Rationale Suicide is a major public health concern and Northern Ireland (NI) has the highest rate of both self-harm and suicide in the UK and Ireland. In order to target prevention strategies effectively it is vital to understand who is most at risk. Those who present with self-harm offer a prime opportunity for intervention. The aim of this study is to examine the risk factors for completed suicide following presentation with self-harm.

Data The Northern Ireland Registry of Self-Harm (NIRSH) collects information on all self-harm and suicide ideation presentations to all Emergency Departments in NI. NIRSH data for the four years 2012–2015 was linked to centralised electronic data relating to primary care, social services and prescribed medication and mortality records.

Methods Initial analyses describing the profile of those who present with self-harm was followed by logistic regression to quantify the likelihood of mortality with adjustment for factors associated with mental ill health and suicide risk.

Results The cohort consisted of all 1,483,435 individuals born or resident in NI from 1st January 1970 until 31^st December 2015 (maximum age in 2015, 45 years). During the follow-up period, 11,371 (0.8%) individuals presented with self-harm and 1,719 (0.1%) died by suicide. Rates of self-harm were equivalent for males and females with highest rates observed in the 18–24 years age group, and more common in deprived than affluent areas (OR=3.34, 95%CI 3.12, 3.57). Rates of self-harm was highest among those who were (or ever had been) in the care of social services (OR=12.06, 95%CI 11.26, 12.93). Most individuals self-harm via self-poisoning with psychotropic medications (70.9%), followed by self-injury with a sharp object (21.6%). Although only 142 (1.3%) of those who presented with self-harm went on to die by suicide, in the unadjusted model those who self-harmed were almost 12 times more likely to die by suicide compared to those who did not present with self-harm (OR=11.79, 95%CI 9.92,14.01). Suicide was more likely in individuals who self-harmed with more violent methods such as strangulation or drowning. After full adjustment for gender, age, social services care, area of residence and deprivation individuals who self-harm are still 7 times more likely to die by suicide (OR=7.05, 95%CI 5.88,8.45).

Further analysis is underway to explore which particular traits and characteristics of those who self-harm are most associated with risk of suicide using cox regression in order in inform intervention targeting.