Health Care Proxy Research Articles

End-of-life decision making for persons with dementia: Proxies' perception of support.

Healthcare proxies need support in making end-of-life decisions for persons with dementia (PWD). This study explored perceptions of support in decision making among proxies of PWD through semi-structured interviews with 20 proxies. Thematic analysis identified three sources of support: family, doctors, and religiosity/spirituality. Family's engagement in care discussions and support for proxies' decisions were viewed helpful while disagreement or criticism, combined with lack of knowledge about PWD's condition and needs, were not. Doctors were viewed supportive when proxies felt doctors respected their opinions and PWD's wishes. Doctor-PWD rapport influenced proxies' views of medical advice from doctors. Although religiosity/spirituality provided guidance and hope, it also presented conflicts when PWD's wishes differed from proxies' beliefs. Therefore, families of PWD should be provided with assistance in reconciling or mediating family conflicts and further education about the illness trajectory as well as risks/benefits of life-sustaining treatments. Assistance should also be provided to address religious/spiritual conflicts.

Efficacy of a Church-Based, Culturally Tailored Program to Promote Completion of Advance Directives Among Asian Americans.

Having an Advance Directive (AD) can help to guide medical decision-making. Asian Americans (AA) are less likely than White Americans to complete an AD. This pilot study investigated the feasibility and efficacy of a church-based intervention to increase knowledge and behavior change related to AD among Chinese and Vietnamese Americans. This study utilized a single group pre- and post-intervention design with 174 participants from 4 churches. Domain assessed: demographics; AD-related knowledge, beliefs, attitudes, and intentions; AD completion; and conversations with a healthcare proxy. Data were analyzed using Chi square and multiple logistic regression techniques. We observed significant increases in participants' AD-related knowledge, intentions, and a gain in supportive beliefs and attitudes about AD, resulting in 71.8% AD completion, and 25.0% having had a proxy conversation. Providing culturally-tailored intervention and step-by-step guidance can help to achieve significant changes in AD related knowledge and behavior in AA church goers.

Abstract 156: Facilitating End-of-Life Decision-Making and Preparedness Planning Among Heart Failure Patients Hospitalized for Advanced Disease

Background: To address the increasingly strident call for better, more informed shared decision-making about treatment decisions and goals of care in patients with end-stage heart failure (HF), we designed and implemented a supportive care intervention for hospitalized patients with a very poor prognosis. Methods: All patients admitted to the advanced HF service at a single hospital over a 6-mo. period were screened for inclusion criteria (deemed non-candidacy for LVAD/transplantation and high likelihood for death or persistently poor quality of life at 6 months, based on the Allen risk score). Enrollees were given a written exercise, The Conversation Starter Kit, to complete with their healthcare proxy before an outpatient visit with a HF nurse practitioner (NP) trained in supportive care techniques. The intervention provided at this visit consisted of 5 domains: disease understanding, symptom control, legal matters, goals of care, and the role of the healthcare proxy. The KCCQ, PEACE Illness Acceptance Scale, Decisional Conflict Scale, and Kaldjian’s Goals of Care were collected at enrollment and 1 month after the intervention, along with 6-month outcomes of readmission and death. Results: Of 77 patients admitted to the service who were not candidates for advanced therapies, 16 met eligibility criteria. Nine declined participation and 7 of a planned 10 were enrolled; the pilot was terminated early due to low enrollment. Only 2 patients completed the intervention; reasons for not completing varied for the 5 remaining patients. (Table) Despite a pre-existing relationship between patients and NPs, rigorous facilitator training, and a comprehensive, evidence-based, outpatient intervention, significant barriers to implementation included patient reluctance/fear, uncertainty surrounding disease progression, time constraints of completing the intervention within a scheduled clinic visit, and lack of efficacy due to patients having too-advanced disease. Conclusions: An intervention to engage very high-risk, advanced HF patients in shared decision-making about end-of-life care failed to be successfully and sustainably implemented. Future work should assess the feasibility and efficacy of designing an early and iterative intervention in a larger population of HF patients with less advanced disease.

Concordance of Advance Care Plans With Inpatient Directives in the Electronic Medical Record for Older Patients Admitted From the Emergency Department

ContextMeasuring What Matters identified quality indicators to examine the percentage of patients with documentation of a surrogate decision maker and preferences for life-sustaining treatments. ObjectivesTo determine the rate of advance care planning in older adults presenting to the emergency department (ED) and translation into medical directives in the electronic medical record (EMR). MethodsA convenience sample of adults 65 years or older was recruited from a large urban ED beginning in January 2012. We administered a baseline interview and survey in English or Spanish, including questions about whether patients had a documented health care proxy or living will. For patients admitted to the hospital who had a health care proxy or living will, chart abstraction was performed to determine whether their advance care preferences were documented in the EMR. ResultsFrom February 2012 to May 2013, 53.8% (367 of 682) of older adults who completed the survey in the ED reported having a health care proxy, and 40.2% (274 of 682) had a living will. Of those admitted to the hospital, only 4% (4 of 94) of patients who said they had a living will had medical directives documented in the EMR. Similarly, only 4% (5 of 115) of patients who had a health care proxy had the person's name or contact information documented in their medical record. ConclusionAbout half of the patients 65 years or older arriving in the ED have done significant advance care planning, but most plans are not recorded in the EMR.

Abstract TP287: Informed Consent for IV tPA in New York State Designated Stroke Centers: A Current Snapshot

Introduction: The New York State Department of Health (NYS DOH) began designation of Stroke Centers in 2004. Investigators sought to ascertain current informed consent procedures in the 0-3 hour and 3-4.5 hour treatment windows. Hypothesis: There is an association between IV tPA informed consent practices and hospital demographics among NYS DOH Stroke Centers across treatment windows. Methods: Hospitals were sent a 13-question on-line survey to determine whether written, verbal or no consent was required within the 0-3 hour and 3-4.5 hour treatment windows. Surveys were conducted in August and September of 2014. Hospital size, academic vs. non-academic and authorizers of consent were obtained. Chi-square tests were used to assess possible association between demographic measures and whether or not consent was required in either treatment window. Results: One hundred and eleven of the 117 Stroke Centers responded (95%). Within the 3-hour treatment window, 28% required written consent compared to 54% that required verbal approval. In the 3-4.5 hour treatment window, 97 (87%) used IV tPA with 59% requiring written consent, compared to 33% requiring verbal discussion only. Combining both treatment windows, 98% accepted a health care proxy or surrogate to give consent in lieu of the patient. Among hospitals with fewer than 500 beds, 86% (70/81) did require some form of written or verbal consent within the 3-hour treatment window, compared to only 70% (21/30) of hospitals with 500+ beds (p < 0.05). Beyond the 3-hour treatment window, there was no significant association between bed size and consent requirement. Fifty five percent of hospitals were academic vs. 45% non-academic. Academic status was not related to form of consent during either time window. Conclusions: The majority of hospitals did not require written consent within the 3-hour treatment window, but did require written consent for the 3-4.5 hour treatment window (2-fold increase in latter window). Smaller hospitals were significantly more likely than larger hospitals to require any form of consent in the 0-3 hour window. Further research should be conducted to determine whether this variability in consent for IV tPA affects clinical outcomes, which may in turn impact health policy and practice.

Development of Personalized Health Messages to Promote Engagement in Advance Care Planning.

To develop and test the acceptability of personalized intervention materials to promote advance care planning (ACP) based on the Transtheoretical Model (TTM), in which readiness to change is a critical organizing construct. Development study creating an expert system delivering TTM-personalized feedback reports and stage-matched brochures with more-general information on ACP and modifications based on participant reviews. Senior centers. Community-living persons aged 65 and older (N = 77). Participant ratings of length, attractiveness, and trustworthiness of and reactions to reports and brochures. The expert system assessed participants' readiness to engage in each of four ACP behaviors: completion of a living will, naming a health care proxy, communication with loved ones about quality vs quantity of life, and communication with clinicians about quality vs quantity of life. The system also assessed pros and cons of engagement and values and beliefs that influence engagement. The system provided individualized feedback based on the assessment, with brochures providing additional general information. Initial participant review indicating unacceptable length led to revision of feedback reports from full-sentence paragraph format to bulleted format. After review, the majority of participants rated the materials as easy to read, trustworthy, providing new information, making them more comfortable reading about ACP, and increasing interest in participating in ACP. Older adults found an expert system individualized feedback report and accompanying brochure to promote ACP engagement to highly acceptable and engaging. Additional research is necessary to examine the effects of these materials on behavior change.

The Living Will. My Personal Viewpoint and Reasons for Introducing it into Italian Legislation

Human dignity, the right to health and life, the freedom of self-determination of the patient regarding treatments, the position of a healthcare proxy and of the doctor – are not so easy to reconcile. By means of the living will and the particular attention of the Legislator to these important themes it is possible to avoid frequent discussions and conflicts, above all when the legislation is non-existent, as in Italy which is still characterized by the lack of solid specific ad hoc legislation. This measure, essentially, is particularly effective in order to prevent ex ante, or to solve ex post, the difficult dilemmas which arise in situations which are characterized by the absence of a capacity for self-determination, safeguarding, forever and in all situations, the dignity and the identity of the human being.

A Case of Iatrogenic Vascular Air Embolism Treated with Bipap and Nitroglycerin

Background: There is little scientific data to guide inter-hospital transport of critically ill patients. This leads to practice variation with how providers communicate healthcare information between facilities and inappropriate communication can lead to significant patient harm. We analyzed data used by accepting providers during hospital transfer to form a rational protocol for information exchange. Methods: We conducted a prospective observational study of inter-hospital patient transfers to our medical intensive care unit (ICU). We recorded data that receiving clinicians requested, and whether that information was available upon admission. Following observation, we asked providers to complete a survey indicating whether a data point was useful for clinical decision making for the particular clinical case they had received. We analyzed the relative frequency of data requests and the discordance between available and requested data. Results: Twenty-five physician-patient interactions were observed with 45 surveys completed by critical care providers. On average, 13 data points were utilized for patients with perceived “mild” illness versus 18 data points for patients with “severe” illness. The most requested data were code status (19/25), blood culture status (19/25), and medications administered to the patient (16/25). Other data points identified as useful were past medical history, vital signs, white blood cell count, hemoglobin, lactate, pH, PaCO2 and chest x-ray findings with minimal variability depending on presumed diagnosis (respiratory failure, sepsis or other). Code status (7/19), arterial/venous blood gas (5/12), lactate (4/10), and medical power of attorney (3/5) were the most frequently unavailable data points when requested. Conclusion: Critical care providers use a small number of data points during the inter-hospital admission process, but many of these are frequently unavailable. A formal structured hand off tool is needed to improve information management during inter-hospital transfer. Such a tool must emphasize resuscitation status, critical labs, and ongoing interventions.

Advance Directives to Withhold Oral Food and Water in Dementia

Euthanasia advocates have recently begun counseling people to create advance directives calling for oral food and water to be withheld if the person reaches a certain stage of dementia. The author shows that these directives are in fact requests for euthanasia, and they leave vulnerable people subject to poor-quality care. Both surrogate decision makers (health care proxies) and Catholic institutions have a moral obligation not to implement such directives, and surrogates, rather than withdrawing as proxies, have a moral obligation to advocate for the life and proper care of the incompetent person. Finally, the author argues that society is morally culpable if it does not strongly resist euthanasia in all its forms. National Catholic Bioethics Quarterly 16.3 (Autumn 2016): 421–434.

Skill-building through Task-Oriented Motor Practice (STOMP) intervention for activities of daily living: study protocol for a randomized, single blinded clinical trial.

Progressive disability in activities of daily living (ADL) is inevitable for people with Alzheimer's disease and related dementias (ADRD). Attempts to slow or prevent ADL disability have been unsuccessful despite making progress in behavioral training methods. Missing from this research is an emphasis on how we maximize a patient's engagement during training and the rigorous examination of implementation protocols (dosing and training methods) which may advantage learning in people with ADRD. Our team addressed this gap with the development of the STOMP (Skill-building through Task-Oriented Motor Practice) intervention which creates methods for obtaining ADL goals that support "personhood" and tests high-intensity protocols that appear to advantage learning and sustained learning over time. Through this study, we aim to evaluate differential outcomes by dose levels as well as assess the moderating effects of attention to task during training. Randomized-controlled trial with 32 participants with dementia assigned to either the original, intensive STOMP protocol (3 hours/day, 5 days/week for 2 weeks) or a less-intensive STOMP protocol (1 hour/day, 2 days/week for 2 weeks) delivered by an occupational therapy assistant in the home. ADL training is delivered using motor learning theory techniques of blocked practice, continuous verbal praise, errorless learning and intense dosing schedules. Inclusion criteria: English speaking, adults 50-80 years old that live with a legally-authorized representative that can provide consent, who can follow a one-step command, have three ADL goals they want to address and can participate in an intense therapy protocol. Exclusions include diagnoses of Creutzfeldt-Jakob Dementia, delirium or receptive/global aphasia. Recruitment will occur through direct mailing, physician referral and media/support group presentations. Blinded occupational therapists will complete baseline, post-intervention and 3-month follow-up assessments in the home. Repeated measures ANOVA and graphs will be used to interpret and display results. Through this protocol, we will examine differential outcomes by dose for the STOMP ADL intervention. Our results will inform dosing parameters for future intervention studies for people with ADRD. ClinicalTrials.gov identifier: NCT02356055. This study protocol was approved by the University of Oklahoma Health Sciences Center Institutional Review Board (#4648) and will be performed in accordance with the Declaration of Helsinki.

Becoming Culturally Competent in the Face of Diversity

Becoming Culturally Competent in the Face of Diversity

Dying well: How equal is end of life care among gynecologic oncology patients?

ObjectiveTo identify disparities in utilization of end of life (EoL) resources by gynecologic oncology (GO) patients. MethodsThis retrospective analysis of the medical records of GO patients treated 1/2007–12/2011 and deceased 1/2012–8/2014 evaluated patient demographics, disease characteristics, and utilization of EoL resources. Chi-square, Fisher's exact test, Mann Whitney and Kruskal–Wallis tests were used for statistical analysis. ResultsOf 189 patients analyzed, 113 (60%) were white, 38 (20%) Hispanic, 31 (16%) black, and seven (4%) Asian. Ninety-five (48%) had ovarian cancer, 51 (26%) uterine, 47 (23%) cervical, seven (3%) vulvar/vaginal. In the last 30days of life (DoL), 18 (10%) had multiple hospital admissions, 10 (5%) admitted to the Intensive Care Unit (ICU), 30 (16%) multiple Emergency Room (ER) visits, 45 (24%) received aggressive medical care and eight (4%) received chemotherapy in the final 14 DoL. Furthermore, 54 (29%) had no Supportive Care referral and 29 (15%) no hospice referral. Only 46 (24%) had a Medical Power of Attorney (PoA) or Living Will (LW) on file.Non-white race was associated with increased odds of dying without hospice (OR 3.07; 95%CI [1.27, 2.46], p=0.013). However, non-white patients who enrolled in hospice did so earlier than white patients (42 v. 27days before death, p=0.054). Non-white patients were also significantly less likely to have PoA/LW documentation (24% v. 76%, p=0.009) even if enrolled in hospice (12% v. 31%, p=0.007). ConclusionsSignificant racial disparities in hospice enrollment and PoA/LW documentation were seen in GO patients. This warrants further study to identify barriers to use of EoL resources.

Improving health care proxy documentation using a web-based interview through a patient portal.

Health care proxy (HCP) documentation is suboptimal. To improve rates of proxy selection and documentation, we sought to develop and evaluate a web-based interview to guide patients in their selection, and to capture their choices in their electronic health record (EHR). We developed and implemented a HCP interview within the patient portal of a large academic health system. We analyzed the experience, together with demographic and clinical factors, of the first 200 patients who used the portal to complete the interview. We invited users to comment about their experience and analyzed their comments using established qualitative methods. From January 20, 2015 to March 13, 2015, 139 of the 200 patients who completed the interview submitted their HCP information for their clinician to review in the EHR. These patients had a median age of 57 years (Inter Quartile Range (IQR) 45-67) and most were healthy. The 99 patients who did not previously have HCP information in their EHR were more likely to complete and then submit their information than the 101 patients who previously had a proxy in their health record (odds ratio 2.4, P = .005). Qualitative analysis identified several ways in which the portal-based interview reminded, encouraged, and facilitated patients to complete their HCP. Patients found our online interview convenient and helpful in facilitating selection and documentation of an HCP. Our study demonstrates that a web-based interview to collect and share a patient's HCP information is both feasible and useful.

Comparison of Perceived and Technical Healthcare Quality in Primary Health Facilities: Implications for a Sustainable National Health Insurance Scheme in Ghana.

BackgroundQuality care in health facilities is critical for a sustainable health insurance system because of its influence on clients’ decisions to participate in health insurance and utilize health services. Exploration of the different dimensions of healthcare quality and their associations will help determine more effective quality improvement interventions and health insurance sustainability strategies, especially in resource constrained countries in Africa where universal access to good quality care remains a challenge.PurposeTo examine the differences in perceptions of clients and health staff on quality healthcare and determine if these perceptions are associated with technical quality proxies in health facilities. Implications of the findings for a sustainable National Health Insurance Scheme (NHIS) in Ghana are also discussed.MethodsThis is a cross-sectional study in two southern regions in Ghana involving 64 primary health facilities: 1,903 households and 324 health staff. Data collection lasted from March to June, 2012. A Wilcoxon-Mann-Whitney test was performed to determine differences in client and health staff perceptions of quality healthcare. Spearman’s rank correlation test was used to ascertain associations between perceived and technical quality care proxies in health facilities, and ordered logistic regression employed to predict the determinants of client and staff-perceived quality healthcare.ResultsNegative association was found between technical quality and client-perceived quality care (coef. = -0.0991, p<0.0001). Significant staff-client perception differences were found in all healthcare quality proxies, suggesting some level of unbalanced commitment to quality improvement and potential information asymmetry between clients and service providers. Overall, the findings suggest that increased efforts towards technical quality care alone will not necessarily translate into better client-perceived quality care and willingness to utilize health services in NHIS-accredited health facilities.ConclusionThere is the need to intensify client education and balanced commitment to technical and perceived quality improvement efforts. This will help enhance client confidence in Ghana’s healthcare system, stimulate active participation in the national health insurance, increase healthcare utilization and ultimately improve public health outcomes.

Discussing goals of care preoperatively in high-risk oncologic surgery.

38 Background: Major oncologic surgery has a high incidence of perioperative morbidity. Despite these risks, most surgeons do not discuss goals of care prior to surgery. This study assessed the feasibility of a preoperative goals of care discussion for high risk oncologic surgery. Methods: Observational study of patients undergoing high risk surgery. Prior to surgery, patients were referred for consultation with a surgeon certified in both Palliative and Critical Care. Consultations followed a semi-structured template (table). Patient and conversation outcomes were assessed, and qualitative analysis of consult notes was performed to identify themes. Results: Twelve patients were approached and 11 participated. Mean time between consultation and surgery was 7 days (1-36). All patients named a health care proxy and expressed goals of care. Most (8/11) specified their preferences for life support. Themes included hope –surgery will return them to independence/home; fear – pain, anxiety; and faith–“it is up to God” as reason for not discussing suboptimal surgical outcomes. One patient declined surgery after discussion and transitioned to hospice. Ten survived to discharge: 6 to home (2 with hospice), 4 to rehab. Conclusions: Preoperative goals of care discussions are feasible for high risk oncologic surgery patients. Eliciting goals of care helped clarify wishes for life sustaining therapy. Goals of care discussions do not destroy hope, but can reveal unrealistic expectations for surgery and outcomes. [Table: see text]

Predictors of referral to outpatient specialty palliative care (SPC) in gynecologic cancer (GC) patients.

159 Background: Prior studies in GC patients have described predictors of inpatient palliative care (PC) consultation, but predictors of outpatient SPC consultation have not been elucidated. We sought to identify factors predictive of referral and associated care outcomes. Methods: We performed a cross-sectional study of GC patients seen in the gynecologic oncology clinic at a comprehensive cancer center over a three month period. As a part of routine care, patients completed a symptom questionnaire. Patients previously seen at the outpatient PC clinic were compared to those who had not with respect to demographics, disease characteristics, symptom scores, and provider factors using univariate statistics. A multivariate model was created to identify independent predictors of referral. Results: 913 patients completed the symptom survey. 76 patients (8%) had been seen in the outpatient PC clinic. Disease factors associated with referral included site (p &lt; 0.01), stage (p &lt; 0.01), evidence of disease (p &lt; 0.01), active treatment (p &lt; 0.01), and time point in the disease trajectory (p &lt; 0.01). Women with moderate to severe pain (p &lt; 0.01), sadness (p = 0.03), distress (p &lt; 0.01), fatigue (p &lt; 0.01), neuropathy (p = 0.03), and sexual dysfunction (p &lt; 0.01) were more likely to have seen PC. Marital status, number of symptoms, and patient provider were also predictive of referral (all p &lt; 0.01). In a multivariate model, site, stage, number of symptoms, moderate to severe sexual dysfunction, and provider were independently associated with referral. Compared to women who had not been referred, patients seen in the PC clinic were more likely to have a health care proxy documented in the electronic medical record (p &lt; 0.01). Among patients with related symptoms, patients referred to PC more often had an opioid prescribed for pain (p &lt; 0.01) and medications prescribed for depression (p &lt; 0.01), anxiety (p = 0.04), insomnia (p &lt; 0.01), and fatigue (p &lt; 0.01). Conclusions: Women with depression, anxiety, insomnia, and fatigue were more likely to receive pharmacologic treatment for these symptoms from a SPC provider. Future research should identify referral triggers for those patients most likely to benefit from outpatient SPC consultation.

Preparedness for End of Life-a Survey of Jerusalem District Nursing Homes.

To evaluate the quality of end-of-life (EOL) care in nursing homes. Survey and semistructured interviews. Jerusalem district nursing homes. Staff members of 28 long-term care and skilled nursing facilities in the Jerusalem area in Israel of various ethnic, religious, and administrative affiliations (N = 207). Qualitative analysis of semistructured interviews and statistical analysis of questionnaires. Most staff members reported that EOL preferences were unknown for more than 90% of residents and that fewer than 10% had a healthcare proxy. Most staff members recalled conducting fewer than five EOL conversations over the past year with residents or family members and could recall fewer than five cases in which a resident was allowed to die in the nursing home. According to staff opinions the prevalence of tube feeding was estimated at greater than 10%, initiated because of aspiration, malnutrition, and understaffing, often against family's preferences. More than 25% of staff members believed that pain management was inadequate. Knowledge about management of chronic pain was poor in half of nurses and nearly one-third of physicians. Most staff would rather not receive the treatments they administered to residents. Nursing homes in Jerusalem lack competency for quality EOL care, and there are multiple psychological, training, and policy challenges to improvement.

PC-FACS

PC-FACS

Advance care planning in cystic fibrosis: Current practices, challenges, and opportunities

BackgroundStudies in cystic fibrosis (CF) report late attention to advance care planning (ACP). The purpose of this study was to examine ACP with patients receiving care at US adult CF care programs. MethodsChart abstraction was used to examine ACP with adults with CF dying from respiratory failure between 2011 and 2013. ResultsWe reviewed 210 deaths among 67 CF care programs. Median age at death was 29years (range 18–73). Median FEV1 in the year preceding death was 33% predicted (range 13–100%); 68% had severe lung disease with FEV1<40% predicted. ACP was documented for 129 (61%), often during hospitalization (61%). Those with ACP had earlier documentation of treatment preferences, before the last month of life (73% v. 35%; p=<0.01). Advance directives were completed by 93% of those with ACP versus 75% without (p<0.01); DNR orders and health care proxy designation occurred more often for those with ACP. Patients awaiting lung transplant had similar rates of ACP as those who were not (67% v. 61%; p=0.55). The frequency of ACP varied significantly among the 29 programs contributing data from four or more deaths. ConclusionsACP in CF often occurs late in the disease course. Important decisions default to surrogates when opportunities for ACP are missed. Provision of ACP varies significantly among adult CF care programs. Careful evaluation of opportunities to enhance ACP and implementation of recommended approaches may lead to better practices in this important aspect of CF care.

Rates and risks for late referral to hospice in patients with primary malignant brain tumors

Primary malignant brain tumors (PMBTs) are devastating malignancies with poor prognosis. Optimizing psychosocial and supportive care is critical, especially in the later stages of disease. This retrospective cohort study compared early versus late hospice enrollment of PMBT patients admitted to the home hospice program of a large urban, not-for-profit home health care agency between 2009 and 2013. Of 160 patients with PMBT followed to death in hospice care, 32 (22.5%) were enrolled within 7 days of death. When compared with patients referred to hospice more than 7 days before death, a greater proportion of those with late referral were bedbound at admission (97.2% vs 61.3%; OR=21.85; 95% CI, 3.42-919.20; P < .001), aphasic (61.1% vs 20.2%; OR = 6.13; 95% CI, 2.59-15.02; P < .001), unresponsive (38.9% vs 4%; OR = 14.76,;95% CI, 4.47-57.98; P < .001), or dyspneic (27.8% vs 9.7%; OR = 21.85; 95% CI, 3.42-10.12; P = .011). In multivariable analysis, male patients who were receiving Medicaid or charitable care and were without a health care proxy were more likely to enroll in hospice within 1 week of death. Late hospice referral in PMBT is common. PMBT patients enrolled late in hospice are severely neurologically debilitated at the time hospice is initiated and therefore may not derive optimal benefit from multidisciplinary hospice care. Men, patients with lower socioeconomic status, and those without a health care proxy may be at risk for late hospice care and may benefit from proactive discussion about end-of-life care in PMBT, but prospective studies are needed.