Abstract Purpose To synthesize the literature in relation to findings of system errors through reviews of suicide deaths in the public mental health system. Data sources A systematic narrative meta-synthesis using the PRISMA methodology was conducted. Study selection All English language articles published between 2000 and 2017 that reported on system errors identified through reviews of suicide deaths were included. Articles that reported on patient factors, contact with General Practitioners or individual cases were excluded. Data extraction Results were extracted and summarized. An overarching coding framework was developed inductively. This coding framework was reapplied to the full data set. Results of data synthesis Fourteen peer reviewed publications were identified. Nine focussed on suicide deaths that occurred in hospital or psychiatric inpatient units. Five studies focussed on suicide deaths while being treated in the community. Vulnerabilities were identified throughout the patient’s journey (i.e. point of entry, transitioning between teams, and point of exit with the service) and centred on information gathering (i.e. inadequate and incomplete risk assessments or lack of family involvement) and information flow (i.e. transitions between different teams). Beyond enhancing policy, guidelines, documentation and regular training for frontline staff there were very limited suggestions as to how systems can make it easier for staff to support their patients. Conclusions There are currently limited studies that have investigated learnings and recommendations. Identifying critical vulnerabilities in systems and to be proactive about these could be one way to develop a highly reliable mental health care system.