‘The Future is Probably Now’: Understanding of illness, uncertainty and end‐of‐life discussions in older adults with heart failure and family caregivers

Abstract Background Earlier end‐of‐life communication is critical for people with heart failure given the uncertainty and high‐risk of mortality in illness. Despite this, end‐of‐life communication is uncommon in heart failure. Left unaddressed, lack of end‐of‐life discussions can lead to discordant care at the end of life. Objective This study explores patients' and caregivers’ understanding of illness, experiences of uncertainty, and perceptions of end‐of‐life discussions in advanced illness. Design Interpretive descriptive qualitative study of older adults with heart failure and family caregivers. Fourteen semi‐structured interviews were conducted with 19 participants in Ontario, Canada. Interviews were transcribed verbatim and content analysis was used to analyse the data. Main results Understanding of illness was shaped by participants’ illness‐related experiences (e.g. symptoms, hospitalizations and self‐care routines) and the ability to adapt to challenges of illness. Participants were knowledgeable of heart failure management, and yet, were limited in their understanding of the consequences of illness. Participants adapted to the challenges of illness which appeared to influence their perception of overall health. Uncertainty reflected participants’ inability to connect manifestations of heart failure as part of the progression of illness towards the end of life. Most participants had not engaged in prior end‐of‐life discussions. Conclusion Detailed knowledge of heart failure management does not necessarily translate to an understanding of the consequences of illness. The ability to adapt to illness‐related challenges may delay older adults and family caregivers from engaging in end‐of‐life discussions. Future research is needed to examine the impact of addressing the consequences of illness in facilitating earlier end‐of‐life communication.

Health Partners, and the Temmy Latner Centre for Palliative Care at Sinai Health System. The findings, interpretations and conclusions reported in this study solely reflect the opinions of the authors and are independent from the funders.

| INTRODUC TI ON
Chronic heart failure is a progressive illness characterized by periodic exacerbations, uncertainty in the illness trajectory, and high symptom burden that can lead to repeated interruptions in quality of life. 1 Heart failure affects over 26 million people globally. 2 Although survival has improved over the last decade, over 50% of patients still die within 5 years of being diagnosed. 3 Despite this evidence, it has been reported that <40% of people with heart failure are aware of their diagnosis. 4 Older adults experience additional challenges, such as with multimorbidity and frailty, that increase the complexity of their illness. The increased complexity further compounds uncertainty in illness. 5,6 The complexity of managing multiple illnesses and its unpredictable course has been found to contribute to a poor understanding of illness. Multiple reviews have found that most patients and caregivers do not receive prognostic information and have unmet information needs. 7-10 A poor understanding of illness can further lead to uncertain experiences for people with advanced illnesses such as heart failure. 11 Clinical interactions that are designed to improve understanding of illness and to identify individual's goals of care are infrequently integrated into chronic disease management. Endof-life communication, which includes goals-of-care discussions, aims to create a shared understanding of an individual's values and preferences among patients, caregivers and clinicians. 12 The integration of end-of-life communication early in the course of illness has been endorsed by multiple professional associations including the Canadian Cardiovascular Society, the American Heart Association, and the European Heart Association. [13][14][15] Yet, end-of-life communication still occurs late in the course of illness, if at all, for people with advanced illness such as heart failure. 16,17 Integration of end-of-life communication into heart failure management is one approach to address the uncertainties of illness and improve understanding. It also has the potential to shift the delivery of care from being disease-oriented to personcentred, thereby providing support to individuals based on their identified goals and values. 18 However, clinicians report lack of understanding of illness among patients and caregivers to be the most important barrier of end-of-life communication. 19 To overcome these barriers in clinical practice, there is a need to better understand patient and caregiver perceptions of end-of-life discussions within the context of their understanding of illness.
This study explores what patients and caregivers understand about their illness and how their understanding relates to their experience of uncertainty and end-of-life discussions in advanced illness.

| Study design and theoretical framework
This qualitative study was conducted following an interpretive description approach. 20 Interpretive description is a pragmatic methodology that aims to generate knowledge that is useful and applicable in clinical practice. 21 Interpretive qualitative research acknowledges the constructed and context-dependent nature in which health-related experiences form. 20,21 This study was further guided by the Reconceptualized Uncertainty in Illness Theory by Merle Mishel. 22 Mishel defines uncertainty as the inability to attach meaning to illness-related events (e.g. symptoms and hospitalizations) that provide the individual with information about their illness. 23 One's ability to process illness-related information depends on their prior experience of similar events and gained familiarity over time. It can also be influenced by their care team, which includes health-care providers and informal caregivers (e.g. family members), as well as their own cognitive capacity to process illness-related information.
These antecedents accumulate to an individual's ability to understand illness-related experiences. Uncertainty in illness develops when there are gaps in understanding of illness. For people with chronic illnesses, as experiences of uncertainty persist over time, the theory posits that prolonged uncertainty will push individuals to adjust to a life with uncertainty. The Reconceptualized Uncertainty in Illness Theory has been used in previous research of older adults with advanced illness. 11

| Setting and participants
Participants were recruited from an outpatient clinic in an academic health sciences centre in a metropolitan city in Ontario, Canada. A multidisciplinary team of clinicians with subspecialty training in heart failure management operates the specialized heart failure clinic. Purposeful sampling technique was used to recruit patients 65 years or older with a clinical diagnosis of advanced heart failure (New York Heart Association Class III/IV) and/or older adults over the age of 80 from the heart failure clinic (see Box 1). We specifically sought patients that meet these criteria as the challenges of illness are amplified in advanced stages of heart failure and as older adults often have multiple chronic conditions that challenge their understanding and management of illness. Patients who had already been referred to palliative care specialists were intentionally excluded.
This decision was made as we purposely sought to sample patients and caregivers earlier in the illness trajectory, before the end-of-life, to explore perception of earlier end-of-life communication in advanced illness, and because current practices, in Ontario, reflect late referral to palliative care for patients with advanced illness. 24 Family caregivers of patients, when available, were invited to participate in the study.
To identify study candidates, the lead author (JI) met two cardiologists. Candidates were sent information letters prior to their appointment to notify them of the study. On the days of candidates' appointments, at the end, the cardiologists sought permission from the patient, and when available their family caregiver, for a researcher to speak to them about the study. For interested candidates, JI explained the study purpose and asked for their participation in approximately an hour-long interview. Participants were given the informed consent form and offered a choice to be interviewed in the clinic or in their homes. In total, 23 candidates were approached, of them, 19 participated, two patients and one caregiver declined, and one candidate could not be approached due to cognitive decline.
After completing 14 interviews with 12 patients and seven family caregivers, saturation of key content was reached as determined by the lack of new data being generated on key phenomena of interest.

| Data collection
The data collection process began by gathering information on participant demographics (e.g. age, sex, living arrangement and chronic conditions). Semi-structured interviews with patients and family car-

| Data analysis
The initial analysis was performed inductively and simultaneously with data collection to exhaust the categories of interest using NVivo11 (QSR International). The in-depth analysis of data began after all interviews had been conducted. Content analysis was used to categorize data according to the specific research questions under inquiry. Content analysis allows for inferences to be made from the data in its context to provide insights on a phenomenon, and it is an analytic technique that is flexible to both inductive and deductive analysis. 25 Given the alignment between the research aim and the Reconceptualized Uncertainty in Illness Theory, a hybrid of conventional content analysis and a directed approach to content analysis was used. Conventional content analysis is typically used to describe and analyse data in the absence of a theoretical framework or when understanding of a phenomenon is limited, while directed approach to content analysis is used to extend conceptually a theoretical framework. 26 The in-depth analytical process began by deductively analysing the data into three broad categories: understanding of illness, uncertainty and end-of-life communication.
This process involved reviewing the transcripts iteratively to code the data as they related to the identified categories. Within these broad categories, the data were analysed inductively and codes were generated to gain a deeper understanding of the phenomena within and across the categories. Regular meetings were held between the lead and senior author (JI and KK) to discuss findings as they developed, and check-in meetings were held with all authors as analysis progressed. To verify the categories and interpretation of the data, the senior author (KK) reviewed three transcripts in depth, reviewed all codes and descriptions, the accompanying excerpts and interpretations, and posed questions to challenge the findings. This step was included to ensure that the interpretations were defensible and to increase the dependability of the findings.

| Ethics
Prior to data collection, the study was approved by the Research Ethics Boards at Mount Sinai Hospital (July 12, 2017). All participants provided consent for their interview to be recorded, transcribed, and for the findings to be shared through presentations and publications.

| Participant characteristics
Fourteen interviews were conducted with 19 participants (12 patients and 7 family caregivers). The mean age of patients was 82.5 years (SD = 6.4); the majority were male (58%), lived with their spouse (58%), and reported a mean of 5 chronic conditions (SD = 2.2) (see Tables 1 and 2). The mean age of caregivers was 67 years (SD = 13.7), and the majority were female (71%) ( Table 3).
The findings are organized into three categories: understanding of illness; illness understanding and experiences of uncertainty; and influence of uncertainty on end-of-life discussions.

| Understanding of illness
Participants' understanding of illness was shaped by two main phenomena: (i) illness-related experiences and (ii) adaptation to illnessrelated challenges.

| Illness-related experiences
Over the course of illness, patients accumulated a multitude of ill-

| Illness understanding and experiences of uncertainty
In spite of participants' knowledge of heart failure management, par-

| D ISCUSS I ON
This study explored what older adults with advanced heart failure and family caregivers understand about their illness and how it relates to experiences of uncertainty and perceptions of end-of-life discussions. Participants' understanding of illness was shaped by illness-related experiences and the ability to adapt to the challenges of illness. Adapting to challenges occurred regularly such that making modifications to living situations had become a normalized process. One example of this process of adaptation is the perception that hospitalizations are a routine part of heart failure management.
This raises a question as to whether the cumulative experience of being discharged from hospitalizations in heart failure distorts one's perception of illness such that patients and caregivers gain a false sense of recovery despite the progressive nature of illness and the threat of mortality. This may be a phenomenon that is unique to people who experience intermittent decline. Our findings corroborate prior studies [27][28][29] ; in a study of older adults with advanced heart failure, Klindworth and colleagues found that older patients do not perceive heart failure as a life-limiting illness. 28 In a longitudinal study of patients with advanced chronic obstructive pulmonary disease, an advanced illness with a similar illness trajectory to heart failure, Pinnock and colleagues found that patients reported having been 'restored to normal and the threat had receded' following an exacerbation. 29 Consequently, the patients with chronic obstructive pulmonary disease felt that the presence of death was not

| Alignment to the reconceptualized uncertainty in illness theory
Several findings from the study align with the Reconceptualized Uncertainty in Illness Theory. It appears that the accumulation of illness-related events (e.g. symptoms and hospitalizations) contributed to participants' understanding of illness-related experiences.
This was most notable in participants' ability to recognize the onset of heart failure exacerbations. Family members and informal caregivers, who provide support to patients, also appeared to reduce uncertainty as they supplemented patients' understanding of illness.
Participants' ability to adapt to illness-related challenges and further integrating the process of adaptation into patients' and caregivers' lives support the notion that individuals adapt to a new way of living upon experiences of prolonged uncertainty.

| Strengths and limitations
This The findings of this study should be interpreted within its limitations. All patients in this study had multiple chronic conditions; however, given the focus on heart failure to understand its unique illness trajectory, it is possible that the complexity of multimorbidity in participants' understanding of illness was not fully captured.
There is also a question regarding the extent to which participants shared their thoughts given the sensitivity of the topic explored.
Furthermore, this study did not try to discern whether or not participants were willing to articulate their understanding or truly lacked understanding of the consequences of illness. Nonetheless, the findings highlight aspects of illness that clinicians may need to proactively address in care, particularly as it can impact patients and caregivers' end-of-life experience. Our sample of caregivers is also small and diverse (i.e. includes spouses and children). A greater sample would be helpful and is needed to discern the different experiences of engaging in end-of-life discussions. The study participants were recruited from one specialized outpatient clinic where clinicians are trained in heart failure management as well as teach patient education to their patients and caregivers. All but one participant was Caucasian, which also does not reflect the entire heart failure population. This is, in part, a reflection of the inclusion criteria that excluded patients who are not fluent in English. How these findings would hold among a culturally diverse sample of patients is unknown, as previous research has found variations in end-of-life preferences in different populations. All of these factors limit the transferability of the study findings.

| CON CLUS ION
This study explored older adults with advanced heart failure and

ACK N OWLED G EM ENTS
We would like to acknowledge and thank Dr. Rob Fowler and Dr.
Patricia Strachan for closely reviewing and commenting on the thesis from which this article originates.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data cannot be shared publicly because of restrictions found in the research ethics approval. Study participants did not provide consent to have transcripts of their interviews to be shared publicly.
Deidentified data for researchers who meet the criteria for access to confidential data may be requested from the Mount Sinai Hospital Research Ethics Board. For more information, please contact Jennifer.im@mail.utoronto.ca.