Gender responsivity in the assessment and treatment of offenders

Abstract

Risk assessment tools used to assess justice-involved women have been predominantly developed using male offenders, with validations for females having mixed results. Female-specific risk assessment tools are not well established, and validations are limited. In terms of treatment, either programmes in various jurisdictions have been developed specifically for females, or addendums are available to assist in the facilitation of programmes initially developed for male offenders. The efficacy of these has been varied with gender-responsive programmes proving to be more successful in the reduction of reoffending if justice-involved women follow a gender-specific pathway to offending than if they follow a non-gender-specific pathway. This paper discusses gender differences in offending and provides an overview of assessment and treatment of justice-involved women in Australia, the UK, New Zealand and Canada. The assessment and treatment practices for justice-involved women are viewed through an evidence-based lens, and opportunities for future development are identified.

1. Introduction

Although the proportion of offenders who are female is marginal across multiple Australian and international jurisdictions, the number of justice-involved women continues to increase in both custodial and community samples, and specifically for violence (de Vogel et al., 2016). Approximately 24.8% of all offenders proceeded against by police in the 2020/21 financial year were female (Australian Bureau of Statistics, ABS, 2022b), and this percentage is slightly less at 22.0% of those serving community orders in Western Australia (WA; higher than the national average of 19.4%; ABS, 2022a). Justice-involved women represent 7.7% of the prison population in Australia (ABS, 2021b), which is higher than comparable jurisdictions including Canada (5.6%), New Zealand (7.4%) and the United Kingdom (England and Wales; 4.6%); however, the USA remains the highest (9.8%; Walmsley, 2017). In comparison, 9.5% of the prison population in WA was female in 2021, with an incarceration rate nearly double that of the national average (ABS, 2022a).

In regards to offence type in Australia, the most prolific offences have consistently been acts intended to cause injury and illicit drug offences since 2017–2018 (ABS, 2022b). WA has the fourth largest population of violent justice-involved women proceeded against by the police in the country (ABS, 2022b), and the majority of justice-involved women in Australia, and specifically Western Australia, are incarcerated for acts intended to cause injury (ABS, 2021b).

The assessment and treatment of justice-involved women have been somewhat debated in the research. Covington (2008) notes that despite an increase in the rate of incarceration for justice-involved women, the criminal justice system has been unable to address this growth or meet the needs of this cohort effectively or in an appropriate manner: that is, to be gender-responsive. Covington (2008) defines gender-responsive as ‘the creation of an environment – through site selection, staff selection, program development, and program content and materials – that reflects an understanding of the realities of women’s and girls’ lives and that addresses and responds to their challenges and strengths’ (p. 377).

Historically, programmes for females were not gender-responsive as they did not address the range and types of needs of women (Joiner, 2011). Numerous researchers have argued the importance of gender responsivity in the assessment and treatment of justice-involved women, highlighting the need for holistic and trauma-informed practices to best meet the needs of the female offending cohort (Bloom et al., 2003; Covington & Bloom, 2007; de Vogel et al., 2019a; Greig, 2014; Joiner, 2011; van Voorhis et al., 2010; Willison & Lutter, 2009).

Developments in the assessment and treatment of justice-involved women continue across Australia and internationally. To the researchers’ knowledge, a consolidation of gender-responsive practices across jurisdictions has not been conducted. As such, this paper seeks to provide an overview of gender-responsive assessment and treatment of justice-involved women in various jurisdictions including Australia, UK, New Zealand and Canada. The international jurisdictions have been selected due to previous sharing of research, assessment practices and treatment programmes for offenders; with two of these jurisdictions having an Indigenous cohort that is similar to that in Australia in terms of proportion and overrepresentation in the criminal justice system. Further, these jurisdictions share a common law system as they are all part of the Commonwealth. The Australian jurisdictions were selected due to the similarities of treatment available for custodial based offenders and/or the scope of publicly available data.

The paper consolidates the information on gender-responsive assessment and treatment practices in these jurisdictions, provides an overview of the evaluations of female-specific treatment for offenders, summarises the findings of risk assessment validity for females for both gender-neutral and gender-specific tools and provides a pragmatic overview of recommendations regarding the future direction of assessment and treatment for justice-involved women.

2. Gender differences in offending and offenders

2.1. Complex trauma

Women’s pathways into prison and subsequent release into the community are qualitatively different to those of men regarding the types and details of offending, sentences, duration of prison terms and the rate of return. The abuse histories of women are significant (Stathopoulos & Quadara, 2014), and childhood victimisation increases the risk of ongoing victimisation in adulthood, particularly for women (Price-Robertson & Rush, 2013), with patterns of childhood abuse (sexual and/or physical) and ongoing abuse. As such, women tend to enter prison with more complex needs and a higher level of need than men (Gower et al., 2022, 2023; Stathopoulos & Quadara, 2014). Studies have demonstrated a clear relationship between early and ongoing victimisation, substance use and recidivism (Cimino et al., 2015; Moustafa et al., 2021; Saxena & Messina, 2021; Taylor, 2015).

Justice-involved women are the minority of those involved with the criminal justice system, and reference to evidence-based practices is largely centred on research conducted with male samples (Gobeil et al., 2016). As the pathways to offending are different for females (Brennan et al., 2010; Daly, 1992), it is likely that the contributing factors to offending are somewhat gender-specific.

Trauma and offending in women are highly intertwined, particularly for violent offenders (Saxena & Messina, 2021). Research suggests that a significant proportion of violent justice-involved women have been victims of various forms of trauma, including physical and sexual abuse, neglect and other adverse childhood experiences (Kruttschnitt et al., 2002; Saxena & Messina, 2021). Additionally, research has suggested that female offenders who have experienced trauma may be less responsive to traditional forms of punishment and rehabilitation, with the prison environment adding to the traumatic experiences of justice-involved women (Stathopoulos & Quadara, 2014).

2.2. Gender-specific factors

Factors identified in the research as being specifically linked to female offending include parenting responsibilities, historical trauma and mental health (Barlow & Weare, 2019; Blanchette & Brown, 2006; Brennan et al., 2010; Covington & Bloom, 2007; de Vogel et al., 2016; Salisbury et al., 2009), with well-established factors such as substance abuse, antisocial peer relationships and negative family/marital connections being more prevalent for adult and juvenile justice-involved females than males (Andrews et al., 2012; Prince et al., 2021). Despite this, recent research did not identify substance use as a significant predictor of reoffending in justice-involved women, noting that although substance use is indeed a risk factor for females, the link to reoffending was stronger for male offenders (de Vogel et al., 2022).

Further, Stephenson et al. (2014), noted that in an English sample of 586 male and juvenile offenders, juvenile females were more likely to have experienced abuse or neglect, attempted suicide, engaged in self-harm and be raised in a disadvantaged household than violent males. This pattern remains similar in adulthood. In a Dutch matched study of 140 justice-involved women, justice-involved women reported more historical and recent abuse and engaged in more self-harming behaviour, and a greater proportion demonstrated borderline personality disorder (BPD) traits than male offenders (Muller & Kempes, 2016). Further, research in Australia and Canada has found the rate of psychiatric disorders is higher amongst incarcerated offenders than that of community members for both males and females (Butler et al., 2006; Derkzen et al., 2013; Tye & Mullen, 2006).

2.3. Commencement of offending behaviour and frequency

In terms of onset and frequency, females commence offending at a later age and tend to engage in less crime throughout the lifespan than males (D’Unger et al., 2002; Ferrante, 2013; Weaver, 2010).

Further, in a Finnish sample of juvenile offenders, offending peaked in late adolescence in males whereas the offending of females tended to remain consistent throughout adolescence, with males 4.4 times more likely to engage in offending than females (Elonheimo et al., 2014). Further, the rate of male offending was greater than females across all crime types, with the smallest difference in property and drug offences (Elonheimo et al., 2014).

In Australia, male juvenile offenders are 2.2 times more likely to offend than their female counterparts, with the smallest differences in theft and violent acts (ABS, 2022b). Further, Western Australia has the highest offender rate for violent acts committed by juvenile offenders in the country (ABS, 2022b).

2.4. Gender differences in violent offending

Gender differences are further evident in violent offending with the onset and subsequent cessation of violent offending occurring earlier in the life trajectory of females than males (Reisig et al., 2006). In a Western Australian sample of offenders incarcerated for violent offences, females had a later onset of offending behaviour (based on convictions) than a matched sample of males, engaged in juvenile crime less frequently, and had less occurrences of violence across the lifespan (Gower et al., 2023). Further, women’s use of violence tends to be more often reactive and relational, results in serious harm less frequently than that of males (de Vogel & de Vries Robbé, 2012b) and involves victims from within the offender’s social circle (de Vogel & Lancel, 2016). Despite this, and as per Babcock et al. (2003), it would seem that the relationship to the victim and nature of the violence (general or within an intimate relationship) influences whether the violence is instrumental or reactive, rather than it simply being a matter of violent women tending to be more emotionally driven in their actions than males. Further, there is more homogeneity in the violent female cohort in terms of risk and need than in males (Gower et al., 2023; Murdoch et al., 2012), with common factors including childhood abuse and violence, normalisation of violence and poor emotional regulation (de Vogel et al., 2016; Garcia-Mansilla et al., 2011; Murdoch et al., 2012).

In a psychiatric sample, Greig (2014) did not find any significant gender differences across rate of violence, severity, type of violence or the frequency in which it occurred. Despite this, it was acknowledged that the females in the sample tended to exhibit more property-based violence during institutionalisation, with males more likely to direct their violence towards a person. In contrast, when focused on physical violence, de Vogel et al. (2016) note that behaviour in a forensic psychiatric facility was more problematic for females, with more self-harm and violence, particularly directed at staff, and higher rates of manipulative behaviour. The type of violence committed by females was different to that of males, with victims more likely to be related to the perpetrator for females.

Aside from the differences across gender, further contributing factors differed between females who were violent towards partners only and those who were violent towards partners and others. Babcock et al. (2003) noted that generally violent females reported more traumatic symptoms yet did not experience significantly more abuse than females who were violent to partners only. They were, however, more likely to have witnessed aggression by their mothers, experienced more physical abuse, caused their partners more physical injury and engaged in violent behaviours more often than those who were violent towards partners only. Generally, violent females had more diagnosed psychological problems than non-violent females, and Babcock et al. (2003) noted that this cohort tended to have been raised and/or socialised to believe that females engaging in violence is acceptable, and this is particularly relevant for Aboriginal offenders. In expansion, Aboriginal girls (and women) are not discouraged from engaging in physical aggression and/or violence simply based on their gender, and they are not taught that males hold dominion over such acts, as per the Western narrative (Bessarab, 2017).

In relation to recidivism, McCoy and Miller (2013) did not find any significant differences in reoffending rates by gender; however, non-violent males were more likely to engage in reoffending involving property-related offences, whereas females more frequently reoffended by engaging in prostitution. The presence of social supports significantly reduced recidivism in females but did not have a significant impact on reoffending rates for males. This indicates that protective factors, or areas of strength are impacted by gender.

3. Risk assessment tools and validity for females

It has been acknowledged that risk assessment tools used to assess justice-involved women have been predominantly developed using male offenders (Bauman et al., 2013; de Vogel et al., 2019a; de Vogel & Nicholls, 2016; Greig, 2014; Salisbury et al., 2009; Wardrop et al., 2019), with subsequent validations for females having mixed results. Researchers have found that the addition of gender-responsive factors to existing risk assessment tools such as the Level of Service Inventory–Revised (LSI–R) increases the validity for justice-involved women (van Voorhis et al., 2010), thereby ensuring the identification of essential treatment needs for women and contributing towards treatment planning.

Assessment tools that have been used to determine the risk and/or needs of offenders in Australia and international jurisdictions include the Level of Service/Risk, Need, Responsivity (LS/RNR), Historical Clinical Risk Management (HCR–20 V3), the Psychopathy Checklist Revised (PCL–R; Geraghty & Woodham, 2015; Gower et al., 2020) and the Violence Risk Scale (VRS; Gower et al., 2020).

Validations of the LS suite of tools have provided evidence that the tool reliably predicts reoffending in justice-involved women generally (Dyck et al., 2018; Hye-Sun, 2010; Manchak et al., 2009; van Voorhis et al., 2010) but the validity for violent justice-involved women has not been consistent, with low predictability (Ostermann & Herrschaft, 2013; Watkins, 2011), higher predictability than that for males (Singh et al., 2011; Stewart, 2011) or lack of significance (Gordon et al., 2015) reported. Females who follow more gendered pathways to crime, as per Daly (1992), do not seem to be accurately assessed for risk. As per Reisig et al. (2006), the LSI–R accurately predicted recidivism for financially motivated females with higher scores equating with higher rates of reoffending. In contrast, there was no significance in the predictability of the LSI–R for recidivism of females who followed gendered pathways to crime. Overestimates of risk were particularly prevalent for women whose offences were drug connected or involved harm to others, or who had been subject to harm themselves (Reisig et al., 2006).

Specifically for violent females, the HCR–20 and PCL–R have been found to be accurate predictors of inpatient aggression/violence (de Vogel & Lancel, 2016; de Vries Robbé et al., 2016; Gray & Snowden, 2016; O'Shea et al., 2014; Strub et al., 2016) and community violence (Coid et al., 2015; Warren et al., 2018) for forensic psychiatric patients and offenders. The single validation study for females and the VRS (Stewart, 2011) identified higher predictability of subsequent violence than the normative sample (males) or other validation studies conducted using the male offending cohort (Lewis et al., 2013).

3.1. Risk levels

Of the primary risk assessment tools used to assess general and violent offending in Australia, only the VRS has specific risk-level ranges for justice-involved women. The LS/RNR has stratified norms for percentiles and recidivism rates associated with the total risk score for females; however, this does not alter the overall label of low, medium or high. Many studies have acknowledged the lower recidivism rates of violent females than males for both general offending (Gray & Snowden, 2016; Hye-Sun, 2010) and violent offending (Coid et al., 2015; Garcia-Mansilla et al., 2011; Hye-Sun, 2010), including Stewart (2011) who subsequently validated the VRS for females and established gender-specific risk-level ranges to address the underestimation of risk for this cohort.

The HCR–20 and PCL–R do not have different scoring structures for females, yet risk levels on the HCR–20 are determined by the clinician in consideration of the number and relevance of factors present, and the degree of intervention/supervision required to mitigate this risk. The cut-off scores for the PCL–R differ by jurisdiction with the choice to either diagnose or limit discussion to psychopathic traits dependent on the jurisdiction. Although the PCL–R has a different cut-off score to indicate psychopathy in females (de Vogel & Lancel, 2016), this remains theoretical and seems restricted to application in research only at this stage.

4. Female-specific risk assessment tools and validity

The addition of gender-responsive factors to existing gender-neutral risk assessment tools improve overall predictive validity (Bauman et al., 2013; van Voorhis et al., 2010; Wardrop et al., 2019). Although the addition of gender-responsive factors, including mental health, self-esteem, self-efficacy, relationships and child and adult victimisation, did not improve on predictability of the LSI–R (Salisbury et al., 2009), participant numbers were limited. When van Voorhis et al. (2010) increased the scope of the research to include a greater number of participants, the supplementation of the LSI–R with gender-responsive factors significantly improved the validity of the tool. Further, the gender-responsive factors that achieved significance were somewhat different for those on probation/parole, incarcerated offenders and released prisoners, suggesting shifting priorities in the factors dependent on the status of the offender. Similar results are evident when gender-responsive risk assessment tools or addendums have been used for justice-involved women, as evidenced by the Female Additional Manual (FAM; de Vogel et al., 2014), Women’s Risk Need Assessment (WRNA; van Voorhis et al., 2010) and the Service Planning Instrument for Women (SPIn–W; Orbis Partners, 2006).

4.1. Female Additional Manual (FAM)

Assessment tools for justice-involved women have taken the form of either addendums/supplements for existing risk assessment tools or standalone risk assessment. The FAM (de Vogel et al., 2014) to be used with the HCR–20, for example, includes female-specific items regarding violence for females plus gender-specific guidelines to consider when scoring standard items. Validations of the FAM have been varied, with Greig (2014) noting that the FAM improved on the predictability of HCR–20 items in a sample of 41 females from a psychiatric facility, although differences failed to achieve significance. The FAM was strong in the prediction of threatening behaviour in females, particularly upon review of the 5-point risk judgement. In a forensic psychiatric sample of 78 female psychiatric patients who were followed up for 11.8 years, the FAM was a strong predictor of general recidivism but failed to achieve significance for the prediction of violent behaviour (de Vogel et al., 2019b).

4.2. Women’s Risk Needs Assessment (WRNA)

In line with the research discussed above, the WRNA specifically measures risk and identifies the treatment needs of justice-involved women as a means of reducing prison misconduct rather than recidivism in the community (van Voorhis et al., 2010). In an extensive validation study with 469 female prisoners from across three states (Missouri, Ohio and Rhode Island), with a follow-up time of 12 months, the WRNA demonstrated strong predictive validity for prison misconduct for justice-involved women. Further, the trailer version of the WRNA (WRNA–T) was designed to supplement existing risk assessment tools such as the LSI–R. When the WRNA-T was combined with the LSI–R in a sample of 210 female prisoners, the overall validity of the LSI–R improved (Bauman et al., 2013).

4.3. Service Planning Instrument for Women (Canada) (SPIn)

The SPIn (Orbis Partners, 2006) adheres to the RNR principles and incorporates strengths, but not simply from opposing sides of the same construct. Instead, it is acknowledged that strengths may indeed be present within the same factor, with criminal peers being a clear example when offenders have both antisocial and prosocial peers within their friendship groups. Offenders can be classed as both high need and high strength within the same factor (Jones & Robinson, 2018). Irrespective of the overall level of risk for an offender, a high amount of strengths leads to lower rates of recidivism (Jones & Robinson, 2018). A specific version for females, the SPIn–W, incorporates 10 gender-responsive factors. In a sample of 274 justice involved females who were rated as medium or high risk on the LSI–R, the SPIn–W demonstrated higher predictive validity than the LSI–R, and the additional factors were found to aid in referring females to services that specialise in trauma (Robinson et al., 2012).

5. Treatment needs for justice-involved women

The research on the treatment needs of justice-involved women has consistently found mental health, substance abuse, victimisation, parenting concerns and relationships to be the primary treatment needs of females (Covington & Bloom, 2007; de Vogel & de Vries Robbé, 2012a; Salisbury et al., 2009), particularly if reductions in recidivism are to occur. Identification of the gender-specific treatment needs alongside the general needs captured in current risk assessment tools would enable service providers to structure treatment that is appropriate and to achieve the best outcomes. Considering prison behaviour, for example, van Voorhis et al. (2010) found that mental health, substance abuse and trauma were particularly important to aid in females’ adjustment to the prison environment. Addressing such treatment needs would likely lead to a reduction of ongoing traumatisation of this cohort along with more positive behaviour in the prison environment, thereby lessening the pressures on custodial based staff.

Although gender-specific factors such as mental health and victimisation are important to consider when providing criminogenic treatment to justice-involved women, the ‘big four’ remain vital. As detailed in Gower et al. (2020), research on the treatment needs of justice-involved women in prison compared to the community indicated that the ‘big four’ were consistent across samples.

In contrast, Salisbury et al. (2009) indicated different treatment needs for females when comparing community and prison samples, and the ‘big four’ were not the strongest predictors for prison misconduct or recidivism. Different treatment needs were identified for managing prison misconduct, with behaviour more problematic if self-efficacy was high, yet this same trait proved to be protective when offenders were released into the community. Childhood victimisation was related to prison misconduct yet this was not a significant predictor of reoffending once offenders were released into the community. Similarly, van Voorhis et al. (2010) found that although the validity of gender-neutral risk assessment tools was strong for females, different treatment priorities were identified for females compared to males. The measurement of gender-responsive treatment needs such as adult victimisation, anger and family support enhanced the validity of the LSI–R, particularly for prisoners. This information suggests that the treatment priorities for justice-involved women differ depending on whether the focus is on prison misconduct or recidivism rates across community and prison populations.

6. Efficacy of gender-responsive treatment

Treatment effects are greater when gender-responsive strategies are incorporated as responsivity factors (Blanchette & Brown, 2006; de Vries Robbé et al., 2015; King, 2011). Further, gender-responsive interventions produce greater improvements in recidivism rates than gender-neutral interventions for substance-related offending (Gobeil et al., 2016; Kubiak et al., 2016) and violent offending (Kubiak et al., 2016; Messina et al., 2016). In a sample of 5109 justice-involved women with substance use issues, those who completed gender-responsive treatment programmes demonstrated lower risk of substance-use-related arrests upon completion in comparison to general alcohol and other drug (AOD) treatment. Further, significant reductions in arrests were maintained two years after treatment (Kissin et al., 2014).

Despite these findings, there is some evidence to suggest that the efficacy of gender-responsive treatment is not as strong for justice-involved women who do not follow a gendered pathway to crime (Day et al., 2015; Gobeil et al., 2016; Saxena et al., 2014). Day et al. (2015) found that gender-responsive treatment resulted in significantly less reoffending for females than traditional programmes, but this was only true for those who had gender-specific risk factors that contributed to problematic behaviour. Although the study focused on juveniles in secure detention, the findings suggest that the treatment of juvenile females is not a case of gender responsivity alone, and that consideration of the risk factors is imperative. In adult justice-involved women, Saxena et al. (2014) found that although gender-responsive treatment resulted in lower risk of reoffending for justice-involved women, this effect was limited for those who did not have a history of trauma. Finally, Gobeil et al. (2016) conducted a meta-analysis on correctional interventions for justice-involved women and concluded that gender-responsive treatments are more effective at reducing the risk of reoffending than general neutral treatment, and this positive impact is greater for those offenders who have followed a gendered pathway to crime.

In discussing gender-responsive treatment, the work of Covington and colleagues must be mentioned. Various gender-responsive programmes have been developed by Covington and colleagues in collaboration with the University of California Los Angeles (Van Voorhis, 2022) and are facilitated in Californian correctional facilities. These programmes (Beyond Trauma; Covington (2016), Helping Women Recover; Covington (2019), Beyond) Violence; Covington (2013) utilise various approaches, including cognitive behaviour therapy (CBT), psychoeducation, art therapy and eastern practices such as mindfulness, to address trauma, self-esteem, abuse and relationships, to name a few (Van Voorhis, 2022). In subsequent evaluations, these programmes have been found to reduce substance use in a sample of 115 female prisoners with histories of abuse (Saxena et al., 2014) and result in lower rates of return to custody than that for standard treatment in a sample of 91 female prisoners incarcerated for violent offences (Messina et al., 2016).

Specifically for violent offenders, gender-responsive treatment has demonstrated significantly better and longer term reductions in recidivism than a gender-neutral violence programme provided in a custodial environment (Kubiak et al., 2016). Overall, female-specific programmes have been found to reduce psychiatric symptoms, thereby improving mental health, promoting psychological well-being, reducing posttraumatic stress disorder (PTSD) symptoms and reducing substance use (Messina et al., 2012). The Beyond Violence programme (Covington, 2013), for example, was developed for females with a history of violence. As it addresses trauma, both perpetration of violence and victimisation from violence are addressed in the programme using the same therapeutic techniques as per Covington’s gender-responsive treatment programmes mentioned above. Preliminary evaluation results have been positive, with significant improvements found in psychological functioning (Messina et al., 2016) and lower recidivism rates than in a group of women who engaged in standard treatment (Kubiak et al., 2016).

Despite these positive findings, in their meta-analytic review of the efficacy of treatment for violent offenders, Papalia et al. (2019) noted that the limited number of studies specifically focused on violent justice-involved women prevent evidence-based conclusions on gender differences. Evidence on the efficacy of programmes for violent females is limited; however, research on the differing needs of female offending continues to be pursued in Australia and elsewhere to contribute to ongoing improvements in the assessment and treatment of justice-involved women internationally and across Australia.

7. Jurisdictional scan on gender-responsive strategies

Various jurisdictions, including Australia, Canada, New Zealand and the UK, have implemented gender-responsive strategies to assess and treat justice-involved women with varying degrees of success. These strategies include female-specific risk assessment tools, gender-responsive programmes and operational practices/initiatives designed to improve the efficacy of the treatment and management of this cohort. Three international and three national jurisdictions were reviewed to highlight gender-responsive practices in these geographical areas as case studies. The specific international jurisdictions were selected as many rehabilitative practices implemented in WA are based on research that originates from the UK, Canada and New Zealand. As the main researcher is based in WA, and the data utilised in previous chapters was sourced from WA prisoners, this state was selected as a case study. For comparison, South Australia and Victoria were selected due to publicly available data, some consistency across assessment and treatment practices in the former, and comprehensive strategies developed for females in Victoria.

7.1. International jurisdictions

7.1.1. Canada

The proportion of the Federal prisoner population in Canada who were female has ranged from 6.2% to 8.4% since the 2013–2014 financial year. There has been a steady reduction in the proportion of the female Federal prisoner population since 2016/2017, with the most recent rate reported as 6.2% (Public Safety Canada Portfolio Corrections Statistics Committee, 2019). The Correctional Service of Canada (CSC) developed five core principles for servicing the needs of incarcerated justice-involved women – empower women offenders, nurture meaningful and responsible choices, foster respect and dignity, create a supportive environment, and promote shared responsibility (Correctional Service of Canada, CSC, 1990). Since this time, the principles have been incorporated into the treatment of justice-involved women to enable this cohort to identify and implement relevant behavioural change (Wardrop et al., 2019). Programmes available to justice-involved women are holistic in nature and adhere to the RNR principles, with an Indigenous stream of programmes available that encompass spirituality and connection to Elders. Although the RNR principles are adhered to across the country, the risk assessment tools used to determine risk and treatment needs vary across Canada, dependent on the province/region (Mugford et al., 2017).

The programme suite includes motivational programmes, medium and high intensity (to address general offending, violence and substance use), rolling programmes, self-management (maintenance) and a female-specific sex offender programme (Appendix), all of which are trauma-informed. Further, the programmes take a scaffolding approach, with the Women’s Engagement Program completed in preparation for more intensive criminogenic programmes and the medium-intensity programme as a pre-requisite for the high-intensity programme for those assessed as high risk of reoffending, with each programme building on the skills of the previous treatment (CSC, 2020a).

The CSC provides comprehensive mental health services to justice-involved women in acknowledgement of the higher prevalence of psychiatric disorders than males with a focus on wellness, participation from the offender, female centred and minimal restrictions when possible. Further, a specific intensive mental health programme is offered to justice-involved women on Dialectical Behaviour Therapy (DBT) to assist females who require specific assistance in managing their emotions and associated behaviours, with a multi-disciplinary team (MDT) providing consistent care rather than treatment being restricted to group hours (CSC, 2020b).

An evaluation of correctional reintegration programmes conducted by CSC identified that justice-involved women who completed programmes tended to engage in less reoffending than non-completers who were eligible for the programme, despite small sample sizes (CSC, 2019). Further, the evaluation noted that eligibility and risk assessments were overridden by clinicians in many cases (CSC, 2019), and this moves away from the RNR model, thereby impacting the potential successes of the programme. In line with this, the programmes for females did not achieve the intended outcomes for those released to the community. Recommendations were made for improvement in the areas of gender responsivity, programme content, scheduling and assessment, and subsequent research into the way in which responsivity practices are incorporated into programme facilitation. Finally, the evaluation focused on a short timeframe for programme participation, thereby limiting the numbers and ability to assess efficacy for different offender types (e.g. violence, sexual, substance use). It is not known whether these recommendations have yet been addressed.

7.1.2. United Kingdom (UK)

As at September 2021, approximately 4.0% of the UK prison population are female, and this is consistent with annual figures from 2020 (Ministry of Justice, 2021b). Offenders are assessed utilising a consistent suite of assessment tools regardless of gender and include the Offender Assessment System (OASys), Offender Group Reconviction Scale (OGRS3), OASys Violence Predictor (OVP), Risk of Serious Recidivism (RSR) and Risk of Serious Harm (RoSH; J. Urquhart, personal communication, 12 November 2020). The OASys was developed by the National Probation Service (NPS) and the Prison Service; it adheres to the ‘What Works’ literature and is a combination of actuarial methods and structured professional judgement (Howard & Dixon, 2012). The OASys determines likelihood of reoffending in consideration of criminogenic needs such as employment, education, substance use, problem thinking and relationship issues.

Treatment programmes are based on CBT, adhere to RNR principles and are accredited programmes as they must meet a set of criteria assessed by an independent panel of experts from various international jurisdictions (Ministry of Justice, 2021a). There are specific programmes for female prisoners (Table A1) that provide the opportunity for offenders with violence and substance use in particular to receive gender-responsive treatment. Although an evaluation of the Choices, Actions, Relationships, Emotions (CARE) programme was completed in 2019 (Wilkinson et al., 2019), the sample size was relatively small, a matched sample was not used, and a measure of recidivism was not included. Despite this, significant improvements were identified in the areas of emotional management, coping skills and specifically anger management, with reductions in prison misconduct evident within the 12 months following programme completion.

In terms of operational processes, a training programme is available for both community and custodial staff that outlines best practice with regards to working with women and how to engage in trauma-informed practice (Ministry of Justice, 2018). In addition, the Female Offender Strategy has been developed to support the management of females in prison including the establishment of mother and baby units and a dedicated women’s psychological service to meet the holistic needs of justice-involved women (Ministry of Justice, 2018).

7.1.3. New Zealand (NZ)

NZ uses an algorithm and models of reconviction (RoC) and reimprisonment (RoI) developed for use with justice-involved men and women (Poels, 2005) to determine probability of recidivism in a five-year follow-up period whereby the offence would be serious enough to result in imprisonment. Although the RoC*RoI works well with justice-involved women, King (2004) noted that the development of risk assessment measures specifically for females would likely have resulted in a different set of factors, although an improvement in predictive accuracy cannot be guaranteed. Additional risk assessment tools are used to determine the risk and treatment needs of violent justice-involved women including the VRS or HCR–20; Level of Service/Case Management Inventory (LS/CMI); and Psychopathy Checklist Revised (PCL–R) or Psychopathy Checklist Screening Version (PCL–SV), with some inconsistency in the selection of risk assessment tools due to a lack of departmental guidelines (Poels, 2005).

For the treatment of justice-involved women, NZ Corrections developed a treatment programme ‘Kowhiritanga’ (making choices) to address standard criminogenic needs, with a focus on emotional and victimisation issues (Department of Corrections, 2009). All rehabilitative programmes in NZ are evaluated annually to determine the efficacy and impact on both reconviction and reimprisonment. Since its inception, Kowhiritanga has consistently demonstrated lower reconviction and reimprisonment rates for programme completers than for a matched sample; however, these results have not achieved significance. The most recent numbers demonstrated a difference of 1.2 percentage points for reimprisonment and 6.4 percentage points for reconviction within 12 months following release from prison (Department of Corrections, 2021).

NZ Corrections launched the Women’s Strategy 2017–2021 in August 2017 and noted three components: intervention and services meet the unique risk and needs of justice-involved women; the management of justice-involved women must be trauma-informed and empowering; and management of women must reflect the importance of relationships to women. Following the launch, the programme suite for females was enhanced, and the ‘Kimihia’ programme (Creating Pathways for Wellbeing), a violence programme specifically for females, was developed. Kimihia was available to those justice-involved women who were not eligible for Kowhiritanga due to complex needs and high risk, as measured by the LS/CMI (Bakker et al., 1999). The programme was developed in collaboration with custodial officers, psychologists, programme staff, cultural advisers and female prisoners who had completed a programme for females (Appleyard, 2018). The programme is trauma-informed, aligns with the central eight factors as per Andrews and Bonta (2010) and is holistic to meet the treatment needs of females including mental health, victimisation, lack of support networks, parental stress, financial issues and employment (Appleyard, 2018). A full evaluation of this programme has not yet been completed.

7.2. Australian jurisdictions

7.2.1. South Australia (SA)

Approximately 7.3% of SA’s custodial population is female (ABS, 2021a). In consultation with a wide network of stakeholders, the Department of Corrections South Australia (DCSSA) has developed a Women’s Action Plan with the key objective to improve the circumstances of justice-involved women to enable them to lead healthy lifestyles, engage in meaningful opportunities, develop prosocial supportive relationships and avoid offending (Department of Corrections South Australia, DCSSA, 2019). The action plan was originally launched in 2015, but is periodically revised and includes specific objectives regarding the delivery of a correctional service for females; facilitation of programmes to meet the diverse needs of females; workforce recruitment, training and development; gender responsivity in policy and planning; and pathways/linkages to community and community services.

For assessment, DCSSA utilises the LS/RNR to assess the risk and needs of justice-involved women, but no other risk assessment tools are used at this stage (G. Mercer, personal communication, 6 November 2020). In terms of programming, DCSSA has developed a new gender and trauma-informed programme for females, ‘Stepping Stones’; however, as this is relatively new, and progress was interrupted by COVID, an evaluation has not been conducted beyond initial interviews with participants from the pilot group (G. Mercer, personal communications, 6 November 2020 and 8 March 2022). The programme consists of 120 hours and was developed in collaboration with Aboriginal Elders, Rehabilitation Programs Branch clinicians, representatives from Women’s Safety Services South Australia and the DCSSA Women’s Advisor. The programme targets key criminogenic needs including substance abuse, relationships and antisocial attitudes but incorporates gender-specific factors including trauma, parenting responsibilities and family and domestic violence.

In addition, DCSSA facilitates a psychological skills programme for women that is based on DBT and targets justice-involved women diagnosed with BPD; this programme is due to be reviewed in 2021–2022 (G. Mercer and J. McFarland, personal communication, 6 November 2020). For Indigenous justice-involved women, DCSSA has developed a programme ‘Aboriginal Women’s Our Way My Choice’ that covers intergenerational trauma and identity as an Indigenous female, and the pilot of this programme commenced in 2021, with a full evaluation expected in 2023 (G. Mercer, personal communication, 8 March 2022).

7.2.2. Victoria

Victoria has the lowest rate of justice-involved women in custody in Australia at 5.8% (ABS, 2021a). The assessment of justice-involved women includes the LS/RNR to determine general risk level and treatment needs (Corrections Victoria, 2020b). Specialised assessments are subsequently conducted for serious violent justice-involved women (HCR–20) and female sex offenders (Risk of Sexual Violence Protocol, RSVP; Adult Parole Board of Victoria, 2020), and it is apparent that these tools differ to those used with male offenders (VRS–SV and Static–99R) due to the lack of validation for use with justice-involved women.

Corrections Victoria has implemented many standards and policies pertaining to the management, assessment and treatment of justice-involved women, including a women’s policy (Corrections Victoria, 2017), standards for the management of female prisoners (Corrections Victoria, 2014) and Women’s Correctional Services Framework (Corrections Victoria, 2007). Further, female-specific programmes have been implemented covering general and violent offending (Table A1), with additional non-clinical programmes that focus on short-term intervention for violence (including family and domestic violence, FDV), positive relationships, victim survivor programmes, parenting programmes and cultural services (Corrections Victoria, 2020a). Programmes specifically for FDV, the Medium Intensity Violence Prevention Program and the high-intensity version of the same, do not seem to have been developed specifically for females; however, female perpetrators are eligible to participate.

Evaluations of the Better Pathways strategy for women has occurred, with findings illustrating that this strategy has had a positive impact on the reduction of recidivism. Unfortunately, evaluations for criminogenic programmes for females in Victoria do not appear to be publicly available.

7.2.3. Western Australia (WA)

WA has the highest rate of justice-involved women in custody at 9.8%, considerably higher than the national average of 7.7% (ABS, 2021a). To determine the risk level and treatment needs of justice-involved women, the LS/RNR and VRS are used, with the Stable 2007 utilised to identify treatment needs for sex offenders in the absence of scoring. Although none of these tools have been developed specifically for justice-involved women, the VRS has female-specific score ranges for each risk level as per the validation study completed by Stewart (2011), and these have been implemented across the state.

Operationally, the philosophy and strategy for managing females in prison is in the final stages of development, with the Women and Young People directorate aims including the provision of gender-responsive, trauma-informed and culturally appropriate services, treatment and support to justice-involved women (Department of Justice, 2021). Initiatives such as Aboriginal language courses to maximise the transition back to the community, the building of a mental health unit within the maximum-security female prison and pathways to employment programmes have been implemented, and these initiatives focus on the development of life skills, traineeships and vocational skills, and education.

The minimum security facility, Boronia, addresses the diverse needs of females in prison, acknowledging that although prison is a punishment, it provides the opportunity to enhance each prisoner’s potential and reunite them with family/community in improved health (Botello, 2017). Boronia operations aim to reflect the responsibilities faced by women, with a strong focus on community and family, achieved through activities such as work in the community and various traineeships.

In terms of criminogenic treatment specifically designed for justice-involved women, this is limited to the Choice, Change and Consequences (CCC) programme developed in New Zealand and encompasses CBT and DBT to address general and substance abuse and violent offending in females (Department of Justice, 2018). To date, an evaluation of CCC has not been completed. The substance use programme available for justice-involved women (Pathways; Wanberg & Milkman, 2006) resulted in higher rates of recidivism for females identified as higher risk of substance use (Office of the Inspector of Custodial Services, OICS, 2014); therefore, OICS recommended the development of an intensive substance abuse programme specifically for females. A modified therapeutic community to address substance use for justice-involved women was established in 2018 with treatment being trauma-informed, holistic and facilitated in response to the specific needs of justice-involved women; however, this has yet to be evaluated in terms of its efficacy at reducing recidivism. Early results indicate a return to custody rate of less than 3% (Department of Justice, 2021).

8. Future of assessment and treatment of justice-involved women

8.1. Risk assessment

Researchers and practitioners continue to strive to improve the current practices associated with the assessment and treatment of justice-involved women, with the debate between gender-neutral and gender-responsive risk assessment ongoing. Although there is support that widely used risk assessment measures such as the LS/RNR, HCR–20 and VRS are valid for use with justice-involved women (Gower et al., 2020), such tools must be administered in an ethical and gender-responsive manner (Olver & Stockdale, 2021). For example, the LS/RNR specifically allows users to capture factors identified as gender-specific such as financial issues, parenting concerns, mental health and trauma. Although these factors are not part of the scored section of the tool, they are no less important in the assessment process. Olver and Stockdale (2021) suggest that specific risk assessment tools for justice-involved women are not necessarily warranted, and that assessment should take a gendered approach with specific guidelines to ensure that the assessment and management of justice-involved women are ethical and responsible. This supports the work of de Vogel et al. (2014) and the development of the Female Additional Manual (FAM; de Vogel et al., 2014), designed as an addendum to the HCR–20, an already widely validated tool, thereby providing a guide for practitioners in the gender-responsive assessment of violent females.

8.2. Treatment

As previously discussed, justice-involved women have gender-specific treatment needs that are linked to recidivism. For example, Andrews et al. (2012) mentioned that although the ‘big four’ have been recognised as the primary predictors of offending, there is more of a ‘big five’ for women, with substance use being a primary need for this cohort. It is therefore likely that substance abuse will need to be addressed in any treatment programme and should aim to help justice-involved women identify the triggers for substance abuse; develop an understanding of the interconnectedness of trauma, mental health, violence and substance abuse; and reduce recidivism (Wright et al., 2012).

Justice-involved women generally have above-average rates of cumulative trauma (physical and sexual abuse) and mental illness, and substance use as a means of managing distress and unmet treatment needs (Bloom & Covington, 2008). Further, they report higher rates of childhood abuse, both physical and sexual, and victimisation as an adult (Ney et al., 2012). Such trauma can lead to a range of negative outcomes, such as low self-esteem, depression and anxiety, which can contribute to criminal behaviour. If the needs of justice-involved women are not met in a trauma-informed manner, their exposure can further compound their issues. It is vital that justice services provide appropriate interventions that address offender behaviour as well as the underlying issues to reduce reoffending.

Turning to WA, none of the programmes currently available to justice-involved women in prison address trauma explicitly, yet they are trauma-informed. Specifically, the modified therapeutic community set up at Wandoo Rehabilitation Prison to target AOD aims to address trauma; however, it is not only the programme itself that targets this need, but rather the therapeutic community as a whole (Cyrenian House, 2018). An evaluation of this programme/facility has not yet been completed.

Justice-involved women in WA present with more criminogenic needs than males irrespective of offence type (general and violent) or Aboriginality (Gower et al., 2022). For violent offenders, this may result in more treatment needs that need to be addressed despite treatment programme duration not being greater than equivalent treatment for males. If programmes are to be holistic, trauma-informed and gender-responsive in their design, then this may not allow time for justice-involved women to address all those needs adequately in a programme. In the absence of updated evaluations, the efficacy of such programmes cannot be unpacked further, and it is not possible to determine whether the programmes currently available to females are effective, or what part programme duration plays in this efficacy.

8.3. Recommendations for future research

8.3.1. Risk assessment

Justice-involved women have gender-specific factors that are associated with criminal behaviour such as historical abuse, mental health, parenting and financial difficulties (Messina & Esparza, 2022). If popular risk assessment tools such as the LS/RNR do not encapsulate these factors, yet are used to determine treatment and services, then it is possible that the primary needs of justice-involved women are not being addressed.

Although the LS suite of tools has generally proven to be valid for females in the justice system (Gower et al., 2020), its validity is not consistent, particularly with Australian violent females (Gordon et al., 2015). As treatment and services provided to offenders are guided by risk assessment, it would be beneficial if future research investigated the validity of the LS/RNR alongside gender-specific factors, as per the work of Stewart (2011), for example. Gender-specific factors may not improve the validity of tools such as the LS/RNR; however, the inclusion of such factors may aid in directing justice-involved women to appropriate treatment and services.

Turning to risk assessment for violent offending, WA utilises the risk scores and associated risk levels specific for females on the VRS, but other Australian jurisdictions are yet to use the VRS with violent justice-involved women (personal communication, E. Ziersch, 3 May 2022). Further, as the VRS–SV has not been validated for females, this cohort is assessed comprehensively for violence using the full VRS. As previously mentioned, Stewart (2011) found strong validity for the VRS and its ability to predict both general and violent reoffending in females despite low base rates of reoffending.

The work of Stewart (2011) continues to be the only known validation of the VRS for justice-involved women. This has likely contributed to the prevention of wider use across Australia, and, as such, it is suggested that future research focuses on validating the VRS for use with justice-involved women across multiple jurisdictions. Aboriginality must also be captured in any validation studies due to the overrepresentation of Australian Aboriginals in the criminal justice system. Further, as recent research has found differences in the VRS items by Aboriginality in a cohort of violent justice-involved women (Gower et al., 2022), validity for Aboriginal Australians should be determined separately to ensure that the VRS items are predictive for all violent offenders regardless of gender and/or culture.

8.3.1. Treatment

Firstly, for all jurisdictions, and as per best practice, regular evaluation of treatment programmes is necessary to ensure efficacy of treatment, inform potential improvements in programmes and maintain a current evidence base that supports ongoing facilitation of such programmes. To date, WA has not as yet released an evaluation of any female-specific programmes; therefore, an evaluation of programmes such as the CCC to determine the efficacy for both Aboriginal and non-Aboriginal women would be beneficial to determine the ongoing treatment needs and demands of justice-involved women. Gower et al. (2023) identified that most violent justice-involved women were high risk on both the LS/RNR and VRS, suggesting that the demand for intervention is predominantly driven by high-risk women with complex needs. The question is then whether programmes such as the CCC are adequately meeting the needs of these women.

If incarcerated violent justice-involved women continue to be predominantly high risk, and the upward trajectory of violence is maintained, any increase in female incarceration rates, particularly violence, will continue to place pressure on correctional systems to meet the needs of this cohort, particularly as those who are high risk are more likely to return to the justice system. As such, there is an opportunity for improvements in the assessment and management of violent justice-involved women in prisons through more evaluation, programme development and research into the specific needs of this cohort.

9. Conclusion

The overclassification of justice-involved women leads to more intensive supervision and more frequent custodial terms and treatment, thereby adding stress onto already stretched resources within justice (Andrews & Bonta, 2010). In the jurisdictions discussed – Canada, UK, NZ, VIC, SA and WA – it is clear that gender-responsive strategies have been implemented in recent years, with work ongoing specifically in the treatment and management of justice-involved women. It is expected that evaluations of various treatment programmes and initiatives are either underway or planned in these jurisdictions, and such evaluations are expected not only to contribute to the global research, but also to ensure efficacy and subsequent ongoing funding.

There do not seem to have been any advances in the area of risk assessment specifically for justice-involved women in these jurisdictions, potentially due to inconsistencies in the international research. If the validity of tools such as the HCR–20, LS/RNR and VRS for justice-involved women is marginally improved by the addition of gender-specific factors (de Vogel et al., 2014; Stewart, 2011) and/or that the prioritisation for each of the central eight factors is different for females, this provides support for the development of a female-specific tool to determine risk, need and responsivity factors. Certainly, the central eight factors may not differ, but what may change is the prioritisation of these factors with greater focus given to peer and familial relationships, for example.

The VRS risk levels for justice-involved women are based on one study, and subsequent validations have not been found. Although approximately 11% of the sample was Aboriginal in the validation study conducted by Stewart (2011), it is not clear whether the VRS is a reliable predictor of violent offending in an Australian Aboriginal sample. Further, validations of female-specific programmes in WA and NZ including the CCC in WA, and Kimihia and Kowhiritanga in NZ, have not been conducted (or not been made publicly available) so the efficacy of these programmes for justice-involved women, and specifically Indigenous women, is not known.

Failing to consider female-specific issues in risk assessment, programme content and/or responsivity factors has an impact on accessibility and efficacy of treatment. Validations of tools with an Australian sample of justice-involved women are needed to ensure that the current suite of risk assessment tools accurately capture the risk and likelihood of reoffending (both general and violence), and specifically the treatment needs to be targeted if reductions in recidivism are to occur.

Appendix.

Table A1.

Overview of the risk assessment tools and interventions available for justice-involved women in various jurisdictions

Jurisdiction	Assessment tools	Criminogenic programmes/interventions	Target group
United Kingdom	Offender Group Reconviction Scale (OGRS3) Offender Assessment System (OASys) Violence Predictor (OVP) Risk of Serious Recidivism (RSR) Risk of Serious Harm (RoSH)	Choices, Actions, Relationships, Emotions (CARE)	Medium- and high-risk violent women
		Control of Violence for Angry Impulsive Drinkers – Group Secure Women (COVAID-GSW)	Violent females under influence of alcohol who are in secure units
Western Australia	Risk of Reoffending – Prison Version (RoR–PV) Level of Service/Risk, Need, Responsivity (LS/RNR) Violence Risk Scale (VRS) Stable 2007 (not scored – clinical information only)	Pathways (100 hr)	High-risk alcohol or drug problem
		Choice, Change and Consequences (CCC) (152.5 hr)	Medium- and high-risk general and violent women
		Moral Reconation Therapy (MRT) in Modified Therapeutic Community	Medium- and high-risk general plus medium- or high-risk specific to general alcohol and other drug (AOD)
Victoria	LS/RNR Historical Clinical Risk Management (HCR–20) Risk of Sexual Violence Protocol (RSVP)	Sea Change for Women (anger management)	Violent women (not family and domestic violence, FDV, perpetrators)
		Moderate Intensity Violence Intervention Program (MIVP)	Medium-risk females who perpetrate FDV
		High Intensity Violence Intervention Program (HIVP)	High-risk females who perpetrate FDV
South Australia	LS/RNR	Stepping Stones	Female offenders
		Psychological Skills Program	Female offenders
		Aboriginal Women’s Our Way My Choice	Aboriginal female offenders
Canada	Vary across Canada dependent on province/region, but there are 20 different risk assessment tools utilised across the country including the LS/RNR and VRS	Women’s Engagement Program (preparatory programme)	Female offenders
		Women Offenders Moderate Intensity Program	Female offenders
		Indigenous Women Offenders Moderate Intensity Program	Indigenous female offenders
		Women Offender High Intensity Program	Female offenders assessed as high risk
		Indigenous Women Offender High Intensity Program	Indigenous female offenders assessed as high risk
		Women Offender Self-management programme	Female sex offenders
		Women’s Sex Offender Program	Female sex offenders assessed as medium or high risk
		Women’s Modular Intervention	Female offenders living in secure units
New Zealand		Motivational programme	Medium- and high-risk females
		Intensive treatment programme – substance use	High-risk females with high AOD needs
		Kowhiritanga	Medium-risk females
		Kimihia	High-risk violent females with complex needs

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Ethical standards

Declaration of conflicts of interest

Menna Gower has declared no conflicts of interest

Frank Morgan has declared no conflicts of interest

Julie Saunders has declared no conflicts of interest

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee [Human Ethics Office of Research, RA/4/1/6720; and Research and Ethics Committee at the Department of Justice, Project ID 363] and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed consent

The waiver of informed consent was approved by the Human Research Ethics Committee (HREC) at the University of Western Australia as the data are retrospective and sourced from pre-existing databases, the research was classed as negligible risk, and confidentiality could not be upheld if informed consent was pursued. Contact with participants was not required at any point of the research.