Big data analytics and predictive risk modelling in social work: Lessons from the reform of child protection services In Aotearoa New Zealand

Paper presented at #husITa16 in Seoul, Korea, 29 June 2016.


Philip Gillingham (University of Queensland, Australia).


Background: “Big data” initiatives that aim to bring together and mine data from multiple databases across government and non-government agencies promise new insights into the aetiology of social problems which may then be used to inform the organization and delivery of social services. In predictive risk modelling, algorithms can be developed using these large volumes of data, which when applied to individuals in the general population, can estimate the likelihood of them experiencing a particular problem. The aim of identifying individuals in this way is to avoid future problems by offering preventative services and supports. Predictive risk modelling is well established in healthcare but has yet to be used in social work. However, an initiative in Aotearoa/New Zealand to develop the Predictive Risk Model (PRM) to prevent child maltreatment has perhaps brought it a step closer to being applied in a social work setting. The use of the PRM has caused much debate, much of which has focused on the ethical problems associated with its development and application. The aim in this paper though is to provide an analysis that addresses the question of what practical lessons have been learned from the process of its development.

Methods: An analysis of the publicly available documents about the development of the PRM was conducted.
Results: The analysis focuses on two key areas, namely the predictive ability of the PRM and the potential insights it provides through what it identified as the main predictors of child maltreatment.
Implications: The implications of this analysis for the reform of child protection services and for the future development of predictive risk modelling in social work more generally will be presented.



To follow


Back to programme

%d bloggers like this: