Brian Taylor is Emeritus Professor of Social Work at Ulster University in Northern Ireland. His 40-year professional social work journey has included practice, management, training, teaching and research. This led him increasingly to focus on decision making, assessment and risk, and the knowledge required to develop practice and manage services. He was principal organiser of the DARE conferences from their beginning in 2010. Brian has supervised about 20 PhD students on related topics, including some jointly with colleagues in communication studies, health care, psychology, law and computer science. He is an author on textbooks for qualifying and post-qualifying studies on these topics, and was recently the lead editor of the Sage Handbook of Decision Making, Assessment and Risk in Social Work. Brian is a Fellow of the UK Academy of Social Sciences; an honorary Senior Fellow of the School for Social Care Research at the National Institute for Health Research, London; an honorary Associate Scientist of the Harding Centre for Risk Literacy at the Max Planck Institute for Human Development, Berlin. He was a founder member of the Board of the European Social Work Research Association (ESWRA), and founder-convenor of the ESWRA Decisions, Assessment and Risk Special Interest Group. (https://pure.ulster.ac.uk/en/persons/brian-taylor).

Decisions, Assessment and Risk: Where Are We Now?
This presentation will focus on the connections between undertaking assessment, making decisions, and the management of risk in social work. It will be a personal reflection on ‘the journey’ of the DARE conferences from their start in 2010, and Brian’s perspective on what we have learned and key topics to explore in the future

Rhema Vaithianathan is a Professor of Economics at Auckland University of Technology (NZ) where she is director of the Centre for Social Data Analytics, a research centre that uses data analytics for social impact. Rhema leads a multidisciplinary team that works with health and human services agencies around the world to develop and deploy research-led machine learning tools. CSDA takes a human-centred approach to data-tools which help agencies serve their communities better. Rhema led development of the world-first Allegheny Family Screening Tool , a decision support tool for child welfare hotline screening tool implemented in 2016 profiled in The New York Times Magazine and Nature News. CSDA continue to develop similar tools for counties in Pennsylvania and Colorado. She also led the development and implementation of the Allegheny Housing Assessment tool, a machine learning decision support tool used in Allegheny County since 2020 to prioritize housing within the homelessness system.

Can using predictive risk models reduce racial bias? Evidence from Child Welfare and Homelessness Systems

Machine learning tools, actuarial tools and predictive risk models (‘algorithms’) can help human systems make better decisions (Kleinberg et al. 2018). As such, algorithms are increasingly promoted as useful complements to human decision making in a wide variety of settings, including bail decisions (Chohlas-Wood 2020), resume screening (Raghavan et al. 2020), health care (Price 2019), and, more recently, to support social workers. As their prevalence grows, so too do concerns that algorithms may entrench or exacerbate existing disparities in system interactions, and in particular disparities across race. However, human bias is also well-established, and has been shown to cause racial disparities in many of society’s institutions.

In this talk, I use case studies of the introduction of predictive risk models in child welfare and homelessness settings where previously social workers had used paper and pen assessment tools or pure clinical judgement. I provide evidence from RCT and quasi-experimental studies to show a remarkable impact of these tools – leading to improved outcomes as well as reduction in racial disparities.  

I discuss the implication of these results for social work – arguing that because social workers are often making decisions in urgent high-stakes settings, where there is potential for partial information (or even information overload), the use algorithms that are able to summarise hundreds of data fields can – if carefully deployed – attenuate against biases inherent in clinical judgement alone.

Cheryl Regehr is a Professor in the Factor-Inwentash Faculty of Social Work, with cross-appointments to the Faculty of Law and the Institute for Medical Sciences at University of Toronto. She served as the Vice-President and Provost for the University of Toronto (2013-2023); Vice-Provost Academic Programs (2009-2013) and Dean of the Factor-Inwentash Faculty of Social Work (2006-2009). She is a Visiting Professor at the National Institutes of Health Research (NIHR) Policy Research Unit in Health and Social Care Workforce at King’s College London (UK). She has been elected as a Fellow of the UK Academy of Social Sciences (2023), Fellow of the American Academy of Social Work and Social Welfare (2022), and as Fellow of the Canadian Institutes of Health Sciences (2023). Professor Regehr’s six books (Oxford University Press, Columbia University Press and University of Toronto Press), and over 160 scholarly articles focus on forensic mental health; trauma and recovery; and stress, trauma and decision-making in high stress professions. Her current funded research projects involve testing a new model for improving professional decision-making in situations of risk and uncertainty; understanding trauma in archivists; and cyber-violence against public health professionals during the COVID pandemic. (https://socialwork.utoronto.ca/profiles/cheryl-regehr/)

Stress, Trauma and Decision-Making in Social Work. Social workers are called upon by society to undertake high-stakes assessments and make decisions regarding future risk of harm. Risk decisions in these situations are fraught: emotions are high; information is often incomplete, inadequate, or misleading; situations can deteriorate or change quickly with little warning; and public expectations and media scrutiny are intense. In the end, despite best intentions, skills, and efforts, the outcomes of these complex decisions are at times tragic. This presentation will discuss research regarding social workers’ exposure to workplace stress and trauma, the impact of stress and trauma on professional decision-making, and the implementation and testing of an educational intervention aimed at improving decision-making in situations of risk and uncertainty.