Objectives The objective of the present study is to describe the patterns of health service use and of prescription claims in the year preceding an offense leading to a verdict of not criminally responsible on account of a mental disorder (NCRMD). Methods Provincial health administrative databases were used to identify medical services, hospitalizations, and ambulatory prescription claims among 1,014 individuals found NCRMD in Québec. Contacts in the year preceding the offense were analyzed using descriptive analyses and latent class analysis. Results Overall, 71.4% of subjects were in contact with services for mental health reasons within a year of their NCRMD offense. Among those that received services and not hospitalized for psychiatric reasons at the time of the offense, 20.7% committed the NCRMD offense within a week of the most recent mental health contact. Among those that had at least one prescription claim for an antipsychotic, 45.8% were not taking any antipsychotic at the time of the offense. Latent class analysis provided a multidimensional representation of mental health service use and showed that 58.4% of subjects had no or very rare contact with services. Conclusions Many forensic patients are likely to have experienced service disruption or discontinuity while in the community, for reasons that may relate to perceived need for care, to service organization, or to the acceptability, availability, and accessibility of services. Given the serious impact of the “forensic” label on the lives of service users, not to mention the increased pressure on resources, the considerable economic costs, and the impact on victims, there is reason to advocate for a greater involvement of mental and physical health service providers in early prevention of violence, which requires reorganizing resources to share the forensic knowledge upstream, before an offense is committed.