Cancer treatment decisions for people living with dementia: Experiences of family carers, a qualitative interview study

Abstract Background As the UK population ages, the prevalence of both dementia and cancer will increase. Family carers of people with dementia who are subsequently diagnosed with cancer are often involved in treatment decisions about cancer. These decisions are uniquely challenging. Objectives To explore the experience of carers involved in cancer treatment decisions for people with dementia. Design A cross‐sectional qualitative interview study with inductive thematic analysis. Setting and Participants Sixteen carers of people with dementia were identified via Primary Care Research Networks and the Join Dementia Research database. Results Three main themes were derived: ‘already at breaking point’, which describes the extreme strain that carers were already under when the cancer diagnosis was made; ‘maintaining the status quo’, which describes how despite the gravity of a cancer diagnosis, avoiding further dementia‐related deterioration was of prime importance; and ‘LPA’, which explores the benefits and frustrations of the use of lasting powers of attorney. Discussion Current services are ill‐equipped to deal with people who have a combination of dementia and cancer. Proxy decisions about cancer care are made in the context of carer stress and exhaustion, which is exacerbated by shortcomings in service provision. Conclusions As the prevalence of comorbid cancer and dementia rises, there is an urgent need to improve services that support carers with proxy health care decision‐making. Patient or Public Contribution The study design was codeveloped with a local dementia‐specific patient and public involvement (PPI) group. A project‐specific PPI group was formed with support from the Alzheimer's Society Research Partnership scheme to provide further bespoke input.


| INTRODUCTION
There are approximately 850,000 people living with dementia in the United Kingdom, with 95% of cases occurring in those over 65. 1 An estimated 700,000 people, usually family members, provide informal care for someone with dementia. 2 The prevalence of dementia in the United Kingdom is predicted to rise dramatically over coming decades as the population ages, with the greatest increase seen among those in the oldest age groups. 2 In addition to dementia, older people are at risk of developing other age-related comorbidities including cancer. A large English cohort study found that the odds of cancer patients also having dementia dramatically increases with age. 3 Recent analysis of UK GP records found that 7.5% of people over 75 with cancer also have dementia. 4 This represents a small but significant proportion of cancer patients. Dementia is associated with worse cancer outcomes across all measures, including all-cause mortality. [5][6][7] This reflects the link between dementia and comorbidities and the direct effect of dementia on survival and cancer-specific outcomes. People with dementia are less likely to partake in screening (for breast, bowel and cervical cancer) and less likely to report early symptoms or signs of cancer, which may instead be picked up at a later stage by a carer, contributing to worse outcomes. Once diagnosed, there is evidence that patients with a diagnosis of cancer and dementia are often treated less aggressively. 8 Any older person with a new cancer diagnosis must navigate a complex set of choices. They must balance their cancer pathology and available treatments against their own perception of their mental and physical condition, frailty and social situation and their personal views and priorities. [9][10][11] People with pre-existing dementia often do not have the capacity to make decisions of this complexity, and in these cases, where available, a family member is required to do so on their behalf.
Research examining the impact that making life-altering decisions by proxy has on family carers is mostly limited to issues regarding dementia diagnosis and subsequent dementia care, including end-of-life care. Proxy decision-making can be highly distressing, and fraught with guilt and uncertainty for carers. A lack of information and emotional support has been highlighted as a particular problem. [12][13][14] The specific experience of proxy decision-making regarding cancer treatment for people with dementia is almost wholly unexplored, yet as the prevalence of both cancer and dementia increases it is a challenge that will confront more and more people.
This study used a qualitative approach to explore the experiences of family carers who have been involved in making cancer treatment decisions on behalf of a relative living with dementia.

| Study design
A cross-sectional qualitative study.

| Participants
Family carers of people with dementia who had received a diagnosis of cancer (excluding squamous and basal cell skin cancers) within the last two years, where the carer had been required to make proxy decisions about cancer testing and treatment on the patient's behalf.

| Recruitment
A purposive sampling approach was used to achieve a maximum variation sample. 15 Participants were identified through two sources: Twenty GP practices across Yorkshire and the Humber that were registered with the National Institute for Health Research (NIHR) Clinical Research Network (CRN) Portfolio performed a database search for patients coded with dementia who had received a cancer diagnosis within the last 2 years. Carers of these patients who were also registered with the practice were sent an invitation letter from the practice outlining the study. Those who returned the reply slip were contacted to arrange an interview. Due to the sensitive nature of the topic, our patient and public involvement (PPI) group felt it was important that only people who wished to take part should be known to researchers, so no information regarding details of the patients contacted by their GP practices was held by the research team.
We, therefore, do not have data about the number of people who refused to take part in the study, or their reasons.
Participants were also recruited via the online NIHR database, Join Dementia Research. Four hundred and thirty volunteers who self-identified as dementia carers living within 50 miles of Sheffield were sent an email describing the study. Those who responded and met the inclusion criteria were contacted to arrange an interview.
Again, this meant data regarding participant refusal could not be collected as only a small number of those contacted would actually have been eligible.
Participants received a £15 shopping voucher to compensate them for their time. Recruitment was carried out between April 2019 and January 2020.

| Patient and public involvement
The study design was codeveloped with a local dementia-specific PPI group. The project idea and methodology were presented at a face-to-face group meeting with three people with dementia and three carers. The recruitment method was discussed in detail and they advised on the wording of the initial invitation letter that was sent. A project-specific PPI group was formed with support from the Alzheimer's Society Research Partnership scheme to provide further bespoke input. This consists of four female former carers of people with dementia. The initial project meeting was held face-to-face and focused on developing the interview topic guide. Following data collection, we met virtually (due to the COVID-19 pandemic) to discuss results and our interpretation of them.

| Ethics and research governance
Ethics committee approval was obtained from an MHRA compliant REC (REC number 18/NW/085). Research governance approval was granted by the University of Sheffield. All participants gave written informed consent.

| Interviews
The interview topic guide was developed with input from the research team and feedback from the project-specific PPI group.
Although no formal pilot interview was undertaken, we allowed concepts generated from initial interviews to inform and adapt the topic guide for subsequent ones. cared for) was also there. Telephone or email contact was made between C. H. and the interviewee before the interview to make arrangements. In some cases, the participant's experiences were briefly discussed but the interview schedule remained he same regardless of how much information was known before the interview. Participants were aware that C. H. was a GP trainee as well as a researcher.
Interviews were audiorecorded and transcribed verbatim. Field notes were made after the interview. Participants were not asked to review the transcripts.

| Analysis
Data were analysed inductively, using a reflexive, thematic analytical framework. 16 Initial analysis happened alongside data collection, so emerging concepts could be explored during subsequent interviews.

| Participants
Sixteen family carers were interviewed (one as a son/granddaughter diad) ( Table 1, participant characteristics). Data saturation was achieved.
All participants had cared for patients whose dementia was sufficiently advanced by the time they were diagnosed with cancer that they were unable to make decisions about treatment independently.
A wide range of decisions relating to cancer diagnosis and treatment were covered in the interviews, from consenting for investigations and treatment (surgery, radiotherapy and chemotherapy) to palliation.

| Themes
Three main themes were identified that describe the experience of  and if he goes the whole ward will be disrupted', but they didn't really take heed, so they saw my dad in full force.
And I was just sat there going 'I told you.'

P5 (female, father had bowel cancer)
Many cited the difficulties of ending up in 'specialty silos', where people with dementia were treated by professionals who only focused on a specialty-specific condition (their cancer), rather than adopting a holistic approach to managing their health.
A friend of mine who was a nurse said, 'the surgeon will be thinking with a surgeon's head on like you know, like "Oh well if I get this out, that will sort that out"'. LPAs were also a common cause of dispute between carers and health care professionals. There was often an apparent lack of knowledge and understanding within medical teams regarding the authority granted to attorneys, particularly if there was a disagreement between professionals and carers over the best course of action.
Dad The findings illustrate that relatives were often already at the limit of their capacity to cope with the burden of caring for someone with dementia before their loved one's cancer diagnosis. Decisionmaking was influenced by a need to maintain their relative's wellbeing and avoid provoking a deterioration in dementia-related behaviour. For some, their situation was so challenging that the cancer diagnosis was a relief that represented a possible escape from dementia.
Services across the NHS are ill-equipped to manage people with dementia who need to access care for other aspects of their health.
Exhausted carers are left trying to contain the situations that result, causing further stress. Carers cited a lack of knowledgeable support from professionals, poor understanding of their situation and failure to view the person with dementia holistically. LPAs were viewed as supportive tools but were frequently a source of conflict with medical staff, highlighting a lack of professional knowledge in this area.

| Strengths and weaknesses
This study is unique within the cancer-dementia sphere because participants were recruited from community settings rather than tertiary centres. This enabled us to capture a wide range of experiences encompassing different stages of dementia and at multiple points along the cancer journey. Data saturation was achieved through a socioeconomically diverse sample across Yorkshire and the Humber.
it provides an in-depth exploration of the views of a very hard-toreach group, which is underrepresented in research. 18 This sample lacks cultural diversity; all participants were White British. This reflects the demographics of the participating CRN GP practices and also that of people who are registered with JDR.
All CRN-registered practices across Yorkshire and the Humber were invited to participate, therefore this may also reflect a lack of capacity for practices who serve more culturally diverse populations to participate in research. As a result, the transferability of results to other cultural settings is limited and the authors recognize that these findings sadly do not address the under-representation of minority ethnic groups within dementia research. 19 It is also acknowledged that, while people affected by dementia were instrumental in the development of this project, people with cancer were not involved in our PPI work. It is possible that some aspects of dealing with cancer specifically were not addressed in as much detail as they could have been. As discussed in Section 2, due to the recruitment method we unfortunately do not have data regarding participant refusal or their reasons.
All interviews were carried out by the same person (a GP trainee), which is likely to have influenced the conduct of the interviews and responses of some participants (although they clearly still felt able to be critical of NHS services). This was addressed by a reflexive approach during data analysis, 20 having a second researcher conduct an independent analysis and by triangulating findings with the research team and the project-specific PPI group.

| Implications for research and practice
People with dementia are some of our most elderly and frail patients, yet when they have to access NHS services for physical health problems they experience a system that is not equipped to provide the holistic care they need. The findings of this study highlight a need to modify cancer care pathways so that they can accommodate the needs of people with dementia, from initial investigations and diagnosis all the way through to treatment and palliative care. Codesign of services with people with dementia and carers has the potential to result in services that better meet their needs.
Clear routes to support for carers are needed from the point of diagnosis throughout the disease process. There is a strong argument for such services being community-based rather than in secondary care, where they are better placed to target the needs of their specific populations. 32 However community services must be adequately and sustainably resourced and funded to provide the quality of support that people living with dementia and their carers so desperately need.
This study found that LPAs for health and welfare currently do not empower relatives to make decisions in the way they are intended. In fact, they were frequently a source of conflict between carers and HCPs. Carers were left feeling they had to fight for the care that they and their relative wanted, when their forward planning in preparing the LPA was intended to make such situations more straightforward. LPAs have existed in the United Kingdom since 2007 as part of the Mental Capacity Act. 33 Research to understand the barriers and facilitators to their appropriate use in health settings may be useful.

| CONCLUSION
The prevalence of comorbid cancer and dementia will continue to rise. Decisions about cancer care for people with dementia are complex and carers making these decisions by proxy must balance competing priorities in stressful and emotive circumstances. Services are currently ill-prepared to support them in this, exacerbating their distress.