What are the factors associated with people with advanced dementia refusing assistance with personal care?

Abstract Background People with dementia sometimes refuse assistance with personal care activities such as washing or dressing. We aimed to investigate the factors associated with refusals of care in advanced dementia. Methods A cross‐sectional study using informant‐based measures. Participants were people with advanced dementia and their caregivers (family carers or care‐home staff) (n = 260, 130 dyads) in the UK. Mixed effects linear models were used to examine the effects of neuropsychiatric behaviours, ability with activities of daily living, professional input, co‐morbidities, psychotropic medications, environment modifications, and caregiver factors including type and training status on refusals of care. The Refusal of Care Informant Scale was used, range 1–13; higher scores indicate more refusal behaviours. Results Higher independence in activities of daily living was associated with less refusal behaviours (coefficient = −0.11, p < 0.001 [95% confidence interval −0.15, −0.07]). Higher agitation was associated with more refusal behaviours (0.11, p < 0.001 [0.06, 0.15]). No other statistically significant differences were found. There was no demonstrable evidence of differences in number of refusals of care between family and care‐home caregivers or between dementia‐trained or ‐untrained caregivers. Conclusions Results suggest refusals of care have similar prevalence regardless of caregiver type (family or care home) or dementia training status, indicating that current dementia training has no impact on refusals of care or may not be implemented as intended. Improving independence in activities of daily living and reducing agitations may help prevent refusals of care. To establish causality, future research should consider embedding these factors into interventions targeting refusal of care.

� In advanced dementia, higher independence in activities of daily living was associated with less refusals and higher agitation was associated with more refusals.
� Caregiver setting (family or care home) or training in dementia (none or some) were not associated with refusals of care.

| INTRODUCTION
There are over 55 million people living with dementia globally, 1 with over 885,000 in the UK. 2 In the UK, 60% of people with dementia are in the advanced stages, 2 which can last up to 5 years. 3 In the advanced stages, people living with dementia require considerable assistance with basic activities of daily living such as going to the toilet, washing, and dressing. 4 Often this assistance with personal care is provided by family carers or care-home staff (henceforth collectively termed as caregivers). 2 Sometimes receiving personal care assistance is not acceptable to the person with dementia and they refuse it. [5][6][7][8] Refusals of care can be distressing for both the person with dementia 9 and their caregiver. For example, family carers have been found to experience psychological distress as refusals provide barriers to completing functional care. 10 Whereas care-home staff can experience feelings of discomfort created by the tension between their caring values and the situations refusals generate. 11 Refusals can lead to caregiver stress and burden 6,12 and have potential negative consequences such as urine burns if adequate care is not completed. 13 Interventions evidenced as reducing refusals of care include playing music during care or using different bathing techniques such as a thermal bath or strip wash. 14 Refusals of care have been linked to factors such as the caregiver's communication style 12,15 ; uninvited direct care, 16 psychotic symptoms such as hallucinations and delusions 17,18 ; depression 8,17,18 ; the person not understanding, being confused or disoriented, 17,19 or inadequate staffing, workload, and time management in care homes. 19 However, there has been limited research investigating factors in combination to identify those associated with refusals of care in dementia. Factors can be categorised as, those related to the: � Person with dementia: dementia sub-type, age, gender, ethnicity, psychotropic medication use, ability in activities of daily living, neuropsychiatric behaviours, co-morbidities, health status.
� Environment: modifications to the bedroom and/or bathroom, professional involvement.
Better understanding of the factors associated with refusals of care in advanced dementia should illuminate the key areas to target to reduce refusals.
In this study, we aimed to examine the factors listed above to determine those associated with refusals of assistance with personal care in advanced dementia in both family (informal) and care-home (formal) settings. We hypothesised that there would be differences in refusals of care between care settings (care home and family) and between caregivers with dementia training and those without (some family caregivers access dementia training and some care-home staff do not). We theorised: (1) that as formal care staff in care homes were professional caregivers they would experience less refusals of care, and (2) that trained caregivers, irrespective of care setting, would experience less refusals of care.
Ethical approval was sought and given from the Queen's Square Research Ethics Committee, London (reference: 251339).

| Study design and participants
A cross-sectional study using informant-based measures was used.
Participants were recruited in dyads: a person with advanced dementia either supported at home or in a care home, aged 65 or over, receiving physical assistance with personal care activities such as washing or dressing and the family carer or care-home staff member who regularly assisted them. Family carers were the primary caregiver for the person with dementia and were over the age of 18. Care-home staff were usually assisting the person with dementia with their care at least six instances a week and were over the age of 18.
To determine the eligibility criteria of advanced dementia, the Frontier Dementia Rating Scale (FRS), 20 was used after consent, but before any other data collection. The FRS is a well validated 30-item dementia staging tool, completed through a caregiver interview, assessing changes in areas such as household chores, selfcare and behaviour.

| Recruitment
People with dementia in England not living in care homes and their family carers were identified from either a local research-team database or a national database called Join Dementia Research, where those registered had expressed an interest in being contacted about research opportunities. Written information was provided, and the first author discussed the research with potential participants.
Care homes in the UK offer 24-h accommodation, meals, and assistance with personal care either with or without qualified nursing staff. Care homes assisting people with dementia in the East of England were identified from the Care Quality Commission 21 database in the public domain. After a letter, follow-up telephone call and visit from the first author, eight care-home managers agreed for their care homes (two with qualified nursing and six without qualified nursing) to be involved in the study. Written information was provided to residents, residents' family members and care staff identified by managers as eligible and who might be interested. The first author was available to discuss the research.
Written consent or advice was obtained for all participants. For people with dementia assessed as lacking the capacity to decide about whether to take part, in line with the Mental Capacity Act of England and Wales 2005, 22 their family members or close friends provided written advice as to whether they would have been likely to decide to take part if they had capacity.

| Sample size calculation
Sample size was calculated a priori using GPower. 23

| Measures
Refusals of care were assessed through the RoCIS, 7 a 13-item, questionnaire developed and validated as part of the first author's programme of research 7 asking whether refusal behaviours such as verbal refusals, clamping the jaw or pushing the caregiver away were present during personal care interactions in the last month (higher scores indicate more refusal behaviours).
Caregiver type (care home or family) and dementia training (yes, no) were assessed with a bespoke questionnaire; these two factors were our hypothesised main exposures of interest.
Care setting was a grouping variable-either a care home or the family setting for each dyad. Each care home was defined as a cluster and all dyads from the family setting were assigned to one (family) cluster.
Demographic details such as age, gender and ethnicity, and psychotropic medication use and health status over the last month were taken using a bespoke questionnaire.

Neuropsychiatric symptoms were assessed using the 12-domain
Neuropsychiatric Inventory (NPI) Questionnaire (NPI for family carers, NPI nursing home version for care-home staff [NPI-NH]). 24,25 The domains constitute items such as anxiety, delusions and disinhibition and are rated by frequency and severity. Total frequency � severity score (0-144) was used, with higher scores indicating higher presence of neuropsychiatric symptoms.
Agitation was assessed using the Cohen Mansfield Agitation Inventory (CMAI), 26 a 29-item measure including items such as screaming, biting, and pushing. Total score was used , with higher scores indicating higher presence of agitation. Comorbidities such as liver disease, heart failure and cancer were assessed using the Charlson Comorbidity 19-item Index. 31 Total score was used with higher score indicating more comorbidities.
Caregiver confidence and training needs were assessed using a bespoke self-report questionnaire assessing confidence and perceived training needs in relation to different personal care activities. Two scores were calculated separately, with higher scores indicating more confidence or higher training needs.
Caregiver management style was assessed using the 28-item caregiver self-report Dementia Management Strategies Scale 32 which assesses caregiver management styles (criticism, active management, and encouragement) used over the previous month. A slightly adapted version was used changing the term 'my relative' to 'the person'. For this study, scores of each style were converted to percentages and the predominant management style coded for each caregiver.

| Analysis
Analysis was conducted in Stata17. 33 BACKHOUSE ET AL.

| Linear mixed model analysis
Screening via univariate linear mixed models suggested that dementia sub-type, gender, ethnicity, and caregiver management may  Our results point to two key factors for interventions aimed at refusals of care to consider: agitation and dependence in ADLs.

| Strengths, limitations, and suggestions for future research
This study provides insight into the factors associated with refusals