Machine learning for social and health data analytics
The State of Queensland has been going through a process to substantially reform its data collection and management systems to improve client outcomes.
The objective of the project is to identify opportunities for data-driven improvements, with new methods of predictive risk modelling, to support the Department to better achieve its goal of improving child safety outcomes.
A team of researchers, led by Professor Rhema Vaithianathan will explore the use of predictive analytics in child welfare to produce an options paper advising how the Department could employ predictive risk modelling to improve child safety outcomes, and develop the design of a new data management system to incorporate best practice decision support functionality.
Predictive risk models are becoming increasingly used by child welfare services around the world as it has been found to improve the accuracy of decisions and reduced disparities. Examples are the Allegheny Family Screening Tool which is used to triage calls to child abuse hotline in Pittsburgh, Pennsylvania, USA; and the Douglas County Decision Aide used to help in collaborative decision-making for Douglas County, Colorado, USA.