What the Evidence Tells Us About Safer Reintegration

This is a sample post published to demonstrate the Bond Blog layout. It will be replaced or edited before launch.

Reintegration works when the conditions for change are present. That is the through line of the desistance literature, of restorative practice, and of the Circles of Support and Accountability (CoSA) model that Bond Trust grew from in 2014. People leaving prison do not change because they are watched more closely. They change when they are connected, when the expectations on them are clear, and when the practical scaffolding of daily life holds long enough for a new identity to take root.

The five conditions below are the ones our practice, our training programme and the discussions at Te Ara Oranga keep returning to. Each one is written as something to act on, not simply something to agree with.

1. Build the relationship before the gate, not after it

The first weeks after release carry the heaviest load: appointments, housing, identification, money, whānau expectations, and the shock of an environment that has moved on without you. A relationship formed in that window starts from a standing start. A relationship formed months earlier, inside, is already load bearing when it is needed most. This is why Fellowship Support Network volunteers begin meeting people in the later stages of a sentence rather than at the gate.

• Start pre-release contact early enough that trust is established, not assumed.
• Carry the same person or people through the release transition wherever possible.
• Plan the practical items, accommodation, ID, transport, health, before release day, not during it.

2. Treat accountability as a feature of relationship

Accountability and support are often framed as opposites, with monitoring on one side and encouragement on the other. The CoSA model treats them as the same act: people are more willing to be honest about risk with someone who has shown up consistently and who has no interest in either rescuing them or catching them out. Accountability grounded in relationship surfaces problems early, while they are still small enough to work with.

• Name expectations plainly at the start of a support relationship.
• Respond to disclosure with curiosity before consequence.
• Keep boundaries visible, so that support never depends on things going well.

3. Strengthen protective factors rather than auditing risk

Risk assessment tells you what to watch. It does not tell you what to build. The protective factors that show up repeatedly in reintegration research are unglamorous and cumulative: routine, stable housing, emotional regulation, purposeful activity, whānau connection, a prosocial identity, and a sense that the future is worth planning for. Every one of them can be worked on directly.

• Choose one or two protective factors to strengthen deliberately, rather than tracking all of them loosely.
• Make routine and belonging explicit goals, not by-products.
• Measure progress by what has been built, not only by what has been avoided.

4. Work in a trauma-informed and dignity-forward way

Much of what looks like non-compliance is a survival strategy that made sense somewhere else. Institutionalisation, shame and disrupted identity shape how people receive help, how they answer questions, and how quickly they withdraw. Trauma-informed practice reduces shame, which in turn increases engagement, and engagement is what makes safer decisions possible. Bond's training programme puts it as curiosity over judgement, boundaries over rescuing, and mana motuhake over dependency.

• Ask what happened to a person before asking what is wrong with them.
• Design contact so that people keep their agency, including the right to decline.
• Train for de-escalation and ethical boundaries as core skills, not optional extras.

5. Reduce the variation the system itself creates

Not all of the difficulty in reintegration comes from the person reintegrating. Inconsistent use of special conditions, unclear expectations, fragmented communication between agencies, and practice that varies by region all add stress and risk that nobody intended. This is a system design problem, and it is fixable. It is also why Bond convened the National Advisory Group on Special Conditions, to support clearer, evidence-based and culturally grounded guidance across the motu.

• Write conditions and expectations in language the person can act on without an interpreter.
• Agree who communicates what, between agencies, before a problem forces the question.
• Escalate unexplained regional variation as a practice issue, not an individual one.

Where this leads

None of these five conditions works alone. Relationships without practical stability wear out; stability without relationship rarely lasts; both are undermined when the system around them is unclear. Taken together they describe the ecosystem that Bond Trust works to strengthen through partnership, training and sector convening, while the Fellowship Support Network carries the direct person-centred mahi.

For more on the initiatives and evidence behind this approach, see our work across Aotearoa, or read our companion post on how we share stories of change. Practitioners, partner agencies and community groups are welcome to get in touch about training and collaboration.

Read more about the evidence base behind our work on our Insights and Evidence page, or see how these principles shape practice in What We Do.