As the wave of evidence based practice gained popularity in the twenty first century, Structured Decision-Making (SDM) Tools were implemented as a method for assessing risk. There are a notable number of cases in which clinical risk assessment failed to accurately predict risk.  SDM Tools were seen as the way to consistently and accurately assess risk to prevent tragedies[1]. This blog focuses on the implementation and resulting validity of the SDM Tools in child protection and policing in Canada.

When frontline social workers and police officers were trained to use SDM Tools, it appears a critical knowledge gap was overlooked. As with many training paradigms, the focus is on the application of the tool and not the research base behind the tool. However, it is the research which explains the why and how the tool was scientifically validated. Honing in on the why and how of SDM Tools, they were validated by analyzing variables in cases of violence and then determining the statistical significance of the variables, which were then determined to be risk factors for violence.

What is being seen in practice today, in the Canadian context of child protection and policing, is that the SDM Tools are being used as an interview template. When the SDM Tools criteria are being used to ask highly leading questions, such as “Have you ever been forced to have sex?” or “Have you ever been bitten?”, it can lead to errors. The fact that a person responds by saying “yes” to these types of questions is not reliable evidence of this risk factor being present. When the “yes” answer is scored on the SDM Tool, it can cause an inflated score resulting in an incorrect classification of risk.

The solution for social workers and police officers is to use memory-based interviewing, such as the StepWise, to gather relevant case information in a non-leading way. This information can then be used to score the SDM Tool. This approach allows reliable data to be collected and assessed against the SDM Tool, rather than using the SDM Tool criteria to lead the interview and potentially gather unreliable data.

[1] R. Karl Hanson. “The Psychological Assessment of Risk for Crime and Violence,” Canadian Psychology 50, no. 3 (2009): 172-182, DOI: 10.1037/a0015726