Unless and until child welfare risk assessment instruments and tools can be made more accurate and reliable, child welfare systems around the country will continue to be plagued by scandals including high rates of families being re-referred for abuse and neglect, and worse: too high an incidence of child fatalities where the families were known to child welfare authorities.
For nearly four decades, child welfare officials, child advocates and elected public officials have been desperately searching for ways to reform child welfare systems. I say desperately because the recent history of child welfare has been characterized by periodic episodes of scandals, media exposés and class action lawsuits followed by infusions of cash into the system(s) to hire more social workers, lower caseloads, provide more staff training and to support the latest programmatic fad. Throwing large amounts of money at these systems usually quiets them down for a while. But, before long, the cycle repeats itself.
In the mid 1970s and 1980s family preservation was the answer. Initially supported and promoted by the Edna McConnell Clark Foundation, and later by the Annie E. Casey Foundation, family preservation programs spread rapidly throughout the country. The supporters of family preservation claimed these programs were effective in keeping families intact and preventing the out-of-home placement of children. They believed in family preservation despite the fact that the concept was not grounded in any known theory and even in the face of mounting evidence that family preservation programs were not living up to expectations and exaggerated claims of success. The supporters, particularly the true believers, were quick to dismiss the critics of family preservation, including researchers whose studies of family preservation programs were not generating the results the supporters had hoped for.
Family preservation has largely been discredited. What credible research that exists does not present a compelling case for such programs.
One of the major problems family preservation programs faced, and continue to face today, is the inability to accurately identify which families could benefit from such programs and be kept intact while ensuring the safety and well being of children. As the research indicates, many children who should have been removed from home in order to ensure their safety were kept with their families and were victims of further maltreatment.
Tragically, some of them, particularly infants and toddlers, were subsequently killed at the hands of their parents or guardians. Aside from the disappointing results from carefully done studies, family preservation took a beating from the scandals that surfaced in the media and from the horror stories presented to state legislative committees.
While the bloom on the family preservation rose was fading, it was gradually replaced by a new fad, Differential Response, more commonly referred to as DR.
DR is simply the creation of a dual or multi-track pathway for managing and providing services to families referred to child welfare agencies for abuse and neglect. Under DR, the lower and, to some extent, moderate risk families are diverted from formal processing and are provided services on a voluntary basis. The assumption being that families who presented a lesser risk would be more amenable to voluntary services than if they were handled in a more official and, presumably, adversarial manner. The high-risk cases would continue to be handled in the traditional manner.
Like family preservation, the supporters of DR are making claims about the effectiveness of DR programs that are not supported by the best available research. And, once again, it appears that one of the major problems confronting DR is the inability to identify which families are at less risk for committing future acts of abuse or neglect and would be appropriate to be handled informally.
I’m not particularly optimistic that more research on family preservation and DR programs will advance our knowledge about how to make these programs more effective. I think that greater progress would be made by focusing more of our attention and research efforts on improving risk assessment in child welfare. It is generally understood that risk assessment protocols and tools that are actuarially based are more accurate and reliable than others.
Despite this, a recent systematic review of risk assessment tools from throughout the world conducted by Dr. Jane Barlow and her colleagues in the United Kingdom concluded that there is “limited evidence about the effectiveness of the available tools,” including those that are actuarially based. The available tools are simply not accurate enough and fall short of what child welfare staff need in order to make truly well-informed decisions.
I’m also somewhat skeptical that traditional social science methods will help in boosting the accuracy of existing child welfare risk assessment instruments. There have been many attempts to improve them over the years, and these efforts have only resulted in marginal improvements in their levels of accuracy.
There is, however, some cause for optimism. There is emerging evidence that risk assessment accuracy and reliability can be significantly improved by using data mining, predictive analytics and machine learning, research methods not typically used by traditionally trained social scientists. For example, a study of juveniles adjudicated in the Philadelphia Family Court that I worked on generated results in the 90 percent range with respect to classification and predictive validity.
A similar study involving approximately 140,000 cases using these methods is now underway in Florida. While not yet completed, the preliminary findings are very encouraging. Also, in Florida, a recent study using these methods was completed that examined child fatalities. And another is underway using predictive analytics and machine learning to predict which youth who age out of foster care are at most risk of becoming homeless.
As I mentioned earlier, family preservation has been discredited because of the scandals and horror stories about what happened to abused and neglected children who were kept at home when they should have been placed into foster care. These children suffered further maltreatment, and some of them were killed. Differential Response has not yet been discredited, but it is on the path to becoming so unless significant advances can be made in our ability to identify which families and children can benefit from this kind of program without keeping the children in harms way.
Ira M. Schwartz is a consultant, former provost of Temple University and former dean of the University of Pennsylvania School of Social Work.