Supporting formerly incarcerated people before and during COVID-19: is socially distanced (re)integration possible?

ABSTRACT Returning to the community after being incarcerated brings many challenges. In Victoria, Australia, a government-funded contract with non-government organisations (NGOs) allocates reintegration workers to assist with the post-release social integration process. In 2020, we interviewed reintegration workers to explore how they performed their roles before and during the COVID-19 restrictions. The key finding was that building rapport to tailor support was the most crucial aspect of practice, which workers could not adequately do without meeting face-to-face. Strengths-based practices, consisting of holistic, trauma-informed interactions, should become enshrined in reintegration job roles and the key performance indicators (KPIs) that measure success of the government contract.


Introduction
When COVID-19 was detected in Australia, not many realised the broad-reaching impact it would have for years to come. For the individuals who were eagerly awaiting their release from custody-looking forward to hugging their family members and (re)engaging with the community-the expectations of what that day would entail changed overnight. In this article we present our findings from interviews with reintegration workersthose tasked with the role of supporting formerly incarcerated people to adjust to life post prison. We explored how they were working with their clients to navigate reintegration during the government restrictions to 'stem the spread' of COVID-19. Government lockdowns in Victoria, Australia from 2020 to 2022 produced some of the most severe social restrictions in the world, such as nightly curfews, monetary fines for visiting relatives without permission and a two-hour limit on outdoor exercise required to be within 5 km of your home. These constraints impacted the reintegration sector of Victoria, and potentially, the reintegration process itself.
This Victorian-based Australian study presents the perspectives of reintegration workers who were employed by non-governmental organisations (NGOs). We conducted Victoria puts each person through a risk assessment prior to release and offers support to those who test as being at high-risk of re-offending. The objective of the Corrections Victoria Reintegration Pathway (CVRP) is to 'reduce risk to the Victorian community' by targeting seven domains 1 (Corrections Victoria, 2019). Since 2014, ReConnect has been the post-release arm of the CVRP, and at the time of this research, the contract was awarded to three NGOs. ReConnect can provide support for one month (targeted) or six months (extended). The duration of support is determined by Corrections Victoria-a decision which is informed by the in-custody risk assessment (Gelb, Liddell, & Martinovic, 2021). Reintegration workers have key performance indicators (KPIs) to fulfil-centred around the seven domains and reported to Corrections Victoria-yet there is no consequence for not meeting these KPIs. At the conclusion of the one/six-month period, workers can apply to extend support if they believe it necessary, and this application may be approved or denied by Corrections Victoria. Reintegration workers have previously expressed concern regarding the time limits placed on support, the rigidity of services provided and constraints on effectively advocating for clients due to disclosure requirements from Corrections Victoria (Carlton & Segrave, 2016;Franich, Sandy, & Stone, 2021).
The dominant model for offender 'rehabilitation' within Corrections Victoria (and most correctional settings in Australia and internationally) is the Risk-Need-Responsivity (RNR) framework (Ziv, 2019). RNR mandates and administers 'treatments' to target certain characteristics believed to predict re-offending (Andrews, Bonta, & Hoge, 1990). Prior research found reintegration workers felt the sector was lacking holistic, tailored, client-centred, trauma-informed approaches (Carlton & Segrave, 2016;Franich et al., 2021;Stone et al., 2017), characteristics which are absent from RNR. However, holistic, tailored, client-centred, trauma-informed practice are key elements of 'strengths-based' approaches, which are currently favoured by community service providers when working with disadvantaged cohorts (Askew et al., 2020;Collinson & Best, 2019). Strengths-based approaches originated from evidence-based case management practice and are characterised by an absence of coerced or mandated 'treatment' while fostering client agency-seen as a critical aspect when working with 'at-risk' populations (Tyler, Heffernan, & Fortune, 2020;Vandevelde et al., 2017). Strengths-based approaches shift the focus from 'treating' factors to target recidivism to a more holistic practice aimed at improving the client's experience of their own life-with the underlying theory that this will reduce re-offending (Ward & Brown, 2004).
While prior studies in Victoria documented how both workers and formerly incarcerated people found the service sector difficult to navigate (Carlton & Segrave, 2016;Franich et al., 2021), little is known about how workers manage these challenges, nor how practice has been altered as a result of COVID-19. It was the aim of this study to explore how workers at state-funded NGOs navigated the dual frameworks to practice before and during the COVID-19 restrictions.

COVID-19: implications on CJS practices
COVID-19 instigated significant social changes worldwide. In Victoria, governmentimposed social restrictions to stem the spread of COVID-19 between 2020 and 2022 were some of the most severe in the world. The restrictions implemented in institutions such as aged-care facilities, hospitals and prisons were particularly severe due to the close proximity of residents and the heightened risk of contracting the virus in closed environments (WHO, 2021). Corrections Victoria increased hygiene practices, mandated testing and masks and reduced the flow of people in and out of prisons as much as possible. All non-custodial staff and volunteers ceased on-site work, and all face-to-face visits were suspended and, where possible, replaced with video visits (Gelb et al., 2021). It should be noted that in 2023, while somewhat relaxed, all of these restrictions are still in place in Victoria-with the exception of the return of face-to-face visits.
Countries such as the United States reduced their prison populations with earlyrelease policies (Miller, Martin, & Topaz, 2022), but in Victoria, judicial practices had the greatest impact on reducing the prison population. Courts in Victoria moved to online hearings beginning in March 2020-for those sites with the technological capabilities (Battersby, 2021). As a result of the courts slowing down and more bail applications being approved, the rate of release for remandees went up while the intake of sentenced people went down (Greener, 2021). Between March and August 2020, Victoria's prison population saw the biggest month-to-month decline in the history of available data (Payne & Hanley, 2020). Although releasing people due to COVID-19 was not a policy, social changes at the time indirectly caused a significant reduction in the number of people incarcerated in Victoria (Greener, 2021). We believed these shifts in state and judicial practices were likely to have impacted the reintegration sector in Victoria, and this research aimed to explore that by interviewing reintegration workers. The research questions were: (1) What are the current definitions of social integration? (a) How does Corrections Victoria define social integration? (b) How do reintegration workers define social integration? (c) What are the ways in which conceptualisations of social integration might have changed since COVID-19? (2) What are the experiences of reintegration workers in providing reintegration support? (3) What are the ways service provision has changed since the COVID-19 restrictions?

Methods
This qualitative study was approached from the paradigm of critical realism, which contends reasons are the causes of actions (Archer, 2007). A core doctrine of critical realism is that human knowledge of reality only captures a fraction of reality (Fletcher, 2017). Critical realism posits an objective reality exists, but knowledge of this is moderated through conceptual frameworks (Danermark, Ekstrom, & Karlsson, 2019). In the context of this research, critical realism situates the perspectives of reintegration workers to be likely influenced by their personal experiences and the structures they work within. This does not mean each worker's insight is an objective depiction; however, each worker's perspective captures an aspect of the reality of reintegration work. Multiple viewpoints will capture a cross-section of social understandings regarding the postrelease integration of formerly incarcerated people and the impact of COVID-19 restrictions upon this process.

Data collection and sample
We recruited participants via email by contacting the three NGOs providing reintegration support in Victoria. Each NGO was sent a flyer explaining the study, which was then distributed by the organisation to reintegration workers. Workers then contacted us if they wished to participate. Two NGOs allowed staff to participate during work hours, and one requested participation occur outside of work hours. Participants chose a time and date that suited them, and either returned the signed participant information and consent form (PICF) or gave recorded verbal consent prior to the interview commencing. Due to the social-distancing restrictions in place at the time, these interviews were conducted remotely-either over the phone or Skype. Participants chose or were assigned a pseudonym. Once the interviews were transcribed, participants received the full transcript and made any desired edits before it was included in the data set.

In-depth interviews
We had eight respondents express interest, and researcher Kate Kennedy conducted interviews of 60-100 minutes between September and October 2020. In-depth interviews allowed participants to use their own language, perceptions and understandings to relay their experiences (Maynes, Pierce, & Laslett, 2008). Narrative interviewing techniques influenced our approach to in-depth interviews. In the context of narrative practices, an interview is seen as a collaborative creation produced by both interviewer and interviewee (Presser & Sandberg, 2017). Instead of being depicted as relaying social 'facts' to the interviewer, interviewees are viewed as purposefully conveying events and their own subjectivity-with the researcher's presence and influence viewed as an essential part of that dynamic (Gubrium & Holstein, 2012). We prioritised building rapport while actively engaging in conversation with participants, allowing them to guide the conversation to topics they saw as important.

Data analysis
We analysed the data using inductive, reflexive thematic analysis. Parallel data collection and analysis consisted of logging preliminary themes and field notes during interviews. This procedure meant initial themes acted as guides to navigate more extensive analysis (Tuckett, 2005). Once interviews and transcriptions were finalised, a thorough analysis of the complete data set was undertaken with data coding following the six-step procedure outlined by Braun and Clarke (2006). A second thorough analysis was conducted with the assistance of two theories: the exclusive society (Young, 1999) and the morphogenetic approach (Archer, 1995). This clarified themes around actuarial criminology and exerting agency within structures respectively. Our goal for analysis was not to pinpoint consensus between participants, nor to produce reliable and replicable coding (Braun & Clarke, 2021). Rather, our aim was to engage thoughtfully and reflexively with the data, and to generate rich and nuanced understandings of the concept under study.

Ethical considerations
The RMIT University College of Design and Social Context Human Ethics Advisory Network (CHEAN) approved this project (NHMRC Code: EC00237). Participants were advised that while confidentiality would be carefully upheld, anonymity could not be guaranteed as the selection pool was small and the topic quite specific. This was stated in the PICF, and participants reviewed and edited transcripts to minimise identifying information.

Limitations
The limitations were the small scale of the study, and the restriction on participation by one of the three NGOs working in the area. This means the cross-section of views was not as diverse as it could have been. It is also possible that those practitioners who volunteered to participate were those most committed to their practice.

Results
To understand what changes the sector endured as a result of the COVID-19 restrictions, it is first necessary to understand what practice was like prior to the pandemic. To this end, we spoke with participants about what aspects of their job they considered most crucial, and what a typical day was like before the government restrictions were introduced. Below we present the results derived from our interviews with participants on both their pre-and post-COVID practice, and this is followed by a discussion of these results. The themes generated from analysis were gaining trust, risk assessment, service referral, stigma, measuring and reporting success and changes to practice from COVID-19.
Gaining trust: 'Show your humanness' Five participants believed most of their clients were not 'integrated' with the community before incarceration, and all participants believed incarceration created further barriers to social integration: 'Being part of the collective and what we experience as society … I don't think many of our [clients] have really had the opportunity' (Patricia). Participants said it took time before clients felt comfortable accepting support: 'There are aspects of prison which make people reluctant to talk about weakness' (Damien). Viewing clients as marginalised influenced practice, and all participants regarded genuine rapport as critical in order 'to bring down some of those walls' (Steve). Five participants discussed prerelease work as vital: 'You really sort of establish that trust in there, and then it extends to the community' (Eugene). Participants said that conducting service brokerage to support post-release integration-their stated job role-was impossible to perform without first building trust and rapport.
Picking clients up at release was viewed as an important part of building that rapport: 'It's definitely one of my favourite days ever working with people; it's a very exciting time' (Freja). Six participants described the purposeful conversations they instigated with clients: 'The first thing we talk about is that basically, now the playing field is equal. "You and I are equals now. Your opinion's as valuable as mine"' (Damien). 'Get them all new clothes if that's what they want … "Who do you wanna be? … You're not that person anymore, you're out"' (Phoebe). David noted the first day is busy-reactivating bank accounts, registering with Centrelink and Medicare, going to VicRoads to check license status: 'You can then summarise back to them "Hey, what a bloody epic day that's been! You've done this, you've got that" … the hair on their arms start lifting up, like "Wow. I've done all that"' (David). All participants were conscious of interactions with their clients and the potential impact these had on rapport-building and client autonomy.
Outreach work was also discussed as important for rapport: 'They see that we actually care, that we actually are making an effort to go out and check up on them' (Steve). In building rapport, half of the participants we interviewed spoke about their car as a counselling space: 'It's a time when people feel the most relaxed … it's not like a clinical space, and you can have really powerful conversations with people' (Phoebe). Damien used this time to identify his client's needs: 'They're not feeling like you've got your eyes on them, and they can talk freely' (Damien). Outreach was a large part of participants' day-to-day work, and all participants viewed it as an important opportunity to connect with clients.
Participants discussed how some of the barriers to building rapport impacted practice. Freja and Damien highlighted the importance of trust: 'Sometimes they have had experiences where they've been let down, or there hasn't been follow through' (Freja). 'I just think they're so not used to people being genuine' (Damien). Freja and Damien overcame this by showing their personalities: 'Crack a joke, be friendly, show your humanness … the justice system is very dry, very impersonal. The humanness is really removed from it' (Freja). 'What I keep discovering works best is when I kind of step outside of the role' (Damien).
Another barrier identified by Phoebe was clients' relationship with the justice system. She commented: 'If you work in a prison, if you work with police, if you're parole, you're all part of the same system that incarcerated them' (Phoebe). Eugene and David overcame this by differentiating themselves from justice employees: 'I'll tell people straight up "Look, I don't work for Corrections Victoria … I'm purely here to help you"' (David). 'We sort of remind them along the way that we're not here to manage any compliance' (Eugene). Damien overcame this barrier by changing from the NGO's form-based assessment during pre-release visits to general conversation: 'Everybody seems to revel in the fact of … "It's not just another person asking me a whole lot of shit questions I've already answered; this is somebody who actually wants to hear what I've got to say"' (Damien).
Risk assessment: 'You do have to take a chance on someone' Participants described their work as influenced by the NGO's frameworks to practice, described as 'trauma-informed practice, narrative practices, and strengths-based practices' (Patricia) delivered with 'a person-centred approach' (Eugene). However, Patricia noted strengths-based practice can be challenging to implement: 'A lot of what happens in custody is risk-averse and compliance driven. And our funders are Corrections, so … [laughs]' (Patricia). Participants received an abridged risk assessment derived from RNR with each referral. This assessment was created by Corrections Victoria and is designed to predict re-offending by assessing risks and needs, however, participants had contrary opinions on its usefulness. Four participants related RNR to the justice system, disassociating it from their practice: 'It's not something that we incorporate a huge amount into our work' (Sandra).
Freja and Patricia looked at the summary version, while Phoebe and Damien ignored it. No worker mentioned accessing the full version: 'I can get on and look on the justice database and find the full version, but I don't really do that … things shift dramatically when people are released' (Patricia). 'It can be a bit dehumanising … to be hypersensitive of the risk that someone poses' (Eugene). Like Eugene, Damien and Phoebe believed their practice would be undermined by prioritising a risk-centric framework: 'Generally, I ignore it. Because I think if I fill my head too much with these ideas of risk, that's all I see' (Damien). 'I've just never let that narrative get anywhere near my work. It plays no role in anything I do' (Phoebe).
Five participants reported that in-custody risk assessments can be inaccurate. Eugene had a client who lived a solitary life but was repeatedly questioned about friends: 'The parole officer was very much focused on his companions, as that was highlighted in custody from that assessment tool as a major risk factor, when in reality, it just wasn't something that was relevant' (Eugene). Participants found parole officers did not give their professional opinion the same validity as an RNR assessment: 'If I was to say, "Oh no, I think that what you're talking about in terms of risk is wrong" they always challenge me and go "No, we know what we've got here"' (Damien). Patricia thought RNR assessments took precedence because they produce a definitive outcome: When we look at someone's needs, that can shift quite significantly in the post-release space … I think it's [RNR] flawed in practice, for sure. I get it, and I think you know, it's great-I think in terms of a measurable tool, sure, everyone likes that, you know? [laughs]. (Patricia) When asked about managing public safety while integrating high-risk clients, Patricia disclosed she found this challenging: Kate, I feel like that's one of, like, that's the hardest bit of my job … We don't want someone sleeping on the streets because that's not why any of us are working in this job, but we also have some sort of obligation to the community. (Patricia) Steve conceptualised this as having three clients, the individual, the program funder and the community: 'We have to kind of look at it from those three different points of views' (Steve). Damien explained he was initially hypervigilant about public safety, but after five years of practice he was now more concerned about the safety of his clients, giving the example of someone 'who was the subject of a story on [a popular current affairs program] and getting death threats. … He may have committed a crime, but this retribution and this sort of outrage is equally as horrible' (Damien). Sandra and David said they increased public safety by supporting people to feel genuinely integrated: 'One of the highest risks of re-offending is when someone feels alone, isolated, and feels that they don't fit in' (David). 'If we can do anything to build their stability, and help them integrate into society, as far as I'm concerned that's going to lower their risks to re-offending' (Sandra). Eugene noted that people cannot transform their lives unless given the space to do so, and felt focusing on risk did not allow for that: 'That's the whole conflict of risk; you do have to take a chance on someone to give them the opportunity to make a change' (Eugene).

Service referral and stigma: 'We don't do your kind of people'
Seven participants reported much of their time went to advocating for clients to have access to services, such as Sandra, who expressed frustration with a psychologist wanting to treat someone at a police station after reading about their risk level: 'This guy's finally opened up about what's going on inside his head and asking for support, and you're going to make him go to a police station, which is the least therapeutic space in the entire world?!' (Sandra).
Freja had trouble finding a doctor 'open to seeing' her clients: 'Even if they're not on suboxone or a methadone program, sometimes they don't want our [clients] sitting in the waiting room' (Freja). Participants described clients being excluded from hotels, rooming houses and caravan parks. Four participants said this created a dilemma regarding how much to reveal: 'I can't lie to save my life … So, what I would generally do is encourage people to apply on their own' (Damien). Phoebe asked housing services to make referrals to rooming houses after being told '"we don't do your kind of people" … most of the time it is the service providers that are the hardest when it comes to that-the stigma' (Phoebe).

Measuring and reporting success: 'That's not ever recorded'
All participants spoke about measuring and reporting successful integration. ReConnect is an outcome-based program, and standardised forms require tangible outcomes for clients. No participant reported any difficulty fulfilling the outcome for 'family/community connectedness': 'That can even be "talking to family" … Normally, in terms of the KPI and how that's measured, that's one of our easiest ones' (Patricia). However, participants did not feel the KPIs captured the process of integration: When you do work with somebody that you feel does become socially reintegrated it is one of the most amazing experiences, but it's also something that's not ever recorded; it's not something that anybody's looking for, it's not a KPI, it's just an experience that you feel personally with the person you're working with. (Phoebe) In addition to the practical outcomes, participants used personal measurements of social integration. Six participants assessed if clients felt safe in the community, such as: 'using public transport for the first time … those little steps are actually the big steps' (Steve). Five participants gauged successful integration on whether the individual felt supported by friends, family and services, such as: 'having a sense of connectedness and community, feeling valued, and also knowing how to participate and how to navigate society' (Patricia). Participants used the rapport they had previously built with clients to determine the presence or absence of these feelings and to gauge progression on this journey: 'It is 100% about your ability to build rapport with somebody, and creating a safe space for them to speak about those things' (Phoebe).
Rapport was also used to tailor interventions for clients. David used one client's love of coffee, taking him to cafés and choosing specific times and places to incrementally increase the number of people present: Until one day, it would've been about three months in, I got a random text message from him saying 'Hey David, guess where I am?' and he sent through a selfie of him in a busy CBD café, absolutely packed chock-full of people, holding his coffee with a big smile on his face. (David) Phoebe took a client for pedicures until she felt safe to continue going on her own: The first day we went for one … I thought she was just going to cry, sitting in the chair. She was shaking. … Afterwards, she couldn't believe how good she felt. … We went a second time, and you know, same experience, she was just really overwhelmed at the start, and then really comfortable. Then a few weeks later, she messaged me saying that she went there by herself-got a pedicure, manicure, and a haircut-and that now she's going to go there regularly. (Phoebe) In line with strengths-based practice, participants saw social integration as a process, with re-offending not necessarily jeopardising that process: Everybody has their own journey … as long as we can steer them back to that main road to take, that is successful in my eyes. And if we can also keep them out of prison, that's also a positive [laughs]. (Steve) 'We may view somebody's success in this role as staying out of prison, but … that's not necessarily the case' (Damien). Yet, participants were conscious that ultimately, only one measurement mattered: 'The justice idea [of social integration] would be them not offending' (Freja).

Changes to practice with COVID-19: 'A totally different type of support'
The Victorian government introduced measures to reduce the spread of COVID-19, including a mask mandate, ceasing community activities and strict limits on social gatherings (DHHS, 2020). Three participants stated the closure of community spaces had been particularly difficult. Sandra reflected it was 'challenging to actually integrate into society, because society's kind of shut down' (Sandra). The groups and activities participants would normally connect clients with became unavailable: 'We have been a bit more limited in the sort of support that we could offer, purely because there's a lot less things people can do in their local community' (Eugene). Seven participants thought the limits on social gatherings had negatively impacted their clients' integration: 'That's what gets them through-being surrounded by family and people and everything … enter[ing] straight back into a community of people who automatically accepted you' (Phoebe).
Due to the social restrictions, pre-release visits shifted to Zoom. Four participants said they could not build as strong a rapport over Zoom: 'You build that connection on a faceto-face interaction' (Eugene). 'All I'm seeing is a one-dimensional picture of somebody … the audio's bad, I'm self-conscious cos I can see my own picture on the screen' (Damien). In contrast, Patricia stated Zoom permitted more pre-release meetings than the usual one or two which occurred prior to COVID, which she thought resulted in a stronger rapport: 'Having a lot more ease to connect with people in custody has meant that we've been able to develop a stronger relationship' (Patricia).
There was strong consensus that losing outreach and release-day pick-up had negatively impacted rapport and subsequently practice. All participants stated interactions with clients had altered under restrictions: 'A big part of the job is actually going out to people and seeing people face-to-face, which is a totally different type of support than just over the phone' (Eugene). 'I am looking forward to the days where we can start doing a lot more face-to-face rather than over the phone and Zoom' (David). Participants could still perform outreach work, but they needed to apply for and receive approval-which took time to organise-so welfare checks were transferred to police, something participants speculated as potentially re-traumatising for their clients. Phoebe was permitted to pick one client up on their release day, but noted they both wore masks: 'It's hard to build a relationship with someone if you don't even really know what they look like' (Phoebe).
Participants also found making referrals during the social restrictions to be time-consuming and stressful: 'trying to talk to the housing service while also being on the phone to mental health, where you'd normally just go and sit with someone and do all of that together' (Patricia). 'Everything's "Oh yeah, just ring up the number and leave a message and we'll get back to you". And it's kind of, who knows?' (Damien).

Discussion
This section discusses the results of the thematic analysis, with the generation of themes guided by the two theoretical frameworks mentioned in the methods section. Jock Young's (1999) work on actuarial social practices in modernity guided understanding around themes related to 'risky' populations being excluded from society, while Margaret Archer's (1995) theory clarified themes relating to workers exerting agency within structures and the ways this changed practice.

Approach to practice: at risk vs risky
When reintegration workers accept a position at a state-funded NGO, they inherit two conceptual frameworks for practice: the strengths-based approach and the risk-based approach. The findings show how these frameworks were experienced by some participants as conflicting. The divergence of frameworks was epitomised in participant definitions of successful integration compared with their description of the 'justice definition'. Participants unanimously endorsed a narrative-desistance definition of successful integration, which states desistance from offending is a process facilitated by external changes in someone's life and how that person perceives and experiences those changes (McNeill, 2006). This was discussed in every interview, and all participants defined social integration as how someone feels. Participants differentiated their own definitions of successful social integration from that of Corrections Victoria, which they noted as defined by recidivism. Nhan, Bowen, and Polzer (2017) documented the same situation in Texas, whereby reintegration workers measured successful integration using a desistance paradigm unrelated to program KPIs or re-offending. This gives some indication of the complex terrain reintegration workers occupy, yet, despite the provision of funding from a state correctional institution, participants did not view their work as administering custodial rehabilitation, but rather, as supporting clients on personal journeys of desistance (McNeill, 2006;Villeneuve, Dufour, & Farrall, 2021).
Sublimating the RNR in favour of a strengths-based approach was influenced by participants' definitions of social integration, which stemmed from internal beliefs informed by professional experiences that their clients were a population to be viewed more as 'at risk' rather than 'risky'. This theme appeared in every interview, with all participants speaking about the complexities and vulnerabilities of clients. It is notable that five of the eight participants spoke only about risks to their clients, such as substance use, threats from the public and mental ill health. Rather than monitoring clients, participants worked to create a space to cultivate post-prison identities, sometimes referred to as assisted desistance in the de-labelling process (Villeneuve et al., 2021;Willis, 2018). Corrections Victoria, however, prioritised risk and public safety and tried to impose this framework on workers in various ways, such as delivering an RNR-based assessment at referral. Yet, in practice, participants experienced little difficulty in determining which paradigm to implement. This is because the strengths-based approach was encouraged by the NGOs and legitimised by workers' personal beliefs regarding what social integration consists of, and how best to support this process.

Risk and public safety: ontological tensions
Although participants reconciled competing frameworks to practice by subordinating risk-based practice in favour of strengths-based practice, the ways they understood and experienced this decision varied. Three participants (Steve, Sandra and David) acknowledged the goal of public safety might appear to conflict with integrating clients deemed high risk; however, they resolved this conflict through the belief that genuine integration would increase public safety. This seems to suggest that for them, implementation of a strengths-based approach would, by proxy, fulfil the objectives of a risk-based approach.
The remaining five participants experienced more tension from what they perceived as philosophical incompatibility between risk-based and strengths-based practices. For example, Eugene spoke about focusing on risk as dehumanising, Freja highlighted the importance of trust and a good relationship, while Damien and Phoebe consciously rejected the risk 'narrative' as a whole. This rejection of the risk 'narrative' is not surprising, given participant reports of risk classification hindering access to services. These perspectives also support the critique of RNR as hindering the development of client/worker trust and rapport by objectifying and pathologizing clients (Ward & Maruna, 2007). It appears that while some participants may have thought that delivering strengths-based practice would ultimately fulfil the aim of risk-based practice, other participants seemed to feel that risk-based practice impeded the delivery of strengths-based practice. Yet, while Eugene, Freja, Damien and Phoebe saw ontological conflict between the two frameworks, they did not seem to wrestle with the ethical implications of subverting one for the other. Patricia, however, found it challenging to implement a strengths-based framework with an adult forensic cohort. It may be worth noting Damien's experience over his years of practice and Patricia's shorter time in the role. Damien's thoughts on risk changed over time, which suggests increased experience affects understandings of risk and ontological conflict with the RNR framework. Damien's shifting perspective-which focused on client safety rather than client risk-echoes Rose's (1998) research on risk assessment/ management in psychiatry. Rose (1998) defined 'public safety' as protecting a particular population by incapacitating individuals from other populations, arguing this obscures the importance of reciprocity and disguises violence perpetrated by the so-called 'community'. Damien's change in focus from client risk to client vulnerabilities reveals reintegration workers' positions towards the risk paradigm are not fixed and can change over time and with experience. As a newer worker to the sector, Patricia may be observing these impacts by describing the RNR as 'flawed in practice'.

Lost connections: technical rapport
The arrival of COVID-19 altered the operation of almost all social structures and institutions, necessitating a substantive response that moved institutions from what Archer (1995) would describe as established stable practice to a state of spontaneous and unstable practice. The reintegration sector was no exception, as the pillars of practice participants held central to their work were destabilised and radically altered. The closure of community spaces and activities severed participants' ability to link clients in with their local community. Prior to the pandemic, participants had facilitated restorative practices to the best of their abilities: ensuring clients were linked with community groups, treated with dignity and respect and encountered as little stigma as possible. Although participants indicated their definition of social integration remained unchanged, it is clear that with the COVID-19 lockdowns, it became impossible for them to implement their previous practices. Community activities were non-existent, there was no way to utilise client strengths to contribute, and the narrative practices participants instituted with clients to facilitate self-worth and identity changes were no longer feasible. All of these were dependent upon face-to-face interactions as they drove together or walked through society side-by-side.
Participants felt the combination of losing pre-release visits, day-of-release pick-up and outreach had negatively impacted their relationships with clients and undermined the element of practice they considered most important. These findings support research regarding online service delivery emerging since COVID-19, such as Liberati et al. (2021) documenting mental health service users experiencing face-to-face delivery as superior, stating building on established relationships over Zoom was comparatively better than trying to create a relationship over Zoom. While it was too early to tell at the time of data collection, participants speculated that losing pre-release face-to-face meetings for establishing rapport and outreach to subsequently build upon rapport had impacted client (re)integration, which is consistent with literature evidencing the role of social relations in both constraining and enabling desistance (Weaver, 2015). All participants expressed passionate feelings of care and empathy for clients when reflecting on the importance of face-to-face interactions, and without this, participants lost the moments of practice they valued most, and which they believed were essential to assisting clients with social integration.
Damien stated what he finds works best is when he 'steps outside the role', however, these aspects-the 'in-between moments' that were seen as the most critical elements of practice-were not technically part of the job role at all. The quality of rapport is not a KPI. The depth of conversations is not reported. The celebrations that occur when someone catches a tram for the first time or has the confidence to get a coffee or go for a pedicure-these achievements were not recognised by the ReConnect funder, Corrections Victoria, as measures of success. Yet, every participant said it was these moments that embodied the process of (re)integration, and it was this aspect which gets lost by the RNR framework shaping the funder's outlook and KPIs.

Remote services: advocating from a distance
Participants felt the inability to see clients face-to-face had effectively hindered their capacity to deliver support in the ways they wanted. The creative methods of referral which had previously circumvented some of the stigma from service providers became impossible to implement. Participants felt the social restrictions had negatively impacted their ability to provide support, which they speculated had impacted client engagement and integration. As Young (1999, p. 22) argues, actuarial risk management combined with exclusion produces a 'spatially and socially segregated deviant other', and this process of labelling and segregation appeared to become compounded with pandemic restrictions. Prior to the pandemic, standard referral pathways were reported as ineffective due to stigma from service providers, yet referrals became even more challenging without the ability to meet face-to-face. The ways participants had previously used advocacy and unconventional service referral methods became unfeasible via telephone, effectively hindering their ability to, in Damien's words, 'step outside the role'. These results show that meaningful, face-to-face interactions are necessary in order to fulfil service brokerage and provide the client-centred, trauma-informed, holistic support that the NGOs and participants felt was so critical to reintegration work. In a post-COVID society where flexible working practices are becoming the 'new normal', it is imperative to preserve the face-to-face interactions in the reintegration sector.

Conclusion and implications
Reintegration workers articulated the practices they found most crucial, namely: building strong, genuine rapport based on trust; having purposeful conversations in casual settings; focusing on incrementally building clients' confidence and self-esteem; and connecting with clients through ongoing, meaningful conversations in order to determine how they felt during their personal journey of (re)integration. These aspects were not specifically part of their role, yet every worker interviewed explained these as central to practice. Service referral consisted more of advocacy, as workers attempted to obtain access to services for their clients without stigma, yet it seemed this goal was rarely achieved. While their job description was one of service brokerage and client success was determined by KPIs, it seems neither of these were an accurate representation of the work which is undertaken in the sector. The most valuable achievements, as Phoebe noted, were 'never recorded'. With the introduction of social restrictions to stem the spread of COVID, these interactions were not only never recorded, they simply no longer occurred. Gone were the car rides used to identify needs, the face-to-face pre-release conversations which established all-important trust and the day-of-release pick up which instilled a sense of self-worth and autonomy. Reintegration workers were stripped back to the bare minimum of their job description-to what was recorded on paper. The implications of these findings are far reaching. Future research is needed to establish whether these practices are consistent throughout the reintegration sector, and whether clients experience this approach to be as beneficial as reintegration workers believe. Research should also be conducted on why RNR appears to dominate correctional practice when Corrections Victoria's Offender Management Framework states RNR is used in conjunction with the Good Lives Model (GLM) and therapeutic jurisprudence. The aspects of practice reintegration workers deemed most crucial were much more in line with GLM and therapeutic jurisprudence. These aspects must become enshrined in the job role and KPIs for the ReConnect program.