Healthcare contacts with self-harm during COVID-19: An e-cohort whole-population-based study using individual-level linked routine electronic health records in Wales, UK, 2016-March 2021

Introduction Reduced rates of help seeking by those who self-harmed during the COVID-19 pandemic have been reported. Objectives To understand changes in healthcare service contacts for self-harm during the COVID-19 pandemic across primary, emergency and secondary care. Methods This retrospective cohort study used routine electronic healthcare data for Wales, United Kingdom, from 2016 to March 14, 2021. Population-based data from primary care, emergency departments and hospital admissions were linked at individual-level. All Welsh residents aged >= 10 years over the study period were included in the study. Primary, emergency and secondary care contacts with self-harm at any time between 2016 and March 14, 2021 were identified. Outcomes were counts, incidence, prevalence and proportion of self-harm contacts relative to all contacts in each and all settings, as well as the proportion of people contacting one or more settings with self-harm. Weekly trends were modelled using generalised estimated equations, with differences between 2020 (to March 2021) and comparison years 2016-2018 (to March 2017-2019) quantified using difference in differences, from which mean rate of odds ratios (mu ROR) across years was reported. Results The study included 3,552,210 individuals over the study period. Self-harm contacts reduced across services in March and December 2020 compared to previous years. Primary care contacts with self-harm reduced disproportionately compared to non-self-harm contacts (mu ROR = 0.7, p<0.05), while their proportion increased in emergency departments during April 2020 (mu ROR = 1.3, p<0.05 in 2/3 comparison years) and hospital admissions during April-May 2020 (mu ROR = 1.2, p<0.05 in 2/3 comparison years). Despite this, those who self-harmed in April 2020 were more likely to be seen in primary care than other settings compared to previous years (mu ROR = 1.2, p<0.05). A lower proportion of those with self-harm contacts in emergency departments were subsequently admitted to hospital in December 2020 compared to previous years (mu ROR = 0.5, p<0.05). Conclusions These findings suggest that those who self-harmed during the COVID-19 pandemic may have been less likely to seek help, and those who did so faced more stringent criteria for admission. Communications encouraging those who self-harm to seek help during pandemics may be beneficial. However, this needs to be supported by maintained provision of mental health services.