A prison sentence is not enough to break the cycle of reoffending if the underlying issues that contribute to it are not addressed. For this reason, the complex process of successfully rehabilitating offenders extends beyond the criminal justice system. Factors such as social disadvantage, unemployment, homelessness, and health and wellbeing also influence a prisoner’s ability to reintegrate into the community.
Rehabilitating and reintegrating prisoners is a core principle of the criminal justice system and a strategic priority for Corrections Victoria (CV), a business unit of the Corrections and Justice Services (CJS) group at the Department of Justice and Community Safety (DJCS).
Ravenhall Correctional Centre (Ravenhall) is Victoria’s newest prison. Privately operated by the GEO Group Australia Pty Ltd (GEO), it was designed to trial new rehabilitation methods that focus on reducing reoffending. To strengthen the prison’s focus on rehabilitation, the state included financially incentivised key performance indicators (KPI) linked to reintegration and recidivism targets in GEO’s contract. This includes KPI 15, which measures reintegration outcomes and KPI 16, which measures rates of reoffending. Among Victorian prisons, performance payments for reintegration and reoffending outcomes are unique to Ravenhall.
Running prisons is a considerable expense for the state and costs are rising. Between December 2012 and December 2019, Victoria’s prison population grew by 58 per cent. More offenders are also returning to prison after their release.
In this audit we assessed if CV and GEO have set up a strategic and operational environment at Ravenhall to help reduce recidivism. We also examined if CV and GEO have developed best practice prisoner management to rehabilitate offenders and reduce recidivism, and if they have effective performance frameworks to evaluate these outcomes.
Changes CV made to Ravenhall’s strategic and operational environment before and after it opened have significantly compromised its ability to achieve its prisoner rehabilitation objectives. CV’s decision to increase the number of places for remand prisoners, and the higher-than-expected proportion of short stay sentenced prisoners, has made GEO’s model for reducing recidivism less relevant to its prisoner population. As fewer prisoners are therefore experiencing the model as GEO originally intended it, it cannot be as effective at reducing Ravenhall’s reoffending rate.
Further, gaps and flaws in CV’s performance and evaluation framework for Ravenhall mean, as yet, it will not be possible for the state to properly learn about the success or otherwise of the unique features of the Ravenhall model on reducing recidivism. The design of KPI 16 means CV cannot fully attribute performance against it to GEO, and the measure for KPI 15 (which is now being revised) is unclear and impractical to implement.
CV also does not have an evaluation framework to assess the actual link, if any, between GEO’s rehabilitation interventions and its recidivism outcomes. The absence of an evaluation framework to understand Ravenhall’s results is a significant missed opportunity.
Ravenhall Correctional Centre
A remand prisoner is a person who has had a charge laid against them but has not had their proceedings finalised. Remand prisoners are not released on bail and await trial or sentencing in prison.
CV made significant changes to Ravenhall’s contract to help ease the system wide impact of growing prisoner numbers. These changes occurred both before and after the prison opened. The changes included introducing remand prisoners, which now make up 52 per cent of Ravenhall’s population, and increasing its number of prisoner places.
As of 31 December 2019, Ravenhall has 1 300 available prisoner places. GEO is waiting for CV to finalise another contract change to increase its number of prisoner places beyond 1 300, which is its design capacity.
CV and GEO agreed in principle that Ravenhall’s focus on rehabilitation and reintegration would be preserved throughout the contract changes. However, these significant changes occurred during Ravenhall’s first two years of operation—a period in which it was still settling and developing its culture.
Rehabilitation and reintegration services
GEO uses an evidenced based model to assess and treat prisoners’ individual risks of reoffending.
GEO designed Ravenhall’s rehabilitation programs for sentenced prisoners who stay at Ravenhall for three months or longer. However:
- half of Ravenhall’s prisoners are remandees
- more than half of all prisoners, on their release, have spent less than three months at Ravenhall.
Data from January 2019 to December 2019 shows that remand, together with sentenced short-stay prisoners (who serve less than three months) made up approximately 70 per cent of Ravenhall’s prisoner population. As a result, GEO’s original model for reducing recidivism is less relevant to its current prisoner population and therefore cannot be as effective at reducing reoffending.
GEO’s post-release model is designed to identify and address prisoners’ individual requirements, particularly those with high community reintegration needs.
From November 2017 (when Ravenhall first received prisoners) to December 2019, remand and short-stay prisoners did not have access to intensive post-release and case management services. This is because CV initially believed that, as part of the contract changes, GEO should provide these services at no extra cost to the state. CV has since recognised that intensive post-release services for remand prisoners was outside of the scope of the contract change. As of December 2019, CV has funded GEO to provide these services for a period of three years.
GEO has multiple assessment tools to identify prisoners’ post release needs. When we reviewed the files of 20 Ravenhall prisoners, we found that GEO has not consistently applied its assessment tools. Some eligible prisoners did not receive timely risk assessments, while other forms of assessments were completed late, or inconsistently.
Comparison to the public system
We compared Ravenhall’s model to the public system to identify any learnings or better practice that could be shared across the system. We found that CV’s Reintegration Pathway, which is used in public prisons, and GEO’s Continuum of Care model are aligned. Both models are based on the same underlying principles and offer similar services. Alignment between these models is important to ensure that all Victorian prisons are integrated and operate as one system.
However, GEO’s Continuum of Care model has some unique features, including:
- continuity of care—former prisoners have access to the same clinicians and staff they engaged with at Ravenhall after their release
- the Bridge Centre—a community reintegration centre where former prisoners can seek post-release support and assistance
- family involvement—the Bridge Centre hosts information nights and provides specialised services to support former prisoners’ families.
It is too early to determine if these features are improving prisoner outcomes compared to the public system. We encourage CV to monitor and compare outcomes as GEO’s model progresses.
Monitoring and evaluating performance
KPI 15 incentivises GEO to use interventions that match prisoners’ post release needs in education and training, employment, housing, alcohol and other drugs (AOD) and mental health.
GEO will receive performance payments of up to $1 million per year for achieving the KPI 15 and 16 targets. This collective $2 million is a small percentage of the overall service payment available to GEO. While KPIs 15 and 16 measure outcomes, several of GEO’s other performance measures also directly relate to rehabilitation and reintegration services.
KPI 15 measures GEO’s success at reintegrating prisoners and KPI 16 measures the rate at which Ravenhall prisoners return to prison within two years. These KPIs are narrow and, on their own, cannot meaningfully measure the prison’s recidivism outcomes.
KPI 15 is not working as intended and has proven difficult for GEO to design and implement. Originally, CV and GEO had different interpretations of the KPI. For example, they did not agree on which prisoners are eligible to be counted. CV and GEO are finalising revisions to KPI 15 based on a new shared understanding of who should be counted and how performance is assessed. Consequently, all KPI 15 results are under review and CV may retrospectively adjust GEO’s past performance payments.
KPI 16 incentivises GEO to achieve a lower return to prison rate than the average rate for other prisons.
GEO is aiming to achieve a 12 per cent lower rate for Ravenhall’s general prisoner population and a 14 per cent lower rate for Aboriginal and Torres Strait Islander offenders.
KPI 16 is not an appropriate measure for determining performance payments because it does not consider the amount of time an offender spends at Ravenhall or how much of their sentence they serve there. As a result, the KPI cannot effectively link prisoners’ reoffending outcomes to Ravenhall’s unique programs and interventions.
Additionally, KPI 15 and 16 do not include or measure remand prisoners. The substantial increase in remand prisoner numbers at Ravenhall therefore limits CV’s ability to assess the success of GEO’s Continuum of Care model.
CV has no plan to evaluate Ravenhall’s outcomes beyond the KPI measures, despite the prison having a key objective to trial new methods to reduce reoffending. This means that if Ravenhall has a lower rate of recidivism than other prisons, CV does not have an evaluation framework capable of understanding the links between Ravenhall’s specific interventions and its results.
CV regularly conducts research and evaluation projects about reintegration and reoffending measures. However, it does not have any system wide evaluations currently planned.
Performance outcomes to date
Beyond KPI 15 and 16, a number of Ravenhall’s other service delivery outcomes (SDO) and KPIs measure outputs relating to reintegration and rehabilitation, such as education and program completion rates. Ravenhall has had mixed performance results for these other measures. As a relatively new prison, we expected that it would take time for it to embed its programs and begin meeting its performance targets. Initially Ravenhall did not meet its performance targets for:
- SDO 14—prisoner engagement in purposeful activity
- SDO 15—vocational education and training
- SDO 23—case management.
GEO has made changes to address these performance issues. While GEO did not achieve its performance targets for SDO 14 and 23 in the second half of 2019, it achieved improved outcomes. Notably, GEO has consistently passed SDO 15 from July 2019 to December 2019 after its benchmark was adjusted. CV and GEO have also revised several SDO and KPI benchmarks and definitions to ensure they are appropriate for Ravenhall’s changed prisoner cohort.
We recommend that the Department of Justice and Community Safety:
1. review, and where necessary revise, KPIs 15 and 16, to:
- ensure they are appropriate measures for determining performance payments
- ensure they are working as intended
- determine if they should be applied to other private prisons in Victoria (see Section 3.2)
2. develop and implement an evaluation framework to assess reoffending outcomes at Ravenhall Correctional Centre, including:
- which interventions contributed to the outcome
- if outcomes differ between cohorts and potential causes
- if outcomes differ for those who have attended the Bridge Centre compared to those who have not
- if outcomes can be causally attributed or correlated to Ravenhall Correctional Centre (see Section 3.3)
3. advise government on the costs and benefits of different mixes of remand and short-stay prisoners compared to longer stay sentenced prisoners at Ravenhall and advise on a level that achieves an optimal balance between meeting demand for prisoner places and supporting Ravenhall to improve recidivism outcomes (see Sections 2.2, 2.3 and 2.4).
Responses to recommendations
We have consulted with DJCS and GEO and we considered their views when reaching our audit conclusions. As required by the Audit Act 1994, we gave a draft copy of this report to those agencies and asked for their submissions or comments.
We also provided a copy of the report to the Department of Premier and Cabinet.
The following is a summary of those responses. The full responses are included in Appendix A:
- DJCS accepted the three recommendations directed to it and provided an action plan detailing how it will address them.
- DJCS noted the context that Ravenhall’s contract changes were made in and outlined CV and GEO’s work to date to refine KPI 15.
- GEO stated that while Ravenhall’s prisoner cohort has changed, it is committed to rehabilitation and reintegration.