“What keeps you awake at night?” my colleague asked me. At the time, I was a child protective services administrator responsible for thousands of at risk children on any given day.   While there were many things that weighed on my mind, worker turnover created many sleepless nights, and days filled with anxiety and fear.

I was reminded of these fears when I watched the recent CNBC segment, Can life as a data point save America’s at-risk children?

When I was an administrator, of all the things that impacted quality work, workforce stability was perhaps the biggest driver of quality outcomes. I would quickly see a direct correlation between workforce stability, case load size, performance outcomes, financial performance, and client satisfaction.   When worker retention was high, so was performance. When worker retention worsened, so did performance. Worker turnover was created by many things with a general understanding that turnover attributed to more turnover.

Formal and informal stay and exit interviews provided some insight including:

• Too much paperwork
• Taking time off only added to work load upon return
• Not making enough impact
• Limited technology and tools needed to perform the job

Recent child protective services work allocation studies have shown that 35-45% of a case worker's time is spent doing administrative duties and that less than 25% of a case worker's time is spent conducting client facing activities.   Technology can help flip those numbers, but I used to encounter a reluctance to embrace innovations that could impact the workforce and the children being served. As someone that has been in the trenches and led large child protective services operations, I truly believe in the power of analytics to support informed decision making that will help improve child welfare outcomes and save lives.

Implementing an operational analytic solution in a state and/or county child welfare system can actually impact worker efficiency and effectiveness by:

• Eliminating worker bias and need for fidelity monitoring.
• Allowing for a more accurate assessment of risk as opposed to many current actuarial models.
• Decreasing time spent on administrative tasks and increasing percentage of time on client facing activities.
• Helping workers prioritize case related tasks.
• Providing caseworkers with better information to help inform decision making on behalf of kids.
• Allowing for an ongoing assessment of risk and safety based on real time or daily data feeds rather than at specific intervals throughout the life of a case (a lot can happen on a case between home visits)
• Helping inform inexperienced workers on risk levels and what factors attributed to risk to help inform placement and service related decisions.
• Truly impacting the ability of a caseworker to have critical case related information available at their fingertips.

In addition, an analytic solution will not increase racial disparity or negatively impact the ability for community organizations to serve at risk children. I have yet to see an analytic model where race and/or ethnicity is a stand-alone or weighted risk factor when determining risk of re-maltreatment. As this Chronicle for Social Change article indicates, it could actually reduce the chance of a low-income household being unfairly characterized as high risk.

In a recent White Paper titled, “A Path Forward: Policy Options For Protecting Children From Child Abuse and Neglect Fatalities”, released by The Commission to Eliminate Child Abuse and Neglect Fatalities, an initial finding was that:

“There is insufficient knowledge about the circumstances of child abuse and neglect fatalities and few proven strategies to prevent child abuse and neglect fatalities”. The report goes on to say, “Although thousands of children die because of abuse or neglect in the U.S. in a given year, unfortunately not much is known about the circumstances of these fatalities. Even less is known about which strategies have been proven to prevent child abuse and neglect fatalities. There is a clear need for a national research agenda on preventing child abuse and neglect fatalities. In addition, there are steps policymakers can take to maximize what we know from existing data.”

The reality is that, as an industry, child protective services has continued to do things the same way for a long time and expect different outcomes. Innovation is a great catch phrase, but often moves at a snail’s pace due to fear of change. Given the fact that the industry is data rich, but analysis poor, now is the time for administrators and policy makers to lean on the power of analytics in an effort to do things differently and expect different and improved outcomes.

The LA County analysis highlighted in the CNBC piece shows the very real promise of analytics to alleviate case worker burden and improve child safety. To borrow the commission’s phrase, this is a “path forward”, technologically. I believe this is just the beginning of the use of analytics in child welfare. Much will be learned along the way, I’m sure, and I welcome your comments on the promise, and perils, of these developments.