Are public health measures and individualised care compatible in the face of a pandemic? A national observational study of bereaved relatives’ experiences during the COVID-19 pandemic

Background: COVID-19 public health restrictions have affected end-of-life care experiences for dying patients and their families. Aim: To explore bereaved relatives’ experiences of quality of care and family support provided during the last days of life; to identify the impact of factors associated with perceived support. Design: A national, observational, open online survey was developed and disseminated via social media, public fora and professional networks (June–September 2020). Validated instruments and purposively designed questions assessed experiences. Analysis used descriptive statistics, logistic regression and thematic analysis of free-text responses. Participants: Individuals (⩾18 years) who had experienced the death of a relative/friend (all care settings) within the United Kingdome during the COVID-19 pandemic. Results: Respondents (n = 278, mean 53.4 years) tended to be female (n = 216, 78%); over half were ‘son/daughter’ (174, 62.6%) to the deceased. Deceased individuals (mean 81.6 years) most frequently died in their ‘usual place of care’ (n = 192, 69.3%). Analysis established five conceptual themes affecting individualised care: (1) public health restrictions compounding the distress of ‘not knowing’; (2) disparate views about support from doctors and nurses; (3) challenges in communication and level of preparedness for the death; (4) delivery of compassionate care; (5) emotional needs and potential impact on grief. Male respondents (OR 2.9, p = 0.03) and those able to visit (OR 2.2, p = 0.04) were independently associated with good perceptions of family support. Conclusion: Despite public health restrictions, individualised care can be enabled by proactive, informative communication; recognising dying in a timely manner and facilitating the ability to be present before death.


Introduction
Following the emergence of COVID-19 in December 2019, the World Health Organisation declared a pandemic on March 11th 2020. 1 Pandemics cause increased demand for healthcare services, including those provided by palliative and end-of-life care. 2 Changing patterns of mortality are recognised with increased deaths occurring within community and hospital settings, due to both COVID-19 and non-COVID related illnesses. 3,4 Staff illness and redeployment impact on the workload and skill mix of staff. Hospice services may have fewer in-patient admissions and need to shift resources to help enhance community care and provision. 3,5 Natural disasters, including pandemics, have a profound impact on end-of-life care, often resulting in a 'stark departure from a palliative care approach'. 6 Restricted visiting policies, fewer end-of-life discussions and the loss of usual death and bereavement rituals can jeopardize typical end-of-life care values and leave patients feeling isolated and fearful. 6 Additionally, the support of the family, one of the two 'pillars' of palliative care, can be compromised. 7 Values focused on individual needs and preferences, such as the wish for family presence and not to die alone, may be limited. 8 Few studies focus on direct service-users views during pandemics and those which have, involved seeking views about the use of telemedicine. 6 It is unlikely that COVID-19 will be the last global pandemic; we must learn from people's experiences of the current crisis. The views of bereaved people are especially pertinent as they provide accounts of both patient care and family support. Additionally, perceptions about quality of the dying experience, along with preparedness for death, are associated with risks of complicated grief. 9

Aim
The aim of this study was to explore the experiences of individuals who had suffered a bereavement within the United Kingdom (UK) during the COVID-19 pandemic.
We present key data relating to quality of care, communication and emotional support. These represent elements within the 'quality of dying experience' which can affect grief and bereavement. 9 Additionally, we examined whether any demographic or clinical factors were independently associated with perceptions of adequate family support in the last days of life.

Survey development
An observational, open online survey was developed using Qualtrics online survey software and tested by the research team. The survey tool was adapted from a measure used in an affiliated, ongoing Horizon 2020 research project (The iLIVE Project: www.iliveproject.eu) to ensure relevance for deaths occurring during the pandemic and be in keeping with culturally appropriate language and practice (Supplemental File 1). The survey contained questions about demographics (deceased individual and respondents), an abbreviated version of the 'Care Of the Dying Evaluation' (CODE™) questionnaire 10 and relevant questions about COVID-19 and its consequences. The internationally used CODE™ questionnaire focuses on quality of care and family support during the last days of life. 11 One key outcome from CODE™ is whether or not respondents perceive themselves to be adequately supported in the last days of life. Free-text space was available for additional comments.
Five members from the online advisory panel of the Clinical Research and Innovation Office at Sheffield

Study population
The questionnaire was aimed at individuals (⩾ 18 years) who had experienced the death of a family member, within the UK, during the initial part of the COVID-19 pandemic (March-September 2021) and were able to provide informed on-line consent. There were no additional inclusion or exclusion criteria. For clarity and simplicity, 'bereaved relatives' or 'family member' are used as collective terms, also representing bereaved friends and neighbours.

Recruitment
Awareness about the study and circulation of the online survey was undertaken via local and national bereavement organisations (CRUSE, National Bereavement Alliance), as well as UK professional, public and charitable networks related to palliative care and minority ethnic communities, for example, Association of Palliative Medicine of Great Britain and Ireland, All Ireland Institute of Hospice and Palliative Care, National Society of Allied and Independent Funeral Directors and Black Thrive.
Organisations used email distribution lists, newsletters and/or social media; to promote accrual, the survey link was advertised via social media, press releases and personal contacts. Survey instructions asked recipients to forward links to others who may be interested in taking part, enabling snowball sampling and maximum survey reach. Data was collected between 1st June 2020 and 30th September 2020. Responses were anonymous unless respondents supplied their name and email address in the optional final field to indicate interest in further research. Additionally, respondents were invited to participate in an in-depth qualitative interview about their experiences; the results are published elsewhere. 12

Data analysis
Quantitative data were analysed using descriptive statistics within SPSS (version 26). Associations between specific respondent and deceased individual characteristics and care settings (independent variables) and whether respondents perceived themselves to be adequately supported (dependant variable) were analysed using univariate and multivariate logistic regression analyses.
Qualitative analysis of the free text data was conducted using Braun and Clarkes' 13 principles of inductive thematic analysis. The analysis was conducted by CRM and SRM, who familiarised themselves and independently coded the data, then developed themes. They collaboratively reviewed and revised the themes, in conjunction with the quantitative data, and with the wider research team through critical dialogue. This approach was undertaken to enhance and illustrate study findings. 14

Ethical considerations
The study protocol was approved by the University of Liverpool Central Research Ethics Committee (Ref: 7761). Potential participants viewed the 'Participant Information Sheet' on-line, which outlined details including the study purpose, details of data storage and use and contact details for organisations which could provide bereavement support. All respondents provided informed, online consent. Due to the sensitive nature of the survey content, there was no forced response/requirement for participants to answer all questions. Additionally, a specified time period between the death and completion of the survey was not stated, allowing individuals to decide the 'right time' for them.

Response rate
From 384 potential respondents who accessed the survey, 79 (20.6%) consented to participate in the study but did not complete the questionnaire. From the remaining 305 respondents, 27 completed demographic details only (non-respondents) and 278/384 (72.4%) completed the questionnaire (respondents). There was no statistically significant difference between respondents and nonrespondents for gender (p = 0.25) or mean age (p = 0.1). The deceased individuals from the respondents' group were predominately older in age compared with the nonrespondents' group (mean age 81.6 years (SD 12.2) vs 68.8 years (SD 18.1), p = 0.001). Free-text comments were provided by 34% (n = 104) of respondents.

Demographics for respondents and the deceased individuals
The mean age of respondents was 53.4 years (median 55.0 years, range 19-68 years). All except one respondent was from a White British ethnic group. Over threequarters identified as female (216, 78.0%) and over half were the 'son/daughter' of the deceased individual (157, 56.5%).
The majority of deceased individuals died in England (179, 69.1%) and within their 'usual place of care' (192, 69.1%). Over half of the deaths occurred in a nursing care or residential home (162, 58.5%) ( Table 1). Just under 10% (26, 9.4%) had died within a specific COVID-19 hospital ward. Almost a third of all deceased individuals (82, 32.0%) were thought to have COVID-19. Many of the deceased individuals had a chronic illness, the most common being dementia (148, 53.2%).

Themes
Using the quantitative survey data and the qualitative verbatim free-text responses, five conceptual themes were identified, linking to aspects of individualised care, under which aggregated results and analysis are presented. The themes include: (1) public health restrictions compounding the distress of 'not knowing'; (2) disparate views about support from doctors and nurses; (3) challenges in communication and level of preparedness for the death; (4) delivery of compassionate care; (5) emotional needs and potential impact on grief.

Theme 1. Public health restrictions compounding the distress of 'not knowing'
Pandemic restrictions meant that over half (157, 56.5%) of the respondents were unable to visit during the last days of life and a third (92, 33.7%) were unable to have any online contact (e.g. through Skype) ( Table 1). Respondents clearly identified profound distress due to the impact of the restrictions, expressing sadness, guilt, anxiety, fear and feeling 'cheated'. '

Theme 2. Disparate views about support from doctors and nurses
Respondents reported a greater level of confidence and trust in nursing staff compared with doctors (Table 2). Additionally, respondents indicated nursing staff more frequently had provided care with respect and dignity. For example, almost half of respondents (119, 48.4%) perceived that nursing staff had treated their family member with dignity and respect 'all of the time'. The same question asked of doctors received less than a third of responses (75, 31.1%) although for both groups, there was a high proportion of 'don't know' responses.
Respondents experiencing the death of a family member within the community setting shared differing perceptions about the level of support provided by nursing staff and doctors. They reported an absence of the usual continuity of care and visits from their General Practitioner (GP).

Theme 3. Challenges in communication and level of preparedness for the death
Almost half of respondents perceived their family members' care and treatment was definitely affected due to COVID-19 (103, 44.6%) ( Table 2). Nearly a fifth of respondents did not receive a meaningful explanation about their family members' condition and/or treatment (49, 19.9%) and over a third did not perceive themselves to be involved in decisions about care or treatment (96, 37.9%). Communication with health and social care teams was reliant on remote methods, usually via telephone, which created a disconnect between families and the relevant healthcare teams. Those not identified as the immediate next-of-kin were dependent on 'second-hand' information. Respondents reported being frustrated by not receiving desired and required information about their   days of life (Table 3). Some respondents (71, 30.1%) perceived the emotional support provided by the healthcare teams to be 'poor'; also reflected in perceptions about the support available at the actual time of death (Table 3).
Respondents reported there was a lack of regard for meeting families' needs and providing psychosocial support. This predominately related to the issue of social isolation, enforced by pandemic restrictions, limiting access to provide direct support to their relative, being able to visit and the opportunity to view the deceased individual.
'However he had covid 19 symptoms and this meant that we were not able to access hep (sic) and support. We cared for him almost completely on out (sic) own.. . ..Even his body was treated disrespectfully by the funeral director when he was taken form (sic) the house, and we were not allowed to see him again.' (Respondent 57, daughter-in-law, own home) Respondents provided their initial reflections about the impact of the bereavement and potential issues which may arise in their grief. Experiences were described as 'traumatic'; some being uncertain whether they would 'ever get over it'; or struggling to make sense of their situation.

Predictive factors for adequate family support in the last days of life
In order to assess whether key demographic or clinical factors independently predicted perceived support, univariate and multivariate regression analyses were conducted. The variables chosen (respondents' age, gender, relationship to deceased, ability to visit, presence of dementia and place of death) are recognised risk factors for complicated grief (gender, relationship to deceased) 15 or are indirectly associated with levels of preparedness and support (age, ability to visit, presence of dementia, place of death) which have impact on psychological outcomes in bereavement. 15 The 'relationship to deceased' was categorised into 'husband/wife/partner', 'son/daughter', 'grandson/granddaughter', with other relationships grouped as 'other'.
The variables univariately associated with perceived support were respondent gender and ability to visit (Table 4). Men were almost three times more likely to report being adequately supported compared to women (adjusted OR 2.9, CI 1.44-5.94, p = 0.03). Those able to visit were twice as likely to report being adequately supported during the last days of life, compared to those unable to visit (adjusted OR 2.2, CI 1.3-3.75, p = 0.04) ( Table 5).

Main findings
Being unable to visit during the last days of life affected preparedness for the death, perceived levels of emotional  Missing values shown (in italics), but not included in percentages.  18 In order to provide holistic patient updates, information to family members include aspects of medical care and personal care, such as details and discussion surrounding individual preferences and wishes. Our qualitative study (linked to this survey), 12 details relatives' experiences when a family member was dying and reported how crucial it was to maintain connection, virtually or otherwise. The current study highlights this, especially within the care home setting, where restrictions had a pronounced impact on emotional well-being and the use of technology was not without issues. This current study provides a more inclusive landscape of acute and community settings (including deaths at home, nursing homes and hospices) and particularly highlights the disparate views perceived between doctors and nurses.
Recognition of the dying phase of illness is challenging, particularly for those with dementia, where disease trajectory is complex and variable. 19 The rapid and unpredictable dying phase of COVID-19 posed new challenges for clinical teams in recognising that death may be approaching. However, where local leadership anticipated changes and enabled timely visits, this had a profoundly positive affect. Further education and research focused on enabling health and social care staff to recognise dying and feel confident to talk honestly with relatives about this, should be prioritised.
Men were significantly more likely to feel adequately supported. This finding may relate to females being more prone to disclose emotional challenges in their coping. 20 Additionally, female caregivers may subjectively experience higher distress and perceive less support from others, including the clinical team. 21 The loss of the usual support networks, this may have compounded distress levels, to which future research could explore.

Strengths and limitations
This research represents the first UK study of bereaved relatives' views about quality of care and family support provided during the last days of life and highlights the challenges and emotions experienced during the COVID-19 pandemic. COVID-19 studies from the Netherlands, United States and China, [22][23][24][25] have explored acute grief and the psychological impact during the bereavement period. This study, however, closely examines the quality of the dying experience which can have subsequent impact on grief.
Our study has limitations. Firstly, the convenience sample (with no specific sample size calculated) limits generalisability. Despite efforts to disseminate widely, respondents tended to be relatives of older individuals dying in community settings. All but one respondent was from a White British ethnic background, limiting the representativeness of the sample. Secondly, for

Conclusion
Though challenging, public health measures in response to a pandemic can be actioned in ways that maintain high quality of end-of-life care for patients and their families. Care services must acknowledge the significance of individual deaths and ensure active approaches to support those who are bereaved. This may achieve headway in healing the barriers between what was required to limit the spread of the virus and the subsequent compromise on individualised care. Recognising dying, communicating holistic aspects of care and enabling visits whenever is possible, should continue to be priorities, during the ongoing pandemic. These approaches should be allied with risk stratification measures to help identify those in most need of bereavement support, including the identification of those unable to visit.