Association Between Referral-to-Death Interval and Location of Death of Patients Referred to a Hospital-Based Specialist Palliative Care Service

Transcription

1 Vol. 46 No. 2 August 2013 Journal of Pain and Symptom Management 173 Original Article Association Between Referral-to-Death Interval and Location of Death of Patients Referred to a Hospital-Based Specialist Palliative Care Service Jissy Vijo Poulose, MBBS, MMed (Public Health), Grad Cert Pall Care (Flinders), Young Kyung Do, MD, MPH, PhD, and Patricia Soek Hui Neo, MBBS, MRCP, MPH Lien Centre for Palliative Care (J.V.P.); Program in Health Services and Systems Research (Y.K.D.), Duke-NUS Graduate Medical School Singapore, Singapore; and Department of Palliative Medicine (P.S.H.N.), National Cancer Centre Singapore, Singapore Abstract Context. The time interval between palliative care referral and death may play a role in determining the last place of care and location of death of patients referred to palliative care teams. Objectives. To examine the association between the referral-to-death interval and location of death of patients referred to a hospital-based palliative care service in Singapore. Methods. A retrospective analysis of data from a palliative care service s administrative database was performed. Individual patient s referral-to-death interval was calculated using the date of first contact with the service and date of death. Multinomial regression analysis was done to determine the influence of referral-to-death interval in predicting death at home and in an inpatient hospice facility compared with death in hospital, separately by gender. Results. Of 842 patients, 52% were female and 56% were aged 65 years or older. Terminal cancer was the diagnosis for most patients (86%). Three hundred ninety patients (46%) died outside the hospital setting. A referral-to-death interval of $30 days (as opposed to <30 days) was associated with an increased likelihood of dying at home (odds ratio [OR] 2.21, 95% CI 1.34e3.67 for males and OR 3.33, 95% CI 2.07e5.35 for females) or in an inpatient hospice facility (OR 2.02, 95% CI 1.13e3.60 for males and OR 2.69, 95% CI 1.55e4.66 for females) compared with death in hospital. Male patients age, ethnicity, and marital status were found to be the contributing factors in predicting death at home. Conclusion. Longer referral-to-death interval was associated with death outside the hospital for patients enrolled in a hospital-based service. The study highlights the importance of early referral in predicting the last place of care and location of death of palliative care patients. J Pain Symptom Manage 2013;46:173e181. Ó 2013 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. Address correspondence to: Young Kyung Do, MD, MPH, PhD, Program in Health Services and Systems Research, Duke-NUS Graduate Medical School Ó 2013 U.S. Cancer Pain Relief Committee. Published by Elsevier Inc. All rights reserved. Singapore, 8 College Road, Singapore , Singapore. young.do@duke-nus.edu.sg Accepted for publication: August 6, /$ - see front matter

2 174 Poulose et al. Vol. 46 No. 2 August 2013 Key Words Palliative care, hospital, timing of referral, place of death, Singapore Introduction Enabling patients to spend their last days of life at the place of their choice is one of the important goals of hospice and palliative care services. Previous research has demonstrated that the quality of life at the end of life (EOL) has a strong association with the place of terminal care and location of death. 1e4 Death at home or in an inpatient hospice is widely considered to be a marker of good quality EOL care. 5e11 Therefore, the location of death is often used as an outcome measure to evaluate palliative care services. 12,13 Studies involving palliative care patients have suggested that palliative care referrals might reduce unnecessary hospital stays and hospital deaths. 1,14e16 However, less is known about the influence of timing of referral to palliative care in determining the location of death of patients referred to hospital palliative care teams. Several studies have previously examined the time interval between palliative care referral and death as an indicator of timing of referral but not in the context of place of death per se. 17e19 A recent Japanese study reported timing of referral from hospital to home palliative care as the key determinant of death at home for patients with advanced cancer. 20 Similarly, a study from the U.S. highlighted the role of hospital palliative care teams in discharge planning and transfer of patients to their preferred location of terminal care, indicating that the time interval between palliative care referral and death has implications for palliative care outcomes, including the location of death. 21 Discharge planning is considered an essential element of palliative care consultation in the hospital setting, although alleviation of acute distress is the primary goal when the reason for referral is management of complex symptoms. In Singapore, palliative care services were started in 1985 as hospice home care and inpatient hospice facilities run by charities mainly to care for terminally ill cancer patients. Currently, palliative care services are available in all major public hospitals in Singapore, and community-based palliative care is provided solely by voluntary welfare organizations providing hospice care in patients homes and inpatient hospice facilities. 22,23 Singapore is a small nation with no rural-urban differentiation and has a well-developed health care infrastructure. Hospitals, residential areas, and community-based care facilities are located not far away from each other and are easily accessible in terms of geographical location. Patients enrolled in the hospital palliative care services are referred by their primary physicians from other disciplines (i.e., oncology, geriatrics, renal medicine, cardiology, neurology, and respiratory medicine). These patients are often transferred to community-based palliative care services after stabilization of acute physical distress, but some patients may die in the hospital during the first consultation period itself as a result of worsening physical condition associated with advanced illness. 24 A demographic bulletin published by the Registry of Births and Deaths has reported the latest data on the proportion of deaths occurring in hospital (60%), at home (26%), and other places (13%, comprising deaths occurring in nursing homes, clinics, charitable institutions, and inpatient hospices) in the general population in Singapore. 25 Two recent population-based studies examined the place of death and factors influencing the place of death among cancer patients and the elderly in Singapore, specifically investigating the influence of sociodemographics and diagnosis. 26,27 So far, there is no published literature examining the place of death or factors influencing the place of death of patients referred to palliative care specialists in Singapore. Considering these issues together, we conducted a study on patients enrolled in a hospital-based palliative care service with the place of death as the outcome of interest. This study specifically aimed to determine the association between referral-to-death interval and location of death of patients who were referred to a hospital-based consultative service in Singapore. Referral-to-death interval was used as

3 Vol. 46 No. 2 August 2013 Referral-to-Death Interval and Location of Death 175 a proxy indicator of timeliness of referral to palliative care, and the role of factors such as age, gender, ethnicity, marital status, and diagnosis in determining the location of death of patients also was considered in our analysis. Methods Setting and Study Design The study involved retrospective analysis of data on patients referred to the palliative care service located at the National Cancer Centre Singapore. This is a stand-alone department, with no inpatient palliative care beds, providing consultative services in the hospital setting. Referrals are made by doctors working in the cancer center and adjacent tertiary care hospital mainly for terminal care, management of difficult symptoms, and discharge planning. Most referrals to the service come from the hospital wards (about 86%), with the rest coming from specialist outpatient clinics at the cancer center. A team comprising five consultants (doctors with specialist certification in palliative medicine), three nurses, and a few junior doctors (trainees in palliative medicine and regular medical officers) share the daily workload with input from medical social workers working in the cancer center. The palliative care specialists also provide regular outpatient follow-up for patients enrolled in the service. Although the service does not provide home care, arranging referrals to community-based palliative care services is an integral part of consultation. Data Collection and Statistical Analysis All patients referred to the service from January 1, 2007, to December 31, 2007, and seen by the palliative care specialists were selected to be included in the study (N ¼ 939). Data on the date of first contact with the palliative care service, date of death, and place of death were extracted from the service s administrative database. Information on the date of death and location of death was originally provided by the community hospice nurse/family member on request by the service after a patient s death. Data on age at death, gender, ethnicity, marital status, and diagnosis also were extracted from the database to be included in the analysis. A total of 842 patients (89.6% of the original sample) were included in the final analysis after excluding subjects with missing data for the study variables. The information on the location of death was missing for 6% of the patients who, therefore, were excluded from the analysis. Patients whose deaths occurred in locations other than hospital, home, and inpatient hospice (i.e., nursing homes and other places) constituted only 1% of the sample and were likewise excluded. The final analysis was restricted to the three major ethnic groups in Singapore (i.e., Chinese, Malays, and Indians), excluding patients belonging to the Other ethnic groups (0.6%). Ethics approval to conduct this study was obtained from the National University of Singapore Institutional Review Board. Statistical Analysis The dependent variable (location of death) was defined as a multinomial variable representing three categories: home (patient s home), inpatient hospice, and hospital (reference category). Referral-to-death interval was defined as the interval between the first contact with the service and date of death. We calculated the referral-to-death interval for all patients, using the date of first contact with the specialist palliative care service and date of death. Statistical analysis included simple descriptive statistics, cross-tabulations, and multiple regression analysis. Given that our dependent variable (location of death) included three categories, we used a multinomial logistic regression method for analysis. We categorized the referral-to-death intervals into two groups (<30 days and $30 days), with <30 days being the reference category. Chi-square tests were used to determine associations between the dependent variable and each predictor variable; a two-sided P-value #0.05 was considered statistically significant. Patients were stratified by gender before the analysis to account for potential differences between men and women (e.g., cancer types) and explore if there were significant differences between men and women in terms of the influence of sociodemographics and diagnosisdfactors known to predict the place of death in the general population. Odds ratios (ORs) with 95% CIs were generated for each variable included in the regression analysis. All statistical tests were

4 176 Poulose et al. Vol. 46 No. 2 August 2013 performed using IBM SPSS Statistics version 20 (IBM SPSS Inc., Armonk, NY). Results The descriptive profile of the study population is presented in Table 1. Terminal cancer was the diagnosis for most patients (86.6%). More men than women were married at the time of admission (75.3% vs. 55.1%). There was no statistically significant association between gender and location of death (P-value for Chi-square test ¼ 0.27). The median referral-to-death interval was 29 days. Univariate analyses showed significant association Table 1 Profile of Patients (N ¼ 842) Variables n (%) Median Range Demographics Age (years) 67 13e101 Age group (years) < (43.6) $ (56.4) Gender Male 405 (48.1) Female 437 (51.9) Ethnicity Chinese 734 (87.2) Malay 69 (8.2) Indian 39 (4.6) Marital status Married 546 (64.8) Not married a 296 (35.2) Diagnosis Cancer Lung 138 (16.4) Colorectal 130 (15.4) Breast 76 (9.0) Stomach 49 (5.8) Hematological 32 (3.8) Prostate 24 (2.9) Pancreas 38 (4.5) Nasopharynx 28 (3.3) Liver 44 (5.2) Other cancers 170 (20.2) Noncancer 113 (13.5) Referral-to-death interval (days) 29 1e1235 First contact with 180 (21.4) PC <7 days before death First contact with 40 (4.8) PC >365 days before death Location of death Hospital 452 (53.7) Home 241 (28.6) Inpatient hospice 149 (17.7) PC ¼ palliative care service. a Not married ¼ had no spouse at the time of referral (widowed, divorced, separated, or never married). between the referral-to-death interval and location of death. Results from the multinomial logistic regression analysis of dying at home and in a community-based inpatient hospice compared with death in a hospital are presented by gender in Table 2. A referral-to-death interval of $30 days (compared with <30 days) was associated with death at home (OR 2.21, 95% CI 1.34e3.67 for males and OR 3.33, 95% CI 2.07e5.35 for females) and in an inpatient hospice facility (OR 2.02, 95% CI 1.13e3.60 for males and OR 2.69, 95% CI 1.55e4.66 for females) compared with death in a hospital. For male patients, age at death being $65 years (OR 2.36; 95% CI 1.40e3.97), Malay ethnicity (OR 3.13; 95% CI 1.26e7.74), and having a spouse at the time of referral (OR 1.87; 95% CI 1.01e3.47) were associated with a higher likelihood of dying at home, whereas for females, the corresponding associations were in the same direction but statistically insignificant at the 5% level. A diagnosis of stomach cancer (OR 3.01, 95% CI 1.07e8.40 for males and OR 3.48, 95% CI 1.08e11.24 for females) was independently associated with a higher likelihood of dying at home. For females, a diagnosis of pancreatic cancer (OR 6.23; 95% CI 1.64e23.57) and colorectal cancer (OR 2.62; 95% CI 1.05e6.52) also was associated with an increased likelihood of death at home, whereas for male patients, a diagnosis of nasopharyngeal cancer (OR 3.62; 95% CI 1.08e12.11) was statistically significantly associated with a higher likelihood of dying in an inpatient hospice. Discussion Our analysis showed that the time interval between palliative care referral and death had statistically significant and relatively strong independent association with the location of death of patients referred to the hospitalbased consultation service. Longer referral-todeath interval was associated with an increased likelihood of dying at home or in an inpatient hospice compared with death in the hospital; these results were consistent for both males and females. For male patients, sociodemographic factors such as age, ethnicity, and marital status also were found to have some

5 Vol. 46 No. 2 August 2013 Referral-to-Death Interval and Location of Death 177 Table 2 Multinomial Regression Analysis of Factors Associated With Location of Death, Stratified by Gender Males (n ¼ 405) Females (n ¼ 437) Home a Inpatient Hospice a Home a Inpatient Hospice a Variables OR (95% CI) OR (95% CI) OR (95% CI) OR (95% CI) Referral-to-death interval (days) < $ (1.34e3.67) 2.02 (1.13e3.60) 3.33 (2.07e5.35) 2.69 (1.55e4.66) Sociodemographics Age at death (years) < $ (1.40e3.97) 1.21 (0.68e2.17) 1.59 (0.96e2.64) 1.45 (0.81e2.57) Ethnicity Chinese Malay 3.13 (1.26e7.74) 0.19 (0.02e1.61) 1.10 (0.50e2.38) 0.69 (0.24e1.98) Indian 0.51 (0.13e1.91) 0.49 (0.10e2.31) 0.50 (0.16e1.56) 0.54 (0.13e2.12) Marital status Not married Married 1.87 (1.01e3.47) 0.85 (0.45e1.61) 1.09 (0.68e1.73) 0.88 (0.51e1.52) Diagnosis Cancer Lung Colorectal 1.00 (0.45e2.21) 0.47 (0.17e1.23) 2.62 (1.05e6.52) 1.14 (0.42e3.11) Breast d d 1.63 (0.65e4.08) 1.67 (0.68e4.11) Stomach 3.01 (1.07e8.40) 0.23 (0.02e2.00) 3.48 (1.08e11.24) 1.80 (0.50e6.44) Hematological 0.20 (0.02e1.99) 0.36 (0.04e3.28) 2.06 (0.63e6.79) 0.83 (0.19e3.54) Prostate 1.59 (0.56e4.48) 0.43 (0.08e2.21) d d Pancreas 2.40 (0.76e7.63) 1.05 (0.24e4.49) 6.23 (1.64e23.57) 2.45 (0.54e11.04) Nasopharynx 1.26 (0.31e5.16) 3.62 (1.08e12.11) 1.94 (0.35e10.5) 0.60 (0.06e6.14) Liver 0.69 (0.21e2.23) 0.81 (0.25e2.60) 2.61 (0.63e10.81) 0.95 (0.16e5.57) Other cancers 1.23 (0.54e2.79) 1.07 (0.45e2.51) 1.58 (0.69e3.62) 0.97 (0.42e2.28) Noncancer 0.78 (0.31e1.96) 0.96 (0.37e2.47) 1.87 (0.74e4.75) 0.37 (0.10e1.32) OR ¼ odds ratio; CI ¼ confidence interval. a Hospital is the reference category. influence on the location of death. These factors did not show any statistically significant impact in determining the location of death among females. Diagnosis of certain cancer types also was associated with a higher likelihood of dying outside the hospital, a consistent finding in the literature. 27e29 Referral-to-death interval was identified as a key determinant of death at home and in an inpatient hospice facility for patients referred to this service. As our results indicate, the period after referral to the hospital palliative care team had a strong influence in deciding the last place of care and place of death of patients referred to the service. Although this consultation service was primarily based in the hospital setting, a fairly good proportion of patients (46.3%) died outside the hospital, either at home or in an inpatient hospice, highlighting the role of the hospital palliative care teams in enabling patients to spend their last days of life outside the hospital setting. Transferring patients to the community-based care setting is an outcome of proper discharge planning by the palliative care teams. Our results suggest that this can be successfully achieved if patients are referred to palliative care when prognosis is in terms of months rather than days or hours. Discharge planning involves several processes, such as communication of prognosis, family conferences, clarification of goals of care, and reassessment of preferences for location of death. 21 Communication between the hospital team and community-based palliative care services also needs to be initiated concurrently in planning transfer from hospital to home hospice care or inpatient hospice facilities (depending on the availability of beds and capacity of the community hospice care services). All these are timeconsuming processes that make the period after referral a crucial phase of palliative care consultation in the hospital setting. Many previous studies have pointed out the importance of the timing of referral in achieving the goals of palliative care consultation.

6 178 Poulose et al. Vol. 46 No. 2 August 2013 Late referrals have been consistently reported as a barrier to accomplish the desired benefits of palliative care. 17,19,30e36 In our study, 21.4% of patients were referred to palliative care less than seven days before death. A referral-todeath interval of less than seven days is considered to be late referral to palliative care, based on the perception of family members and physicians regarding the optimum period needed for achieving the goals of hospice care. 33,37,38 In the study conducted by Schockett et al., 33 the median referral-to-death interval for patients who were referred too late (as perceived by bereaved family members) was 21 days. In another study, the proportion of families who said that they benefited from hospice care services was highest for patients who had a length of stay of approximately 90 days. 38 A recently published randomized controlled trial provided evidence of the benefits of early referral to palliative care to the quality of life and mood of patients with metastatic lung cancer. This controlled trial also demonstrated the association between palliative care and a reduction in aggressive health care interventions at EOL. 39 However, in view of the wide variations in trajectories of lifelimiting illnesses and the difficulties involved in prognostication, the determination of optimum timing of palliative care referral remains a difficult issue for referring physicians. In addition, the availability of health care resources (i.e., manpower, hospice beds, health care financing, and so forth) and referring physicians attitudes toward palliative care (including knowledge and beliefs about the benefits of palliative care) also determine referral practices. 40e44 Promoting need-based referrals (rather than prognosis-based) to palliative care may help in overcoming some of these issues. 45 In two previous population-based studies on the place of death in Singapore, Malay ethnicity, female gender, and old age were identified as the predictors of death at home. 26,27 In our study (involving only palliative care patients), we did not observe any statistically significant differences in the occurrence of home deaths between men and women when we included gender as an explanatory variable in the analysis, before stratification. Nonetheless, similar to the findings of the two previous studies on the place of death in Singapore, Malay ethnicity and old age were identified as independent predictors of death at home for male patients in our study, whereas for female patients, these factors did not show any statistically significant impact (in both population-based studies, these findings were attributed to the religious values of Malay patients [99% of Malays are Muslims] and the cultural preferences of the elderly in Asian countries). In addition, we found that men who were married at the time of referral had an increased likelihood of dying at home, highlighting the role of spousal support in enhancing terminal care at home. Studies that examined the influence of sociodemographics on the place of death of palliative care patients had reported that admission to hospice (for palliative care) reduces, but not completely eliminates, the effect of sociodemographics on the place of death. 46,47 Again, these are not easily quantifiable factors, and some of these variables may even interact with each other. 48 Patient preferences for EOL care also may change as death approaches. 49 Systematic reviews of studies on the place of death have shown that the place of death of cancer patients and other terminally ill patients is determined by an array of factors, including individual, disease-related, and health care system-related variables. 29 Patient and family preferences, the availability of health care resources, social support, and home care input have been identified as some of the important factors predicting death at home. 28,29,50 Studies on patients referred to palliative care teams have identified factors such as patient and caregiver preferences, level of caregiver support, home-based palliative care, and social support for the caregivers, as well as the duration and frequency of contact with palliative care teams, as important determinants of location of patients death. 51,52 Early referral is still essential for the assessment of these factors to enable patients and families to make informed decisions regarding the place of terminal care and death. Our study has several limitations. It primarily involved analysis of previously collected data on palliative care patients referred to a single service in Singapore using the limited data extracted from the service s administrative database. Patient/family preferences on the place of death, performance status of patients at the time of referral, number of follow-up

7 Vol. 46 No. 2 August 2013 Referral-to-Death Interval and Location of Death 179 visits after discharge, and details of the palliative care input from the community-based palliative care services were not recorded in the database and could not be included in our analytical model. Similarly, factors such as the availability of caregivers for supporting patients suitable for home hospice care and the availability of beds in the inpatient hospice facilities were not accounted for. Although it could be assumed that most patients who were discharged alive were transferred to community-based services, we did not have information on the actual quantity of home care input the patients received from the community hospice services or the number of followup visits made by the patients. We now assume that patients with longer referral-to-death intervals benefited from community-based palliative care and outpatient follow-up by the team compared with those with shorter referral-todeath intervals. Because the information on the performance status of the patients at the time of referral was not captured in the database, we could not determine clearly if the patients were actively dying at the time of first contact with the service to confirm that patients with shorter referral-to-death intervals were referred too late. Finally, as a result of the sociodemographic, cultural, and health care system differences between Singapore and other countries, the findings of this study cannot be generalized or directly compared with the findings of studies conducted elsewhere. This study adds to the growing body of evidence surrounding the relationship between the timing of referral to palliative care and location of death of patients receiving palliative care. Our study used data from the largest hospital-based palliative care service for cancer patients in Singapore. Consistent with the findings of previous studies from the West, our study findings emphasized the importance of timely initiation of referral in achieving the goals of palliative care, especially for effective discharge planning of patients referred to palliative care services in the hospital setting. 18,19,33,38,39 The influence of age, ethnicity, marital status, and diagnosis on the location of death of palliative care patients also was taken into consideration in our analysis. These findings have implications for both policy and practice in the arena of EOL care. At a policy level, our findings contribute to establishing the evidence needed for the allocation of adequate resources for terminally ill patients, to allow them to spend their last days in the community setting. Above all, our study helps in creating awareness among health care professionals about the role of early referral to palliative care in addressing the needs and preferences of patients and families facing life-limiting illnesses. Larger prospective studies are to be conducted to establish the influence of early referral in determining the last place of care and location of death of patients receiving palliative care. Disclosures and Acknowledgments Dr. Do was supported by an A*STAR infrastructure grant to the Program in Health Services and Systems Research, Duke-NUS Graduate Medical School Singapore. The authors declare no conflicts of interest. The authors thank Associate Professor Cynthia Goh, Senior Consultant, Department of Palliative Medicine, National Cancer Centre Singapore, for permission to use the department s data in conducting this study. Dr. Chetna Malhotra and Associate Professor Angelique Chan at Duke-NUS provided helpful comments on an earlier version of the manuscript. Mary Ann Bautista at Duke-NUS provided assistance with revision. References 1. Houttekier D, Cohen J, Van den Block L, Bossuyt N, Deliens L. Involvement of palliative care services strongly predicts place of death in Belgium. J Palliat Med 2010;13:1461e Mezey M, Dubler NN, Mitty E, Brody AA. What impact do setting and transitions have on the quality of life at the end of life and the quality of the dying process? Gerontologist 2002;42:54e Teno JM, Clarridge BR, Casey V, et al. Family perspectives on end-of-life care at the last place of care. JAMA 2004;291:88e Wright AA, Keating NL, Balboni TA, et al. Place of death: correlations with quality of life of patients with cancer and predictors of bereaved caregivers mental health. J Clin Oncol 2010;28:4457e Lee A, Pang WS. Preferred place of deathda local study of cancer patients and their relatives. Singapore Med J 1998;39:447e450.

8 180 Poulose et al. Vol. 46 No. 2 August Higginson IJ, Sen-Gupta GJ. Place of care in advanced cancer: a qualitative systematic literature review of patient preferences. J Palliat Med 2000;3: 287e Clark D. Between hope and acceptance: the medicalisation of dying. BMJ 2002;324:905e Tang ST. When death is imminent: where terminally ill patients with cancer prefer to die and why. Cancer Nurs 2003;26:245e Beccaro M, Costantini M, Giorgi Rossi P, et al. Actual and preferred place of death of cancer patients. Results from the Italian survey of the dying of cancer (ISDOC). J Epidemiol Community Health 2006;60:412e Yao CA, Hu WY, Lai YF, et al. Does dying at home influence the good death of terminal cancer patients? J Pain Symptom Manage 2007;34: 497e Casarett DJ, Quill TE. I m not ready for hospice : strategies for timely and effective hospice discussions. Ann Intern Med 2007;146:443e Macleod U. Place of death: is home always best? Br J Hosp Med 2011;72:441e Agar M, Currow DC, Shelby-James TM, et al. Preference for place of care and place of death in palliative care: are these different questions? Palliat Med 2008;22:787e Cohen J, Houttekier D, Chambaere K, Bilsen J, Deliens L. The use of palliative care services associated with better dying circumstances. Results from an epidemiological population-based study in the Brussels metropolitan region. J Pain Symptom Manage 2011;42:839e Costantini M, Higginson IJ, Boni L, et al. Effect of a palliative home care team on hospital admissions among patients with advanced cancer. Palliat Med 2003;17:315e Bruera E, Neumann CM, Gagnon B, et al. Edmonton Regional Palliative Care Program: impact on patterns of terminal cancer care. CMAJ 1999; 161:290e Cheng WW, Willey J, Palmer JL, Zhang T, Bruera E. Interval between palliative care referral and death among patients treated at a comprehensive cancer center. J Palliat Med 2005;8:1025e Osta BE, Palmer JL, Paraskevopoulos T, et al. Interval between first palliative care consult and death in patients diagnosed with advanced cancer at a comprehensive cancer center. J Palliat Med 2008; 11:51e Costantini M, Toscani F, Gallucci M, et al. Terminal cancer patients and timing of referral to palliative care: a multicenter prospective cohort study. J Pain Symptom Manage 1999;18:243e Fukui S, Fujita J, Tsujimura M, et al. Late referrals to home palliative care service affecting death at home in advanced cancer patients in Japan: a nationwide survey. Ann Oncol 2011;22:2113e Fromme EK, Bascom PB, Smith MD, et al. Survival, mortality, and location of death for patients seen by a hospital-based palliative care team. J Palliat Med 2006;9:903e Goh CR, Shaw RJ. Evolution of a hospice home care service in Singapore. Ann Acad Med Singapore 1994;23:275e Hong PW, O Connor M. Development of palliative care in Singapore: an overview. Singapore Nurs J 2009;36:48e Radha Krishna LK, Poulose JV, Tan BS, Goh C. Opioid use amongst cancer patients at the end of life. Ann Acad Med Singapore 2010;39:790e Registry of Births and Deaths, Immigration and Checkpoints Authority, Singapore. Singapore Demographic Bulletin January Available from %20Releases/SDB/SDB_January% pdf. Accessed July 12, Beng AK, Fong CW, Shum E, et al. Where the elderly die: the influence of socio-demographic factors and cause of death on people dying at home. Ann Acad Med Singapore 2009;38:676e Hong CY, Chow KY, Poulose J, et al. Place of death and its determinants for patients with cancer in Singapore: an analysis of data from the Singapore Cancer Registry, J Palliat Med 2011;14: 1128e Tang ST, McCorkle R. Determinants of place of death for terminal cancer patients. Cancer Invest 2001;19:165e Gomes B, Higginson IJ. Factors influencing death at home in terminally ill patients with cancer: systematic review. BMJ 2006;332:515e Morita T, Miyashita M, Tsuneto S, Sato K, Shima Y. Late referrals to palliative care units in Japan: nationwide follow-up survey and effects of palliative care team involvement after the Cancer Control Act. J Pain Symptom Manage 2009;38: 191e Tang ST, Chen ML, Huang EW, et al. Hospice utilization in Taiwan by cancer patients who died between 2000 and J Pain Symptom Manage 2007;33:446e Christakis NA, Escarce JJ. Survival of Medicare patients after enrollment in hospice programs. N Engl J Med 1996;335:172e Schockett ER, Teno JM, Miller SC, Stuart B. Late referral to hospice and bereaved family member perception of quality of end-of-life care. J Pain Symptom Manage 2005;30:400e Hui D, Elsayem A, Palla S, et al. Discharge outcomes and survival of patients with advanced cancer admitted to an acute palliative care unit at a comprehensive cancer center. J Palliat Med 2010;13:49e57.

9 Vol. 46 No. 2 August 2013 Referral-to-Death Interval and Location of Death Baek YJ, Shin DW, Choi JY, et al. Late referral to palliative care services in Korea. J Pain Symptom Manage 2011;41:692e Morita T, Akechi T, Ikenaga M, et al. Late referrals to specialized palliative care service in Japan. J Clin Oncol 2005;23:2637e Christakis NA, Iwashyna TJ. Attitude and selfreported practice regarding prognostication in a national sample of internists. Arch Intern Med 1998; 158:2389e Rickerson E, Harrold J, Kapo J, Carroll JT, Casarett D. Timing of hospice referral and families perceptions of services: are earlier hospice referrals better? J Am Geriatr Soc 2005;53:819e Temel JS, Greer JA, Muzikansky A, et al. Early palliative care for patients with metastatic nonsmall-cell lung cancer. N Engl J Med 2010;363: 733e Brickner L, Scannell K, Marquet S, Ackerson L. Barriers to hospice care and referrals: survey of physicians knowledge, attitudes, and perceptions in a health maintenance organization. J Palliat Med 2004;7:411e Bradley EH, Cramer LD, Bogardus ST Jr, et al. Physicians ratings of their knowledge, attitudes, and end-of-life-care practices. Acad Med 2002;77: 305e Johnson CE, Girgis A, Paul CL, Currow DC. Cancer specialists palliative care referral practices and perceptions: results of a national survey. Palliat Med 2008;22:51e Rodriguez KL, Barnato AE, Arnold RM. Perceptions and utilization of palliative care services in acute care hospitals. J Palliat Med 2007;10:99e Johnson CB, Slaninka SC. Barriers to accessing hospice services before a late terminal stage. Death Stud 1999;23:225e Waller A, Girgis A, Johnson C, et al. Facilitating needs based cancer care for people with a chronic disease: evaluation of an intervention using a multicentre interrupted time series design. BMC Palliat Care 2010;9: Moinpour CM, Polissar L. Factors affecting place of death of hospice and non-hospice cancer patients. Am J Public Health 1989;79:1549e Johnson KS, Kuchibhatala M, Sloane RJ, et al. Ethnic differences in the place of death of elderly hospice enrollees. J Am Geriatr Soc 2005;53: 2209e Wachterman MW, Sommers BD. The impact of gender and marital status on end-of-life care: evidence from the National Mortality Follow-Back Survey. J Palliat Med 2006;9:343e Gerrard R, Campbell J, Minton O, et al. Achieving the preferred place of care for hospitalized patients at the end of life. Palliat Med 2011;25: 333e Murray MA, Fiset V, Young S, Kryworuchko J. Where the dying live: a systematic review of determinants of place of end-of-life cancer care. Oncol Nurs Forum 2009;36:69e McWhinney IR, Bass MJ, Orr V. Factors associated with location of death (home or hospital) of patients referred to a palliative care team. CMAJ 1995;152:361e Alonso-Babarro A, Bruera E, Varela-Cerdeira M, et al. Can this patient be discharged home? Factors associated with at-home death among patients with cancer. J Clin Oncol 2011;29:1159e1167.