Risk assessment tools are widely used in psychiatric hospitals and criminal justice systems to help predict violent behaviour and inform sentencing and release decisions. Yet their predictive accuracy remains uncertain and expert opinion is divided, according to a new systematic review published in the BMJ.

An international team of researchers led by Seena Fazel from Oxford University, set out to investigate the predictive validity of tools commonly used to assess the risk of violence, sexual, and criminal behaviour.

The team carried out a comprehensive search for replication studies published up until 1st Jan 2011 and then conducted a systematic review and tabular meta-analysis following PRISMA guidelines.

They analysed risk assessments conducted on 24,827 people from 13 countries including the UK and the US. Of these 24% offended over an average of 50 months.

They investigated potential sources of heterogeneity using meta-regression and subgroup analyses.

Here’s what they found:

• The negative predictive value of the risk assessment tools was good:
• 91% of people who were predicted not to violently offend, did not go on to offend
• However, the positive predictive value was no better than chance:
• 41% of individuals judged to be at moderate or high risk by violence risk assessment tools went on to violently offend
• 23% of those judged to be at moderate or high risk by sexual risk assessment tools went on to sexually offend
• 52% of those judged to be at moderate or high risk of committing any offence went on to offend

The authors concluded:

Although risk assessment tools are widely used in clinical and criminal justice settings, their predictive accuracy varies depending on how they are used. They seem to identify low risk individuals with high levels of accuracy, but their use as sole determinants of detention, sentencing, and release is not supported by the current evidence. Further research is needed to examine their contribution to treatment and management.

Link

Fazel, S. et al Use of risk assessment instruments to predict violence and antisocial behaviour in 73 samples involving 24 827 people: systematic review and meta-analysis. BMJ 2012;345:e4692.