Join us this week on the New Zealand Initiative podcast! Our host, Dr Eric Crampton chats with Professor Rhema Vaithianathan from Auckland University of Technology. Rhema gives us a peek into her pioneering work on data-tools, designed with a human touch to better serve communities.

In this fascinating chat, they delve into Rhema's trailblazing project, the Allegheny Family Screening Tool. This unique decision-aid tool is a global first, introduced in 2016 to support child welfare hotline screening. The duo also touch on the equally impressive Allegheny Housing Assessment tool, another innovative machine-learning tool used since 2020 for prioritising housing within the homelessness system.

Rhema reveals the power of predictive risk modelling in the US child protection system, where her team used administrative data to help improve outcomes for children. Listen to how this model uses past case data to guide operators in their decisions regarding potential abuse cases.

In the US, this approach worked to substantially reduce child hospitalisation. It could be done in New Zealand as well. In fact, the work started here. But New Zealand’s politics and public service has a very difficult time with new approaches. American localism means that one innovative county can try it out, and demonstrate the benefits to others.

All this and more in our enlightening podcast. Stay tuned for Part Two, where the conversation on AI and predictive models in child protection systems, and on encouraging public sector innovation more broadly, continues!

Podcast: Data analytics for social impact (part one)

The New Zealand Initiative