Post-traumatic stress disorder and hazardous alcohol use among incarcerated adults: associations with criminogenic needs and reoffending in New South Wales, Australia Open Access

Adults involved in the criminal justice system show strikingly high rates of trauma and post-traumatic stress disorder (PTSD). PTSD can be a debilitating mental health condition characterised by intrusive re-experiencing of trauma, avoidance of trauma-related stimuli, negative alterations in cognition and mood and heightened arousal and reactivity, all of which can lead to significant functional impairment (American Psychiatric Association, 2013). Compared to the lifetime prevalence of PTSD of 1.2% for males and 2.7% for females in the general population, the lifetime prevalence of PTSD in prisoners globally is estimated to be 18% and 40% in male and female inmates, respectively (Baranyi et al., 2018). In Australia, 21% of incarcerated adults were found to meet diagnostic criteria for PTSD (Butler et al., 2005). Rates of trauma exposure and mental disorders are even higher among First Nations people in custody in Australia, due to the structural discrimination, colonisation, racism and intergenerational trauma experienced (Rose et al., 2019; Marr et al., 2025). Further, once incarcerated, 89–97% will experience one or more potentially traumatic events during incarceration, significantly increasing the likelihood of developing PTSD (Piper and Berle, 2019).

A growing body of research highlights the complex relationship between trauma, PTSD and criminal behaviour. While most individuals with PTSD do not engage in criminal behaviour, population-based studies have found that PTSD is associated with an elevated risk of later justice involvement, even after accounting for shared familial and psychiatric factors (Paulino et al., 2023). Both trauma and PTSD can affect the stress-response, memory and reward systems, leading to alterations in emotional regulation and reactivity, threat perception, blunted response to natural rewards, enhanced rewarding effects of psychoactive substances, difficulty with inhibiting impulses and behaviours and in decision-making (Lovallo, 2013; Yau and Potenza, 2013; Puetz and McCrory, 2015; Svingen, 2023). These neurobiological and psychological changes may engender a cascade of effects that increase vulnerability to a variety of behavioural and social challenges, including contact with the criminal justice system (Ardino, 2012). Importantly, these outcomes are shaped not only by individual experiences but also by broader structural factors, including social disadvantage, systemic inequities and limited access to specific trauma-informed care. Recognising these intersecting and overlapping pathways are essential for developing compassionate, evidence-based interventions that support recovery and reduce justice involvement.

Despite the high prevalence in prison populations, the impact of PTSD on recidivism remains unclear. Few studies have longitudinally investigated whether PTSD influences the risk of reoffending post-release from custody (Facer-Irwin et al., 2019). Research examining the effects of trauma, and particularly PTSD, on reoffending behaviours is sparse and often limited to smaller samples with specific criminal justice and demographic characteristics, limiting generalisability (Facer-Irwin et al., 2019). However, among these selected samples, PTSD has been found to be associated with repeat offending. For example, in a sample of men attending an interpersonal violence treatment program, higher scores of PTSD and substance use measures were found to be associated with greater general violence recidivism (Miles-McLean et al., 2019). Likewise, among those experiencing a mental disorder, PTSD increased the likelihood of arrest and serious recidivism (Sadeh and McNiel, 2014).

It is also unclear whether PTSD might be directly related to reoffending, or whether known co-occurrence with substance use might account for this relationship. It is well-known that substance use, particularly alcohol use, frequently co-occurs with PTSD and criminal behaviour (Mills et al., 2006; Boden et al., 2013; Debell et al., 2014). Individuals with PTSD may use alcohol as a coping mechanism to manage symptoms such as hyperarousal, intrusive memories and emotional distress (Hawn et al., 2020). However, alcohol use can impair judgement, increase impulsivity and contribute to aggressive or risky behaviours, potentially heightening the risk of reoffending, particularly among people with co-occurring PTSD (Barrett et al., 2014). Hazardous alcohol use is independently associated with offending, especially violent and impulsive behaviours (Felson and Staff, 2010; Boden et al., 2012; Boden et al., 2013; Lawler et al., 2022).

Beyond direct behavioural effects, PTSD and hazardous alcohol use may also influence offending indirectly by exacerbating other, well-established risk factors for reoffending, known as criminogenic needs. Criminogenic needs are dynamic risk factors that contribute to criminal behaviour, across domains such as substance use, emotional functioning, social relationships, attitudes towards crime and history of antisocial behaviour (Bonta and Andrews, 2007). Both PTSD and alcohol use problems have been associated with other criminogenic needs, including poorer health, housing instability, educational and employment difficulties and social disadvantage (Zielinski, 2009; Hardcastle et al., 2018; Zielinski et al., 2024). Thus, there are a myriad of factors that co-occur with PTSD that can increase the risk of reoffending. Understanding these relationships is critical for tailoring rehabilitation and wellbeing programs and improving outcomes for individuals in custody.

The relationships between PTSD, alcohol use, criminogenic needs and reoffending among those involved in the NSW criminal justice system remain unclear. Therefore, this study aims to (1) determine the prevalence of trauma exposure, PTSD, hazardous alcohol use and the co-occurrence of PTSD and hazardous alcohol use within individuals in custody in NSW, using the 2015 Network Patient Health Survey; and (2) examine the association between PTSD, harmful alcohol use and their interaction with criminogenic needs and reoffending.

This study utilised a linked data set comprising health, correctional and reoffending data for a representative sample of adults in custody in New South Wales (NSW), Australia. The data set was created through the linkage of three primary sources: the 2015 Network Patient Health Survey (NPHS) administered by Justice Health and Forensic Mental Health Network (Justice Health NSW), the Offender Information Management System (OIMS) maintained by Corrective Services NSW (CSNSW) and the Reoffending Database (ROD) held by the NSW Bureau of Crime Statistics and Research (BOCSAR). Ethical approval for the study was obtained from the Justice Health Human Research Ethics Committee (2020/ETH03134), CSNSW Ethics Committee (D2021/1240552) and the Aboriginal Health and Medical Research Council Human Research Ethics Committee (1775 / 21).

The cohort comprised a purposive sample of 1,131 individuals drawn from the 2015 NPHS, which employed stratified random sampling to recruit people in custody across NSW. Of the 1,493 individuals invited to participate in the 2015 NPHS, 76% consented, and 1,033 participants provided consent for their data to be used in future data linkage studies. The sampling strategy of the 2015 NPHS intentionally oversampled women and Aboriginal and Torres Strait Islander peoples to ensure adequate representation of these key demographic groups. Further details on the sampling methodology and cohort characteristics are available elsewhere (Justice Health and Forensic Mental Health Network, 2015a; Justice Health and Forensic Mental Health Network, 2015b).

The NPHS is a comprehensive, cross-sectional survey conducted periodically since 1996 to assess the health status of people in custody in NSW. The 2015 iteration included face-to-face interviews with 1,131 participants, through which interviewers collected data on physical and mental health, health behaviours and service use. For this study, only data from participants who consented to linkage (n = 1,033) were included. Variables extracted included demographic characteristics, trauma history, PTSD symptoms and alcohol use (see measures section).

OIMS is the central administrative system used by CSNSW to manage individuals in custody or under community supervision. It supports service continuity across custodial and community settings, particularly for individuals with complex needs. For this study, data extracted from OIMS included Level of Service Inventory-Revised (LSI-R) scores and subdomains, and custodial episode and release dates.

The ROD is a longitudinal database maintained by BOCSAR that contains court outcome data for all individuals convicted of criminal offences in NSW since 1994. It includes detailed records of court appearances across all court levels, offence types, penalties and reoffending events. For this study, data were extracted on offence type, date and outcome.

Participants were asked whether they had ever experienced, witnessed or were confronted with a traumatic event in their life. Those who responded “Yes” were then asked to indicate which, from a list of 13 traumatic events (e.g. direct combat experience in a war, a life-threatening accident, fire, flood or other natural disaster) they had experienced. They could also indicate that they did not wish to disclose the type of event, or that they had experienced another trauma type not mentioned.

Participants who endorsed experiencing a traumatic event completed the PTSD-8 (Hansen et al., 2010) to assess meeting current (past two weeks) PTSD diagnosis. The PTSD-8 comprises of eight items asking participants the extent to which they have been impacted by PTSD symptoms across three symptom clusters; intrusion (e.g., “recurrent thoughts or memories of the event”), avoidance (e.g., “avoiding activities that remind you of the event”) and hypervigilance (e.g., “feeling jumpy, easily startled” and “feeling on guard”), in the past two weeks. Responses are scored on a 4-point Likert scale ranging from, 0, “not at all” to 4, “most of the time”. Participants who endorsed exposure to one or more traumatic events and met the criteria for all three symptom clusters based on symptoms experienced in the two weeks prior to being interviewed were coded as meeting current PTSD diagnosis.

Participants’ alcohol use was assessed through the Alcohol Use Disorders Identification Test (AUDIT) (Babor et al., 2001). The AUDIT is a ten-item screening tool developed to assess current alcohol consumption, drinking behaviours and alcohol-related problems. As alcohol consumption in NSW correctional centres is prohibited, items were amended to ask participants to respond to the items in relation to their alcohol use in the 12 months prior to incarceration. Items are scored on a five-point Likert scale ranging from 0 to 4 with higher scores indicating more harmful alcohol use. Scores were summed and a binary variable created indicating the presence of hazardous or dependent alcohol use for those scoring eight or more, as per established thresholds (Saunders et al., 1993; Conigrave et al., 1995). Participants who responded that they had not consumed alcohol in the year prior to incarceration were not asked the remaining AUDIT questions but were included as not meeting criteria for hazardous or dependent alcohol use in the 12-months prior to incarceration.

Participants’ risk of reoffending and criminogenic needs was assessed through the LSI-R (Andrews and Bonta, 2000). The LSI-R is an actuarial risk assessment tool comprised of 54 items that can be summed to predict recidivism risk, with higher scores indicating higher risk for reoffence. Items are grouped into 10 subscales: Criminal History, Education/Employment, Finances, Family/Marital, Accommodations, Leisure/Recreation, Companions, Alcohol/Drug, Emotional/Personal and Attitude/Orientation.

Reoffending outcomes were derived from linked administrative records and defined using offence dates and Australian and New Zealand Standard Offence Classification (ANZSOC) codes. Reoffending was defined as the occurrence of any proven offence, with an offence date falling within the first 12 months of an individual’s “free time” following release from custody. Free time was calculated from the end date of the index custodial episode, as recorded in court finalisation data. Participants were coded as having reoffended if the offence date occurred between 0 and 364 days (i.e. less than 12 months) after their custodial end date and coded as having not reoffended if their first reoffence was after 364 days or if they had no reoffence charges post-release. Participants who had not been released from custody or had less than 12 months since release at the time of analysis were excluded, as they had not accrued a full 12-month period at liberty in which to reoffend. Only charges where participants were found guilty (i.e. excluding charges where the outcome was missing, not guilty, a mental health dismissal, withdrawn, or otherwise disposed of) and originally discharged because of bail, a good behaviour bond, a parole order, release to immigration or intensive correctional order or sentence expiry were retained. Proven charges for 151 breach custodial orders or 152 breach of community-based orders were excluded. This left 907 people in our subsample for our re-offending analyses. Further, to examine serious instances of reoffending, a measure of violent reoffending was created which was restricted to offences involving physical violence or threat thereof, including homicide, assault, robbery and selected offences against the person (ANZSOC codes 111–132, 211–213, 291, 311–312, 511–532, 611–612).

Means and standard deviations (continuous variables) and counts and percentages (categorical variables) were computed to obtain descriptive sample statistics. Group differences in demographic characteristics between those with current PTSD only vs current PTSD and hazardous alcohol use were examined using chi-square tests for categorical variables and one-way ANOVAs for continuous variables.

A series of two-way ANOVAs were conducted to examine the main effects of current PTSD and hazardous alcohol use on LSI-R scores, their interaction and repeated to examine these effects on LSI-R criminogenic needs subscales. These predictors were treated as binary variables and were not mutually exclusive, allowing for the assessment of their independent contributions to each outcome.

We used logistic regression models to examine the relationship between current PTSD, hazardous or dependent drinking in the 12 months before prison and their interaction in predicting both any and violent reoffending, controlling for age, sex and LSI-R criminal history score. This subscale assesses an offender’s past criminal behaviour, incorporating factors such as the number and type of prior offences. Survival analyses were run using a Cox regression model to examine whether current PTSD, hazardous or dependent drinking in the 12 months before prison and the interaction between the two were associated with time until any reoffence and again with time until violent reoffence, controlling for age, sex and LSI-R criminal history score. For people who reoffended, this survival period was the number of days between release and reoffending date. For those who did not reoffend, it reflects the number of days between release and the data censoring date, in this case the data extraction date of 30 Nov 2021.

In regression models, age and LSI-R criminal history scores were centred by subtracting their respective means. Due to a non-linear relationship identified via generalised additive modelling (edf ≈ 3.78), LSI-R criminal history score was log-transformed (log(score + 1)) to approximate linearity with the log-odds, then centred by subtracting the mean of the log-transformed values. The model was fitted using the glm function in R with a binomial family and logit link, with complete cases analysed after excluding missing data.

Our total sample comprised of 1,033 adult participants incarcerated across NSW correctional centres. As shown in Table 1, two thirds were male (66.9%, n = 691), and the average age of participants was 35.57 (SD = 11.60). In their lifetime, over two thirds of participants had experienced, witnessed or been confronted with a traumatic event (n = 662, 66.9%). Of these, roughly 13% (n = 88) did not wish to disclose further information. Of the remaining 574 participants who had experienced a traumatic event, the most commonly experienced type of trauma was witnessing someone being badly injured or killed, experienced by about half of participants (49.9%, n = 287, Table 1). The most common responses to “other” traumatic event included the death of a close family member or friend (73, 11.0%), including suicide of a close friend, family member or child (13, 2%) and death of a child (not by suicide; 16, 2.4%). These experiences were severe, often including witnessing the death or finding the person after death. On average, participants experienced 2.13 (SD = 1.71) types of traumatic events. Nearly half of the sample (428, 43.8%) reported hazardous or dependent drinking in the 12 months prior to prison. Similarly, among participants who met current PTSD criteria, almost half (46, 46.0%) reported hazardous or dependent drinking in the 12 months prior to prison.

Key demographic characteristics did not differ between those meeting criteria for current PTSD only compared to those meeting criteria for current PTSD and hazardous alcohol use in the 12-months prior to incarceration, except for sex, χ²(1) = 4.40, p = 0.036, with a higher proportion of males in the PTSD and hazardous alcohol use group (67.4%) compared to the PTSD-only group (44.4%). No significant associations were found for age, Aboriginal and Torres Strait Islander status, country of birth, language spoken at home, first detention age, current custodial status, educational qualifications, employment status, juvenile justice history, the number of times they had been incarcerated or the age at which they were first incarcerated.

Both current PTSD, F(1, 754) = 12.88, p < 0.001 and hazardous alcohol use, F(1, 754) = 9.28, p = 0.002, were associated with total LSI-R scores. However, the interaction between PTSD and hazardous alcohol use was not significant, F(1, 754) = 0.001, p = 0.971, suggesting no combined effect of these variables on LSIR scores.

For LSI-R criminogenic subscales, current PTSD was significantly associated with higher scores in Criminal History (F(1,774) = 4.78, p = 0.029), Education/Employment (F(1,774) = 5.40, p = 0.020), Finance (F(1,773) = 5.58, p = 0.019), Family/Marital (F(1,774) = 4.82, p = 0.029), Accommodation (F(1,774) = 13.23, p < 0.001), Substance Use (F(1,774) = 7.41, p = 0.007) and Emotional/Personal (F(1,774) = 23.63, p < 0.001), compared to no PTSD. Hazardous alcohol use was significantly associated with higher scores on Education/Employment (F(1,774) = 6.38, p = 0.012) and Substance Use (F(1,774) = 92.87, p < 0.001), but not with other domains. No significant effects were found for either predictor on Leisure, Companionship or Attitudes scores, nor were there any significant interactions between current PTSD and hazardous alcohol use on any LSI-R subscales.

Results of the two logistic regression models examining the relationship between current PTSD, hazardous or dependent drinking in the 12 months before prison and their interaction on reoffending, controlling for age, sex and LSI-R criminal history score, are shown in Table 2. Age and criminal history score were significantly associated with reoffending within 12-months post-release. Younger individuals and those with greater criminal history scores had higher risk of both any and violent reoffending. Hazardous alcohol use was associated with greater odds of violent reoffending only. Other variables, including PTSD, sex and the interaction between PTSD and hazardous alcohol use, were not significant in either model.

In both Cox proportional hazards models, age and LSI-R criminal history score emerged as strong and statistically significant predictors of reoffending and violent reoffending (Table 3). Specifically, each additional year of age was associated with a reduced hazard of reoffending (any: HR = 0.955, p < 0.001; violent: HR = 0.954, p < 0.001), indicating that older individuals reoffended more slowly. In contrast, higher LSI-R criminal history scores substantially increased the hazard (any: HR = 4.217, p < 0.001; violent: HR = 2.982, p < 0.001), suggesting that individuals with more extensive criminal histories reoffended more quickly. For violent reoffending, hazardous alcohol use was associated with a significantly higher hazard (HR = 1.380, p = 0.007), indicating a shorter time to re-offend among those exhibiting hazardous alcohol use in the 12-months prior to incarceration. Neither PTSD nor its interaction with hazardous alcohol use significantly predicted reoffending in either model.

This study provides a comprehensive examination of the prevalence and impact of PTSD among incarcerated individuals in NSW. Consistent with prior research, we found high rates of trauma exposure, PTSD and hazardous alcohol use, with nearly half of those meeting PTSD criteria also reporting hazardous alcohol use prior to incarceration. Importantly, both PTSD and hazardous alcohol use were associated with elevated criminogenic needs, particularly in domains related to education/employment and substance use, with PTSD further associated with elevated criminogenic needs on criminal history, finance, family/marital, accommodation and emotional/personal domains. However, these elevated risk profiles did not directly translate into increased likelihood or faster time to reoffend within 12-months post-release. Younger age and higher criminal history scores were robust predictors of both any and violent reoffending, with individuals in these groups reoffending more quickly following release. Hazardous alcohol use was associated with greater odds of, and a shorter time to violent reoffending. We did not find evidence of any association between PTSD and reoffending, nor any evidence that associations between PTSD and criminogenic needs or reoffending were dependent on alcohol use.

Over two thirds of participants in our sample endorsed exposure to at least one traumatic event and of those, 16% met the criteria for current (past two-week) PTSD diagnosis. This was generally higher than estimates from meta-analyses and umbrella reviews globally, which found the prevalence of current (past six-month) PTSD in prisoners to be 10% (Favril et al., 2024), or 6% in male prisoners and 21% in female prisoners, respectively (Baranyi et al., 2018). Demographic comparisons revealed few differences between those meeting criteria for current PTSD alone vs current PTSD and hazardous alcohol use, with the exception of sex whereby a higher proportion of males were in the PTSD and hazardous alcohol use group compared to the PTSD-only group. Notably, 44% of participants reported hazardous alcohol use in the 12-months leading up to custody, far exceeding general Australian population rates of 22% (O'Brien et al., 2020). Of concern, 28% of the sample met criteria for alcohol dependence. Global meta-analytic estimates of alcohol use disorder have found a prevalence of 24%, although this was based on clinical examination or interviews using validated diagnostic instruments (i.e. excluding self-screening measures like the AUDIT) (Fazel et al., 2017). Taken together, and contrasted with estimates in general populations, these findings underscore the pervasiveness of trauma, PTSD and hazardous alcohol use in incarcerated populations, highlighting ongoing need for consideration and treatment of trauma and alcohol use problems in forensic settings. Interventions such as Seeking Safety are promising avenues for addressing co-occurring PTSD and substance use in custodial settings (Barrett et al., 2015).

Of primary interest to the current study, we investigated whether PTSD or hazardous alcohol use significantly impacted the likelihood of any reoffending or violent reoffending. We did not find evidence that current PTSD was associated with any re-offending or violent reoffending either within a year following release or time-to-event analysis. This was unexpected given PTSD was associated with significantly higher LSI-R scores, suggesting elevated criminogenic risk. In contrast, hazardous alcohol use was associated with greater odds of violent reoffending and a shorter time to violent reoffending post-release. These findings suggest that while PTSD may contribute to broader criminogenic needs, it may not directly predict reoffending behaviour in the short term, over and above what is accounted for by actuarial proxies of risk (LSI-R criminal history score) included in the models. Conversely, hazardous alcohol use appears to be a more proximal risk factor for violent reoffending. It is possible that protective factors not captured in the current data set, such as stable housing, social support or engagement in treatment, may buffer the impact of PTSD on reoffending (Baldry et al., 2006). In addition, the period to capture re-offending was 12-months, which may have not been enough time to capture re-offending behaviour, and both PTSD and alcohol use may have fluctuated between the time of assessment and release, due to treatment or other unmeasured factors, potentially influencing outcomes. Overall, these findings highlight the importance of targeting alcohol use in reoffending prevention and suggest that PTSD may confer risk for other criminogenic needs which may be the ultimate target to reduce reoffending behaviours.

In our sample, age and criminal history scores were found to be significant predictors of reoffending behaviours. This aligns with previous research highlighting age as a significant risk factor for criminogenic risk (Piquero et al., 2013; Spruit et al., 2017; Yukhnenko et al., 2020). Likewise, prior history of justice involvement has consistently been one of the strongest and most reliable predictors of future offending, reflecting entrenched behavioural patterns and systemic factors that may perpetuate cycles of incarceration (Weatherburn and Ramsay, 2018). These findings underscore the importance of considering stable demographic and historical factors in risk assessment and intervention, while also recognising that dynamic needs and contextual influences may shape individual trajectories in more nuanced ways. To reduce reoffending and future criminal justice contact among people in prison, additional resources for interventions should be targeted to younger offenders with more extensive criminal histories, such as evidence-based programs built on risk-needs-responsivity principles (Pooley, 2020).

Given the high prevalence of PTSD in prison populations, it is vital to support the development and implementation of trauma-aware treatment and rehabilitation programs. Even in the absence of a direct association between PTSD and reoffending, the standard of care provided in custodial settings should match that available in the community, and thus individuals in custody must receive appropriate mental health care. Addressing PTSD is essential for improving overall wellbeing and reducing the broader health and social disparities faced by justice-involved populations. It is also critical to recognise that PTSD symptoms such as hypervigilance, emotional numbing and intrusive re-experiencing can significantly impair an individual’s ability to engage with standard rehabilitation programs (Miles-McLean et al., 2019). These symptoms can interfere with learning, interpersonal trust and the ability to reflect on behaviour, which are core components of many rehabilitative interventions. When programs are not trauma-informed, they may inadvertently increase the risk of re-traumatisation, especially if they rely on confrontational methods or fail to create a psychologically safe environment. In contrast, trauma-informed approaches are designed to foster safety, empowerment and emotional regulation, and have been shown to be particularly beneficial for individuals with more severe trauma histories (Messina and Schepps, 2021). Emerging evidence also suggests that such programs may reduce PTSD symptoms and improve engagement, although further research is needed to establish their long-term effectiveness (Malik et al., 2021). Several encouraging initiatives are currently being implemented within Corrective Services NSW, including staff training in trauma-informed practice and the Victim Support Scheme, which provides counselling to victims of crime and abuse in prisons and the community, though more work is needed greater access required.

Moreover, in the Australian context, it is essential to recognise the ongoing impacts of colonisation and intergenerational trauma experienced by Aboriginal and Torres Strait Islander peoples, which contribute to their persistent overrepresentation in the criminal justice system. These systemic issues are compounded by structural racism, over-policing and social disadvantage, which not only increase the likelihood of justice involvement but also shape the nature and severity of trauma experienced. Future research and program development must prioritise culturally safe, trauma-aware practices that are responsive to the unique histories, strengths and needs of First Nations communities and actively work to dismantle the systemic barriers that perpetuate cycles of trauma and incarceration. Research has shown that culturally appropriate interventions, such as those delivered through Indigenous-led mental health services, can lead to improved outcomes for First Nations people (Dudgeon et al., 2014). One example is Queensland’s Indigenous Mental Health Intervention Program (IMHIP), a First Nations-led, multidisciplinary initiative that implements a culturally informed model of care in prison, identifying mental health and wellbeing needs beyond Western conceptualisations and assisting individuals in their transition back to community life (Dale et al., 2023). To ensure the success and sustainability of such programs, adequate funding and resourcing are essential. Reducing the number of First Nations people with mental health conditions, including PTSD and hazardous alcohol use, in prison must be a national priority, involving diverting individuals with mental illness away from the justice system, providing adequate and accessible community-based treatment and ensuring that mental health care in the community is culturally informed, safe and accessible.

There are several strengths to the current study. Firstly, it benefits from a large and purposively sampled population, stratified to reflect key demographics of the NSW prison population. In addition, the use of data linkage with administrative databases reduces reliance on self-report, thereby minimising bias and providing a more accurate and objective measure of recidivism. However, findings should be interpreted in light of several limitations. Although the sample was purposively recruited, results may not be generalisable to prison populations beyond NSW. Future research should aim to replicate these findings in diverse cohorts. Second, analyses were limited to reoffending within the first 12 months following release. While this period is critical for re-entry, given that approximately 40% of released adults in NSW reoffend within this timeframe (NSW Bureau of Crime Statistics and Research, 2025), it may not capture longer-term patterns of recidivism or delayed reoffending. In addition, hazardous alcohol use in the current study refers to a 12-month period prior to custody and may not reflect alcohol use patterns during our time period for reoffending. Furthermore, the measures used to assess trauma and PTSD, whereas brief and practical, have not been formally validated in prison populations, which may affect the accuracy of prevalence estimates. Importantly, the health data used to assess PTSD and alcohol use were collected in 2015. While these measures remain relevant, changes in prison health services, substance use trends and broader social contexts may limit the generalisability of findings to current-day populations. Future studies should consider more recent data to validate and extend these findings. Furthermore, the study did not include information on access to or engagement with treatment services (e.g. mental health or substance use interventions) during or after incarceration. This limits our ability to account for the potential mitigating effects of treatment on reoffending outcomes and criminogenic needs. Finally, although a substantial proportion of participants (43.5%) identified as Aboriginal and Torres Strait Islander, the measures used may not have adequately captured the complexity of intergenerational, systemic and ongoing trauma experienced by First Nations peoples. This highlights the need for culturally responsive tools and approaches that reflect the lived and living experiences of First Nations communities.

This study contributes to a growing body of evidence on the prevalence and impact of PTSD within incarcerated populations, highlighting its association with criminogenic needs but not with short-term reoffending. In contrast, hazardous alcohol use emerged as a significant predictor of violent reoffending, both in terms of increased odds and shorter time to reoffence. The findings underscore the importance of trauma-informed rehabilitation programs that are responsive to the psychological and emotional needs of individuals in custody, but also to substance use issues that may elevate risk. Future research should continue to explore the long-term effects of alcohol use and PTSD on justice outcomes, validate assessment tools for use in custodial settings and evaluate the effectiveness of integrated, trauma- and substance-informed interventions in reducing recidivism and promoting long-term recovery.

The authors wish to thank Justice Health and Forensic Mental Health Network staff, Corrective Services NSW staff, and NSW Bureau of Crime Statistics and Research (BOCSAR) staff for their involvement in data collection as part of this research project.