Ethical Evaluation of the Predict-Align-Prevent Program.
University of Auckland.
Published: December 2018
This report offers an ethical analysis of the Predict-Align-Prevent Program (PAP) to prevent child maltreatment. It begins by explaining the PAP has three phases: during a ‘predict’ phase, PAP uses geospatial predictive risk modeling to identify high-risk geographical locations based on environmental features; during an ‘align’ phase, PAP aims to use the predictive information about the relative locations of future child maltreatment events and proximate risk and protective factors to identify opportunities to work strategically with communities and providers to align services, education, and resources to locations where they are most likely to reach children at risk; and during a ‘prevent’ phase, PAP aims to generate baseline data and to actively surveil risk, protective, and outcomes data in high-risk areas to measure the effectiveness of particular implementations of prevention programs in those areas, and to inform future prevention efforts. It notes PAP differs from many social policy uses of predictive analytics in that it is place- rather than individual- or family-based. Ethical issues are discussed that address the data underpinning the Predict phase of PAP that uses existing reports of child maltreatment to produce its original maps that could lead to increased surveillance and reports in those areas, increased surveillance as potentially intrinsically wrong, the lack of detail on implementation decisions during the Align and Prevent phases that could have potential ethical implications, the need for significant barriers to the recovery of address level data, stigmatization of geographical areas, pushing marginalized families beyond the reach of services, indirect discrimination, transparency, consent, the need for education of those involved in implementing the PAP program, false positives and false negatives, and the use of Euclidean Distance to measure exposure to risk and protective factors.
ethics; program evaluation; program descriptions; child abuse; prevention; primary prevention; risk assessment; community characteristics; predictor variables; predictive analytics