On Monday, September 26, 2016, six-year-old Zymere Perkins perished, later to be found covered in bruises. His grave injuries and years of abuse were heaped on by his parents. Countless times, others bore witness to the abuse of the poor child, yet nothing was done. He was certainly not the first to experience such gruesome treatment and he won’t be the last.
A New York State Office of Children and Family Services report — young Perkins was from Harlem — told the story of 10 other children who died in the 12 weeks leading up to Perkins’ own death. All involved were flagged or highlighted several times for abuse and maltreatment complaints.
The major takeaway here is that there were signs and they were documented — but they went unnoticed. While it’s easy to sit on the sidelines and claim negligence on the part of investigators, the issue is more likely caused by overloaded teams and personnel. The same seemingly insurmountable burdens stretch all the way to childcare centers and foster homes, which are also overloaded.
Something clearly needs to be done. But what? Enter modern technology.
Big Data and Predictive Analytics Platforms Can Ease Investigative Burdens
Believe it or not, modern technologies can be used not just to identify potential risks but also take action before something terrible occurs.
Machine learning, big data and predictive analytics systems are designed to sort through, filter and categorize incoming data. When set up appropriately, they work autonomously — that is, without direction or external input.
To ensure the process remains reliable and efficient, however, developers need to craft the unique experiences necessary. In the case of child protective services, it would likely take some time, possibly years, for developers to come up with an accurate system. That doesn’t mean it’s not worth doing, however.
Marc Cherna, Director of the Allegheny County Department of Human Services, was responsible for launching initiatives using computerized systems to “clean up” his district in Pittsburgh, Pennsylvania. Essentially, he established a multi-tiered team approach to filtering and providing urgency to incoming abuse reports.
Phone calls and reports that came in were screened by a special team that categorized each report based on “indicators,” ultimately helping them establish a risk score. Those indicators included a parents’ age, race, criminal history, past reports and marital and welfare status. If a case is flagged as high risk, it’s sent to an investigator for further research.
Now imagine if this entire system, up until the investigations process, was automated by predictive analytics and big data? Not only would it cut down operational costs and make workloads much smaller for investigators and child protective teams, but it would also make them more efficient and faster to respond to potential risks.
Already, teams in Florida are using predictive analytics to head potential problems up at the pass. The goal is to protect and serve those who matter most: the affected victims and children. So it’s definitely possible, probable and merely a matter of time before a more widespread system is adopted by agencies across the country.
There are some caveats, of course.
Big Data Hurdles for Child Protection
The Illinois Department of Children and Family Services actually decided to end an existing program that utilized computer data mining and predictive analytics to identify potential risks. The department’s Director, Beverly “B.J.” Walker, said they dropped the system because “it didn't seem to be predicting much.”
That alone highlights one incredibly important point. If the systems, which can be costly, are not implemented properly or efficiently, there’s almost nothing to be gained by using them. The burden of proof for whether or not these systems work sits on the shoulders of the developers and system engineers.
It also requires a streamlined and direct collaboration with child protective investigators and professionals. The appropriate information and knowledge must actually be put to use.
There’s another looming issue, however, and one which is sure to raise a few eyebrows. In today’s landscape, security and data protection are major concerns. What happens when data storage and transfer systems are breached by an outside party? Hundreds if not thousands of sensitive records and details could become compromised.
While promising, it’s clear that big data and predictive analytics have a way to go before they can be truly beneficial on a massive scale.