Using Technology to Match Children AEA Presentation 05-18-17

Using Technology to Match Children AEA Presentation 05-18-17

Using technology to match children – AEA presentation 05-18-17

1:15-2:45 p.m. Pompeian Ballroom 1 &2

STEWARDS OF CHANGE INSTITUTE (SOCI) PRESENTATION

By Adam Pertman (NCAP President, Senior Consultant to SOCI)

The technology I’m going to discuss today is cognitive computing, and my remarks are drawn from a white paper I co-wrote for SOCI commissioned by IBM, which has the most-famous cognitive computer – Watson. IBM is obviously a self-interested party, but they provided no input or influence in researching or writing the paper, which is titled “Improving Processes and Practices in Child Welfare: Is Cognitive Computing Part of the Solution?” Here’s a copy; I have several; it’s also on

Without a long explanation, basically what cognitive computing can do is consume huge amounts of information – from handwritten notes to charts and graphs to audio files, internal policy guidelines to research on best practices. Then it assimilates that information, learns what you want from it and provides guidance with which people – child welfare workers, for instance, can make better-informed decisions. Including PREDICTIVE ANALYTICS. The idea is NOT for the computer to tell you what to do, but to improve your own ability to make judgements based on evidence and where it points.

CC, for the first time, provides a toolset that allows us to access information that was once considered to be unattainable, and to get it in a very pragmatic, friendly and familiar interface. It’s not a cure-all by any means, but it offers a unique and powerful set of tools that could be used to address many of the issues and problems faced by child welfare, from 1. better ways to detect and prevent child abuse to 2. determining which supports and services might best help a specific child/family that has specific needs to 3. Better training and retention of workers and, of course to 4. improved matching to optimize the prospects of success for children, youth and their families.

All these numbered items, of course, improve matching, as would increased ability to be more efficient and effective by enhancing interoperability and information-sharing, as well as incorporating the social determinants.

Better information, training and tools clearly also could help address issues relating to disproportionality, including in matching/placement, as well as to figure out how to tackle the broader systemic problems.

Best possible placement is so important because every time a child is moved from home to home, the impact is traumatic – as is the breakup of a family – so better tools are needed to minimize the number and duration of placements by better recruiting and training of families, as well as by improving the placement of individual children with specific families who can best meet their needs.

The benefits for matching/placement could be substantial if patterns are found in the data pointing to how to better coordinate child characteristics (or other factors such as birth family traits or reason for removal) with specific foster or adoptive parents, or specific types of parents. A related analysis might identify the types of training and/or support for both the child and the resource family that could increase their prospects for success, again enabling workers to make better decisions. Given that placement with kin is often the first choice, cognitive tools might also be useful for locating relatives who are not yet known to the agency or who have not been identified by already-identified members of the child’s family. Cognitive could help improve:

  • The effectiveness of specific services for families and children.
  • Whether children are at a higher risk in the future.
  • Which family might be the best placement (ex., kin or not).
  • Factors to include in deciding a placement or reunification.
  • The most potentially successful assignment of a child for adoption or guardianship.

For placement and other aspects, suggest demonstration projects utilizing cognitive tools.

As data is increasingly shared among agencies and other organizations, questions are increasingly raised about which confidentiality/privacy laws and policies apply in various situations. Again, privacy concerns are critical in child placement. Documents explaining those laws and policies could be absorbed by a cognitive system, with the guidance of legal experts to help in the training, so that personnel at all levels could then ask questions about specific types of data and circumstances – and could receive recommendations about what data can or cannot be released to whom and when.

A cognitive system could review all available records relating to a family and suggest what information needed to be examined by the caseworker to make an informed judgment, and could automatically update the risk and safety calculations whenever data changed. The system could also prepare a summary of case data for particular purposes, such as a list of items to be confirmed during a home visit. So home studies improved by CC.

In the white paper, we conclude that cognitive computing could be a transformative technology for the child welfare field.

PAGE - 1