Prison suicide/self-inflicted death is an international public health crisis, harming stakeholders including bereaved families, prisoners, prison staff and death investigators. England and Wales' record prison suicide numbers in 2016 cost at least £400 million. Death rates are an indicator of prison safety, and unsafe prisons mean unsafe societies. I present four case studies of people with very severe mental illness who were remanded to prison from police and/or court custody and went on to take their own lives in prison. I use publicly available data from Ombudsman and Coronial death investigations in England and Wales, highlighting that these accessible sources could be more widely mobilized to reduce the substantial harms and costs of prisoner deaths. Case studies include three men (Lewis Francis, Jason Basalat and Dean Saunders) and one woman (Sarah Reed) who took their own lives between January 2016 and April 2017. All four people were clearly very mentally unwell at the time of their alleged offense and remand to prison. I develop the concept of “risky remands” to highlight that people with very severe mental illness being remanded to prison is a particularly problematic practice. I highlight the implications of people with very severe mental illness transitioning into prison in the first place, arguing that being remanded to prison is not an acceptable or safe pathway into healthcare. I illustrate that police custody suites and courts may lack awareness of mechanisms and/ or the practical ability to transfer ill detainees charged with a serious crime to mental health facilities for assessment and/ or treatment. My analysis amplifies and extends recent Criminal Justice Joint Inspection findings that it is unacceptable to use prisons as a “place of safety,” and that the Department of Health and Social Care, NHS England and the Welsh Government must increase the supply of medium and high secure beds. Moreover, Ombudsman investigations did not engage with the remand transition, effectively legitimizing this risky practice for very ill people. As such, my analysis also counters the apparent “problem of implementation” in prison oversight, instead questioning what reviewers recommend, based on which evidence.