Engagement in Advance Care Planning: The Association With Cognitive Impairment and Care Preferences

Abstract Although there has been growing evidence that Advance care planning (ACP) benefits people with cognitive impairment nearing death, our understanding about this issue is still limited. This study examines whether cognitive impairment is associated with ACP engagement and end-of-life care preferences among older adults in the U.S. Using data from the 2012 National Health and Trends Study (n=1798, aged 65 to 101), we identified four levels of ACP engagement: None (28%), Informal ACP conversation only (12%), Formal ACP only (14%), and Both informal and formal ACP (46%). Older adults with None showed the highest prevalence of having cognitive impairment (17%), followed by those with Formal ACP only (15%) and the other two (6%, 6%). The results of Multinomial Logistic Regression showed that, compared to those without, respondents with cognitive impairment had 143% increased relative risk of having None (RR = 2.43, CI: 1.58-3.73) and 81% increased relative risk of completing Formal ACP only (RR = 1.81, CI: 1.11-2.95) relative to completing Both informal and formal ACP. In addition, respondents with None were more likely to prefer to receive all treatments available nearing death than those with any ACP engagement. Achieving high quality care at the end of life can be more challenging for older adults with cognitive impairment and their family caregivers due to the limited capacity. Although encouraged, informal ACP conversation with loved ones does not necessarily occur before the formal ACP, especially, for those with cognitive impairment. Therefore, they may merit more attention such as early ACP engagement.


CONVERSATIONS ABOUT DEATH AND DYING, END-OF-LIFE CARE PLANS AND PREFERENCES BETWEEN AGING PARENTS AND ADULT CHILDREN
Hyo Jung Lee, and Jacobbina Jin Wen Ng, Nanyang Technological University, Singapore, Singapore This study aims to investigate whether attitude and perception on late-life death and dying, end-of-life care plans and preferences could be better understood from current values shared between aging parents and their adult children in the multi-cultural city-bound country, Singapore. We are in the process of interviewing 20 aging parent-adult child dyads. Up to date, six semi-structured interviews were completed and transcribed. We performed Content analysis to analyze the transcripts. Preliminary findings showed that both aging parents and adult children rarely discussed this issue, although parents had their own plans or preferences. The major barriers against open conversations about death and dying of aging parents include: the perception of not-yet time to talk about this issue (without knowing when the right time is) and tendency to have conversations about death in tandem with finances, but not death itself. Although specific end-of-life care plans or arrangements were not thought out thoroughly, aging parents expressed a high level of trust and reliance on close family members' decisions regarding their end-of-life care. They tended to agree on joint decision-making process within family, even though adult children had no or unmatched ideas about their aging parents' end-of-life wishes. This did not necessarily align with previous findings in Western countries, underscoring individuals' control over their own death and dying process. Open conversation within family, family-involved advance care planning, or joint decision-making processes may be warranted to promote quality of life and death in older Singaporeans and well-being of their family members of all ages.

END-OF-LIFE DECISION MAKING IN CHRONIC LIFE-LIMITING DISEASE: A CONCEPT ANALYSIS AND CONCEPTUAL MODEL
Kristin Levoy, 1 Elise Tarbi, 1 and Joseph De Santis, 2 1. University of Pennsylvania,Philadelphia,Pennsylvania,United States,2. University of Miami,Coral Gables,Florida,United States Most deaths from chronic life-limiting diseases are preceded by end-of-life decisions, yet patients and caregivers are ill-equipped to make them. The lack of a common vocabulary surrounding end-of-life decision making and the paucity of conceptual models that explicate its components hamper improvements in clinical practice and research. Walker and Avant's method for concept analysis was used to investigate uses of "end-of-life decision making" in the literature in order to identify its components (antecedents, attributes, consequences), describe stakeholder roles (patients, family/ caregivers, health care providers), and develop a conceptual model. An iterative search strategy resulted in 143 included sources. These encompassed 1,127,095 patients (primarily older adults), 3,384 family/caregivers, and 588 healthcare providers. Evidence revealed that end-of-life decision making is a process, not a discrete event. This begins with preparation (antecedents), which involves the designation of a decision maker and iterative patient, family/caregiver, and healthcare provider communication across the chronic illness. These preparatory processes inform end-of-life decisions, which possess three attributes: 1) serial choices in the terminal illness phase that are 2) weighed in terms of their potential outcomes 3) through patient and family/caregiver collaboration. The components impact patients' death experiences, caregivers' bereavement, and healthcare systems (outcomes). The resulting conceptual model highlights the larger context of preparation (beyond advance care planning) and the prominent role of caregivers in the end-of-life decision making process. Enhanced measurement must account for the dose, content, and quality of the preparation and decision components that collectively contribute to outcomes, which holds implications for practice improvements and research.

ENGAGEMENT IN ADVANCE CARE PLANNING: THE ASSOCIATION WITH COGNITIVE IMPAIRMENT AND CARE PREFERENCES
Hyo Jung Lee, 1 and Giyeon Kim, 2 1. Nanyang Technological University, Singapore, Singapore, 2. Chung-Ang University, Seoul, Republic of Korea Although there has been growing evidence that Advance care planning (ACP) benefits people with cognitive impairment nearing death, our understanding about this issue is still limited. This study examines whether cognitive impairment is associated with ACP engagement and end-of-life GSA 2020 Annual Scientific Meeting Innovation in Aging, 2020, Vol. 4, No. S1 care preferences among older adults in the U.S. Using data from the 2012 National Health and Trends Study (n=1798, aged 65 to 101), we identified four levels of ACP engagement: None (28%), Informal ACP conversation only (12%), Formal ACP only (14%), and Both informal and formal ACP (46%). Older adults with None showed the highest prevalence of having cognitive impairment (17%), followed by those with Formal ACP only (15%) and the other two (6%, 6%). The results of Multinomial Logistic Regression showed that, compared to those without, respondents with cognitive impairment had 143% increased relative risk of having None (RR = 2.43, CI: 1.58-3.73) and 81% increased relative risk of completing Formal ACP only (RR = 1.81, CI: 1.11-2.95) relative to completing Both informal and formal ACP. In addition, respondents with None were more likely to prefer to receive all treatments available nearing death than those with any ACP engagement. Achieving high quality care at the end of life can be more challenging for older adults with cognitive impairment and their family caregivers due to the limited capacity. Although encouraged, informal ACP conversation with loved ones does not necessarily occur before the formal ACP, especially, for those with cognitive impairment. Therefore, they may merit more attention such as early ACP engagement.

. Chapters Health System, Tampa, Florida, United States
Stroke is the second leading cause of death globally among people aged 60+, yet only 9% of hospice decedents have a primary diagnosis of stroke and little research has examined their end-of-life care experiences. Late referral and admission to hospice is an indicator of poor end of-life-care quality. This project identified factors predicting short stays in hospice, defined as a length of stay (LOS) of 14 days or less, via chart review of 100 hospice patients with a primary diagnosis of stroke. Of the 98 patients with complete data, 89% died in hospice; 11% were live discharges. Most patients were female, married, and referred to hospice from a hospital. Only 21% of patients entered hospice with a completed advance directive. Approximately 75% (n=73) of patients had a short LOS. Binary logistic regression indicated that gender, race, marital status, and having an advance directive at hospice admission were unassociated with LOS. Controlling for demographics, patients referred from home were 74% less likely to have a short LOS (OR=.26, CI=.08-.83) compared to those referred from a hospital, with a mean LOS of 44 and 13 days, respectively. Including Palliative Performance Scale (PPS) score in the model attenuated the effect of referral location. For every ten percent increase in PPS, participants were 85% less likely to have a short LOS (OR=.15, CI=.07-.32). Advance care planning should be more widely promoted among patients at high risk for stroke. There are opportunities for earlier referral to hospice for patients with a diagnosis of stroke.

HOME HOSPICE VERSUS OTHER PLACES OF DEATH AND CAREGIVER STRESSES Elizabeth Luth, and Holly Prigerson, Weill Cornell Medical College, Larchmont, New York, United States
As of 2017, more individuals in the U.S. die at home than in any other location. Hospice care was designed to provide support for people who are dying and their families. However, dying persons may have rapidly emerging needs that home hospice does not immediately meet, thereby, exposing family members to be "first responders." Thus, home death may result in distress and burden for dying individuals' family members, even when hospice is involved. This study uses multivariable regression analysis to explore the relationship between place of death and stressful end-of-life experiences in a sample of 185 patients with advanced cancer. We also analyze which end-of-life experiences are associated with death location. Compared to home hospice death, we found dying in a hospital was associated with fewer caregiver exposures to, and reports of fearfulness and helplessness in response to, stressful end-of-life events. Compared to home hospice death, hospital death was associated with decreased frequency of choking, falls, confusion/delirium, and feeling the patient has had enough. It was also associated with less fearfulness about choking and falls and less hopelessness about falls, the patient having enough, and thinking the patient was dead. Our results suggest home death with hospice care, often involves undesirable experiences that result in more caregiver fear and helplessness than dying in a hospital without hospice care. Research is needed to understand how to best support family members through stressful end-of-life experiences even when supported by home hospice services. Reflection on memorialization may differentially influence nomination and narration of self-defining memories across the lifespan, including the extent to which positive character strengths (i.e., virtues) are represented. We investigated characteristics of self-defining memories across adulthood and in the context of being memorialized after death. Young, middle-age, and older adults were randomly assigned to narrate a memorialization-based self-defining memory (a memory to describe you after you are gone; n = 103) or current self-defining memory (a memory to describe your current self; n = 99). Participants rated qualities of their memory (e.g., personal significance) and the extent to which the memory represents them as virtuous (e.g., courageous, empathic, etc.). There were no age or condition differences in personal significance of the memory narrative (p = 0.43). However, there was an age-by-condition interaction for representations of virtue described in the self-defining memory, F(2, 199) = 3.94, p = 0.002. Young adults rated their selfdefining memories as more virtuous in the memorialization condition than in the current self condition (p = 0.001).