Time Of Hospice Enrollment Research Articles

Homestead together: Pediatric palliative care telehealth support for rural children with cancer during home-based end-of-life care.

Children with terminal cancer and their families describe a preference for home-based end-of-life care. Inadequate support outside of the hospital is a limiting factor in home location feasibility, particularly in rural regions lacking pediatric-trained hospice providers. The purpose of this longitudinal palliative telehealth support pilot study was to explore physical and emotional symptom burden and family impact assessments for children with terminal cancer receiving home based-hospice care. Each child received standard of care home-based hospice care from an adult-trained rural hospice team with the inclusion of telehealth pediatric palliative care visits at a scheduled minimum of every 14days. Eleven children (mean age 11.9years) received pediatric palliative telehealth visits a minimum of every 14days, with an average of 4.8 additional telehealth visits initiated by the family. Average time from enrollment to death was 21.6days (range 4-95). Children self-reported higher physical symptom prevalence than parents or hospice nurses perceived the child was experiencing at time of hospice enrollment with underrecognition of the child's emotional burden. At the time of hospice enrollment, family impact was reported by family caregivers as 46.4/100 (SD 18.7), with noted trend of improved family function while receiving home hospice care with telehealth support. All children remained at home for end-of-life care. Pediatric palliative care telehealth combined with adult-trained rural hospice providers may be utilized to support pediatric oncology patients and their family caregivers as part of longitudinal home-based hospice care.

A Pilot Study of the Effects of COMPLETE: A Communication Plan Early Through End of Life, on End-of-Life Outcomes in Children With Cancer

ContextMost children with cancer die in hospital settings, without hospice, and many suffer from high-intensity medical interventions and pain at end of life (EOL). ObjectivesTo examine the effects of COMPLETE: a communication plan early through EOL to increase hospice enrollment in children with cancer at EOL. MethodsThis is a two-phase, single-arm, two-center, and prospective pilot study of hospice enrollment in children with cancer whose parents received COMPLETE. COMPLETE is a series of medical doctor (MD)/registered nurse (RN)-guided discussions of goals of care using visual aids that begin at diagnosis. COMPLETE training for MD/RNs in Phase II was revised to increase their use of empathy. Preintervention/postintervention measurements for child include: time of hospice enrollment, pain, high-intensity medical interventions at EOL, and location of death; and for parent the following: uncertainty and hope. ResultsTwenty-one parents of 18 children enrolled in the study, and 13 children were followed through EOL. At EOL, 11 (84.6%) died on home hospice or inpatient hospice, and only two (15%) received high-intensity medical interventions. Similar to published findings in the initial 13 parents enrolled in Phase I, parents in Phase II (n = 7) had improvement in hope and uncertainty, and child pain was decreased. Revised training resulted in significant improvement in MD/RN (N = 6) use of empathy (11% in Phase I vs. 100% in Phase II; P = 0.001). ConclusionCOMPLETE resulted in increased hospice enrollment in children with cancer at EOL compared with historical controls. In preanalysis/postanalysis, COMPLETE decreased child pain while supporting hope and reducing uncertainty in their parents.

High-risk communication at the time of hospice enrollment: Standardizing pediatric hospital to hospice sign out.

High-risk communication at the time of hospice enrollment: Standardizing pediatric hospital to hospice sign out.

Frequency and Risk Factors for Live Discharge from Hospice.

To report frequencies and associated risk factors for 4 distinct causes of live discharge from hospice. Retrospective cohort study using electronic medical records of hospice patients who received care from a large urban not-for-profit hospice agency in New York City during a 3-year period between 2013 and 2015 (n = 9,190). Roughly one in five hospice patients were discharged alive (21%; n = 1911). Acute hospitalization was the most frequent reason for live discharge (42% of all live discharges; n = 802). Additional reasons included elective revocation to resume disease-directed treatments (18%; n = 343), disqualification (14%; n = 271), and service transfers or moves (26%; n = 495). Multinomial logistic regression analyses revealed that risk for acute hospitalization was higher among younger patients (age AOR = 0.98 [95% CI = 0.98-0.99] P < .01), racial/ethnic minorities (Hispanic AOR = 2.23 [CI = 1.82-2.73] P < .001; African American OR = 2.46 [CI = 2.00-3.03] P < .001; Asian/other OR = 1.63 [CI = 1.25-2.11] P < .001), and patients without advance directives (AOR = 1.41 [95% CI = 0.98-0.99] P < .001). Disqualification occurred much more frequently among patients with non-cancer diagnoses, including dementia (AOR = 13.14 [95% CI = 7.96-21.61] P < .001) and pulmonary disease (AOR = 11.68 [95% CI = 6.58-20.74] P < .001). Transfers and service moves were more common among Hispanics (AOR = 1.56 [95% CI = 1.45-2.34] P < .001), African Americans (AOR = 1.35 [95% CI = 1.03-1.79] P < .05), patients without a primary caregiver (AOR = 1.35 [95% CI = 1.09-1.67] P < .001), and those without advance directives (AOR = 1.30 [95% CI = 1.07-1.58] P < .01). Further research into factors that underlie live discharge events, especially acute hospitalization, is warranted given their cost and burden for patients/families. Hospices should develop strategies to address acute medical crises and thoroughly evaluate patients' suitability, unmet needs, and knowledge about end-of-life issues at the time of hospice enrollment, especially for those with non-cancer diagnoses.

Can Hospices Predict which Patients Will Die within Six Months?

To determine whether it is possible to predict, at the time of hospice enrollment, which patients will die within 6 months. Electronic health record-based retrospective cohort study. Patients admitted to 10 hospices in the CHOICE network (Coalition of Hospices Organized to Investigate Comparative Effectiveness). Hospice patients. Mortality at 6 months following hospice admission. Among 126,620 patients admitted to 10 hospices, 118,532 (93.6%) died within 6 months. In a multivariable logistic regression model, five characteristics were independent predictors of 6-month mortality. For instance, patients younger than 65 years were less likely to die within 6 months (odds ratio [OR] 0.64; 95% confidence interval [CI] 0.45-0.91; p=0.014). Conversely, male patients were more likely to die within 6 months (OR 1.47; 95% CI 1.05-2.02; p=;0.036). After adjusting for other variables in this model, there were several subgroups with a low probability of 6-month probability (e.g., stroke and Palliative Performance Scale [PPS] score=50; adjusted probability of 6-month mortality=39.4%; 95% CI: 13.9%-72.5%). However, 95% confidence intervals of these 6-month mortality predictions extended above 50%. Hospices might use several variables to identify patients with a relatively low risk for 6-month mortality and who therefore may become ineligible to continue hospice services if they fail to show significant disease progression.

Are advance directives associated with better hospice care?

To describe individuals with advance directives at the time of hospice enrollment and to determine whether they have patterns of care and outcomes that are different from those of individuals without advance directives. Electronic health record-based retrospective cohort study with propensity score-adjusted analysis. Three hospice programs in the United States. Individuals admitted to hospice between January 1, 2008, and May 15, 2012 (N = 49,370). Timing of hospice enrollment before death, rates of voluntary withdrawal from hospice, and site of death. Most participants (35,968, 73%) had advance directives at the time of hospice enrollment. These participants were enrolled in hospice longer (median 29 vs 15 days) and had longer survival times before death (adjusted hazard ratio = 0.62; 95% confidence interval (CI) 0.58-0.66; P < .001). They were less likely to die within the first week after hospice enrollment (24.3% vs 33.2%; adjusted odds ratio (aOR) = 0.83, 95% CI = 0.78-0.88; P < .001). Participants with advance directives were less likely to leave hospice voluntarily (2.2% vs 3.4%; aOR = 0.82, 95% CI = 0.74-0.90; P = .003) and more likely to die at home or in a nursing home than in an inpatient unit (15.3% vs 25.8%; aOR = 0.82, 95% CI = 0.77-0.87; P < .001). Participants with advance directives were enrolled in hospice for a longer period of time before death than those without and were more likely to die in the setting of their choice.

Do Advance Directives Improve Hospice Care? (S701)

Objectives 1. Learn about and understand benefits that advance directives offer patients in hospice care. 2. Discuss advance directives in patients with different demographic variables to help them identify with these groups. This will help them in promoting advance care planning. Background. A general consensus supports the use of advance directives to ensure that patients with serious illnesses receive care that is consistent with their preferences. However, it is not known how the presence of an advance directive influences patients’ care during later stages of illness after enrollment in hospice. Research Objectives. To define the population of patients who have an advance directive at the time of hospice enrollment and to determine whether these patients have patterns of care and outcomes that are different than those without advance directives. Methods. We examined the timing of enrollment in hospice relative to death. We used kaplan meier survival curves with the log rank test to evaluate unadjusted differences in survival and we used a cox proportional hazards model to examine differences in survival between patients with and without advance directives after adjusting for the propensity score. We examined the proportion of patients with short (<1 week) and very short (<1 day) length of stay in hospice using bivariate analysis after propensity score adjustment. We used a similar survival analysis approach to compare rates of voluntary withdrawal from hospice, censoring for death. Finally, we restricted the sample to patients who died in hospice (excluding withdrawals) and examined the association between advance directives and site of death (home or nursing home vs. hospital or inpatient hospice unit) using both bivariate analysis and propensity score adjustment. Results. Among 49,370 patients, 35,968 (73%) had advance directives at the time of hospice enrollment. These patients had a longer course of hospice care prior to death (median 29 vs. 15 days, adjusted hazard ratio: 0.62; 95% CI 0.580.66; p<0.001), and were less likely to enroll during the last week of life (24.3% vs. 33.2%; adjusted OR: 0.83, 95% CI 0.78-0.88; p<0.001). Patients with advance directives were also less likely to leave hospice voluntarily (2.2% vs. 3.4%; adjusted OR: 0.82, 95% CI 0.74-0.90; p1⁄40.003) and more likely to die in a private home or nursing home vs. an inpatient unit (15.3% vs. 25.8%; adjusted OR: 0.82, 95% CI 0.77-0.87; p<0.001). Conclusions. Patients with advance directives were enrolled in hospice for a longer period of time prior to their death and have outcomes, such as death at home, which are more consistent with the principles of hospice care. Implications for research, policy, or practice. It is important to understand whether, and how, advance directives shape trajectories of care in hospice. With this knowledge, clinicians will be better able to serve their patients and their community at large. Hopefully, this will be a platform to help clinicians realize other advantages of advance directives and will help to motivate them to promote them in their hospices as well as in their community.

Can Hospices Predict Which Patients Are Likely to Live More than Six Months? (FR421-B)

1.Discuss factors among dementia and debility patients more likely to lead to long survival in hospice patients.2.Discuss 6-month survival statistics among patients with cancer based upon admission performance status. Some studies have analyzed characteristics associated with long stays in subgroups of hospice admissions but there have been no studies that have determined predictors of long stays in the overall hospice population. To determine whether it is possible to predict, at the time of hospice enrollment, patients with a high likelihood of surviving more than 6 months. Data were extracted from the electronic records of participating CHOICE network hospices for patients admitted between January 2008 and May 2012. A multivariable regression model was used to identify independent predictors of better 6-month survival. Among the 117,188 patients whose status (living vs. dead) at 6 months was known, 15,129 (11%) were still alive at 6 months. A strong predictor of increased 6-month mortality was admission from a hospital (OR=1.42, 95% CI=1.36, 1.48, P=<.001). Debility (OR=-1.29, 95% CI=-1.34, -1.23, P=<.001) and dementia patients (OR=-1.42, 95% CI=-1.48, 1.36, P=<.001) were less likely to die within 6 months. The greatest odds of 6-month mortality was for patients admitted with PPS scores of 30 (OR=-1.83, 95% CI=-1.92, -1.74, P=<.001). Even for patients in the best prognostic group (patients admitted with dementia and PPS 60% or greater), however, odds of surviving 6 months was only 48%. Although there are several variables that could be used to identify those patients who are more likely than others to survive for 6 months, even those patients with the best prognosis still meet Medicare's prognostic criteria for eligibility.

Risk factors for complicated grief (CG) and utilization of a hospice bereavement program.

e19588 Background: The Medicare hospice benefit mandates that bereavement services be provided to a surviving caregiver for a period of one year following a patient’s (pt) death. Identifying caregivers at risk for CG is essential so that appropriate interventions can be provided. Methods: Our objective was to identify specific features predicting for CG in surviving caregivers of pts with a malignancy enrolled in an academic based hospice program and determine utilization of Medicare’s bereavement services. Hospice charts for 65 patients whose caregivers were consecutively discharged from bereavement follow-up between January 2009 and December 2011 were reviewed. Assessments for CG features were performed with use of chaplain and social worker visits, and phone calls from a trained bereavement volunteer to caregivers at 6-8 wks and 6 months after a pt’s death. A discharge assessment was completed at 13 months. Results: Fifty (77%) hospice pts had a cancer (ca) diagnosis and tended to have a shorter average length of hospice enrollment than non-ca pts (34 vs 60 days, p=0.11). There was no difference in risk of caregiver CG between ca vs non-ca diagnoses (p=0.12). Twenty-one (42%) of ca pt caregivers were identified at risk for CG. Ca pts ≤ 50 yrs at the time of hospice enrollment were more likely to have caregivers identified at risk for CG compared to those ≥ 50 yrs (3, 100% vs 18, 38%, p=0.07). At risk ca caregivers had a lower average pt age at time of hospice enrollment than lower risk caregivers (mean pt age 67 vs 76 yrs, p=0.03). Eight ca caregivers (38%) identified at risk for CG received intensive bereavement counseling; 13 (62%) subsequently received routine bereavement follow up; none were lost to follow up. Length of enrollment, caregiver employment status and caregiver relationship did not predict for risk of CG. No ca caregivers required intensive bereavement services beyond 12 months post pt death. Conclusions: Young patient age at time of hospice enrollment is associated with an increased risk of caregiver CG.Intensive bereavement counseling was required by 38% of ca caregivers, and none required &gt;1 yr of service. Further study needs to be done to see if caregivers are receiving adequate bereavement support through the hospice benefit.

Cholinesterase Inhibitor and N-Methyl-D-Aspartic Acid Receptor Antagonist Use in Older Adults with End-Stage Dementia: A Survey of Hospice Medical Directors

Cholinesterase inhibitors and N-methyl-D-aspartic acid (NMDA) receptor antagonists are Food and Drug Administration (FDA) approved for the treatment of moderate to severe Alzheimer's disease. As dementia progresses to the end stage and patients become hospice-eligible, clinicians consider whether or not to continue these therapies without the benefit of scientific evidence. We sought to describe hospice medical directors practice patterns and experiences in the use and discontinuation of cholinesterase inhibitors and NMDA receptor antagonists in hospice patients that meet the Medicare hospice criteria for dementia. Mail survey of hospice medical directors from a random sample from the National Hospice and Palliative Care Organization. Of the 413 eligible participants, 152 completed surveys were returned, yielding a response rate of 37%. Of the respondents, 75% and 33% reported that at least 20% of their patients were taking a cholinesterase inhibitor or memantine, respectively, at the time of hospice admission. The majority of respondents do not consider these therapies effective in persons with end-stage dementia, however, a subset believe that these medications improved patient outcomes including stabilization of cognition (22%), decrease in challenging behaviors (28%), and maintenance of patient function (22%) as well as caregiver outcomes namely reduced caregiver burden (20%) and improved caregiver quality of life (20%). While 80% of respondents recommended discontinuing these therapies to families at the time of hospice enrollment, 72% of respondents reported that families experienced difficulty stopping these therapies. A subset of respondents observed accelerated cognitive (30%) and functional decline (26%) or emergence of challenging behaviors (32%) with medication discontinuation. The findings from this survey indicate that cholinesterase inhibitors and/or NMDA receptor antagonists are prescribed for a subset of patients with advanced dementia and that a proportion of hospice medical directors report clinical benefit from the ongoing use of these agents. In addition, physician preferences for discontinuing these therapies are frequently at odds with the wishes of family members. Prospective studies are needed to evaluate the clinical impact of the discontinuation of these therapies on patient and caregiver outcomes.

Acetylcholinesterase Inhibitor and N-Methyl-D-Aspartic Acid Receptor Antagonist Use among Hospice Enrollees with a Primary Diagnosis of Dementia

To describe acetylcholinesterase inhibitor (AChEI) and memantine use among persons over the age of 65 admitted to hospice with a primary diagnosis of dementia and identify patient and hospice program characteristics associated with the use of these agents. Retrospective, cross-sectional study. Administrative database of a national hospice pharmacy provider. A total of 10,065 persons with end-stage dementia admitted to one of 441 U.S. hospices in 2004. The frequency of AChEI and memantine use was determined and utilized as the unit of analysis for bivariate and multivariate comparisons with patient and hospice program characteristics. Twenty-one percent (2148/10,065) of patients were prescribed AChEI and/or memantine therapy at the time of hospice enrollment. Of these, 49.5% were prescribed donepezil. Odds of receiving AChEI and/or memantine therapy were less likely if the patient was female, (odds ratio [OR] 0.68, 0.62-0.76), died while enrolled in hospice (OR 0.75, 0.67-0.85), received care at home (0.80, 0.71-0.89), or had a hospice length of stay (LOS) less than 7 days (0.53, 0.45-0.62). Patients who had a LOS of at least 60 days were significantly more likely to have received such therapies (OR 1.41 [1.24-1.60] for 61-180 days and 1.33 [1.15-1.54] for over 180 days). A notable number of hospice enrollees with a primary diagnosis of dementia were prescribed AChEI and/or NMDA receptor antagonist therapy. Studies are needed to better define the role of these agents as well as the impact of medication discontinuation in persons with end-stage dementia.

Length of Survival of Patients with Cancer in Hospice: A Retrospective Analysis of Patients Treated at a Major Cancer Center Versus Other Practice Settings

This is a retrospective study of the length of survival (LOS) in hospice of patients with cancer treated at a major cancer center compared to other treatment sites. Of 670 patients, the 185 (28%) treated at a major cancer center had unique characteristics, including higher median Palliative Performance Score (PPS) at the time of hospice enrollment (45 versus 40, p = 0.009), and longer median LOS in hospice (35 versus 21 days, p = 0.02: log rank test). Additional variables that predicted longer LOS were higher PPS, Medicare or Medicaid, self-referral, unmarried status, and non-executed advance directives. After adjusting survival for PPS with a Cox proportional hazard model, the hazard ratio for PPS remained statistically significant (95% confidence interval [CI]: 0.95-0.97] while that for the treatment site was not (95% CI: 0.73-1.04]. The performance status, and not the treatment site, was the dominant predictor of the LOS of patients with cancer in hospice.

Timing of Hospice Referral and Families' Perceptions of Services: Are Earlier Hospice Referrals Better?

To determine whether families of patients who enroll in hospice near the end of life believe that they receive less benefit from hospice services than families of patients who enroll earlier. Semistructured interviews at the time of hospice enrollment and 1 month after the patient's death. This study was conducted at the Hospice of Lancaster County, Lancaster, PA, over a 12-month period spanning 2003-2004. First-time hospice admissions. Interviews assessed the anticipated helpfulness of six hospice services (enrollment interviews) and the actual helpfulness of the same services (follow-up interviews). Length of stay in hospice was associated with families' reports of the number of services provided (Spearman rho=0.34, P<.001), and with the mean helpfulness ratings for the services that were provided (Spearman rho=0.34, P<.001), although even with short lengths of stay, most families reported that they received the services that they anticipated and that those services were helpful. These results indicate that families feel they receive greater benefits from longer lengths of stay in hospice. Future efforts to define an "optimal" length of stay in hospice should consider patients' and families' perceptions of the benefits that hospice offers.

Length of survival (LOS) of cancer patients in hospice and prior enrollment in phase I studies

8158 Background: Short LOS in hospice might be an important measure of poor quality of life for terminally ill cancer patients. It is unknown whether enrollment of cancer patients in phase I studies at the end of life delays their referral to hospice and adversely affects their LOS in hospice. Methods: 180 adult cancer patients treated at Roswell Park Cancer Institute and admitted to a community hospice program between 1/1/02–6/30/02 were included in this exploratory analysis. We retrospectively examined the possible variables associated with short LOS of cancer patients in hospice at a major cancer center, including prior enrollment in phase I studies. Survival analysis was performed with Kaplan-Meier estimation, log-rank comparison, and univariate and multivariate Cox models. Results: The median LOS in hospice was 35 days for all patients and 49 days for the 7% of patients who were previously enrolled in phase I studies. Of all patients, 27% died within two weeks and 8% lived longer than 6 months. Variables associated with short LOS were male gender [p=0.06], executed advance directives [p=0.06], insurance other than Medicare/Medicaid [p=0.07], and poor performance status at the time of hospice enrollment, as assessed by Karnofsky Performance Scale [p= 0.02] or Palliative Performance Scale (PPS) [p< 0.001]. Prior enrollment in phase I studies was not associated with short LOS in hospice [Hazard Ratio: 1.00 & p= 0.99]. In a multivariate Cox model, low PPS and male gender predicted a short LOS in hospice [p<0.001 & p=0.01, respectively]. The PPS and gender separated patients into four risk groups with different LOS in hospice [p<0.001]: males with low PPS had the shortest LOS, while females with high PPS had the longest one. Conclusions: The performance status of cancer patients at the time of hospice enrollment is a strong predictor of the LOS in hospice. The association between male gender and short LOS in hospice, as seen in other studies, needs to be explored further. Prior enrollment of cancer patients in phase I studies did not adversely affect their LOS in hospice. Author Disclosure Employment or Leadership Consultant or Advisory Stock Ownership Honoraria Research Funding Expert Testimony Other Remuneration State University of New York at Buffalo

Length of survival (LOS) of cancer patients in hospice and prior enrollment in phase I studies

8158 Background: Short LOS in hospice might be an important measure of poor quality of life for terminally ill cancer patients. It is unknown whether enrollment of cancer patients in phase I studies at the end of life delays their referral to hospice and adversely affects their LOS in hospice. Methods: 180 adult cancer patients treated at Roswell Park Cancer Institute and admitted to a community hospice program between 1/1/02–6/30/02 were included in this exploratory analysis. We retrospectively examined the possible variables associated with short LOS of cancer patients in hospice at a major cancer center, including prior enrollment in phase I studies. Survival analysis was performed with Kaplan-Meier estimation, log-rank comparison, and univariate and multivariate Cox models. Results: The median LOS in hospice was 35 days for all patients and 49 days for the 7% of patients who were previously enrolled in phase I studies. Of all patients, 27% died within two weeks and 8% lived longer than 6 months. Variables associated with short LOS were male gender [p=0.06], executed advance directives [p=0.06], insurance other than Medicare/Medicaid [p=0.07], and poor performance status at the time of hospice enrollment, as assessed by Karnofsky Performance Scale [p= 0.02] or Palliative Performance Scale (PPS) [p< 0.001]. Prior enrollment in phase I studies was not associated with short LOS in hospice [Hazard Ratio: 1.00 & p= 0.99]. In a multivariate Cox model, low PPS and male gender predicted a short LOS in hospice [p<0.001 & p=0.01, respectively]. The PPS and gender separated patients into four risk groups with different LOS in hospice [p<0.001]: males with low PPS had the shortest LOS, while females with high PPS had the longest one. Conclusions: The performance status of cancer patients at the time of hospice enrollment is a strong predictor of the LOS in hospice. The association between male gender and short LOS in hospice, as seen in other studies, needs to be explored further. Prior enrollment of cancer patients in phase I studies did not adversely affect their LOS in hospice. Author Disclosure Employment or Leadership Consultant or Advisory Stock Ownership Honoraria Research Funding Expert Testimony Other Remuneration State University of New York at Buffalo

Medicare hospice referral criteria for patients with amyotrophic lateral sclerosis: a need for improvement.

We conducted a retrospective review of 97 consecutive patients with amyotrophic lateral sclerosis (ALS) who were accepted into hospice care from a tertiary ALS center. Five patients met Medicare criteria at time of hospice enrollment. The mean number of hospice days was 85 (range, 1-534). All but 2 patients met hospice criteria proposed by the Columbia University ALS group. The present Medicare hospice criteria should be changed to reflect the reality of patients dying of ALS.

Spousal Illness Burden Is Associated with Delayed Use of Hospice Care in Terminally Ill Patients

Patients' families are important in hospice not only because they themselves might benefit from it, but also because they influence whether and how hospice is used to begin with. We hypothesized that, among the married, people with less capable spouses (in this case, those who are less healthy) should be less able to use home hospice care and should enroll in hospice later in the course of their illness, closer to death. We tested this hypothesis by evaluating the impact of spousal health on the timing of patients' use of hospice in a select group of 517 couples, a total of 1034 patients, for whom we have health data for both partners from Medicare claims records. We evaluated cohort survival using the Kaplan-Meier method and Cox regression, and we measured illness burden using the Charlson comorbidity score. Among the probands, 45.6% had a primary cancer diagnosis, 44.5% were female, 7.5% were nonwhite, their mean age was 80.0 years (+/-7.6 SD), and their mean Charlson score was 3.1 +/- 3.0. The probands' median survival after hospice enrollment was 55 days. Spouses were similarly composed demographically, but were less likely to have a cancer primary diagnosis (35.4% had cancer) and were slightly younger than the probands; their mean Charlson score was 2.9 +/- 2.9. Each additional point in the Charlson score of the spouse, controlling for measured characteristics of the proband, is associated with a 5.1% increase in the risk of death after enrollment. That is, after controlling for the characteristics of hospice patients themselves, those with sicker spouses are enrolled later. Compared with patients with the healthiest spouses, those with moderately sick and very sick spouses had considerably later enrollment into hospice, with the median survival decreasing from 111 to 49 to 22 days across the groups. Our data suggest that there may be a separate role of spousal health in affecting the time of hospice enrollment, and that patients' social support may affect not just their health status to begin with, but also their use of health care in general.