# Associations Between Supported Accommodation and Health and Re-offending Outcomes: a Retrospective Data Linkage Study

## TL;DR

A study in Sydney found that a supported accommodation program reduced criminal charges among men released from prison, but had limited effects on health outcomes and reincarceration.

## Contribution

This study provides empirical evidence on the impact of a specific supported accommodation program on criminal and health outcomes post-prison release.

## Key findings

- Attending Rainbow Lodge reduced criminal charges by 44%.
- No significant association was found between attending Rainbow Lodge and emergency department presentations or ambulance attendance.
- There was no evidence that attending Rainbow Lodge affected reincarceration rates.

## Abstract

Following release from prison, housing and health issues form a complex and mutually reinforcing dynamic, increasing reincarceration risk. Supported accommodation aims to mitigate these post-release challenges. We describe the impact of attending Rainbow Lodge (RL), a post-release supported accommodation service for men in Sydney, Australia, on criminal justice and emergency health outcomes. Our retrospective cohort study using linked administrative data includes 415 individuals referred to RL between January 2015 and October 2020. Outcomes of interest were rates of criminal charges, emergency department (ED) presentations and ambulance attendance; and time to first reincarceration, criminal charge, ED presentation and ambulance attendance. The exposure of interest was attending RL; covariates included demographic characteristics, release year and prior criminal justice and emergency health contact. Those who attended RL (n = 170, 41%) more commonly identified as Aboriginal or Torres Strait Islander (52% vs 41%; p = 0.025). There was strong evidence that attending RL reduced the incidence criminal charges (adjusted rate ratio [ARR] = 0.56; 95% confidence interval [CI] 0.340.86; p = 0.009). Absolute rates indicate a weak protective effect of RL attendance on ED presentation and ambulance attendance; however, adjusted analyses indicated no evidence of an association between attending RL and rates of ED presentations (ARR = 0.88; 95% CI = 0.65–1.21), or ambulance attendance (ARR = 0.82; 95% CI = 0.57–1.18). There was no evidence of an association between attending RL and time to first reincarceration, charge, ED presentation or ambulance attendance. Greater detail about reasons for emergency health service contact and other self-report outcome measures may better inform how supported accommodation is meeting its intended aims.

The online version contains supplementary material available at 10.1007/s11524-023-00824-w.