Data on Risk Algorithm Used By Sensemaking Alerts for Frontline Engagement and Response
Ministry of Social and Family DevelopmentSpeakers
Summary
This question concerns whether there are published data on the Sensemaking Alerts for Frontline Engagement and Response (SAFER) risk algorithm and plans to incorporate additional databases, as asked by Dr Neo Kok Beng. Minister for Social and Family Development Masagos Zulkifli B M M explained that the algorithm was developed internally by the Ministry of Social and Family Development using Government datasets, actual child protection cases, and professional knowledge, rather than adapted from published algorithms. The Ministry regularly monitors, validates, and refines the algorithm to effectively balance the risks of missed cases against unnecessary alerts. Minister Masagos Zulkifli B M M emphasised that SAFER merely supports professional judgement, with trained professionals reviewing all flagged concerns before deciding on follow-up actions. He added that the Ministry will continue exploring additional datasets to enhance SAFER, carefully assessing them against child protection needs, proportionality, and data governance safeguards.
Transcript
72 Dr Neo Kok Beng asked the Minister for Social and Family Development (a) whether there are any published papers or data on the risk algorithm used by the Sensemaking Alerts for Frontline Engagement and Response (SAFER) and its effectiveness; and (b) whether there are any plans to integrate additional Government or private databases into SAFER to make it more comprehensive.
Mr Masagos Zulkifli B M M: The risk algorithm used by the Sensemaking Alerts for Frontline Engagement and Response (SAFER) Unit was developed by the Ministry of Social and Family Development (MSF) based on relevant Government datasets and profiles of actual child protection cases, and professional practice knowledge. It was designed to reflect risk factors and patterns observed in real-world cases and was not adapted from any published algorithm.
MSF will regularly monitor, validate and refine the algorithm to ensure it continues to perform as intended and remains fit for purpose. This includes assessing its effectiveness in identifying cases of concern while appropriately balancing the risks of unnecessary alerts and missed cases. I want to emphasise that SAFER supports but does not replace professional judgement. Potential concerns surfaced by SAFER are reviewed by trained professionals against the available case information before any follow-up action is determined.
MSF will continue to explore additional datasets that can enhance SAFER's effectiveness. Any inclusion of new datasets would be carefully assessed based on factors, such as child protection needs, proportionality and data governance safeguards.