Friday, April 28, 2017

SDM Not the Root of System Failure in Yonatan’s Death

I read with great interest Daniel Heimpel’s column, Yonatan’s Death Spurs Deep Look at How L.A. Gauges Child Abuse Risk. Yonatan Aguilar had been the subject of four prior reports to DCFS through 2012. Yet, he was left in his home without any supervision from the Los Angeles Department of Child and Family Services (DCFS). Records obtained by the Los Angeles Times revealed that Yonatan was kept in closets for three years before he died.
Heimpel’s article focused on criticisms of the Structured Decision Making (SDM) assessment tools, currently used by more than half the states, for allowing this tragedy to occur. While I agree with critics that the new predictive analytics tools are preferable to SDM, the facts of this case are such that better risk assessment tools would not have prevented this tragedy.
In order to understand what went wrong, it is important to understand that every investigation involving child maltreatment in Los Angeles involves three separate decisions. Only two of these decisions involve SDM.
The first decision that has to be made in any investigation is the safety assessment, which is part of the SDM suite of instruments. The safety assessment, as described in California’s SDM Policy Manual, asks whether the child is in “present danger of immediate, serious harm.”
As Heimpel reported, Yonatan was found to be “safe” in all four cases; that is, he was not deemed to be in imminent danger of serious harm. This assessment may well have been correct, as Yonatan may not have been seriously harmed for months or years following this assessment.
Second, there is the decision about whether to substantiate the allegation that was made in the hotline phone call that prompted the investigation. This decision is based on an “allegation assessment,” described in Los Angeles’ Child Welfare Policy Manual. This assessment is somewhat similar to a police investigation and does not involve SDM. According to the Los Angeles Times review of the record, all four allegations involving Yonatan were found to be either inconclusive (insufficient evidence of abuse or neglect) or unfounded (false).
Finally, the investigative worker must complete a “risk assessment” using another SDM tool. The risk assessment attempts to gauge “the likelihood that a family will maltreat their child in the next 18 to 24 months.” When the risk level is high or very high, SDM recommends that a case be opened.
Four times, Yonatan was found to be at high risk of future maltreatment. But a case was never opened.
Based upon my reading of the Los Angeles Times reporting on the case records, it appears to me that there are two major areas where the system may have failed Yonatan.
First, some or all of four investigations into allegations involving Yonatan may have been flawed. When Yonatan was four, a school nurse called police because he had several two- and three-inch scratches on his face, according to case records reviewed by the Los Angeles Times. Yonatan told police that his mother was angry at him for getting in trouble at school and that she had slapped and scratched him. But the investigator chose to believe his mother, who denied hitting him and said he might have gotten the scratches because he slept on the floor.
In March 2012, two reports came from Yonatan’s school within the span of four days. The first report related that Yonatan came to school with a black eye and gave conflicting reports of how he got it. The second report alleged that Yonatan came to school dirty most of the time and was always hungry, grabbing and hoarding food from the cafeteria — an obvious red flag. Yet, no allegations were substantiated.
There is evidence that many investigations conducted by DCFS are flawed. A devastating report submitted in 2012 by a special investigative unit found that failures in the “front end” contributed to 13 out of 15 child fatalities it reviewed. The investigators cited “the poor quality of factual information that served as the foundation for the assessments and subsequent poor decisions,” often rooted in a failure to interview key people with knowledge of the child’s situation.
The second way in which the system may have failed Yonatan is the failure to open a case after any of the four investigations as recommended by SDM, even though he was found to be at high risk each time. In order to understand this, it is important to understand one fact.
As an agency spokesperson confirmed, if the allegations of maltreatment are not substantiated, a case cannot be opened without the parent’s permission unless the county chooses to take the parent to court. I do not know how often this happens in a case where the child is not removed, but it certainly did not happen in Yonatan’s case.
One cannot imagine that Yonatan’s mother would have been receptive to the idea of a DCFS case. Case records reviewed by the Los Angeles Times indicate that Aguilar told a social worker that she was tired of social workers coming to the house and was withdrawing him from an after-school program to avoid further problems.
For the next three years, Yonatan did not go to school. Instead, he remained hidden in locked closets and sedated on liquid sleeping aids.
It is my belief that Yonatan’s death did not stem from a failure of SDM. I prefer the new predictive analytics tools because they are less vulnerable to manipulation and bring in data from outside sources rather than relying on parent reports. But using a “predictive analytics” tool would not have changed this tragic case,  because SDM accurately classified him as high risk.
Rather, Yonatan’s death stems from flawed policies and practices. These may include investigations that fail to get at the truth of what occurred. And they certainly include the inability to protect children at risk unless past maltreatment can be documented.
This column was published in the Chronicle of Social Change on November 7, 2016.

No comments:

Post a Comment