Background: Experiences of end-of-life care and early bereavement during the COVID-19 pandemic are poorly understood. Aim: To identify clinical and demographic risk factors for sub-optimal end-of-life care and pandemic-related challenges prior to death and in early bereavement, to inform clinical practice, policy and bereavement support. Design: Online national survey of adults bereaved in the UK (deaths between 16 March 2020 and 2 January 2021), recruited via media, social media, national associations and organisations. Setting/participants: 711 participants, mean age 49.5 (SD 12.9, range 18–90). 628 (88.6%) were female. Mean age of the deceased was 72.2 (SD 16.1, range miscarriage to 102 years). 311 (43.8%) deaths were from confirmed/suspected COVID-19. Results: Deaths in hospital/care home increased the likelihood of poorer experiences at the end of life; for example, being unable to visit or say goodbye as wanted ( p < 0.001). COVID-19 was also associated with worse experiences before and after death; for example, feeling unsupported by healthcare professionals ( p < 0.001), social isolation/loneliness (OR = 0.439; 95% CI: 0.261–0.739), and limited contact with relatives/friends (OR = 0.465; 95% CI: 0.254–0.852). Expected deaths were associated with a higher likelihood of positive end-of-life care experiences. The deceased being a partner or child also increased the likelihood of positive experiences, however being a bereaved partner strongly increased odds of social isolation/loneliness, for example, OR = 0.092 (95% CI: 0.028–0.297) partner versus distant family member. Conclusions: Four clear risk factors were found for poorer end-of-life care and pandemic-related challenges in bereavement: place, cause and expectedness of death, and relationship to the deceased.

Changes made as a result of publishing processes such as copy-editing, formatting and page numbers may not be reflected in this version. For the definitive version of this publication, please refer to the published source. You are advised to consult the publisher's version if you wish to cite this paper.

No abstract