Suicide is a prevalent and international problem which has substantive economic and psychological consequences. This has led to governments placing prevention of suicide as a priority on healthcare agendas. Recognition has been given to vulnerable groups in society that have been identified as being at particularly high risk of self-harm and suicide. This includes those in contact with mental health and forensic services. There is a great deal of literature that has considered the risk factors, processes and mechanisms associated with suicide. Comparatively only a small amount of literature has looked at the concept of suicidality within restricted samples such as psychiatric inpatients and prisoners. This may be as a consequence of extensive ethical and procedural processes that are involved in conducting research in such settings. This results in it being necessary to continually make generalisations from community based literature, meaning that factors relating specifically to such settings may be overlooked or underestimated.In the first paper, the initial sections consider existing risk assessments and models of suicidality. Predominantly being structured around static risk factors, means they are often criticised for lacking predictive utility and specificity. Literature examining dynamic psychosocial factors of suicidality in restricted samples was reviewed and 20 articles were identified. A wide range of dynamic correlates are presented. These form a theoretical model of suicidality specific to restricted samples. The clinical and theoretical implications are discussed in terms of risk assessment procedures and adapting and shaping interventions in accordance with the findings. Developing risk assessments around more dynamic factors will allow for greater sensitivity and prediction of those at greatest risk of imminent harm.The second, empirical paper supports the promotion of recovery focused practice and explores the relationship between suicidality and perceived personal agency in patients in secure mental health settings; Personal agency having previously been suggested as conferring resilience to suicidality. Psychometric measures and experience sampling methodology were utilised to examine the relationship. Perceptions of personal agency were found to confer resilience against suicidality. Change in perceptions of personal agency was not associated with suicidality but the overall level of personal agency was. Implications for service delivery are discussed with emphasis given to fostering perceptions of agency, control and self-efficacy and promoting inclusion, empowerment and person centred care.The final paper provides a personal and a critical reflection on the research process. It highlights and discusses clinical and theoretical strengths and limitations of the two papers and considers the methodological processes of both papers in more detail. Further reflections on how practice could be adapted in line with the findings are given. Future directions for research within secure settings are considered, in the hope of maintaining the drive for research with this vulnerable and often overlooked population.