Suicide And Self-Harm In Women With Mental Disorders During Pregnancy And The Year After Birth

Background: Despite the strong links between mental disorder, self-harm and suicide, there is little prospective data on risk factors for later suicide in women who experience perinatal mental disorders. We aimed to examine later suicide in a cohort of women in contact with a mental healthcare provider perinatally and whether self-harm increased suicide risk. Methods: Via data-linkage, de-identified service-user electronic healthcare records, national hospital episode statistics and mortality data generated a clinical cohort of women who gave birth in hospital and were in contact with a mental healthcare provider in London, UK, perinatally. Using Natural Language Processing and structured field extraction, we identified clinical, socio-demographic characteristics, self-harm, and suicide. Outcomes: Among 5204 women, crude rate of suicide 9·69/100 000 person-years (95% CI 3·64 - 25·82) within one year of delivery; 27·56 (15·26 - 49·77) within two, and 55·12 (36·29 - 83·71) during total follow up time (39,912·9 person-years). Characteristics of women who did and did not die by suicide were similar apart from illness severity indicators, which were more common in women who died by suicide. In a Cox regression model, lifetime self-harm was not associated with later suicide. Suicides occurred most frequently in the second year after delivery, most commonly by violent means. Suicides beyond two years were most commonly by poisoning. Interpretation: This study provides support for the extension of perinatal mental healthcare to two years post-delivery, particularly for women with severe mental disorders. Funding : The National Institute for Health Research, King's College London Biomedical Research Council, Health Foundation, Academy of Medical Sciences. Declaration of Interests: None. Ethics Approval Statement: CRIS has pre-existing ethical approval via the Oxfordshire Research Ethics Committee C (ref 18/SC/0372). The use of HES data is included in the CRIS REC approval and the BRC also have approval through the Health Research Authority’s support under Section 251 of the NHS Act 2006 (Ref: ECC 3- 04(f)/2011). Linkage between CRIS and ONS Mortality data is approved via the Health Research Authority’s support under Section 251 of the NHS Act 2006. Individual CRIS projects require ethical approval from the CRIS Oversight Committee (approval 16-069), which is chaired by a service-user and member of the SLAM BRC Stakeholder Participation theme.