only in two tables and the text briefly summarizing them.

In the past, CFRC had recommended data sharing between agencies to improve coordination of services for the most troubled families, who are often involved with multiple agencies. In its 2016 report, reflecting the period just before I joined the panel, the CFRC recommended that the District “should allocate funding for the implementation and utilization of DC Cross Connect throughout the human services and public services cluster agencies” in order to better meet the needs of vulnerable children and families. (The recommendation was directed to the Department of Human Services (DHS), which did not have jurisdiction over the other agencies included in the recommendation, and DHS did not respond to that part of the recommendation.) Cross-Connect is an effort to integrate care between DHS, the Department of Behavioral Health, and CFSA, incuding the sharing of data. In 2022, I became aware that a similar proposal for a citywide database to track information on anyone served by DC government agencies is a key element of the Gun Violence Reduction Strategic Plan prepared for the District by the National Institute for Criminal Justice Reform, and I suggested that we might consider such a recommendation. My suggestion resulted in a presentation by the CJCC but not a new recommendation for sharing data between agencies in the District. This new ideological climate, where there is great suspicion that data sharing can be used against marginalized populations, rather than to protect their most vulnerable members, was not fertile soil for such a recommendation.

It is unfortunate that I cannot relate specific details behind the generalities that I have reported here, except those taken from published annual reports. Strict rules around the confidentiality of meetings and information shared govern the operations of CFRC. Before every meeting, members sign a confidentiality agreement promising not to disclose any information discussed during the meeting. Those rules are clearly excessive. The panel is given case histories with no names provided. These case studies can and should be available to the public (with the redaction of any information that could give away the identity of the families.) The public deserves to know that the funds it spends on child protection often fails to protection children. Hiding this information merely protects the agencies involved. That’s why I’m hoping that the DC Council will pass legislation allowing the release of the summaries provided to CFRC (with redaction of any information that would clearly give away the identity of the decedents and their families.)

In his preface to the CFRC’s 2018 report, Chief Medical Examiner Roger Mitchell stated that “the CFRC is moving toward being a leading voice in the prevention of child fatalities in the District of Columbia.” But until committee members are willing to put the needs of children first, CFRC will never be such a leading voice in preventing child fatalities in the District. Now that I am off the CFRC, I hope that other members will be courageous enough to stand up for the rights of children to be safe and well cared for, even at the risk of becoming a gadfly–which was clearly the reason for my removal.


  1. These presentations focused on: (a) Criminal Justice Coordination Committee on DC’s Gun Violence Prevention Plan and its work to implement it; and (b) the US Attorney’s ATTEND program to reduce school truancy.
  2. There was no on-boarding or training when I entered the pane, so it took me at least a year to realize that I was eligible to join the IMRT as well as the CFRT. Once I understood that CFRC members are eligible to participate on both teams, I began attending the IMRT meetings as well.
  3. This includes only those cases reviewed in full by the IMRT; this information was not available for those that were included only as part of a statistical review, which is used as a way of studying the deaths of most infants who died of natural causes. Many IMR cases are reviewed statistically not individually; for example 14 out of the 51 cases reviewed in 2019, (the last normal year before Covid) were reviewed statistically. In 2020, during which the committee missed six months of case reviews, 29 of the 47 cases reviewed were statistical reviews of infant natural deaths.

Using algorithms in child welfare: promise, confusion and controversy

Source: Vaithianathan, et. al, Allegheny Family Screening Tool v2, https://www.alleghenycountyanalytics.us/wp-content/uploads/2019/05/Methodology-V2-from-16-ACDHS-26_PredictiveRisk_Package_050119_FINAL-7.pdf

The use of algorithms developed through machine learning for the purpose of improving human decisionmaking is becoming more common in child welfare and in other areas of government, like criminal justice. These tools are often supported as a way to reduce racial and other biases by those making decisions about how individuals will be treated. But opponents have raised concerns that algorithms will increase bias because they are developed using data on systems that are already known to exhibit racial or other disparities. Early research suggests that algorithms can identify the highest-risk children referred to child protective services while reducing racial disparities. But many questions remain about how these tools work in practice and whether their effectiveness will be limited by the mandate to avoid reinforcing racial disparities in child welfare.

The Allegheny Family Screening Tool (AFST), implemented in 2016, is the first algorithm (of a type known as a Predictive Risk Model) to be used in decisions about the screening of referrals by a child protective services hotline. When a call (known as a “referral”) comes into the hotline in Allegheny County (which includes the city of Pittsburgh), the intake worker (or hotline screener) who takes the call must decide whether to screen it in for investigation or screen it out as not relevant to child abuse or neglect. A referral that alleges abuse or severe neglect is automatically forwarded for investigation. For other calls, the screener reviews the information provided by the caller, as well as information on the family’s previous interactions with the Office of Children Youth and Families (CYF) and other agencies. The screener also runs the AFST, which generates a risk score for each child that is used to supplement the professional judgment of the screener and their supervisor.

The AFST was developed to help hotline screeners decide whether a maltreatment referral warrants an in-person investigation, with the hope of improving the quality and consistency of screening decisions.1 The designers of the tool, leading child welfare researchers Rhema Vaithianaithan and Emily Putnam-Hornstein, sought to change the focus of screening decisions to the risk of future harm to the child rather than whether a referral meets the current definition of child maltreatment. In doing so, they sought to reduce both false negatives, or referrals that are screened out when maltreatment was present, and false positives, or referrals screened in where no maltreatment was present. The current version of the AFST uses data on all family members from past referrals and interactions with CYF as well as from the courts, jail, juvenile probation, behavioral health systems, and the child’s birth record to generate a risk score between one and 20 for each child included in a referral.2 The score represents the estimated risk that the child will experience a court-ordered removal from their home in the next two years. which serves as a proxy for serious abuse or neglect.3 Scores indicating a risk of 17 or higher with at least one child aged 16 or under are labeled as “high risk” and recommended to be screened in; approval from a supervisor is required to override this recommendation. Referrals with a risk score of less than 11 and no children under 12 are displayed as “low risk” and recommended for screening out, but the screener can override this recommendation without supervisory approval. For other referrals, the score is used to inform the screener’s decision, in consultation with their supervisor. The score is not seen by those who later investigate the allegations that are screened in, or by anyone else outside the intake unit.

Allegheny County commissioned an independent study by Goldhaber-Fiebert and Prince (2018) of the effect of the original AFST in the 15 to 17 months following full implementation in 2016.4 That study found a “moderate” increase in “screening accuracy,” which the researchers defined for screened-in reports as whether further action (the opening of a new service case or connection with an existing case) was taken by CYF or whether there was another referral within 60 days after the referral wass screened in. Screened-out referrals were deemed “accurate” if a child had no referrals for two months. The researchers found that the number of children being screened in “accurately” increased from about 358 to about 381 per month, or a monthly increase of roughly 24 children. (There are upper and lower bounds provided for all these numbers.) But part of this effect disappeared over time. The number of children being screened out “accurately” actually decreased slightly. The researchers also found that use of the algorithm brought about a halt in the downward trend of screening referrals in for investigaion and no “large or consistent” differences in outcomes across racial or ethnic groups. These results were somewhat disappointing to those who were hoping for a larger impact on accuracy, but also failed to support critics’ fears that the algorithm would increase racial disparities in investigations.5

In 2020, Vaithianathan, Putnam-Hornstein et al. published a study that they had done to validate the AFST by comparing risk scores to hospital injury encounters for children who had been reported to CPS. They took a large sample of 83,311 referrals for 47,305 children in Allegheny County between 2010 and 2016 (before AFST was implemented) and calculated an AFST score for each of them. They linked these children’s records with medical records from the Children’s Hospital of Pittsburgh, the sole provider of secondary care for children in the area. The researchers found large differences in the chances of an injury encounter for children depending on their risk levels. Plotting the risk level (from one to 20) against the chance of an injury hospitalization, the researchers found “a clear association between any-cause injury encounters and risk ventile, with an increase in the gradient for those scoring 17 and higher.” For children in the highest five percent risk level, their rate of an injury encounter for any cause was 14.5 per 100, compared with 4.9 per 100 for children classified as low-risk by the algorithm, who were in the bottom half of the risk distribution. For abuse-related injury encounters, the rate for high-risk children was 2.0 per 100, compared to 0.2 per 100 for low-risk children. And for suicide, the rates for the two groups were 1.0 per thousand compared to 0.1 per thousand.6

As the researchers explain in the 2020 paper, the AFST, a model that was developed to predict foster care placement is able to predict injury harm as captured in data on medical encounters. This is particularly significant because it is harm to children that they really wanted to predict, not placement in foster care, which is only a proxy for such harm. As Dee Wilson stated in his March 2023 commentary, “the 5% of highest risk children had an any-cause injury rate almost three times higher than 50% of the lowest risk children and a rate of abusive injury and self-harm and suicide 10 times the rate of the lowest risk children! AFST is a powerful algorithm when applied to one of the most important safety outcomes in child protection.” The lack of attention to this result by the media and child welfare leaders is disappointing. The extent to which this is due to poor communication by the authors and others, the complexity of the issue or the unwillingness of the child welfare establishment to receive any information suggesting the utility of predictive risk modeling, is unknown.

In an analysis that has not yet been published, Prindle et al.7 built a predictive risk model for San Diego County that was based on CPS data alone, in the absence of a data warehouse. They found a similarly strong relationship between risk scores calculated by the model and hospital encounters due to child maltreatment. Specifically, they found that “children classified by the PRM in the top 10% of risk of future foster care placement had rates of medical encounters for official maltreatment roughly 5 times those of children classified in the bottom 50% of risk.”

Rittenhouse, Putman-Hornstein and Vaithianathan (2022), in an article that is currently undergoing peer review, report finding that among all referrals, the AFST had no significant effect on racial disparities in screening decisions. And among the referrals with the highest risk scores, the AFST significantly reduced Black-White disparities. That result suggests that the high risk protocol for these referrals (requiring an investigation unless the supervisor agrees that it is not needed) plays a role in reducing racial disparities. The researchers also found that the AFST reduced Black-White disparities in case openings and home removal rates for investigated referrals. The reason for this is not clear. The authors speculate that the reduction in screening disparities among the highest-risk group might have played a role, or perhaps that within the lower-risk score groups, screeners might be shifting towards screening in Black and White children with similar risks of foster care placement.8

Hao-Fei Cheng et al. (2022), using the data from the original evaluation, compared the results of the AFST reported in the initial evaluation with the results that would have been obtained by the use of the algorithm alone, without input from screeners and their supervisors. That data, as already discussed, showed that racial disparities in screening did not increase with the use of the AFST. But Cheng et al. found that workers’ decisions reduced the disparities in screen-in rates for Black and White children from 20 percent based on the recommendations of the algorithm alone to nine percent with the workers’ input. This is a strictly theoretical result, since the algorithm was never used nor meant to be used without worker judgment. From sitting with workers as they discussed their cases and interviewing them about their use of the tool, the researchers concluded that screeners adjusted for what they saw as limitations of the AFST (such as the failure to consider the nature of the original referral) and that some consciously tried to reduce racial disparities. The researchers also found that workers’ judgments, while producing lower disparities, were also less accurate than the algorithm’s recommendations. It is not surprising that accuracy is associated with greater racial disparities; evidence indicates that the incidence of maltreatment is considerably higher among Black children than White children due to the disparities in their social and economic characteristics, which in turn reflect America’s history of slavery and racism.

Grimon and Mills (2022) studied the use in Colorado of an algorithm that was similar to the AFST but used only child wefare system data. They ran a randomized trial to compare decisions made by hotline teams that had access to the tool and those that did not. The study found that “giving workers access to the tool reduced child injury hospitalizations by 32 percent” and “considerably” narrowed racial disparities. Surprisingly, though, teams with access to the tool were more likely to choose to investigate children predicted to be low-risk, and less likely to refer for investigation those considered to be high-risk, than workers without access to the tool. Based on text analysis of discussion notes, the authors speculated that access to the tool might have allowed teams to focus on other features of the referral that are not included in the algorithm, such as the nature of the allegation itself. This counterintuitive impact on teams’ decisions is confusing and even disconcerting, since the entire purpose of the tool is to identify the children at highest risk.

Taken together, the studies of algorithmic tools used in screening of child maltreatment reports show that these tools by themselves are very good at assessing risk. When the tools actually implemented by human beings, the results are more confusing and we have fewer studies on which to rely. The initial evaluation of AFST shows a modest improvement in screening accuracy. The results of the trial of a similar tool in Colorado suggest that it achieved predictive success by doing the opposite of what was intended. Studies also find that the tools in practice do not increase racial disparities and may even decrease them. One study, however suggests that there may be a tradeoff between accuracy and the reduction of disparities, with workers disregarding the algorithm’s recommendations in order to reduce disparities at the expense of accuracy.

In addition to Allegheny County, one county in Colorado (Douglas County) is using an algorithm developed by the team of Putnam-Hornstein and Vaithanathian and another (Larimer County) is currently testing such a tool. Los Angeles County is piloting a Risk Stratification Tool designed by the same team that is being used to support supervisors in their management of investigations that are already open. It is designed to “identify investigations that may not have immediate safety concerns, but are at risk of future system involvement.” These investigations are recommended for “enhanced support.” This pilot was implemented with the hope of preventing more tragic incidents after three high-profile deaths of children by abuse whose families had had numerous interactions with the county’s child welfare agency.

Unfortunately, media outlets such as the Associated Press and the Los Angeles Times have published articles that are replete with misinformation, ignoring the promising research findings and the confusing ones as well. Both of these outlets misrepresented the study by Cheng et al., suggesting that the AFST increased racial disparities in screening. In its latest piece, the AP questioned the idea of screening in parents with mental illness, cognitive disabilities, or any “factors that parents cannnot control.” But whether or not parents can control a factor has nothing to do with its relevance to the risk to a child. These biased accounts by the press, as well as by orgahizations iike the American Civil Liberties Union, may be having an impact on government actions. Oregon stopped using an algorithm to help make screening decisions, a decision for which the AP appears eager to take credit. The AP and the PBS NewsHour along with other outlets have also reported that the Justice Department is investigating Allegheny County’s use of the AFST to determine whether it discriminates against people with disabilities or other protected groups.

Early research suggests that algorithmic tools used in child welfare have the potential to identify the children who most need protection. In practice, they seem to be capable of improving the accuracy of screening decisions without increasing racial disparities. But whether the kind of striking accuracy obtained by using the algorithms alone can be obtained without actually increasing racial disparities, given the underlying differential rates of abuse and neglect, is unknown. With the current climate that values eliminating racial disparities over the protection of children (and especially Black children), it is clear that such a tradeoff will not be considered.


  1. See Rittenhouse, K., Putnam-Hornstein, E. & Vaithianathan, R., “Algorithms, Humans and Racial Disparities in Child Protective Services: Evidence from the Allegheny Family Screening Tool,” (2022), available in full at https://krittenh.github.io/katherine-rittenhouse.com/Rittenhouse_Algorithms.pdf, for a fuller description of the screening process.
  2. These are available from Allegheny County’s unique Data Warehouse, which brings together data from a wide variety of sources.
  3. As the developers explained in a paper describing their methodology, a proxy is needed because there was no practical way to measure actual harm to children and use it to develop the algorithm; abuse and neglect data were not available and the number of adverse events like fatalities and near-fatalities would be too small. https://analytics.alleghenycounty.us/2019/05/01/developing-predictive-risk-models-support-child-maltreatment-hotline-screening-decisions/
  4. See Jeremy D. Goldhaber-Fiebert, PhD and Lea Prince, PhD, Impact Evaluation of a Predictive Risk Modeling Tool for Allegheny County’s Child Welfare Office, March 20, 2019, available from https://www.alleghenycountyanalytics.us/wpcontent/uploads/2019/05/Impact-Evaluation-from-16-ACDHS-26_PredictiveRisk_Package_050119_FINAL-6.pdf.
  5. Putnam-Hornstein contends that the results were more promising than they appear to a lay audience. She contends that the algorithm’s ability to achieve the same accuracy with (messy) real-time data as it obtained with (cleaned) historical research data was a victory in itself. She emphasizes that the size of the effect was reduced by the practices surrounding the model rather than the algorithm itself.
  6. There were no differences in rates of cancer encounters by risk level, which were assessed as a “placebo.”
  7. John Prindle, et al, “Validating a Predictive Risk Model for Child Abuse and Neglect using Medical Encounter Data.” Unpublished paper provided by Emily Putnam-Hornstein, March 25, 2023.
  8. Email from Katherine RIttenhouse to author, March 22, 2023.

Neglect as poverty: the myth that won’t go away

Image: consumerhealthday.com

It’s one of those myths that won’t go away and instead is gathering steam–the idea that parents who are found to be neglectful by child welfare agencies are really just poor people being judged for their inability to provide sufficient material support to their children. It doesn’t matter how much evidence is cited against it. The myth continues because it is an essential part of the narrative that is currently dominant in the child welfare arena. Nevertheless it’s been over a year since my last attempt to shed some light on this issue, and some new research has become available, thus it seems a good time to revisit the topic.

It’s Time to Stop Confusing Poverty With Neglect, exhorted Jerry Milner, Children’s Bureau Commissioner and his special assistant David Kelly back in January, 2020, in a typical statement of this myth. “Most of the reasons for child welfare involvement fall into what we call “neglect” rather than physical abuse or exploitation. Our most recent child maltreatment data tell us that 60 percent of victims have a finding of neglect only…More times than not, poverty and struggles to meet the basic, concrete needs of a family are a part of the equation in all types of neglect.” Miller and Kelly now sell their expertise at Family Integrity & Justice Works, an arm of the Public Knowledge consulting firm which has the goal of “replacing child welfare.”

Media outlets have taken this story and run with it. Here is the Philadelphia Inquirer: “A common misunderstanding is that the leading reason kids are taken into the foster care system is because of physical or sexual abuse. But that accounts for only one of six cases. Children far more often are removed from their homes for ‘neglect,’ which often amounts to symptoms of poverty, like food insecurity or unstable housing.”

The Biden Administration has endorsed the idea that most neglect findings reflect nothing but poverty. The Administration on Children and Families (ACF) has solicited applications for a grant of between one and two million dollars “to support the development and national dissemination of best practices to strengthen the capacity of child abuse hotline staff to distinguish between poverty and willful neglect.”

There is no federal definition of child neglect, and state definitions vary. In contrast to abuse, it is usually defined as an act of omission rather than comission. According to the Child Welfare Information Gateway, neglect is “commonly defined in state law as the failure of a parent or other person with responsibility for the child to provide needed food, clothing, shelter, medical care, or supervision to the degree that the child’s health, safety and well-being are threatened with harm.” The most commonly recognized categories of neglect include physical neglect or failure to provide for basic physical needs, failure to provide adequate supervision , educational neglect or failure to educate the child as required by law, and medical neglect.

There is no dispute that more children are found to be neglected than abused. Based on data collected by the federal government and published in Child Maltreatment 2020, three-quarters (76.1 percent) of the children found to be victims of maltreatment in 2020 were found to be neglected. A total of 16.5 percent were found to be physically abused, 9.4 percent were found to be sexually abused, and six percent were found to be victims of some other type of maltreatment.* Of the children who were removed and placed in foster care, according to the 2020 AFCARS Report, 63 percent had neglect listed as a circumstance associated with the child’s removal, compared to 12 percent with physical abuse and four percent with sexual abuse.

But the idea that neglect findings represent nothing but poverty is questionable. First, the neglect deniers would probably agree that most poor parents do not neglect their children but instead find a way to meet their needs, relying on charity, extra work, or subordinating their own wants to the needs of their children. When poor children are deprived of food, clothing or adequate housing, other factors such as substance abuse, mental illness and domestic violence are often involved. Second, more than half of the states exempt from the definition of neglect any deprivation that is due to the lack of financial means of the parents. Third, the definition of neglect is not confined to the failure to provide adequate food, clothing or shelter but instead includes other acts of omission, such as failure to protect a child from dangerous caregivers, or failure to ensure that children go to school and get needed medical care. Lack of supervision, a common form of neglect, can reflect poverty when parents feel they must rely on inadequate arrangements in order to go to work; we just don’t know the degree to which neglect findings reflect such decisions by parents.

But until now we did not have quantitative data concerning the types of neglect being investigated or the importance of risk factors like substance abuse and mental illness. A recent study from California, the nation’s most populous state, begins to fill this data gap. Palmer and colleagues used a representative sample of 295 neglect investigations that took place in California in 2017. They found that only 14 percent of the investigations involved physical neglect–the deprivation of food, clothing, and housing that is most closely connected with poverty. The most common types of neglect that were investigated were inadequate supervision, investigated in 44 percent of the cases, and failure to protect (leaving the child in the care of a known abuser or failure to intervene with known abuse), in 29 percent of cases. Moreover almost all (99 percent) of the investigations of physical neglect included concerns related to substance use, domestic violence, or mental illness; or they involved another type of maltreatment such as physical or sexual abuse or an additional neglect allegation. Thus, the authors conclude that almost no parent was investigated for material deprivation alone, although it is true that they did not separate out any lack of supervision cases that involved the inability to obtain adequate childcare for work or other necessary activities.

The evidence from California is very suggestive, but as the authors caution, it is possible that other states receive more reports that focus on unmet material needs, are less likely to screen out such reports, or emphasize them more during the investigation. This is possible because California, according to a recent study of state neglect definitions, is one of five states that have adopted an “expanded” definition of child neglect, including more neglect types and allowing for the threat of harm, rather than actual harm, in neglect findings. Studies similar to the Palmer study from other states with more limited neglect definitions would be useful.

While the California study is not sufficient to negate the presumption that findings of neglect represent nothing more than poverty, it is important to note that there are no studies supporting this viewpoint. So why does the myth that child welfare treats poverty as neglect persist despite the lack of evidence supporting it, and the many reasons for skepticism? It persists because it supports the narrative and associated policy prescriptions of the child welfare establishment today–child welfare leaders, administrators, legislatures, and influential funders like Casey Family Programs. The dominant narrative describes a racist family policing system that persecutes people only because they are Black, Indigenous or poor. The policy prescriptions involve radically shrinking or even abolishing child welfare systems.

According to the prevailing view, if omissions that are labeled neglect are strictly due to poverty, there is no need to intervene with social services or child removal. Instead, governments should provide economic benefits to neglectful parents. There is a body of research suggesting that economic support for families does help reduce maltreatment, perhaps not only by helping parents meet their children’s financial needs, but also enabling them to provide better childcare and improving parents’ mental health through stress reduction. Independent of their impact on maltreatment, I strongly support increases in the safety net for families and children. But available information suggests that it will take more than financial assistance to cure neglect in most cases. Improved economic supports will not be a replacement for services to help parents address challenges with substance abuse, domestic violence, mental health, and parenting, and for child removal when there is no other option.

What can be done to alleviate the confusion and misinformation around child neglect and poverty? Collecting better data from the states would be helpful. In its annual Child Maltreatment reports, the Children’s Bureau uses data from the National Child Abuse and Neglect Data System (NCANDS). When reporting on the type of maltreatment alleged and then found, states must pick up to four out of eight categories, including physical abuse, “neglect or deprivation of necessities,” medical neglect, sexual abuse, psychological or emotional maltreatment, sex trafficking, no alleged maltreatment, other or “unknown or missing.” It is not clear whether the “neglect” category is supposed to indicate all types of neglect or just those involving “deprivation of necessities,” but there is no way for states to clarify what they mean or to distinguish between the most common types of neglect. The same problem exists with the AFCARS data used to compile federal reports on foster care and adoption.

Clearly, a reform of the data elements that states are required to submit is needed so that resesarchers can see the types of neglect that are being alleged and found for each child. However, such an improvement would not substitute for careful research like the California study cited above because it will never be possible to rely on the thoroughness of database entries by overworked social workers. We cannot be sure they will enter all of the applicable categories, for many reasons, including that not all the applicable categories may be substantiated for a particular case. Moreover, while states are required to report on some caregiver risk factors contributing to abuse and neglect, such as alcohol and drug abuse, emotional disturbance and domestic violence, these seem to be vastly understated by the social workers who enter these factors in state databases. For example, only 26.4 percent of caregivers of maltreated children were found to have the risk factor of drug abuse and only 36 percent of removals involved parental drug abuse, according to federal data. Yet anecdotal reports from states and localities tend to indicate a much higher percentage of cases that involve substance abuse.

Thus, a reform of data collection might help, but would not solve the problem, especially considering that that many child welfare leaders and funders seem inclined to maintain the hypothesis that CPS confuses poverty with neglect. Ideally, the federal government and other funders would support more studies like that of Palmer et al, and more academics would consider performing such studies.

The myth that CPS confuses neglect with poverty is pernicious because, like other myths currently prevalent in child welfare, it runs the risk of hurting abused and neglected children. It is being used to justify dismantling child protective services, eliminating mandatory reporting, or more modest proposals to hamper these critical protections for children. The federal government should improve data collection on child neglect and associated risk factors as well as supporting additional research to provide more accurate estimates of their prevalance.

*According to the report’s authors, “other” could be anything that does not fit into the categories offered by the Child Abuse and Neglect Reporting System and includes threatened abuse and neglect, drug addiction, and lack of supervision according to state comments submitted with the data.

Did child maltreatment fall under COVID-19?

As the Covid-19 pandemic took hold, stay-at-home orders were declared, and school buildings closed, many child advocates voiced fears that child abuse and neglect would increase but would remain unreported as children were locked in with their maltreaters. But some newly available data has led to a spate of commentaries announcing triumphantly that rather than increase, child maltreatment has actually decreased during the pandemic, suggesting to some that we may not need a child welfare system after all. In fact, while the data provides no definitive evidence of either an increase or decline in child maltreatment, there are some concerning indicators from emergency room visits, teen self-reports, and domestic violence data that there may have been an increase in child abuse and neglect after Covid-19 closed in.

There are many reasons to think that the Covid-19 pandemic and our nation’s response to it would have led to a spike in child abuse and neglect. Research indicates that income loss, increased stress, and increased drug abuse and mental illness among parents (all associated with the pandemic) are all risk factors for child abuse and neglect.* On the other hand, the expansion of mutual aid networks and the influx of new government assistance programs with few strings attached may have protected children against abuse and neglect. Data on hotline calls, emergency room visits, child fatalities, teen self-reports of abuse, and domestic violence are being cited as indicators of what happened to maltreatment during the pandemic. I examine the evidence below.

Child maltreatment referrals

As soon as stay-at-home orders were imposed, child advocates warned of the likely drop in reports to child abuse hotlines as children vanished into their homes. And indeed, this is exactly what happened. Individual jurisdictions began reporting large drops in reports starting in April 2020 But national data did not become available until the publication of Child Maltreatment 2020, the compendium and analysis of data the US Children’s Bureau received from states for the fiscal year ending September 30, 2020. According to the report, there were 484,152 screened-in referrals (reports to hotlines) between April and June 2020, following the declaration of emergencies at the national and local levels and the closure of most schools buildings and subsequent transition to virtual operation. This compares to the 627,338 referrals in the same period of 2019–a decrease of 22.8 percent.** For July through September, referrals decreased from 553,199 in 2019 to 446,900, or 19.2 percent. So even in the summer when schools are mostly out anyway, referrals decreased.***

Despite the concerns among child advocates about the drop in hotline calls as a natural consequence of lockdowns and school closures, some parent advocates, such as Robert Sege and Allison Stephens writing in JAMA Pediatrics, have argued that these decreases in hotline calls show that “child physical abuse did not increase during the pandemic.”**** Similarly, In her article entitled An Unintended Abolition: Family Regulation During the COVID-19 Crisis, Anna Arons argues that the decline in hotline reports during the first three months of pandemic shutdowns in New York City relative to the same period the previous year reflects an actual decline in maltreatment rather than the predictable effects of lockdowns and school closures.

Interpreting the decline in hotline reports to suggest a decline in child maltreatment during the pandemic is either naive or disingenuous. The drop in reports was predicted by experts as soon as schools shut down because school personnel make the largest share of reports in a normal year–about 21 percent in FY 2019. The number of reports from school personnel dropped by 58.4 percent in the spring quarter and by 73.5 percent from July through September.** Exhibit 7-B from Child Maltreatment 2020 shows the drastic decline in reports from school personnel, as well as smaller decreases in reports from medical and social services professionals. To claim that this drop in reports reflects reduced abuse and neglect is to disregard the most obvious explanation-that children were seeing less of teachers and other adults who might report signs of abuse or neglect.

Source: Child Maltreatment 2020, https://www.acf.hhs.gov/sites/default/files/documents/cb/cm2020.pdf

In her article about New York City, Anna Arons cites the absence of an oft-predicted surge of child maltreatment reports when schools reopened in September 2020. Far from such a surge, she states, reports did increase, but only “at a rate in line with the typical increase in a non-pandemic fall, rather than a more dramatic leap.” But the grounds for predicting a surge in reports are far from clear. First, only 25 percent of New York City children returned to school buildings in September, as Arons reports. Moreover, is not obvious that the concept of a backlog makes sense in reference to abuse and neglect reports, as it does with tax returns, for example. Bruises may heal, a hungry child may be fed when there is money in the house; living situations may change. Many of the most troubled families are the subject of multiple reports of maltreatment over the course of a year; a child who would have been reported in the spring and again in the fall will not necessarily receive an “extra” report in the fall.*****

Emergency room visits for suspected maltreatment

As the pandemic closed in, child advocates feared that hospital emergency rooms would see an influx of maltreatment-related injuries among children. To address this question, Elizabeth Swedo and colleagues at the Center for Disease Control and Prevention used a platform that provides information on approximately 73 percent of all Emergency Department (ED) visits in the United States. The authors did not find the increase that these advocates feared, reporting that the total number of ED visits related to child abuse and neglect decreased sharply during the early part of the pandemic as compared to the analogous period in 2019, though ED visits for all causes increased even more during that period. Despite the decreases in the number of ED visits for maltreatment, the number of such visits ending in hospitalization stayed the same, which suggests there was no decrease in maltreatment severe enough to result in hospitalization.

Using an administrative database from 52 U.S. children’s hospitals, Kaiser et al. found a sharp decline in all ED visits and hospital admissions, and in visits and admissions for child physical abuse (not including admissions related to sexual abuse or neglect) during the first six months of the pandemic period compared to previous years. Moreover, they found no increase in the severity of the child physical abuse cases resulting in ED visits or hospitalizations. They concluded that coronavirus aid programs and eviction protections might have resulted in reductions in child physical abuse.

To disentangle the effects of reduced healthcare usage during the pandemic changing levels of child maltreatment, Maassel et al. looked at hospitalizations for abusive head trauma (AHT), arguing that it is more difficult for caregivers to forego medical care for such life-threatening injuries. They found a significant decrease in admissions for AHT among 49 children’s hospitals during the COVID pandemic compared to the three previous years.****** They hypothesize that the marked increase in job losses for women, along with more adults working from home, may have led to more children being cared for by two or more caregivers, and specifically fewer being cared for by sole male caregivers, who are the most common perpetrators of AHT.

Swedo et al’s finding that the number of ED visits for abuse or neglect that ended in hospitalization stayed the same contrasts with Kaiser et al and Maassel et al’s findings that hospitalizations for child abuse (and specifically) AHT declined during the early period of the pandemic. One explanation may be that abuse decreased but neglect did not; it may also be relevant that Swedo et al used a different database than did the other two teams. More research is needed to explain these differences.

Child Fatalities

One might argue that child maltreatment fatalities are best indicator of maltreatment rates during the pandemic because fatalities are less likely to avoid being reported than non-fatal maltreatment. Child Maltreatment 2020 contains estimates of child fatalities due to abuse and neglect from all states but Massachusetts, plus the District of Columbia and Puerto Rico. These jurisdictions reported a total of 1,750 fatalities, for a population rate of 2.38 per 100,000 children, compared to 1,825 or 2.50 per 100,000 children in FFY 2019. But to say that the maltreatment fatality rate went down in 2020 as compared to 2019 would be incorrect, because the fatalities counted in one year did not necessarily occur in that year. Rather, the authors indicate that “the child fatality count in this report reflects the federal fiscal year … in which the deaths are determined as due to maltreatment,” which may be different from the year the child actually died.” Such determinations may come a year or more after the fatality occurred. So it is not possible to make inferences from this small decrease in the child maltreatment fatality rate in FY 2020. Moreover, it is not not implausible that the pandemic affected reporting, so that year-to-year comparisons between pandemic years and non-pandemic years are particularly problematic.

Teen Self-Reports of Abuse

Results from a nationwide survey of 7,705 high school students conducted in the first half of 2021 and reported by the New York Times revealed disturbing indications that abuse, at least of teens, increased during the pandemic. Over half (55.1 percent) of adolescents reported being emotionally abused by a parent, and more than one in 10 (11.3 percent) reported being physically abused by a parent. Black students reported the highest rate of physical abuse by a parent–15 percent, compared to 9.8 percent for White students. Students who identified as lesbian, gay or bisexual, and those who identified as “other or questioning” experienced the highest rate of emotional abuse (74.4 percent and 75.9 percent respectively). Female students were more likely to experience emotional abuse by a parent than male students (62.8 percent vs. 46.8 percent). While using a different sampling frame, methodology and wording, a survey of a nationally representative sample of children aged 14 to 17 conducted in 2011 (as quoted by the authors of the new survey) found much lower estimates of abuse–13.9 percent for emotional abuse by a caregiver in the past year and 5.5 percent for physical abuse. The change in these percentages, even if accurate, is not necessarily due to the pandemic, but it is a troubling indicator nonetheless.

Trends in Domestic Violence

Domestic violence is highly correlated with child abuse and neglect, and the same risk factors, heightened by the pandemic, contribute to both of these problems. A systematic review of 12 US studies, most including multiple cities, concluded that domestic violence incidents in the US increased by slightly over eight percent after jurisdictions imposed stay-at-home orders. The authors speculated that the increase in domestic violence was driven by factors such as increased unemployment and financial insecurity and stress associated with childcare and homeschooling–the same factors that might contribute to increased child maltreatment.

I have written often about the propensity for wishful thinking in child welfare, whether it relates to home visiting programs, “race-blind removals,” or other programs and issues. Unfortunately, this propensity is on full display in the commentaries that try to portray reduced calls to child abuse hotlines as showing that child maltreatment did not rise during the pandemic. But it is certainly true that emergency room and hospitalization data do not provide evidence of a surge in child abuse and neglect, and there are even some suggestions that abuse may have declined perhaps due to fewer children being left alone with male caregivers. Overall, the data we have so far do not conclusively demonstrate that maltreatment rose or fell. Some children who lived through this period will eventually share their memories of life at home during the period. But these memories of course will be impossible to generalize. We may never know what really happened to maltreatment during the covid-19 pandemic.

This commentary was revised on May 18, and May 19, 2022 to incorporate new findings on ER visits and hospitalizations by Kaiser et al and Maassel et al.

*How neglect would be affected by a pandemic is somewhat less straightforward than with abuse. Many neglect cases involve lack of supervision, which may have increased with parents leaving children alone to work, with schools and childcares closed. Increased drug and alcohol abuse by parents might have also increased the occurrence of neglect. On the other hand, with more parents unemployed or working at home, lack of supervision may have become less prevalent during the pandemic.

**Unfortunately, the Bureau did not provide the total number of referrals including those screened in and screened out, by quarter. For the whole year the report shows that 54.2% of referrals were screened in, compared to 54.5% in FY 2019.

***The continued suppression of hotline calls could be due to fewer children in summer camps, summer schools, and childcare, as well as fewer attending health appointments and family gatherings in the first summer of the pandemic.

****It is not clear why Sege and Stephens refer to physical abuse only, as they data they discuss concern all types of child maltreatment,

*****However, it is interesting that even in September 2022, when almost all NYC children returned to school, reports did not return to their 2022 level. There are several reasons this could be the case, including a decline in child maltreatment and a decrease in reporting due to changes in messaging coming from ACS and advocates.

******Maassel and colleagues compared AHT admissions between March 11 and September 30 in 2020 to admissions during the same period over the previous three years.

Child Abuse Prevention Month becomes “Family Strengthening Month” in two states: will more follow?

I’m not a big fan of these months, days, and weeks dedicated to specific causes, whether they be Social Work Month, Child Abuse Prevention Month or Foster Care Month. These days, weeks, and months often allow us to feel good by paying lip service to a group or a the cause on social media without taking any concrete steps to help the group or address the problem. But when states begin renaming Child Abuse Prevention Month, there is reason to ask whether this change is a significant reflection of a changed child welfare zeitgeist.

Ronald Reagan declared April to be National Child Abuse Prevention Month in 1983, and the designation soon took hold around the country, with public and private agencies displaying blue pinwheels, sharing information about child maltreatment, and urging the public to get involved in preventing child abuse and neglect. But no longer is that the case in Utah, where April has been renamed Family Strengthening Month, or Montana, where it has been declared Strengthening Families Month.

In Utah, a document called, a bit confusingly: Family Strengthening Month: A Toolkit for 2022 Child Abuse and Neglect Prevention Month, begins with an attempt to answer the question, “Why Family Strengthening Month?” Diane Moore, the director of Utah’s Division of Child and Family Services (DCFS) starts by asserting that “focusing on…. an individual family’s failure ignores any societal or economic influence, and the potential for communities to take action to strengthen families to safely care for their own children.” This statement is confusing. Almost every commentator in the field recognizes that socioeconomic factors influence child abuse and neglect. And asking communities to support families has been a focus of child abuse prevention month on the federal level for some time.

Moore goes on to state that 55% of confirmed allegations are related to some type of neglect in Utah. The preponderance of neglect is often used by left-of-center leaders and commentators as support for the argument that child protection agencies are finding parents guilty of neglect when the real problem is poverty. But Utah is a Republican state, and Moore is not about to blame child maltreatment on poverty. Instead, she states that “High stress, substance abuse, social isolation, and lack of support for parents are among the most common risk factors associated with child abuse and neglect.” Not a word about poverty, unless “lack of support for parents” is an euphemism for it. So it’s not clear what purpose is served by the mention of neglect, or what the “economic influences” mentioned in the first sentence might be.

Moore goes on to say that “When we truly care about the safety and well being of children, then we must equally care about the safety and well being of the adults in those children’s lives.” This statement is questionable as well. Children are more vulnerable than adults, especially the youngest children, and the power imbalance is huge. Moreover, children are our future, and will parent the next set of children. Most parents put the needs of their children before their own needs, so why wouldn’t society do the same? That being said, I agree that parents must be safe and well if they are to keep their children safe and well. But if I have to choose between the well-being of a child and that of an abusive or neglectful parent, I’ll go with the wellbeing of the child any day.

Finally, Moore concludes that “We want to do more in Utah than just prevent abuse and neglect. We want to back away from that line of crisis by leaning in as communities and neighbors in order to ensure that every family has the resources and support they need to be truly successful.” More than “just” prevent abuse and neglect? If that were easy, I’d certainly be happy to aim for more, but I think we are a long way from doing that.

So Utah’s justification of the name-change depends on a set of vague and questionable statements. Then what is the real reason to take the focus off child maltreatment and replace it with “strengthening families”? This change is certainly in tune with the current climate n child welfare. We are supposed to lead with family strengths rather than weaknesses, prioritize keeping families together and minimize government intrusion in family life. If those are the priorities, child abuse and neglect prevention may have to take a back seat. We might even be willing to tolerate more abuse and neglect in order to keep families together – a bit of collateral damage, so to speak. The social worker and supervisor working with Noah Cuatro‘s family wanted to concentrate on its strengths, not its weaknesses. So they ignored the signs of abuse, and Noah was killed by his parents. Collateral damage.

It is interesting that two red states were the first to drop the “Child Abuse Prevention Month” designation. As a child advocate, I have been more critical of Democratic leaders and commentators, because they have tended to be more extreme, with statements equating neglect with poverty proposals like abolishing the “family policing system.” But I’ve been equally hard on the Trump and the Biden appointees to the Administration on Children and Families, because their views are essentially the same. And that is because child welfare is an issue where both sides of the aisle often agree on what I think are terrible policies. The focus on parents’ rights rather than children’s needs jibes with the Left’s focus on racism as the cause of almost everything and its reluctance to punish parents who may be victims of poverty. For the right, parents’ rights have always been important: keep your government out of my family, except when it comes to abortion and birth control. That’s how Left and Right could agree on the Family First Act, a terrible bill which transferred the costs of necessary group care to states while paying lip service to family preservation by offering funding for services that were already funded from other sources.

In Texas, Democrats and Republicans agreed in the 2021 legislative session on a slate of reforms designed to restrict CPS intervention into the lives of families. These laws were pushed by a coalition of strange bedfellows indeed: “abolitionists” who want to abolish child welfare along with police and prisons, with conservative groups intent on reducing government intrusion into families.

So it turns out that two “red” states were the first to rename Child Abuse Prevention month to focus on strengthening families. But next to follow suit may be one or more blue states that are eager to demonstrate their progressive bona fides. Who will be the next? Stay tuned.

The power of wishful thinking revisited: the improbable growth of Healthy Families America

I have written before about the power of wishful thinking and how it causes people to disregard research and real-life results. In that earlier commentary, I discussed the successful promotion of a practice called race-blind removals based on data from an article by a scholar who now denies knowledge of their provenance, and which have been shown to be inaccurate. A program called Healthy Families America (HFA), which currently serves over 70,000 families per year according to its website, offers another example of the power of wishful thinking. This program has become the centerpiece of the nation’s oldest and largest charity dedicated to the prevention of child abuse, even though the program has failed to demonstrate its utility in preventing child maltreatment. This organization, now called Prevent Child Abuse America, launched HFA based on weak evidence that a program in Hawaii called Healthy Start Program (HSP) could prevent child maltreatment. The first experimental study of HSP found no impact on child maltreatment but did nothing to derail the launch of HFA. Studies of HFA programs around the country have found little evidence of reductions in child maltreatment, but the program has continued to grow and now serves more families than any other home visiting program. The story of HFA is a lesson in the power of wishful thinking and the failure of evidence (or lack thereof) to counteract it.

As told in a helpful history of home visiting, all modern programs can trace their origins to Henry Kempe, whose book, The Battered Child, brought about the recognition of child maltreatment as a national problem. To address child abuse, Kempe called for universal prevention through a network of home health visitors. Inspired by Kempe, modern home visiting began with Hawaii’s implementation of the Healthy Start Project (HSP) in 1975. As described in the 1999 evaluation by Duggan and colleagues, HSP was developed by the Hawaii Family Stress Center (HFSC) on the island of Oahu. It had two components: early identification (at the birthing hospital) of families with newborns at risk of child abuse and neglect and home visiting by trained paraprofessionals for those families classified as at-risk who agreed to participate. This initial program was never evaluated, but anecdotal information suggested it was successful in promoting effective parenting, and six similar programs were established on neighboring islands.

As described by Duggan et al., the Hawaii Legislature authorized a three-year pilot program focusing on one neighborhood in Oahu, which began in 1985. There was no control group in the pilot study, and the researchers used CPS reports and changes in family stress in participating families to measure program effectiveness. During the three-year pilot, there were few reports of physical abuse, neglect or imminent harm for program participants. Because evaluations of other home visiting programs had found much higher rates of reported maltreatment in comparison group families, these results were viewed as evidence that the program had a positive impact. According to Duggan and her co-authors, “The pilot study results might have been given too much weight, given the lack of a control group and the short period of follow-up for most families.” Nevertheless, the results of this unpublished study were enough evidence for the Legislature to expand HSP throughout Hawaii starting in 1989.

Home visiting in general was gathering steam in the 1980s and early 1990’s. In 1990, the U.S. General Accounting Office (GAO) issued a report promoting home visitation as a “promising early intervention strategy for at-risk families.” In its summary of research evidence, GAO focused mostly on health and developmental benefits for children, rather than maltreatment prevention. In 1991, the U.S. Advisory Board on Child Abuse and Neglect issued a report recommending a pilot of universal voluntary neonatal home visitation, stating that the efficacy of home visiting as a preventive measure was “already well-established.” The Board cited the results of a federally-funded demonstration begun 17 years earlier as well as the the nurse home visitation program started by David Olds in 1977. But HSP was not mentioned.

Despite the lack of a rigorous evaluation, the National Committee to Prevent Child Abuse (NCPCA, now called Prevent Child Abuse America) the nation’s “oldest and largest organization committed to preventing child abuse and neglect before it happens,” had become interested in using HSP as the nucleus of a national home visiting program. But first, NCPCA conducted a one-year randomized trial of HSP, as described by Duggan et al. The trial suffered from severe methodological limitations, including “less than ideal followup,” differential dropout rates in the program and control groups, the failure to blind interviewers to experimental or control status, and reliance on program staff rather than researchers to measure some outcomes. Nevertheless, the trial concluded that HSP reduced child maltreatment, and this apparently gave NCPCA the assurance it needed to invest in the model.

NCPCA launched Healthy Families America in 1992, with financial support from Ronald MacDonald House Charities, arranging visits to 22 states by Hawaii Family Stress Center Staff. The “theory of change,” or theoretical basis for the program, as quoted by Duggan et al, started with the targeting to all newborns and their parents, which allows for diversified service options determined by individual need. Also part of the theory was a commitment to change at the individual and community levels. Rather than impose a single service model, HFA contained a set of critical elements, which included the prenatal initiation of services and the assessment of all new parents. A network was launched to bring together researchers doing experimental and quasi-experimental studies of HFA programs.

Unlike NCPCA, The Hawaii Department of Health recognized the limitations of both the pilot study and the NCPCA study and initiated a more rigorous evaluation of HSP in 1994. This was a randomized controlled trial, with at-risk families identified at the hospital and randomly assigned to the experimental and control groups. In 1999 the results of the Evaluation of Hawaii’s Healthy Start Program were released as part of an issue of the Future of Children journal containing evaluations of six different home visiting models.  No overall positive program impact emerged after two years of service in terms of child maltreatment (according to maternal reports and child protective services reports). Early HFA evaluation results, published in the same issue, also failed to find effects on abuse and neglect in three randomized trials, which included the HSP evaluation discussed above and another Hawaii HSP study.

David Olds had had begun testing his Nurse Home Visiting Program in 1977 and already had long-term results on the program in Elmira, NY, as well as shorter-term results for a replication in Memphis, Tenn. That program, now known as Nurse Family Partnership, was very different from HFA. It was restricted to first-time teenage mothers and the home visitors were nurses rather than paraprofessionals. The nurses followed detailed protocols for each visit. The researchers found that among low-income unmarried women (but not other participants), the program reduced the rate of childhood injuries and ingestions of hazardous substances that could be associated with child abuse or neglect. Follow-up of the Elmira group when the children were 15 found that the nurse-visited mothers were significantly less likely to have at least one substantiated report of maltreatment. These results are particularly impressive because they overrode a tendency for nurse-visited families to be reported for maltreatment by their nurse visitors. The researchers concluded that the use of nurses, rather than paraprofessionals, was key to the success of the program. In their analysis of all six studies published in the volume, Deanna Gomby et al. concluded that while the HFA and HSP evaluations showed some change in maternal attitudes and self-reported behaviors related to abuse and neglect, only the Nurse Home Visiting Program showed impacts on abuse and neglect other than from self-reports.

Gomby and her co-authors also concluded that the results of the six home visiting evaluations were discouraging for those who had high hopes for home visiting for solving an array of problems. All the programs “struggled to enroll, engage and retain families.” Program benefits generally accrued to only a subset of enrolled families and were often quite modest. The authors explained the disappointing results by concluding that human behavior is hard to change, particularly when problems are community-wide. They recommended that “any new expansion of home visiting programs be reassessed in light of the findings presented in this journal issue” and stated that home visiting services are “best funded as part of a broad set of services for families and children.”

But the home visiting juggernaut was already in motion nationwide. And NCPCA had already made HFA its centerpiece program. Home visiting grew, and HFA grew with it. In 2010, Congress created the Maternal, Infant and Early Childhood Home Visiting Program (MIECHV), which was re-authorized in 2018 with funding of $400 million per year through FY 2022. According to the HFA website, HFA is the model most frequently implemented with MIECHV dollars. Home visiting programs can also receive funding through Medicaid, Title IVB and IVE of the Social Security Act, and many other funding sources.

The infusion of funding for HFA research by NCPCA initiative set in motion a multitude of research projects (both randomized trials and less rigorous studies) that continues to result in publications. Nevertheless, HFA research has yet to find solid evidence that these programs have an impact on child maltreatment: The California Evidence-Based Clearinghouse for Child Welfare (CEBC), the pre-eminent child welfare program clearinghouse, reviewed 19 research reports on HFA. It gave the program a rating of “4” on a scale of 1 to 5 for prevention of child abuse and neglect, meaning the evidence fails to demonstrate that the HFA has an effect on abuse and neglect. HFA did receive a rating of 1 for “child well-being,” based on its impacts on outcomes like physical health, child development, and school readiness. In contrast, Nurse Family Partnership was rated as “1,” “well-supported by the research evidence, for the prevention of child abuse and neglect, as well as for child well-being.

Despite the lack of evidence of its impact on maltreatment, HFA received a rating of “Well Supported” from the new clearinghouse established by the Family First Prevention Services Act (“Family First”) to determine whether a program can receive federal funding under Title IV-E of the Social Security Act. To get such a rating, the program must show improved outcomes based on at least two randomized trials or rigorous quasi-experimental studies. But these outcomes could be any sort of “important child and parent outcome,” (not just child abuse or neglect) and there is no standard for how to measure each outcome. Based on its review of all HFA studies that met their criteria for inclusion, the Clearinghouse found 23 favorable effects, 212 findings of no effect, and four unfavorable effects across 16 outcomes. This included five favorable effects on child safety based on parents’ self-reports of maltreatment, with no favorable effects on other measures of child safety. Self-report is generally frowned upon as a measure of child maltreatment, for obvious reasons. A positive impact of HFA might reflect that participants in HFA were more eager than control group members to provide the “right answer” to questions about maltreatment.

The “well-supported” rating from the Title IV-E clearinghouse opened up a new source of funding for HFA. Passage of Family First as Title VII of the Bipartisan Budget Act of 2018, allowed states to spend Title IV-E funds on programs on services to families with a child welfare in-home case. To take advantage of this new demand, HFA announced in September 2018 that families referred by the child welfare system were now able to enroll until 24 months of age. To serve these families, HFA introduced special child welfare protocols, with limited evidence that that the program was effective for parents who had already abused or neglected their children.* The program had already departed from its initial mission of screening all families with newborns in a geographic area. Even without the child welfare protocols, each program can choose its own admission criteria and there is no universal screening; potential participants are generally referred by health or child welfare agencies, who often can choose between several home visiting programs when referring a client.

Another part of HFA’s original theory of change was a “dual commitment to change at the individual and community levels.” As described by Daro and Harding in their 1999 evaluation of HSA, this meant that HFA “must move beyond direct efforts to help families and begin to serve as a catalyst for reshaping existing child welfare and health care efforts and improving coordination among other prevention and family support initiatives.” This vision has clearly gone by the wayside as HFA has become one choice in a menu of home visiting programs offered by local jurisdictions. Far from trying to enhance and coordinate available community offerings, HFA is busy trying to maximize its share of the pie through its public relations effort, exemplified by the self-promotional statements on its website.

It is disappointing that Prevent Child Abuse America (“Prevent Child Abuse,” formerly NCPCA), an organization that defines its mission as child abuse prevention, decided to fund HFA before it was proven to prevent child maltreatment and without apparently considering other approaches also being tested at the time. And it is concerning that the organization continued with this commitment even after the disappointing evaluations of 1999 might have led them to diversify their investment beyond HFA or even beyond home visiting or to focus more on advocacy rather than services. And finally, that Prevent Child Abuse continues to use charitable contributions made for the prevention of child abuse and neglect to fund a program that has not been proven after 40 years to accomplish this goal, raises serious ethical questions. Twenty-two of the 40 staff listed on the Prevent Child Abuse website have positions with Healthy Families America. Perhaps the charity has backed itself into a corner; it would be difficult to escape this commitment without serious repercussions.

Some federal administrators do not seem to be much more interested in evaluation results than Prevent Child Abuse. The legislation authorizing MCHIEV required a randomized controlled trial (RCT), which may provide useful information on the relative merits of these programs in addressing different outcomes. But strangely, HHS indicated in a response to a critique from the Straight Talk on Evidence Blog that it is not interested in a “horse race” between the models but rather is interested in assessing home visiting in general. This odd statement is an indicator of the kind of thinking that allowed Prevent Child Abuse to invest in HFA for 40 years despite the lack of evidence that it does “Prevent Child Abuse.”

The story of Healthy Families America is not an unusual one. My discussion of the Homebuilders program could also be called “the power of wishful thinking.” Such stories are all too frequent. They show us how wishful thinking can drive leaders to disregard research, especially after they have made a premature decision to commit to one program or course of action.

*One study of Healthy Families New York, published in 2018, looked at a subgroup of 104 mothers who already had a substantiated CPS report, and found a decrease in abuse and neglect among the mothers who were in the experimental group. However, the sample was small and was not planned in advance, so the authors recommend further testing home visiting programs as prevention of repeat maltreatment for child welfare-involved mothers.

When parents’ rights trump children’s needs

Photo by Filipe Leme on Pexels.com

Sometimes it seems like basic humanity and common sense get lost in the scramble to affirm parents’ rights at all cost. Nowhere was this more clear than in a quote from Aysha Shomburg, the former New York City child welfare official who was appointed by President Biden to head the Children’s Bureau. As quoted in The Imprint, Schomberg cited a 15-year-old father facing a termination of parental rights as evidence for the need to eliminate the timelines imposed by the Adoption and Safe Families Act. Speaking of this teen dad, Schomberg said, “That stays in my mind and makes me think, how many young fathers are out there and maybe want to take care of their children, but are maybe up against this timeline?”

After picking my jaw up off the floor, I wondered whether Schomberg thought a fifteen-year old was actually capable of parenting an infant, or whether she thinks ASFA should be amended so a child can stay in foster care as many years as it takes for the parent to grow up.

Schomburg’s statement reminded me of one of the saddest cases I carried as a social worker in the District of Columbia’s foster care system. A two-month-old (I’ll call him “Shawn”) came into care when he was removed from his teenage mother (“Shameka”) after she swung him wildly in his carseat and then stalked off in a temper from a home for teen mothers, abandoning her son there. Shawn was placed with one of the best foster families I have ever known–“the Smiths,” a couple who was Black like Shawn and had raised their own children and fostered numerous others. They quickly fell in love with Shawn and gave him the kind of parenting that textbooks envision. Mrs. Smith stayed home with Shawn all day, talking to him, playing with him, and loving him, until the Smiths placed him in a carefully-chosen early childhood education setting at the age of two. Shawn was the center of Mr. and Mrs. Smith’s lives and part of their extended family of children and grandchildren. I’ll never forget that when he fell in love with trains, they found every train-related toy, game or event.

As the months and then years rolled by, Shawn’s mother stopped visiting him. She had named a father for Shawn, but a paternity test came back negative. Shawn’s goal was changed to adoption with the Smiths and I imagined the happy life awaiting him in their loving home. But one day, Shameka admitted that she had lied about the name of the biological father for the sake of revenge against him. She named the real father, and the paternity test was positive. The father (“Antonio”) soon showed up at the agency, a pleasant seventeen-year-old who was delighted to meet his adorable young son. Shawn’s birth father lived with his parents and siblings in subsidized housing and relied on government assistance. Shawn’s grandfather was excited about the new family member. He told me that two of his older sons also had children as teenagers, and that becoming fathers is what made them actually grow up, finish high school, and get jobs.

The Smiths were devastated, but I assured them that the court would not rip a two-year-old away from the only parents he had ever known. But then I talked to the agency attorney and realized there was no question in her mind that the agency had to change the goal to “reunification” with the father, a perfect stranger. And that is exactly what happened. The goal was changed and the Smiths had to bring Shawn to the agency for progressively longer visits with his birth father. At one visit, Mr. Smith was heard crying in the bathroom.

I am glad I was no longer at the agency when Shawn went ‘home’ with his father. But I’ll never forget the day I ran into Shawn’s Guardian ad Litem, the attorney appointed to represent him in court. “We ruined his life,” she told me. She had visited him often in the months following his return home, and and reported that his new household was chaotic, with none of the routine and predictability so crucial for growing children. And we will never know the effects of being ripped away from the Smiths after two years of security and attachment.

I thought about Shawn when I read Aysha Shomburg’s post. I wondered whether Schomburg cared more about the fifteen year-old than about his son. It was not about the infant’s future. It was about the father’s rights. And indeed, most child welfare officials would say Schomburg was correct in not speculating about the child’s future. Child welfare agencies are not in the business of choosing the best parent, just ensuring that the birth parents can provide the minimal acceptable care. But what about the attachment that Shawn had developed over two years with the Smiths? The importance of attachment, and the consequences of disrupting it for a young child, is why the timelines were included in ASFA–the timelines that Schomberg wants to eliminate. So attachment – and the trauma of disrupting it – does not seem to be a significant issue for her.

Schomburg’s citation of a fifteen year old father as an argument against permanency timelines is an illustration of what’s wrong with mainstream child welfare thinking today. It’s all about parents’ rights, while the most basic of children’s needs are disregarded. It is based on an idealized vision of families rather than the way they really are. It’s the kind of thinking that allowed a child named Noah Cuatro to die when the Los Angeles Department of Children and Family Services told social workers to emphasize his family’s strengths more than its weaknesses. We must stop using that kind of thinking to prescribe our actions toward our most vulnerable citizens–our youngest children.

New data show drop in foster care numbers during pandemic

Source: US Children’s Bureau, AFCARS Report $28, https://www.acf.hhs.gov/sites/default/files/documents/cb/afcarsreport28.pdf

A long-awaited report from the federal government shows that most states saw a decrease in their foster care population during the fiscal year ending September 30, 2020, which included the onset of the COVID-19 pandemic. Both entries to foster care and exits from it declined in Fiscal Year (FY) 2020 compared to the previous fiscal year. These results are not surprising. Stay-at-home orders and school closures beginning in March 2021 resulted in a sharp drop in reports to child abuse hotlines, which in turn presumably brought about the reduction in children entering foster care. At the other end of the foster care pipeline, court shutdowns and a slow transition to virtual operations prolonged foster care stays for many youths. One result that is surprising, however, is the lack of a major decrease in children aging out of foster care, despite the widespread concern about young people being forced out of foster care during a pandemic.

Ever since the COVID-19 pandemic resulted in lockdowns and shut down schools around the country, child welfare researchers have been speculating about the pandemic’s impact on the number of children in foster care. While many states have released data on foster care caseloads following the onset of the pandemic, it was not until November 19, 2021 that the federal Children’s Bureau of the Administration of Children and Families (ACF) released the data it received from the 50 states, the District of Columbia and Puerto Rico for Fiscal Year 2020, which ended more than a year ago on September 30, 2020. The pandemic’s lockdowns and school closures began in the sixth month of the fiscal year, March 2020, so its effects should have been felt during approximately seven months, or slightly over half of the year. The data summarized here are drawn from the Adoption and Foster Care Analysis System (AFCARS) report for Fiscal Year 2020 compared to the 2019 report as well as an analysis of trends in foster care and adoption between FY 2011 and FY 2020. State by state data are taken from an Excel spreadsheet available on the ACF website.

The nation’s foster care population declined from 426,566 on September 30, 2020 to 407,493 children on September 30, 2021. That is a decline of 19,073 or 4.47 percent. According to the Children’s Bureau, this is the largest decrease in the past decade, and the lowest number of children in foster care since FY 2014.* Forty-one states plus Washington DC and Puerto Rico had an overall decrease in their foster care population, with only seven states seeing an increase. The seven states with increases were Arizona, Arkansas, Illinois, Maine, Nebraska, North Dakota and West Virginia. The change in a state’s foster care population depends on the number of entries and the number of exits from foster care. And indeed both entries and exits fell to historic lows in FY 2020. The reduction in entries was even greater than the fall in exits, which was why the number of children in foster care declined rather than increasing.

Entries into foster care fell drastically around the country, from 252,352 in FY 2019 to 216,838 in FY 2020 – a decrease of 14 percent. This was the lowest number of foster care entries since AFCARS data collection began 20 years ago. Foster care entries dropped in all but three states – Arkansas, Illinois, and North Dakota. These three states were also among the seven states with increased total foster care caseloads. It is not surprising that entries into foster care dropped in the wake of pandemic stay-at-home orders and school closings. While we are still waiting for the release of national data on child maltreatment reports in the wake of the pandemic, which are included in a different Children’s Bureau publication, media stories from almost every state indicate that calls to child abuse hotlines fell dramatically. This drop in calls would have led to a fall in investigations and likely a decline in the number of children removed from their homes. Monthly data analyzed by the Children’s Bureau drives home the impact of the Covid-19 pandemic on foster care entries. More than half of the decrease in entries was accounted for by the drops in March, April, and May, immediately following the onset of stay-at-home orders, which were later relaxed or removed, as well as school closures.

Source: Trends in Foster Care and Adoption, FY 2011-FY 2020, https://www.acf.hhs.gov/sites/default/files/documents/cb/trends_fostercare_adoption_11thru20.pdf

Reasons for entry into foster care in FY 2020 remained about the same proportionally as in the previous year, with 64 percent entering for a reason categorized as “neglect,” 35 percent for parental drug abuse, 13 percent for physical abuse, nine percent for housing related reasons and smaller percentages for parental incarceration, parental alcohol abuse, and sexual abuse. (A child may enter foster care for more than one reason, so the percentages add up to more than 100.)

Exits from foster care also decreased nationwide from 249,675 in FY 2019 to 224,396 in FY 2020 – a decrease of 10 percent – a large decrease but not as big as the decrease in entries, which explains why foster care numbers decreased nationwide. Only six states had an increase in foster care exits: Alaska, Illinois, North Carolina, Rhode island, South Dakota and Tennessee. Along with the decrease in exits, the mean time in care rose only slightly from 20.0 to 20.5 months in care, while the median rose from 15.5 to 15.9 months in care. Again, it is not surprising that the pandemic would lead to reduced exits from foster care. In order to reunify with their children, most parents are required to participate in services such as therapy and drug treatment, to obtain new housing, or to do other things that are contingent on assistance from government or private agencies. Child welfare agency staff and courts are also involved the process of exiting from foster care due to reunification, adoption, or guardianship. All of these systems were disrupted by the pandemic and took time to adjust to virtual operations. Monthly data shows that about 68 percent of the decrease in exits was accounted for by the first three months of the pandemic, when agencies and courts were struggling to transition to virtual operations. It is encouraging that the number of exits was approaching normal by September 2020; it will be interesting to see if the number of exits was higher than normal in the early months of FY 2021.

Source: Trends in Foster Care and Adoption, FY 2011-FY 2020, https://www.acf.hhs.gov/sites/default/files/documents/cb/trends_fostercare_adoption_11thru20.pdf

Most exits from foster care occur through family reunification, adoption, guardianship, and emancipation. The proportions exiting for each reason in FY 2020 remained similar to the previous year, while the total number of exits dropped, as shown in Table 3 below. Children exiting through reunification were 48 percent of the young people exiting foster care in FY 2020, and the number of children exiting through reunification dropped by 8.3 percent from FY 2019. Children exiting through adoption were 26 percent of those leaving foster care, and the number of children exiting through adoption fell by 12.6 percent. Exits to guardianships fell by 11 percent and other less frequent reasons for exit fell as well. The drop in reunifications, adoptions and guardianships is not surprising given court delays and also the likely pause in other agency activities during the pandemic. However, nine states did see an increase in children exiting through adoption.

Table 3

Reasons for Exit from Foster Care, FY 2019 and FY 2020

Exit ReasonFY 2019
FY 2019
FY 2020
FY 2020
Living with another relative15,4226%12,4636%2,95919%
Transfer to another agency2,7261%2,2631%46317%
Death of Child3850%3600%256%
Source: US Children’s Bureau, AFCARS Report $28, https://www.acf.hhs.gov/sites/default/files/documents/cb/afcarsreport28.pdf

It is surprising that the number of foster care exits due to emancipation or “aging out” of foster care fell only slightly, to 20,010 in FY 2020 from 20,445 in FY 2019, making emancipations a slightly higher percentage of exits in FY 2020–8.9 percent, vs. 8.2 percent in FY 2019. There has been widespread concern about youth aging out of foster care during the pandemic, and a federal moratorium on emancipations was passed after the fiscal year ended. At least two jurisdictions, California and the District of Columbia, allowed youth to remain in care past their twenty-first birthdays due to the pandemic. It is surprising that this policy in California, with 50,737 youth in care or 12.45 percent of the nation’s foster youth on September 30, 2020, did not result in a bigger drop in emancipation exits nationwide. California’s foster care extension took effect on April 17, 2020 through an executive order by the Governor and was later expanded through the state budget to June 30, 2021. And indeed, data from California via the Child Welfare Indicators Project show that the number of youth exiting through emancipation dropped by over 1,000 from 3,618 in FY 2019 to 2,615 in FY 2020. Since total emancipation exits dropped by only 435 nationwide, it appears that the number of youth exiting care through emancipation outside of California actually increased. This raises concern about the fate of those young people who aged out of care during the first seven months of the pandemic.

In December 2020 (after the Fiscal Year was already over), Congress passed the Supporting Foster Youth and Families Through the Pandemic Act (P.L. 116-260), which banned states from allowing a child to age out of foster care before October 1, 2021, allowed youth who have exited foster care during the pandemic to return to care and added federal funding for this purpose. But this occurred after the end of FY 2020 so it did not affect the numbers for that year. Moreover, The Imprint reported in March 2021 that many states were not offering youth the option to stay in care despite the legislation, raising fears that the number of emancipations in FY 2021 may not have been much lower than the number for FY 2020.

Among the other data included in the AFCARS report, terminations of parental rights decreased by 11.2 percent in FY 2020. This is not surprising given the court shutdowns and delays. Perhaps this decline in TPR’s explains why the number of children waiting to be adopted actually decreased from 123,809 to 117,470, contrary to what might be expected from the decrease in adoptions.

It is disconcerting that some child welfare leaders and media outlets are portraying the reductions in foster care caseloads during FY 2020 as a beneficial byproduct of the pandemic. Despite the fact that maltreatment reports dropped by about half after the pandemic struck, Commissioner David Hansell of New York City’s Administration for Children’s Services told the Imprint that “It was just as likely that the pandemic was ‘a very positive thing’ for children, who were able to spend more time with their parents.” Based on an interview with Connecticut’s Commissioner of Children and Families, an NBC reporter stated that ‘With the pandemic, the last two years have been difficult, but something positive has also happened during that time span. Today, there are fewer kids in foster care in Connecticut.”

Even In normal times, I take issue with using reductions in foster care numbers as an indicator of success. Certainly if foster care placements can be reduced without increasing harm to children, that is a good thing. But in the wake of the pandemic, we know that many children were isolated from adults other than their parents due to stay-at-home orders and school closures, and we have seen a drastic decline in calls to child abuse hotlines. Thus, it is likely that some children were left in unsafe situations. Moreover, the pandemic caused increased stress to many parents, which may have led to increased maltreatment, as some evidence is beginning to show. So when Oregon’s Deputy Director of Child Welfare Practice and Programming told a reporter that “Even though we had fewer calls, the right calls were coming in and we got to the children who needed us,” one wonders how she knows that was the case, and whether her statement reflects wishful thinking rather than actual information.**

There have been many predictions of an onslaught of calls to child protective services hotlines once children returned to school. And indeed, there have been reports of a surge of calls after schools re-opened in Arizona, Kentucky, upstate New York, and other places, but we will have to wait another year for the national data on CPS reports and foster care entries after pandemic closures lifted.

The FY 2020 data on foster care around the country provided in the long-awaited AFCARS report contains few surprises. As expected by many, foster care entries and exits both fell in the first year of the pandemic. Since entries fell more than exits, the total number of children in foster care fell by over four percent. These numbers raise concerns regarding children who remained in unsafe homes and those who stayed in foster care too long due to agency and court delays. The one surprise was a concerning one: the lack of a major pandemic impact on the number of youth aging out of care. The second pandemic fiscal year has already come and gone, but it will be another year before we can get a national picture of how child welfare systems adjusted to operating during a pandemic.

*Our percentages are slightly different from those in the federal Trends report because the Children’s Bureau calculated their percentages based on numbers rounded to the nearest thousand.

*There is evidence that maltreatment referrals from school personnel are less likely to be substantiated than reports from other groups, and this may reflect their tendency to make referrals that do not rise to the level of maltreatment, perhaps out of concern to comply with mandatory reporting requirements. Data from the first three months of the pandemic shared in a webinar showed that referrals which had a lower risk score (measured by predictive analytics) tended to drop off more than referrals with a higher risk score. However as I pointed out in an earlier post, that low-risk referrals dropped off more does not mean that high-risk referrals were not lost as well.

Child maltreatment, home schooling, and an organization in need of support

The Turpin family has been in the public eye once again after NBC broadcast Diane Sawyer’s interview with two of the victims rescued from the “House of Horrors” in Perris, California on January 24, 2018. In riveting footage, Jordan Turpin describes how as a 17-year-old she escaped through a window and called the police on a de-activated cell phone which her parents did not know she had. Never having left the house by herself or spoken to a stranger, she managed to convince a sheriff’s deputy with cellphone photos of her sisters in chains. “If something happened to me, at least I died trying,” Jordan told Sawyer, stating her parents would have killed her if they had caught her. Body camera footage shows deputies walking through the trash-filled house and finding Jordan’s 12 siblings, all but the youngest stunted by malnutrition, one in chains and two others with bruised wrists from chains that had been removed and hidden while the deputies were knocking on the door. Louise and David Turpin have pleaded guilty to multiple counts of cruelty to a dependent adult abuse, false imprisonment, child abuse, and torture, and have been sentenced to 25 years in prison.

David and Louise Turpin were able to hide their extreme abuse and neglect behind the facade of a “private school” operating out of their home. Calling their home a private school is one of the options for homeschooling parents in California. These “schools” are not monitored or inspected aside from an annual fire inspection for those with six or more students, but city officials in the aftermath of the rescue could find no record that such an inspection was ever conducted on the trash-filled and hazardous Turpin home.

California is not atypical in its minimal regulation of homeschooling. As William and Mary’s James Dwyer stated at a 2021 Homeschooling Summit sponsored by Harvard Law School and described here), twelve states require nothing of homeschooling parents, not even notification to the school district; another 15 or so require notification only. The other half of states have some requirements, such as that the parent have a high school degree, that certain subjects be taught, or that students be assessed requirements, but these are generally not reviewed or enforced in a meaningful way. Moreover, no state requires that a state employee or contractor set eyes on the child once homeschooling is approved.

Clearly, the Turpins could not have gotten away with such severe abuse if the children had been in school. Teachers would have seen the extreme malnutrition of the children and the marks from chains and beatings, and the children would have been able to disclose what was happening to them. Education personnel make more child abuse reports than any other group; they made 21 percent of calls to child abuse hotlines in 2019. So it is not surprising that a disproportionate number of the horrific abuse deaths that make the news (such as the Hart children, Natalie Finn in Iowa, Matthew Tirado in Massachusetts and Adrian Jones in Kansas), involved parents who hid behind the guise of homeschooling, even though schooling rarely took place in these homes.

We have no systematic data about the association of homeschooling with child maltreatment due to data limitations. But there are some troubling reports. Child abuse pediatrician Barbara Knox studied 28 children who were victims of abuse so severe that it merits the definition of torture. In most of these cases, the children were kept out of school; about 29 percent were never enrolled in school and another 49 percent were removed from school, allegedly for homeschooling, often after a CPS report was made by education personnel. Connecticut’s Office of the Child Advocate found that of children withdrawn to be homeschooled between 2013 and 2016, 36 percent had at least one prior accepted report for suspected abuse or neglect to the Department of Children’s Services, and the majority of these families had multiple prior reports for suspected maltreatment. The Coalition for Responsible Home Education maintains a database called Homeschooling’s Invisible Children, which includes 454 cases of severe and fatal child abuse in homeschool settings in the United States since the year 1986. Since these are only the cases that made it into the media and were found by CRHE, there may be many more.

Data from the National Center for Education Statistics indicates that the percentage of Americn children who were homeschooled rose dramatically from 1.7 percent in 1999 to 3.4 percent in 2011-2012, then decreased slightly through 2019. There is some anecdotal and statistical evidence that homeschooling rose considerably during the pandemic but no definitive data as of yet; we also do not know how many children will return to school buildings when the pandemic recedes.

While abusive and neglectful parents are likely a very small minority of those who homeschool, the lobbies that represent them oppose any regulation of homeschooling, arguing that the vast majority of homeschooling parents should not be punished for the actions of a small minority. Homeschool parents who oppose regulation are represented by strong lobbies in both state capitals and at the national level. Homeschooling’s national lobby, the Home School Legal Defense Association (HSLDA) resembles the National Rifle Association in the single-minded passion of its members and its surplus of legal resources. Payment of annual dues of $130 to $144 per year buys free legal defense and representation in court for members. Adults who grew up homeschooled reported at the Harvard summit that their parents kept the organization’s telephone number on their refrigerators to be called as soon as CPS showed up at the door. HSLDA sends out email blasts to its members that can result in a barrage of phone calls that can swamp legislators’ offices and even in-person threats and harassment of state legislators, as an investigation by Pro Publica found. 

There are strong homeschool lobbies at the state level as well. In the aftermath of the Turpin case, California Assemblyman Jose Medina introduced a bill that would require a fire inspection for all private schools, including those with five or fewer students. Due to a “massive outcry” from the homeschooling community, the the inspection requirement was eliminated, leaving a bill that required nothing but identification of homeschooling families by name and address. When the eviscerated bill was scheduled for a hearing, hundreds or perhaps thousands of homeschooling families poured into the capitol building, testifying for three hours. No committee member even moved to approve the bill, and it died that day.

Playing David to HSLDA’s Goliath is a mighty little group called the Coalition for Responsible Home Education. CRHE’s mission is to “empower homeschooled children by educating the public and advocating for child-centered, evidence-based policy and practices for families and professionals.” Among its many recommendations, CRHE has several that are designed to protect homeschooled children against abuse and neglect. These include prohibiting homeschooling by parents who have committed offenses that would disqualify them from teaching school, requiring that students be assessed annually by trained mandatory reporters, and flagging certain at-risk children (such as those in families with a history of child protective services involvement) for additional protections and support.

CRHE was launched in 2013 by a group of homeschool alumni who had met through a network of blogs and Facebook groups. In the past seven years, The Coalition does more with less than any other organization I know. As described on its website, CRHE has driven media coverage of the need for homeschooling oversight; conducted extensive research; developed a set of policy recommendations, advocated for homeschooling oversight in over a dozen states and territories and helped craft successful legislation in Georgia; created a comprehensive suite of resources for homeschooling parents and students; and written a bill of rights for homeschooled children. This has all been done with unpaid staff, including its executive director, and contract workers. Now, CRHE is trying to raise funds to pay a part-time executive director next year, with the hope to grow further in the future.

In my research, I have been surprised at the paucity of organizations that advocate for better protection of children from abuse and neglect, a topic that I hope to address in a future post. While CRHE’s focus is limited to home-schooled children, this is a group that is particularly vulnerable, and evidence suggests that these children are disproportionately represented among the most egregious cases of abuse. For this reason, and in light of CHRE’s extraordinary passion and productivity, I cannot think of an organization more deserving of support by those who care about child maltreatment.