Clinicians are integral to end-of-life (EOL) care decision-making and play a critical role in shaping the experiences of older adults and their families during the final days of life. This qualitative interpretative meta-synthesis (QIMS) examined the existing published qualitative literature to synthesize the experiences and perspectives among clinicians in the provision of EOL care to older adults. The search yielded 2822 potentially relevant articles, and 137 duplicates were eliminated resulting in a final sample of six articles. Four new themes were developed: Theme 1: Ambiguity and negotiation in end-of-life communication and decision making; Theme 2: Provider Emotional and Ethical Burden; Theme 3: Tensions with age-based treatment and perceived worth; and Theme 4: Systemic Constraints and Institutional Barriers in End-of-life care. The findings highlight that to mitigate and improve the EOL care experience for older adults, it is imperative to implement system-level reforms that support proactive communication, reduce implicit age-related assumptions, and provide ethical solutions for clinicians.
Background Patient-reported outcomes (PROs) are crucial for accurate patient assessments. In palliative care, the Integrated Palliative care Outcome Scale (IPOS) is a widely used measure of these outcomes. Nevertheless, manually entering PRO data into electronic medical records or databases can be burdensome. Thus, a previous study developed models to automatically extract IPOS scores from clinical conversation data using speech recognition and traditional natural language processing. However, further improvements in performance are required. This study examined the accuracy of extracting IPOS scores from clinical conversation data using the latest large language models (LLMs). Methods We collected conversation data related to IPOS from 100 patients receiving home care at a clinic in Japan between February and May 2023. Two LLMs, Gemini and Copilot, were used to extract IPOS scores. Twenty percent of the data were used to optimize the input instructions for the LLMs, and the remaining 80% were used for performance evaluation. The primary performance metric was the macro-F1 score. Results The mean macro-F1 scores ± standard deviation (SD) for Gemini were 0.96 ± 0.08 for physical symptoms, 0.91 ± 0.11 for emotional symptoms, and 0.84 ± 0.05 for communication/practical issues. For Copilot, the corresponding scores were 0.94 ± 0.08, 0.86 ± 0.10, and 0.85 ± 0.12, respectively. Conclusion These findings suggest the technical feasibility of LLMs for extracting IPOS scores from clinical conversation data, although further studies are needed to confirm this finding. Integrating speech recognition and LLMs may have significant potential to reduce the burden of sharing PRO-related information.
BackgroundTransgender and gender-diverse individuals experience higher mortality rates and elevated incidence of serious illness compared to cisgender populations. Despite these disparities, discrimination and stigma continue to create barriers to high-quality hospice and palliative care.AimsThis scoping review identified the unmet needs of gender-diverse adults receiving hospice and palliative care in the United States.MethodsFollowing PRISMA-ScR guidelines, we systematically searched seven databases between May and June 2025. Inclusion criteria required: English-language publications from US settings, exclusive focus on gender-diverse adults, and examination of unmet needs in serious illness, hospice, or palliative care contexts. Two reviewers independently screened titles, abstracts, and full texts.ResultsFive publications met eligibility criteria. A descriptive thematic analysis identified four categories of unmet needs: (1) competent healthcare professionals with gender-affirming training; (2) holistic, inclusive care delivery models; (3) family and chosen family engagement protocols; and (4) protective institutional policies and legal frameworks. Critical limitations include a small sample size; predominance of healthcare professional perspectives rather than transgender patient voices; reliance on descriptive and narrative-based approaches; and absence of intersectional analyses examining race, class, and disability.ConclusionsThis review exposes substantial under-resourcing of research on gender-diverse patients' end-of-life needs. Urgent priorities include systematic competency training for hospice and palliative care professionals, development of evidence-based clinical guidelines, advocacy for protective legal frameworks, and most critically, conducting robust qualitative and quantitative research centering transgender patients' lived experiences and priorities in serious illness care.
BackgroundCare partners of people with dementia have many responsibilities, including providing care at end-of-life (EOL). Discussions about EOL are often delayed until late in the course of disease, if they occur at all.ObjectiveLearn about the communication and support that care partners need regarding EOL.MethodsWe conducted 27 interviews with care partners and analyzed these data using the Information, Motivation, and Behavioral Skills Model. We analyzed interview transcripts to deductively develop a codebook, and then collapsed codes until inductive themes emerged which aligned under the Information, Motivation, and Behavioral Skills Model.ResultsWe found that clinician-initiated discussions and a positive perception of the clinician-care partner relationship contributed to care partners' willingness to communicate. Care partners usually waited for clinicians to bring up EOL topics and felt apprehensive about these discussions, which hampered their ability to communicate effectively about EOL. Information-seeking was constrained by uncertainty about timing, lack of private discussion opportunities, and perceived limited personalization. Despite recognizing their support needs, care partners found few opportunities to discuss EOL due to the clinician's focus on immediate clinical concerns.ConclusionsEarly clinician-initiated communication and relationship building could improve EOL care. In particular, developing opportunities for private discussions between care partners and clinicians may reduce apprehension and facilitate meaningful EOL conversations.
ObjectivesEmergency physicians frequently encounter patients with serious illness, yet training in palliative care competencies remains inconsistent during emergency medicine (EM) residency. Simulation-based education may provide a method for integrating palliative care principles into EM training. The objective of this study was to qualitatively examine EM residents' perceptions of the intersection between palliative care and emergency medicine following participation in palliative care-focused simulation scenarios.MethodsWe conducted a qualitative pilot study using thematic analysis of semi-structured interviews with EM residents who participated in palliative care-focused simulation scenarios at an urban Level I trauma center. Post-simulation interviews explored residents' prior experiences, perceived knowledge gaps, emotional responses, and challenges related to palliative care in the emergency department. Interviews were transcribed and analyzed inductively by a multidisciplinary research team.ResultsEleven of 51 eligible residents (21%) participated. Primary thematic domains included perceived knowledge deficits, emotional responses to care delivery, and prior clinical experiences. Residents reported uncertainty regarding core palliative care competencies, including interpretation of advance directives, documentation of do-not-resuscitate/do-not-intubate orders, and clinical decision-making during peri-arrest and resuscitation scenarios. Emotional responses included hesitancy and discomfort during high-stakes decision-making, particularly in the setting of limited information.ConclusionsEmergency medicine residents reported frequent exposure to palliative care issues in clinical practice but variable preparation and confidence in navigating these encounters. Palliative-focused simulation elicited perceived knowledge gaps and emotional responses relevant to emergency care delivery. Simulation-based education may offer a structured approach to integrating palliative care competencies into EM training.
Introduction Patients often express a desire to discuss their end of life with physicians, yet patient-provider communication regarding this topic is often poor. Insufficient communication may prevent dying older adults from receiving care that aligns with their preferences. Physicians and informal caregivers may assess the quality of their end-of-life communications differently; however, it is unclear whether these misalignments undermine their quality of care. Methods We conducted semi-structured focus groups with 19 physicians who routinely provide end-of-life care and 21 recently bereaved informal caregivers of older adults. We analyzed transcripts using an inductive, line-by-line coding approach to generate themes grounded in participant experiences. Results We identified two overarching themes highlighting differences in physician and caregiver perspectives regarding end-of-life communication: roles and responsibilities; and perceived barriers to end-of-life planning. Physicians attributed ineffective communication to patients’ poor health literacy, and felt that in-depth end-of-life planning conversations were outside their scope of practice. Bereaved caregivers attributed poor communication to their fear of challenging medical authority, and limited information exchange with physicians. Physicians emphasized financial barriers to end-of-life planning, whereas caregivers emphasized their unmet needs for physician guidance. Conclusions Physicians and bereaved caregivers differed in how end-of-life communication is experienced, evaluated, and acted upon, which exacerbated pre-existing informational barriers for caregivers and created an additional obstacle to achieving goal-concordant care. Physicians’ overestimation of patients’ baseline end-of-life knowledge coupled with caregivers’ unaddressed informational needs and difficulty challenging medical authority at the end of life may limit the quality and delivery of end-of-life care that aligns with patient preferences.
ObjectivesHospitalists are responsible for every facet of inpatient care. Given the wide-reaching nature of their role, it is imperative hospitalists are equipped with skills to successfully navigate serious illness conversations. However, there is no standard communication training that clinicians receive. The aim was to implement and evaluate perceptions of a serious illness conversation program.MethodsThe study was conducted at one academic medical center with hospital medicine physicians. They participated in an 8-hour interactive workshop rooted in adult learning theory, delivered via large group didactics and demonstrations, and incorporated deliberate practice in small groups with simulated patients. Standard cases were adapted to reflect real-world clinical scenarios hospitalists routinely encounter.ResultsOf the 103 respondents, 90 (87.4%) strongly agreed or agreed the workshop was a good investment of their time. For recommending this program to their peers, 93 (90.3%) strongly agreed or agreed. Similarly, 99 (96.1%) strongly agreed or agreed they would use the communication skills in their daily practice. The written feedback was overall positive, citing the role play skills practice was helpful.ConclusionsThe program was well-received and became a required component for new hire divisional onboarding. Future steps include measuring the longitudinal impact on patients and clinicians.
PurposeFamily caregivers of terminal cancer patients experience anticipatory grief. Blended narrative nursing interventions remain insufficiently evaluated in palliative care. This single-group pre-post study examined changes in anticipatory grief following a four-week blended narrative nursing intervention and secondarily explored caregiver characteristics associated with the magnitude of change.MethodsSixty-three caregivers were recruited from five Chinese hospice and palliative care institutions; 55 completed the intervention and were included in the pre-post analyses. The intervention combined daily asynchronous online narrative activities with weekly in-person group workshops facilitated by nurses. Anticipatory grief was assessed using the 27-item Anticipatory Grief Scale at baseline (T0) and post-intervention (T1). Pre-post changes were examined using paired-samples t tests with Bonferroni correction and the absolute values of within-subject Cohen's dz (|dz|). Exploratory stepwise multiple linear regression examined caregiver characteristics associated with reduction scores. Engagement was summarized.ResultsStatistically significant pre-post reductions were observed across all seven anticipatory grief subdomains (all P < .001), with absolute within-subject Cohen's dz values ranging from 0.58 for anxiety to 2.08 for anger. Mean online completion was 78.5%; mean workshop attendance was 3.6 of 4 sessions. Higher caregiver-patient intimacy was associated with smaller reductions in five subdomains, whereas extraversion, religious affiliation, and socioeconomic characteristics were associated with larger reductions in specific domains.ConclusionsSubstantial pre-post reductions in anticipatory grief were observed following the intervention. However, the single-group design precludes causal attribution. Controlled studies with longer follow-up and prespecified moderator analyses are needed to evaluate intervention effectiveness and response heterogeneity.
Climate change is a defining global health challenge of our time, acting as a threat multiplier that erodes decades of health gains and worsens social inequities. People living with serious illness, the population served by Hospice and Palliative Care (HaPC) professionals, are particularly vulnerable to these impacts. While leading health organizations have acknowledged the public health implications of climate change, HaPC has remained largely absent from this discourse. This narrative review and framework proposal situates the field's competencies within the growing climate and health literature and identifies avenues through which HaPC can contribute to addressing climate change. We argue that HaPC professionals - drawing on strengths in serious illness communication, interdisciplinary collaboration, and person-centered care - are uniquely positioned to advance equitable climate adaptation and mitigation at the individual, organizational, and societal levels. Through climate advocacy and climate-conscious care, HaPC professionals can protect patients' health while helping shape healthcare's broader response to climate change.
BackgroundAdvance care planning (ACP) aligns treatment pathways with patient preferences. Many acute care patients are admitted without advance care plans (AC-Plans) which may result in patients receiving care misaligned with their preferences. Accordingly, ACP is needed within hospitals, however, implementation barriers exist.ObjectiveTo evaluate an ACP program developed by a 711-bed hospital to overcome many of the barriers to implementing hospital-based ACP. Success was measured as the percentage of patients with AC-Plans at discharge over the four-year study.MethodsThe hospital's ACP program components included: primary palliative care, simulation-based training, and ACP and AC-Plan training and education. The sample consisted of patients aged 65 years and older who were admitted without documented AC-Plans, at the hospital, between November 1st 2021 to October 31st 2025. Primary evaluations consisted of examining yearly trends in the percentage of patients with AC-Plans at discharge with Kendall's tau and testing association between admission year and the likelihood of having an AC-Plan at discharge with a logistic regression.ResultsThe sample consisted of 25,189 patients. The percentage of patients with ACPs at discharge increased from 4.4% to 19.1% from the first to fourth program year. The odds of having an AC-Plan at discharge in years 2, three, and four were 3.44 (P < 0.001), 5.87 (P < 0.001), and 8.97 (P < 0.001) times higher than the odds of having one during the first program year.ConclusionThe growth and sustainability of the hospital's ACP program demonstrates that barriers to ACP within hospitals can be overcome and successful programs implemented.
IntroductionJuvenile Huntington's disease (JHD) is a rare, severe neurodegenerative disorder with early onset, high symptom burden, and a predictable trajectory toward advanced disability and premature death. Despite this, goals-of-care (GOC) conversations and advance care planning are often delayed, leading to reactive, crisis-driven decisions and increased emotional burden for families and clinicians.Case PresentationWe present a 30-year-old woman with genetically confirmed JHD, diagnosed at age 14, who declined progressively over more than 15 years. Despite longstanding neurological follow-up and repeated hospital admissions, palliative care referral occurred only 20 days before death. At evaluation she had complete dependence, profound cognitive impairment, severe dystonia, advanced dysphagia with gastrostomy, and recurrent aspiration. The absence of prior GOC discussions created uncertainty regarding hospital transfer, antibiotics, artificial nutrition and hydration, and palliative sedation. After retrospective review, interdisciplinary discussion, family meetings, and ethics consultation, the palliative care team established a comfort-focused plan that avoided further transfers, limited nonbeneficial interventions, and prioritized symptom control. Subcutaneous opioids, benzodiazepines, and anticholinergics relieved distress, rigidity, and secretions. The patient remained in the nursing facility until death.DiscussionLate palliative care integration limited anticipatory planning and shifted complex decisions into a stage with little room for deliberation, when cognitive and clinical deterioration constrained meaningful participation.ConclusionsJHD offers a useful model for early palliative care integration. Earlier outpatient referral and closer neurology-palliative care collaboration may enable longitudinal GOC conversations, anticipate complex decisions, and reduce the clinical, emotional, and ethical burden of advanced disease.
Background Xerostomia and thirst are among the most prevalent and distressing symptoms in palliative care, yet the effectiveness of relieving interventions in this population remains uncertain. Objective To evaluate the effectiveness of symptomatic interventions on the intensity of xerostomia and thirst in adults receiving palliative and end-of-life care. Design Systematic review with synthesis without meta-analysis (SWiM), reported per PRISMA 2020. Methods We searched five databases from inception to June 2026 for randomized controlled trials. Two reviewers independently selected studies, extracted data, and assessed risk of bias (RoB 2). Given the small number of studies and incomplete reporting of dispersion measures, data were not imputed; results were synthesized narratively and structured by comparator type. Results Six trials (235 participants) were included. No outcome had sufficient data for meta-analysis. Of five trials evaluable for xerostomia, two favored the intervention, three showed no clear difference, and none favored the comparator. Benefit clustered in comparisons against usual care, whereas placebo-controlled trials showed no advantage of the active agent. Risk of bias was high in three trials and raised concerns in the remainder; certainty of evidence was very low. Conclusions Evidence is insufficient for quantitative synthesis and of very low certainty. Symptom improvement appears more attributable to structured oral care than to any specific agent. Larger, better-reported trials are needed, and thirst remains markedly understudied.
While falls among older adults have been studied extensively in various medical settings, less is known about falls occurring during home hospice care when care is provided by informal caregivers. A retrospective mixed methods design was used to examine patients' and caregivers' experiences with falls, including fall frequency and characteristics, contributing factors, and fall consequences and corrective measures. A data registry of 535 hospice patients who received care between 2004-2020 was queried to identify patients that experienced one or more falls during their stay in one of 4 residential care homes (RCHs). A total of 81 patients (15.1%) experienced one or more falls, and there were 119 falls across 13,860 days of care or 8.6 falls per 1000 days of care. The majority of patients who fell (65%) did so just once, and for most of the falls there were either no injuries (50%) or minor injuries (44%). Only two falls (6%) resulted in hospitalization. In the 24-hour period prior to patients' falls, opiates were shown to be the most frequently administered medication followed by benzodiazepines, with higher rates of falls among patients who consumed both of these medications in combination. Most falls occurred in the bedroom or bathroom and were related to toileting needs. Results suggest that informal caregivers may benefit from additional training when supporting hospice patients' toileting needs, especially when they have received opiates and benzodiazepines. Results illustrate caregivers' efforts to balance fall risk with patients' desire for autonomy as a central element of home hospice care.
Background Suffering in advanced illness is multidimensional and often includes existential and spiritual distress. Brief interventions that can be tolerated at the bedside may be useful in palliative care, but ultra-brief gratitude-focused mindfulness has not been formally evaluated in this setting. Objective To determine whether a single 5-minute mindfulness-of-gratitude practice reduces suffering and improves spiritual well-being among palliative care inpatients compared with attention-matched supportive listening. Methods In this parallel-group, open-label randomized controlled study, adult palliative care inpatients with an Overall Suffering Score ≥4/10 were randomized 1:1 to a scripted 5-minute mindfulness-of-gratitude practice or a 5-minute supportive-listening conversation, both delivered by the same trained clinician. Overall suffering, total suffering, and spiritual well-being measured using the Functional Assessment of Chronic Illness Therapy–Spiritual Well-Being Scale (FACIT-Sp12) were assessed immediately before and after the session. Results Forty patients were randomized, with 20 in each arm, and all completed the protocol. Compared with supportive listening, the mindfulness-of-gratitude arm showed greater reductions in overall suffering (median change −2.5 vs −1.0; P < 0.001) and total suffering (−2.0 vs −1.0; P = 0.037), and greater improvement in FACIT-Sp12 score (+2.0 vs 0.0; P = 0.001). No adverse events were reported. Conclusions A single 5-minute mindfulness-of-gratitude practice produced immediate reductions in suffering and improvements in spiritual well-being. This brief, low-risk intervention may be a practical adjunct to routine inpatient palliative care, although durability and generalizability require further study.
BackgroundAccurate prognostication during the actively dying phase remains a major challenge in palliative care, particularly in home hospice settings where bedside assessment plays a central role. Neurological deterioration preceding death may provide clinically useful indicators of impending death; however, its temporal progression has not been well characterized.MethodsWe conducted a prospective observational study of terminal cancer patients receiving home hospice care between April 2022 and March 2025. A total of 236 patients who died at home were included. Neurological signs-including newly developed neuropathic pain, urinary retention, loss of facial expression, dysphagia, neck rigidity, head rotation, seizures, impaired consciousness, open mouth posture, hyperpyrexia, apnea episodes, and pupillary dilation-were prospectively recorded. The timing of each neurological sign relative to death was documented.ResultsNeurological signs appeared in a recognizable temporal sequence preceding death. Urinary retention appeared earliest, with a median onset of 4 days before death (interquartile range [IQR], 1-7 days), followed by newly developed neuropathic pain at 3 days (IQR, 1-5 days). Loss of facial expression, dysphagia, neck rigidity, head rotation, seizures, and apnea episodes typically occurred approximately 2 days before death. Impaired consciousness, open mouth posture, hyperpyrexia, and pupillary dilation occurred closer to death, with median onset times of 1 day before death. Loss of facial expression was the most frequent neurological sign (96.6%).ConclusionsNeurological deterioration follows a recognizable temporal pattern in terminal cancer patients receiving home hospice care. Systematic assessment of these neurological signs may support bedside prognostication, facilitate anticipatory guidance, and help families prepare for the dying process at home.
Background Research has shown that people who consider themselves religious have a lower advance directive (AD) completion rate. However, advance care planning (ACP) includes a broader range of activities that are important in determining medical treatments. Little has been published on the relationship between spirituality, religion, and engagement in ACP activities. Objective We examined whether religiosity and spirituality are associated with ACP engagement. Methods This is a secondary analysis of interview data collected at baseline during a randomized controlled trial of a nurse-led ACP intervention. Measures included the 5-item Duke University Religious Index (DUREL), a measure of spirituality, and the 15-question ACP Engagement Survey, which examines engagement in multiple aspects of ACP. Results We found a significant association between higher scores on each domain of the DUREL and higher ACP Engagement scores. In separate multivariable models adjusting for demographic factors, the relationship between greater religiosity (higher total DUREL scores) and greater ACP Engagement scores was significant ( P < .0001 for each model). Identifying as spiritual was significantly associated with greater ACP Engagement scores ( P = .0038) in adjusted multivariate models. Conclusions In contrast to prior findings of lower AD completion in more religious individuals, this study found that higher religiosity and spirituality are associated with higher ACP engagement. These results suggest ACP opportunities offered for religious congregations or spirituality focused communities may be especially successful.
IntroductionHospital readmissions in the US place a considerable burden on patients and their caregivers. Our study will investigate whether spiritual intervention during a patient's hospital stay contributes to lower rates of readmissions 30 days after hospital discharge.MethodsTwo hundred patients receiving palliative care will be randomized into control (n = 100) and intervention groups (n = 100). The control group will receive a standard spiritual support visit, while the intervention group will receive a spiritual intervention consisting of active listening, compassionate presence, assessment of post-discharge resources, and a supportive phone call 5 to 7 days after discharge. The primary outcome will be hospital readmission within 30 days of discharge. Board-certified chaplains will be allowed to use their individual skillsets but within the standardized ACA spiritual care model. Training chaplains to assess patients in the same way will help with this discrepancy as the same form will be used.DiscussionWe hypothesize that patients who receive spiritual intervention with a supportive phone call will experience fewer hospital readmissions. Spiritual intervention will instill.confidence in those patients at risk of readmission by assuring them that the hospital staff remain concerned about their well-being even after discharge.
ObjectiveThis systematic review synthesizes randomized controlled trial evidence on the effectiveness of nurse-led ACP interventions for patients with advanced cancer.MethodsWe searched PubMed, Embase, CINAHL, Web of Science, and the Cochrane Central Register from inception to January 2026. Eligible studies were randomized controlled trials (RCTs) evaluating nurse-led ACP interventions in adults with advanced or metastatic cancer. Two reviewers independently screened records, extracted data, and assessed risk of bias using the Cochrane Risk of Bias 2 (RoB 2) tool. A narrative synthesis was performed due to clinical and methodological heterogeneity precluding meta-analysis.ResultsSix studies (7 publications; N = 2816) met inclusion criteria, conducted in the United States, South Korea, and six European countries. Nurse-led ACP interventions consistently increased advance directive completion (adjusted odds ratios 2.5-5.3) and end-of-life conversation rates vs usual care. One trial showed facilitated nurse-led ACP outperformed patient-directed approaches in ACP engagement (adjusted difference 0.25; 95% CI 0.10-0.40; P = .001). Quality of life outcomes were mixed; most trials showed no significant difference at 3 months, though higher-intensity interventions trended toward greater benefit. Overall risk of bias was low to moderate.ConclusionsNurse-led ACP interventions are feasible and effective in improving ACP engagement and end-of-life outcomes for patients with advanced cancer. Successful implementation requires standardized training, adequate intervention intensity, and systematic follow-up. Future trials should use validated patient-reported ACP measures and explore optimal timing and delivery methods.
BackgroundPatients are increasingly using cannabis products to manage a wide range of symptoms; however, few studies have examined the symptom burden of patients with cancer receiving palliative care who self-report using cannabis products.ObjectivesThe purpose of this investigation was to compare the symptom burden of palliative care patients with cancer who self-reported using cannabis products vs those who did not.DesignRetrospective study.Setting/Participants100 patients who self-reported using cannabis products and 300 patients who did not were randomly selected as the case and control groups, respectively. All participants were evaluated during their first consultation at a supportive clinic in a tertiary cancer center in the United States between January 1 and December 31, 2024.MeasurementsData collected included self-reported cannabis use, demographics, primary cancer diagnosis, cancer treatment, symptoms, performance status, and medications.ResultsThe proportion of patients who self-reported using cannabis products was 11.14%. Multivariate logistic regression analysis showed that female sex, older age, and Black or African American race were associated with lower odds of self-reported cannabis use. In contrast, living alone, higher levels of pain, anxiety, and appetite disturbance, were associated with increased odds of cannabis use.ConclusionsPatients with higher symptom burden are more likely to report cannabinoid use. Universal screening accompanied by patient education is important in palliative care.
Hospice has become an important component of End-of-life care in the United States. The use of hospice, however, remains subject to many forces, including the patient, providers, families and cultural beliefs. One consequence of these forces is significant variability of survival of patients in hospice. In a recent study, the authors reported a median survival of approximately 24 days after enrollment. However, a significant minority of patients survive for more than 180 days (the target set by CMS for the Medicare hospice benefit). We examine the factors contributing to the variability of hospice patient survival, including demographics, diagnoses, care setting, and patient drug prescriptions. We apply Cox proportional hazards models to assess the effect of covariates on time-to-event (in this case death in hospice). A high hazard ratio is associated with a shorter life expectancy. In addition to these factors we evaluated medication exposure variables. Medications are strongly associated with the mortality hazard: analgesic use is associated with a nearly three-fold higher hazard (HR = 2.977), and anxiolytic use with a roughly two-fold higher hazard (HR = 1.910) when compared with the group without any anxiolytic use. We provide a detailed understanding of medication use near the end of life that serve as an indicator that medication trajectories can serve as an additional indicator of symptom burden and individualized care planning.
