Prisons are a hotspot for blood-borne virus transmission. To control outbreaks of these infections, it is essential to have an in-depth understanding of risk behaviours of inmates and the impact of existing control measures, and also to apply sensitive methods for detection of new infections occurring within prisons. This thesis describes studies in these three domains to understand the spread of hepatitis C infection in the prisons in New South Wales (NSW). The clinical data and blood samples for this work originated from two prospective studies conducted in the NSW prisons: the Hepatitis C Incidence and Transmission Study in prisons (HITS-p) (2005- 2014; n=590) and the Surveillance and Treatment of Prisoners with Hepatitis C (SToP-C) (2015-2019; n=3691). The first study identified a high rate of incident infections in an at-risk cohort and documented outcomes (spontaneous clearance, chronic infection, and re-infection or superinfection), while the second study demonstrated the impact of scale up of direct acting antiviral (DAA) treatment in reducing incidence – that is treatment as prevention. The first project utilised qualitative data obtained from audiotaped interviews with inmates in the HITS-p study to understand contexts and concerns regarding violence and HCV transmission in prison. Concerns regarding violence were identified at the individual level during blood contact; triggering factors such as drug debt were identified at the network level; racial influence at the community level; and legislation such as delayed parole for violence impacted at the policy level. For the subsequent projects near-full length HCV genomes for genotypes 1a and 3a cases from both cohorts were sequenced with Oxford Nanopore Technology (ONT) using previously published protocols. A total of 211 genotype 1a sequences and 282 genotype 3a sequences were generated. Of these, 28 1a and 63 3a sequences were from samples collected within 6 months of the estimated date of infection and are hence referred to as acute infection sequences. Acute infection sequences from samples collected during 2005 – 2015 (the pre-DAA era) and 2016-2019 (the post-DAA era) were used to model the temporal evolution in the size of infected population using previously published Bayesian evolutionary analysis methods. The effective population size modelled with the genotype 3a infection samples, demonstrated a 21% reduction in the size of infected population in 2019 compared to 2014. By contrast, the trend in the genotype 1a samples was static. The SToP-C sequences (from both acute and chronically infected subjects) were used to identify molecularly related infections (clusters), before within-host viral variants were further characterised within these clusters to identify likely direct transmission events (defined as phylogenetic intermingling of within-host variants between two or more subjects). For genotype 1a, there were 131 sequences which formed 51 clusters, and for genotype 3a, 140 sequences formed 61 clusters. Each cluster had 2-4 sequences. Among these, 41 genotype 1a and 39 genotype 3a clusters were analysed for minor variants. Evidence of a direct transmission of a within-host variant was observed in only one cluster. Subjects in 45 (40%) clusters were from the same prison providing preliminary epidemiological support for the transmissions. These studies highlight the high-risk context for HCV transmissions in the prison setting, and the utility of molecular epidemiological tools for surveillance in this closed setting.