Knowledge is power, as they say.
And knowledge — in the form of data on individuals — can help keep children safe from harm. As an example, one of Georgia’s managed care organizations, CareSource, recently mined data regarding its child and adult members’ healthcare usage to “identify families who may be most at risk for negative health outcomes and potential family separation through foster care placement.” The insurer then offered to families identified as “at-risk” voluntary services including preventive medical and behavioral health care, counseling and case management. Researchers reported a reduction in both emergency room visits and foster care placements among those families.
Consider that the risk factors for child abuse include having a parent in the home with mental health and substance abuse issues and having a parent with a history of domestic violence. Consider also that children who suffer serious or even fatal child abuse often have medical histories of unexplained bruises or other injuries documented in prior medical records. Doesn’t common sense suggest that giving child protective services access to a family’s criminal, behavioral health, and medical records would result in a better determination of whether a given child is in serious risk of harm?
Such were the motivations of officials in Allegheny County, Pennsylvania several years ago when they launched an effort to use existing data from medical, justice, public benefits, and other databases to help screen allegations of child abuse and determine which calls to the child abuse hotline needed to be prioritized. Such efforts have come under fire from those who see them as stereotyping families.
As we’ve reported previously, the ACLU, the Associated Press, and now the US Justice Department have raised concerns that the use of predictive analytics in child protection screening activities is discriminatory.
As the ACLU argues:
“Ultimately, we find that [Allegheny County’s] stated goal to `make decisions based on as much information as possible’ raises concerns and comes at the expense of already impacted and marginalized communities, risking further perpetuation of systemic racism and oppression.”
By including factors such as whether the family receives federal benefits or whether a family member has prior involvement with the juvenile or adult justice systems, the organization says, Allegheny’s model unfairly identifies children in black families as at higher risk of abuse.
The Justice Department’s investigation of the predictive model apparently focuses on whether Allegheny’s algorithms put parents with disabilities at a disadvantage, as diagnoses of parental behavioral and mental health disorders are considered in determining risk to the child. The Associated Press followed up on this issue recently, reporting on a CPS investigation of a family with a young child whose parents suffered from developmental disabilities:
Over time, Allegheny’s tool has tracked if members of the family have diagnoses for schizophrenia or mood disorders. It’s also measured if parents or other children in the household have disabilities, by noting whether any family members received Supplemental Security Income, a federal benefit for people with disabilities. The county said that it factors in SSI payments in part because children with disabilities are more likely to be abused or neglected.
Here’s the dilemma: both views may be true. Data-fueled social surveillance can amount to an invasion of privacy or, worse, lead to a “social credit system” such as the one China is building. At the same time, any of us who have worked in child welfare have seen tragic child abuse that occurred because of what the agency didn’t know about a person’s situation and background.
Should a parent’s developmental disability prevent them from bonding with and caring for their newborn or result in a removal of the child? No. But an agency’s “flag” of that case might mean the difference between the family getting the supportive services they need to thrive and a terrible tragedy. There has to be a way to use the data tools we have while being mindful of our responsibilities to address each family’s strengths and weaknesses.
In Georgia this week, the Legislature is considering a bill to improve mental health care. Part of it calls for better data collection and data sharing among agencies, which has led to a rash of conspiracy theories that the data will be used to monitor individuals or seize their guns. In the midst of these debates, is it surprising that policymakers struggle with finding the right balance?
News update coming Friday!