PURPOSE
This study aims to review international literature systematically to estimate the prevalence of homelessness among incarcerated persons at the time of imprisonment and the time of discharge.

DESIGN/METHODOLOGY/APPROACH
A systematic review methodology was used to identify quantitative observational studies that looked at the prevalence of homelessness at the time of imprisonment, or up to 30 days prior to that point (initial homelessness), and at the time of discharge from prisons. Studies reported in English from inception to 11 September 2019 were searched for using eight databases (PsycInfo, Medline, Embase, CINAHL, PsycArticles, Scopus, Web of Science and the Campbell Collaboration), in addition to grey literature. Studies were screened independently by three researchers. Results of studies meeting inclusion criteria were meta-analysed using a random effects model to generate pooled prevalence data.

FINDINGS
A total of 18 out of 2,131 studies met the inclusion criteria. All studies originated from the USA, Canada, UK, Ireland or Australia. The estimated prevalence of initial homelessness was 23.41% and at time of discharge was 29.94%. Substantial heterogeneity was observed among studies.

ORIGINALITY/VALUE
People in prisons are over twenty times more likely to be homeless than those in the general population. This is likely attributable to a range of health and social factors. Studies in this analysis suggest higher rates of homelessness in minority populations and among those with mental illnesses and neurodevelopmental disorders. While there was significant heterogeneity among studies, the results highlight the global burden of this issue and a clear necessity for targeted interventions to address homelessness in this population.