By Danielle Harris and The Hon Justice Peter Applegarth AM*

How effective is the system of preventative detention for sex offenders that has operated in Australia for nearly 20 years? 

How do we define and measure effectiveness?

Might evidence-based research propose more effective ways to devote the resources that the system currently consumes so as to ensure community safety and encourage desistance?

These are some of the questions being explored in our research, which is supported by the Australasian Institute of Judicial Administration (AIJA). 

A system that has grown and absorbs judicial and other resources

In 2003 Queensland was the first Australian jurisdiction to enact a system for sex offenders to continue to be detained in custody after serving their sentences, or to be released into the community under strict supervision orders. 

The controversial system was found to be constitutional in Fardon.[1] Similar schemes have been enacted in other jurisdictions, sometimes being extended to violent, non-sexual offenders.

A system that was targeted at “the worst of the worst” sexual predators and expected to apply to a dozen or so recidivist paedophiles has extended to capture a wide range of offenders.  These include many Indigenous men whose typical pathway to sexual offending is to commit opportunistic sex crimes when intoxicated by alcohol or other substances.

If released at the end of their sentence on a supervision order, rather than being held on a continuing detention order, these men are subject to 40 or so conditions including abstinence from alcohol, restrictions on where they can move and electronic monitoring.  Many conditions are not directly related to the commission of a sexual offence. They are intended to act like a “canary in a coalmine” and to detect anti-social activities that might signal deterioration in their welfare and an elevated risk of sexual offending.

As a result, courts are burdened with numerous contravention hearings, often for minor breaches, with the respondent returned to custody for months while forensic psychiatrists report to the court about whether the risk of serious sexual offending has changed or warrants further release into the community under a new supervision order.

Courts also have a heavy workload of assessing new cases that come into the system (many of which involve a single serious sexual offence that was committed when the offender was a teenager), conducting annual reviews of continuing detention orders, and managing complex issues of where individuals with anti-social personality disorders and other mental conditions can be accommodated in the community under supervision orders.

In terms of judicial administration, courts like the Supreme Court of Queensland provide what amounts to a very expensive accommodation service for individuals with complex needs for support and supervision.  A major and recurring issue is finding safe, suitable and affordable accommodation that will ensure community safety and provide an opportunity for a troubled individual to lead a more productive life.

A system that was expected to deal with “approximately a dozen or so very, very serious offenders” [2] now has to manage a couple of hundred complex cases at any one time.

The need for an evidence-based assessment of the system

Much has been written about the jurisprudence of preventative detention for sex offenders, violent offenders, terrorists, and others who are kept in custody or closely supervised in the community long after their original sentences have been served. Most research relies on the rates of officially reported recidivism and some studies have included interviews with participants in the system and other stakeholders. 

Academics and judges have questioned the validity of actuarial instruments that are used by forensic psychiatrists and psychologists to assess risk of serious sexual recidivism.[3] Key concerns include the application to a contemporary Australian population of tools that were created and validated on very different North American populations many decades ago.

Some research reviewed an incomplete set of publicly available judgments about sexual reoffences and contraventions of a sexual nature.[4] It suggests that among the many contraventions of supervision orders that have been before the courts over many years, only a few (8.9%) have involved serious sexual offending.

This creates an impression of success and confirms the views of judges that the system has been effective in reducing serious sexual offences. What is missing from this and other analyses is an acknowledgement that the rates of serious sexual recidivism are reliably quite low across time and space. In multiple Western jurisdictions and consistently over at least the last 50 years, officially recorded sexual recidivism rates hover around 8-12%.[5] Desistance from crime is a natural human process and is the modal outcome observed for formally incarcerated people, even among those who have committed serious sexual crime.

If the current system is effective in reducing what otherwise would have been the rate of serious sexual offending, then this raises the issue of why it has been effective. Also, could that effect have been achieved at a lesser financial, social, and individual cost?

The answer to that question depends on a rigorous analysis of data. It also calls for a sophisticated understanding of how “effectiveness” is assessed.

The many individuals subject to post-release community supervision orders have different histories of offending and present different risks of re-offending. For some recidivist paedophiles, reducing the risk of serious sexual re-offending requires a raft of conditions that deny access to potential victims and closely monitors their compliance with orders. For others, whose offending might be much more opportunistic (such as offenders without an underlying sexual deviance) their risk of reoffending can be mitigated by imposing situational controls such as restricting or denying access to alcohol and drugs and providing pro-social supports and counselling.[6]

The research project so far

With the assistance of fellow researcher, Juliet Davis, we have undertaken a systematic literature review with a particular focus on establishing and refining what is meant by “effectiveness”.  Naturally, effectiveness can primarily be measured by reducing or preventing recidivism. But the purposes of the statutory schemes include rehabilitation of offenders as well as protection of the community.  The absence of known recidivism may be one measure of rehabilitation or community safety.  However, one needs to consider the broader implications of desistance research. For example, measures of success go beyond the absence of recidivism and can include indicators of wellbeing and quality of life, involvement in pro-social activities such as employment and education, commitment to interpersonal relationships (including with friends and family) and attachment to community, culture, or country.

Another component of measuring effectiveness is to identify any negative unintended consequences. Our initial field work and analysis describes a problem that may be described as “churn”.  This concerns individuals who breach their supervision orders in minor ways that are neither sexual nor violent and whose contravention proceedings consume much of the system’s finite resources.

These issues highlight the need to identify, on the basis of evidence, why some individuals breach their supervision orders.  Equally important, is why many do not. Rather than be solely concerned with the failures, we could learn much more about effectiveness by examining the kinds of supervision strategies and support services that work in general and for certain types of offending.

What works and what could work better?

The next phase of the research will involve more intensive analysis of data available through Griffith University’s Social Analytics Laboratory, as well as case law and engagement with participants in the field, including forensic psychiatrists who play a major role in the system.  

This initial phase of the research will focus on Queensland, but we hope to expand the work and engagement to other Australian jurisdictions. Other states and territories have things to learn from Queensland’s 20-year experience in this field about what works and what does not.  We all have things to learn about how similar systems operate in different places.

Legislation may be essentially the same in terms of its purpose and requirements. However, its administration may be very different.

Evidence-based research of what works and what does not can inform decisions about the best use of finite court and other public resources to encourage rehabilitation and ensure public safety.


[1] Fardon v Attorney-General for the State of Queensland (2004) 223 CLR 575; [2004] HCA 46.

[2] Queensland, Parliamentary Debates, Legislative Assembly, 4 June 2003, 2581 (RJ Welford, Attorney-General and Minister for Justice).

[3] See, for example, Attorney-General for the State of Queensland v Plummer-Dean [2021] QSC 165.

[4] M T Rowlands, G Palk and R McD Young, ‘Recidivism rates of sex offenders managed under the Dangerous Prisoners (Sexual Offenders) Act 2003: an evaluation of actuarial justice’ (2020) 28(2) Psychiatry, Psychology and Law 310.

[5] D Harris, ‘Desistance from sexual offending’ (2021) 23(2) Current Psychiatry Reports DOI: 10.1007/S11920-020-01219-3.

[6] See, for example, Attorney-General for the State of Queensland v Thaiday [2021] QSC 227.