Ah Reum An 1, June-Koo Lee 2, Young Ho Yun 3,4 and Dae Seog Heo 2,3,4. Original Article

Transcription

1 531312PMJ / Palliative MedicineAn et al. research-article2014 Original Article Terminal cancer patients and their primary caregivers attitudes toward hospice/palliative care and their effects on actual utilization: A prospective cohort study Palliative Medicine 1 10 The Author(s) 2014 Reprints and permissions: sagepub.co.uk/journalspermissions.nav DOI: / pmj.sagepub.com Ah Reum An 1, June-Koo Lee 2, Young Ho Yun 3,4 and Dae Seog Heo 2,3,4 Abstract Background: Previous studies on hospice/palliative care indicated that patients socio-demographic factors, disease status, and availability of health-care resources were associated with hospice/palliative care utilization. However, the impact of family caregivers on hospice/palliative care utilization has not been thoroughly investigated. Aim: To evaluate the association between attitudes toward hospice/palliative care of both patients with terminal cancer (defined as progressive, advanced cancer in which the patient will die within months) and their family caregivers and utilization of inpatient hospice/palliative care facilities. Design: A prospective observational cohort study was performed in 12 hospitals in South Korea. Attitude toward hospice/palliative care was assessed immediately after terminal cancer diagnosis. After the patient s death, caregivers were interviewed whether they utilized hospice/palliative care facilities. Participants: A total of 359 patient caregiver dyads completed baseline questionnaires. After the patients death, 257 caregivers were interviewed. Results: At the baseline questionnaire, 137/359 (38.2%) patients and 185/359 (51.5%) of caregivers preferred hospice/palliative care. Preference for hospice/palliative care was associated with awareness of terminal status among both patients (adjusted odds ratio: 1.87, 95% confidence interval: ) and caregivers (adjusted odds ratio: 2.14, 95% confidence interval: ). Religion, metastasis, and poor performance status were also independently associated with patient preference for hospice/palliative care. At the post-bereavement interview, 104/257 (40.5%) caregivers responded that they utilized hospice/palliative care facilities. Caregiver s preferences for hospice/palliative care were significantly associated with actual utilization (adjusted odds ratio: 2.67, 95% confidence interval: ). No patient-related factors were associated with hospice/palliative care utilization. Conclusion: Promoting awareness of prognosis and to improve communication between doctors and families is important for facilitating the use of hospice/palliative care. Keywords Hospice/palliative care, terminal cancer, patient, family caregiver What is already known about the topic? The utilization of hospice/palliative care (HPC) in terminal cancer patients was known as closely related with patients socio-demographic factors, disease status, and availability of health-care resources. However, the impact of primary caregiver in the decision-making of utilization of HPC facilities has not been thoroughly investigated. 1 Department of Family Medicine, Seoul National University Hospital, Seoul, Korea 2 Department of Internal Medicine, Seoul National University Hospital, Seoul, Korea 3 Department of Medicine, Seoul National University College of Medicine, Seoul, Korea 4 Cancer Research Institute, Seoul National University College of Medicine, Seoul, Korea Corresponding author: Young Ho Yun, Department of Medicine, Seoul National University College of Medicine, 103 Daehak-ro, Jongno-gu, Seoul , Korea. lawyun08@gmail.com;lawyun@snu.ac.kr

2 2 Palliative Medicine What this paper adds? In this prospective cohort study, primary caregiver s preference for HPC is a major determinant of HPC utilization among patients with terminal cancer. Implications for practice, theory, or policy Communication with primary caregiver is as important as the communication with the cancer patients for promoting the utilization of HPC. Introduction Hospice/palliative care (HPC) has been a major component in caring of patients with terminal cancer. Utilization of HPC was associated with effective symptom control and improved quality of life, without shortening of survival duration. 1 3 Previous studies showed that utilization of HPC was closely related with patient-related factors including age, gender, marital status, education, and disease status. 4 7 Furthermore, health care related factors such as discussion with physicians and availability of health-care resources were also associated with HPC utilization However, the influence of family caregiver on decision-making of HPC utilization has not been thoroughly investigated. 11 Because surrogate decision-making by family members is not uncommon in patients with terminal cancer, evaluation of the factors related to family caregivers as well as those related to patients will provide us a more comprehensive understanding on the context of HPC utilization In this prospective cohort study, we administered questionnaires to patients with terminal cancer and their caregivers to determine attitudes toward HPC and factors associated with preference for HPC. In addition, we evaluated the relationship between the expressed preferences of patients and caregivers and actual utilization. Patients and methods Study design This study was part of the Study to Understand Risks, Priority, and Issues at End-of-life (SURPRISE), a prospective, multicenter cohort study designed by physicians who have been motivated in promoting quality of life of Korean cancer patients to identify issues important to patients with terminal cancer including HPC, life-sustaining treatment, and advance care planning. 3,16 Participants were recruited in the clinics of investigator physicians between July 2005 and October 2006 from 12 public or private hospitals in South Korea (Seoul National University Hospital, National Cancer Center, Seoul Asan Medical Center, Dongsan Medical Center, Chungnam University Hospital, Gangneung Asan Hospital, Sunlin Hospital, Gyeongsang National University Hospital, Kyunghee University Hospital, Fatima Hospital, Seoul St. Mary s Hospital, and Chonnam National University Hwasun Hospital). Because of the insufficient availability of home hospice in Korea, many terminal cancer patients in this cohort were referred to inpatient HPC facilities for their terminal care. Participants Patient inclusion criteria were as follows: (1) age 18 years, (2) terminal cancer diagnosed by physician, (3) ability to complete questionnaires or communicate with an interviewer, (4) ability to understand the intent of the study, and (5) willingness to provide informed consent. Terminal cancer was defined as a progressive advanced cancer in which the patient will die within months (because of general prostration, refractory to chemotherapy, refusal of further chemotherapy). Patients were enrolled to this study at outpatient or inpatient clinics by clinical judgment of physicians. Patients were excluded for the following reasons: (1) continuing conventional anticancer treatment after enrollment, (2) no evaluable lesions, or (3) changed treatment plan. Patients identified their primary caregivers as the individuals who provided the most assistance. The family caregivers were invited to participate in this study but were ineligible if they were not well enough to fill out questionnaires or were unable to communicate with an interviewer or understand the purpose of this study well enough to provide informed consent. Using these criteria, 619 patients were eligible for the study, but 138 (22.2%) patients declined to participate. The most common reason for declining study participation was lack of time (n = 85), followed by severity of illness (n = 30), and concern about violation of privacy (n = 23). A total of 481 (77.7%) patients participated in this study and completed the baseline questionnaire. Of the 435 eligible caregivers, 54 (12.4%) declined to participate, and the remaining 381 (87.6%) participated in this study. Patients and caregivers were enrolled within days of terminal cancer diagnosis. Measures Baseline clinical information was collected from medical records and included primary site of cancer, date of

3 An et al. 3 Figure 1. CONSORT diagram. diagnosis, presence or absence of metastasis, reason for terminal diagnosis, and Eastern Cooperative Oncology Group (ECOG) performance status. We administered baseline questionnaires to patients and family caregivers at the same time through face-to-face interviews at an outpatient or inpatient facility, within days of the terminal cancer diagnosis. These questionnaires contained items about demographic information (age, sex, relationship of caregiver to patient, level of education, income, and religion), awareness of terminal illness, and attitudes toward HPC. The patient and family caregiver questionnaires were similar and required approximately 20 min to complete. Investigators followed the patients and caregivers in the clinic, and the caregivers were interviewed by telephone 3 months after the patient s death. At these interviews, we determined whether the patients were informed of their terminal status before death and whether they chose to use inpatient HPC facilities or life-sustaining measures. Data analysis A total of 359 patient caregiver dyads who provided complete information about their attitudes toward HPC at baseline were analyzed. A total of 257 (71.6%) caregivers completed the post-bereavement questionnaire; 102 (28.4%) caregivers were lost to follow-up (Figure 1). Missing data were excluded from each statistical analysis on a variable-by-variable basis. Demographic, social, and clinical variables were compared by paired t-test (continuous variables) or chisquared test (categorical variables). Concordance between patients and caregivers attitudes toward HPC was primarily analyzed with descriptive statistic and also evaluated using Cohen s kappa coefficient, which assesses the extent of observer agreement beyond that expected by chance. We evaluated relationships between attitudes toward HPC and other baseline characteristics demographic, clinical characteristics, and awareness of terminal status using simple logistic regression analysis; odds ratios (ORs) were determined for each independent variable. Patient s preference for HPC was analyzed using only patient characteristics and common variables, not using family caregiver s characteristics. Family caregiver s preference was also analyzed on their characteristics and common variables. Multivariate analysis was performed with hierarchical

4 4 Palliative Medicine variable selection method: demographic variables (age and gender), awareness of the terminal status, and other factors found to be significant in univariate analysis were entered into the multiple logistic regression analyses. Next, we examined the associations between baseline attitudes toward HPC and actual utilization using multiple logistic regression models in 257 post-bereavement responders. We also used hierarchical variable selection to choose a final model that included variables that were statistically significant and demographically and clinically relevant. Harrell s concordance index (c-index) and Akaike s information criterion were used to determine goodness-of-fit of the models. Relevant patient-related variables (age, gender, religion, and ECOG performance status; at baseline) were entered in the first model, and caregiver-related variables (age, gender, and religion; at baseline) in the second model. Statistical analysis was performed using SAS version 9.2 (SAS Institute, Cary, NC, USA). Ethical considerations All patients and caregivers provided informed consent to participate in this study. The Institutional Review Boards of all participating hospitals reviewed and approved the study protocol. Results Baseline characteristics of patients and caregivers Table 1 summarizes the baseline demographic, social, and clinical characteristics of patients with terminal cancer and their primary caregivers (359 patient caregiver dyads). Compared with the caregiver group, the patient group consisted of more men (p = 0.035) and was older (p < 0.001). Awareness of the patient s terminal status was significantly higher in the caregiver group (p < 0.001). Attitudes toward HPC Responses to the baseline questionnaire revealed that 137/359 (38.2%) patients and 185/359 (51.5%) caregivers expressed a preference for HPC (Table 2). The most common reasons for not preferring HPC among the participants (222 patients, 174 caregivers) were the lack of sufficient information about HPC (46.4% patients, 40.8% caregivers), patient s disfavor (27.9% patients, 17.8% caregivers), and cost (7.7% patients, 11.5% caregivers). Concordance between patients and caregivers preferences was 48.2%; Cohen s kappa coefficient was , indicating no agreement. There was also no within-dyad correlation in preferences of patients and caregivers (r = 0.03, p = 0.573). Factors associated with preference for HPC We performed logistic regression analysis to evaluate the relationships between preference for HPC and demographic, social, and clinical factors. Univariate analysis showed that preference for HPC was significantly associated with the following patient-related factors: age < 60 years, Christianity, awareness of terminal status, presence of metastatic lesion, poor performance status (ECOG 3 4), and terminal diagnosis for reason other than worsening general condition. Results of univariate analysis showed that preference for HPC among caregivers was significantly associated with graduation from college and awareness of the patient s terminal status (Table 3). In the multivariate analysis, awareness of terminal status was independently associated with a preference for HPC, for both patients and caregivers (Table 4). Among patients, religion, presence of a metastatic lesion, and poor performance status were also independently correlated with a preference for HPC. Relationship between patients and caregivers preferences and actual utilization of HPC Results of the post-bereavement questionnaire showed that 104/257 (40.5%) patient caregiver dyads utilized HPC. The most common reason for not utilizing HPC was lack of sufficient information about this type of care (27.1%), followed by preference for active anticancer treatment (14.3%) and concern about isolating the patient from family members (12.0%). We used multiple logistic regression with hierarchical variable selection to evaluate the relationship between preference for HPC and actual utilization of HPC (Table 5). Harrell s c-index for the final model was 0.67, and Akaike s information criterion statistic was (df = 9). In the final model, the caregiver s preference for HPC was significantly associated with utilization (adjusted odds ratio (aor): 2.67, 95% confidence interval (CI): ). The caregiver s religion was also significant in the final model; Christianity was associated with higher rates of HPC utilization (aor: 1.75, 95% CI: ). No patient-related factors were associated with HPC utilization. Discussion In this study, we identified factors associated with preference for HPC among patients with terminal cancer and their primary caregivers and evaluated the relationship between preferences and actual utilization of HPC. Previous studies have reported that several patient-related factors, such as age, gender, marital status, and education level, and availability of health-care facilities were associated with HPC utilization In contrast, there was no

5 An et al. 5 Table 1. Baseline characteristics of patients and caregivers who participated in this study. Characteristics Patients (n = 359) Caregivers (n = 359) n % n % Age, years* Mean Standard deviation Gender* Male Female Education completed High school College Religion* (missing = 4, 8) Christianity Non-Christianity Marital status (missing = 2, 1) Married Not married Employment status before cancer diagnosis (missing = 6, 3) Employed Unemployed Employment status after cancer diagnosis (missing = 17, 16) Employed Unemployed ECOG performance status NA NA Primary cancer site (missing = 1) Stomach Colon/rectum Lung Breast Liver Other NA NA Metastasis (missing = 4) Present Absent NA NA Reason for terminal diagnosis Refractory to chemotherapy General prostration Refusal of further chemotherapy Adverse effect of chemotherapy Other NA NA Awareness of patient s terminal status* (missing = 2, 1) Aware Unaware Caregiver relationship to patient Spouse Child or daughter-in-law Other NA NA ECOG: Eastern Cooperative Oncology Group; NA: not applicable. *p < 0.05 patients versus caregivers; paired t-test (continuous variable) or chi-squared test (categorical variables). significant association between patient-related factors and HPC utilization in our cohort. However, we found that the preference of the primary caregiver was the major determinant in HPC utilization, suggesting that the caregivers

6 6 Palliative Medicine Table 2. Attitudes toward HPC and reasons for not preferring HPC. Patients (n = 359) Caregivers (n = 359) n % n % Preferred HPC Did not prefer HPC Reason for not preferring HPC Insufficient information Patient s disfavor Caregiver s disfavor Doctor s disfavor Lack of HPC facilities nearby Cost Other HPC: hospice/palliative care. Table 3. Univariate analysis of factors associated with preference for HPC among patients with terminal cancer and their caregivers. Factor Patients (n = 359) Caregivers (n = 359) n Preferred HPC (%) OR 95% CI n Preferred HPC (%) OR 95% CI Age, years Reference = Reference = 1 Gender Male Female Reference = Reference = 1 Education completed College High school Reference = Reference = 1 Religion Christianity Non-Christianity Reference = Reference = 1 Marital status Married Not married Reference = Reference = 1 Employment status before cancer diagnosis Employed Unemployed Reference = Reference = 1 Employment status after cancer diagnosis Employed Unemployed Reference = Reference = 1 Awareness of terminal status Aware Unaware Reference = Reference = 1 Metastasis Present Absent Reference = 1 NA NA NA NA ECOG performance status Reference = 1 NA NA NA NA Reason for terminal status General prostration Other Reference = 1 NA NA NA NA

7 An et al. 7 Table 3. (Continued) Factor Patients (n = 359) Caregivers (n = 359) n Preferred HPC (%) OR 95% CI n Preferred HPC (%) OR 95% CI Relationship to patient Spouse Other NA NA NA NA Reference = 1 Living with patient Yes No NA NA NA NA Reference = 1 Monthly household income, USD < Reference = Reference = 1 Who paid for treatment Patient Other Reference = Reference = 1 Insurance type National medical insurance Medicaid Reference = Reference = 1 HPC: hospice/palliative care; ECOG: Eastern Cooperative Oncology Group; NA: not applicable; OR: odds ratio; CI: confidence interval. Table 4. Multivariate analysis of factors associated with preference for HPC among patients with terminal cancer and their caregivers. Factors* Patients * Caregivers* aor 95% CI p aor 95% CI p Age, years Reference = 1 Reference = 1 Gender Male Female Reference = 1 Reference = 1 Awareness of the patient s terminal status Aware Unaware Reference = 1 Reference = 1 Religion Christianity Non-Christianity Reference = 1 NA Education completed College High school NA Reference = 1 Metastasis Present Absent Reference = 1 NA ECOG performance status Reference = 1 NA Reason for terminal status General prostration Other Reference = 1 HPC: hospice/palliative care; aor: adjusted odds ratio; ECOG: Eastern Cooperative Oncology Group; NA: not applicable; CI: confidence interval. *Multiple logistic regression analysis was performed with hierarchical variable selection to evaluate factors associated with patient/caregiver s preference for HPC. We entered patient/caregiver s age, gender, and awareness of terminal status in the first model and variables that showed statistical significance in the univariate analysis in the second model. This table shows the final model.

8 8 Palliative Medicine Table 5. Results of multiple logistic regression showing relationships between patient and caregiver preferences and the actual utilization of HPC. Factors a aor 95% CI p Preference for HPC Patients (yes vs no) Caregivers (yes vs no) Patient-related factors Age (18 59 years vs 60 years) Gender (female vs male) Religion (Christianity vs non-christianity) ECOG performance status (3 4 vs 0 2) Caregiver-related factors Age (18 59 years vs 60 years) Gender (female vs male) Religion (Christianity vs non-christianity) HPC: hospice/palliative care; aor: adjusted odds ratio; ECOG: Eastern Cooperative Oncology Group; CI: confidence interval. a Multiple logistic regression analysis was performed with hierarchical variable selection to evaluate relationships between patient s and caregiver s preferences and actual utilization of HPC. We adjusted for relevant patient-related variables in the first step and caregiver-related variables in the second step. This table shows the final model. preferences influenced the decision more than the patients own preferences. This finding suggests that communication with the caregiver could be critical to improve the HPC utilization in patients with terminal cancer. Previous studies evaluating HPC utilization provide supporting evidence for this idea. Two cross-sectional studies performed in the United States reported that family caregivers made decisions regarding hospice care in more than 42% of cases. 5,17 Moreover, Hirschman et al. 18 reported that nearly half of patients were not involved in the hospice enrollment decision. These findings emphasize the crucial role of the caregiver in decisions regarding terminal care. Surrogate decisionmaking is especially prevalent at the end of life because patients are frequently unable to make their own decisions. 13,19 However, previous studies have reported that the surrogates decisions often do not adequately reflect the patients beliefs. 20,21 In our study, concordance between patient and caregiver regarding preference for HPC was only 48.2%, suggesting that better communication is needed so that caregivers better understand the patient s attitudes and preferences regarding medical care at the end of life. Physicians can support patient autonomy by facilitating these important discussions between patients and caregivers. The lack of sufficient information still appears to be a strong barrier to HPC utilization. We found that lack of sufficient information was the most common reason for not preferring HPC and failure to utilize HPC in our study population. This finding suggests that advocacy and publicity are needed to help patients and caregivers understand the benefits of HPC. Many comprehensive cancer centers provide detailed information about HPC to their patients and caregivers, and help them connect with HPC facilities. Institutional approaches such as education, consultation, booklets, and networks, as well as the efforts of individual physicians, are needed to help patients and caregivers gain a better understanding of their options for terminal care. 22 This study suggests that frank disclosure of terminal status to both patients and caregivers is a necessary first step in advance care planning. A lack of knowledge about terminal status was associated with not preferring HPC for both patients and caregivers. This finding is consistent with that of a cohort study performed in the United States; patients who recognized that they were terminally ill were more likely to receive symptom-directed care. 23 Concern about distressing patient or caregiver can be a barrier to frank disclosure; however, recent studies have indicated this should not be the case. 16,24 In a previous study of our group, patients who were directly informed of their terminal status had a significantly better quality of life and lower rate of emotional distress than patients who guessed their terminal status from their worsening condition. 16 The presence of metastasis and poor performance status were associated with patient s preference for HPC. This may indicate that patients preference for HPC increases as their disease progresses. This finding is consistent with previous study, which showed that HPC discussions were higher in patients who died within 2 months, compared to patients who lived longer. 6 The caregiver s religion was another important determinant of HPC utilization in our study population. In Korea, hospice care was introduced by Christian missionaries. The first hospice hospital in Korea, Calvary Hospital, was established in 1964 by Catholic nuns. 25 Catholic and Protestant churches were the main supporters of HPC in

9 An et al. 9 Korea until recently. 25 The independent association of Christianity and HPC utilization may be based on this cultural history. There was a large number of attrition of participants at the baseline, and also in the post-bereavement questionnaire. We did not collect information on the patients who refused to participate in this study at the baseline; therefore, we could not estimate whether this attrition could bias the result of this study. On the other hand, the post hoc analysis of preferences toward HPC in the patients/caregivers who were lost in the post-bereavement questionnaire (n = 102) showed that these caregivers tended to less prefer the HPC (preference for HPC in responders, 54.9%; non-responders, 43.1%; p = by chi-squared statistic). This result might be a reflection of these caregivers emotional status after bereavement: a prospective cohort study showed that aggressive care at the end of life was associated with caregiver s worse post-bereavement adjustment. 26 Our study had several limitations. First, since the utilization of HPC is related to the availability of such services, characteristics of the Korean HPC system should be considered. In Korea, patients who wish to receive HPC typically depend on specialized inpatient facilities because home-based hospice care is limited. 27 In contrast, in Western societies, home-based HPC programs are often available. 28 Thus, our results may have underestimated our participants willingness to utilize HPC because previous studies have reported that patients are more likely to prefer home-based HPC. 29,30 Second, the hospitals participating in our studies were among the most active hospitals regarding HPC services in Korea. As a result, the HPC utilization rate in our study (40.5%) was much higher than the Korean average, which was reported as 8.6% in This 40.5% HPC utilization rate was similar to the hospice utilization rate in the United States, according to the annual report of National Hospice and Palliative Care Organization. 32 Thus, the participants of our study may not represent the general population of patients with terminal cancer. Third, the term terminal cancer is widely used in Korea, both in patients and physicians. However, their understanding often varies, which can often result in difficulties in communication. In this study, we defined terminal cancer as a progressive, advanced disease with only little remaining life expectancy. However, this definition also has some ambiguous aspects, which might cause biased inclusion of study patients. Fourth, this was an observational cohort study that did not evaluate therapeutic interventions. Therefore, we did not analyze the effects of supportive care measures. Last, this analysis was based on a relatively old observational cohort. However, this result could provide reasonable evidence on the role of family caregiver in treatment decision-making in palliative care, which was suggested as a research gap in the recent literature. 11 In conclusion, this study showed that caregiver s preferences are important determinants of HPC utilization among patients with terminal cancer. Our findings emphasize the key role of the family caregiver in end-of-life care and the importance of communication between patients and caregivers regarding patient preferences. Acknowledgements We thank biosciencewriters ( for professional editing of this article. Part of this study was presented at the patient and survivor care poster discussion session at the 47th ASCO annual meeting, Chicago, IL, USA (3 7 June 2011). Ah Reum An and June-Koo Lee contributed equally to this work. Declaration of conflicting interests The authors have no conflicts of interest to declare. Funding This work was partly supported by grants (nos and ) from the National Cancer Center. References 1. Bakitas M, Lyons KD, Hegel MT, et al. Effects of a palliative care intervention on clinical outcomes in patients with advanced cancer: the Project ENABLE II randomized controlled trial. JAMA 2009; 302: Rummans TA, Clark MM, Sloan JA, et al. Impacting quality of life for patients with advanced cancer with a structured multidisciplinary intervention: a randomized controlled trial. J Clin Oncol 2006; 24: Yun YH, Lee MK, Kim SY, et al. Impact of awareness of terminal illness and use of palliative care or intensive care unit on the survival of terminally ill patients with cancer: prospective cohort study. J Clin Oncol 2011; 29: Tang ST, Huang E-W, Liu T-W, et al. A population-based study on the determinants of hospice utilization in the last year of life for Taiwanese cancer decedents, Psychooncology 2010; 19: Chen H, Haley WE, Robinson BE, et al. Decisions for hospice care in patients with advanced cancer. J Am Geriatr Soc 2003; 51: Huskamp HA, Keating NL, Malin JL, et al. Discussions with physicians about hospice among patients with metastatic lung cancer. Arch Intern Med 2009; 169: Ciemins EL, Stuart B, Gerber R, et al. An evaluation of the advanced illness management (AIM) program: increasing hospice utilization in the San Francisco Bay area. J Palliat Med 2006; 9: Alonso-Babarro A, Astray-Mochales J, Domínguez-Bejón F, et al. The association between in-patient death, utilization of hospital resources and availability of palliative home care for cancer patients. Palliat Med 2013; 27: Cohen J, Houttekier D, Onwuteaka-Philipsen B, et al. Which patients with cancer die at home? A study of six European

10 10 Palliative Medicine countries using death certificate data. J Clin Oncol 2010; 28: Pritchard RS, Fisher ES, Teno JM, et al. Influence of patient preferences and local health system characteristics on the place of death. SUPPORT investigators. J Am Geriatr Soc 1998; 46: Bélanger E, RodrÍguez C and Groleau D. Shared decision-making in palliative care: a systematic mixed studies review using narrative synthesis. Palliat Med 2011; 25: Berger JT, DeRenzo EG and Schwartz J. Surrogate decision making: reconciling ethical theory and clinical practice. Ann Intern Med 2008; 149: Lee J-K, Keam B, An AR, et al. Surrogate decision-making in Korean patients with advanced cancer: a longitudinal study. Support Care Cancer 2012; 21: Gomes B and Higginson IJ. Factors influencing death at home in terminally ill patients with cancer: systematic review. BMJ 2006; 332: Fukui S, Fukui N and Kawagoe H. Predictors of place of death for Japanese patients with advanced-stage malignant disease in home care settings: a nationwide survey. Cancer 2004; 101: Yun YH, Kwon YC, Lee MK, et al. Experiences and attitudes of patients with terminal cancer and their family caregivers toward the disclosure of terminal illness. J Clin Oncol 2010; 28: Casarett D, Crowley R, Stevenson C, et al. Making difficult decisions about hospice enrollment: what do patients and families want to know? J Am Geriatr Soc 2005; 53: Hirschman KB, Corcoran AM, Straton JB, et al. Advance care planning and hospice enrollment: who really makes the decision to enroll? J Palliat Med 2010; 13: Emanuel EJ and Emanuel LL. Proxy decision making for incompetent patients. An ethical and empirical analysis. JAMA 1992; 267: Shalowitz DI, Garrett-Mayer E and Wendler D. The accuracy of surrogate decision makers: a systematic review. Arch Intern Med 2006; 166: Fagerlin A, Ditto PH, Danks JH, et al. Projection in surrogate decisions about life-sustaining medical treatments. Health Psychol 2001; 20: Rabow MW, Hauser JM and Adams J. Supporting family caregivers at the end of life: they don t know what they don t know. JAMA 2004; 291: Mack JW, Weeks JC, Wright AA, et al. End-of-life discussions, goal attainment, and distress at the end of life: predictors and outcomes of receipt of care consistent with preferences. J Clin Oncol 2010; 28: Yun YH, Lee CG, Kim SY, et al. The attitudes of cancer patients and their families toward the disclosure of terminal illness. J Clin Oncol 2004; 22: Glass A, Chen L-K, Hwang E, et al. A cross-cultural comparison of hospice development in Japan, South Korea, and Taiwan. J Cross Cult Gerontol 2010; 25: Wright AA, Zhang B, Ray A, et al. Associations between end-of-life discussions, patient mental health, medical care near death, and caregiver bereavement adjustment. JAMA 2008; 300: Kyung MH, Jang YM, Han KH, et al. Current status and activation plan of hospice palliative care in Korea based on hospice palliative care facilities survey. Korean J Hosp Palliat Care 2010; 13: Virnig BA, Ma H, Hartman LK, et al. Access to home-based hospice care for rural populations: identification of areas lacking service. J Palliat Med 2006; 9: Dunlop RJ, Davies RJ and Hockley JM. Preferred versus actual place of death: a hospital palliative care support team experience. Palliat Med 1989; 3: 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: Shin DW, Choi JY, Nam BH, et al. The current status of utilization of palliative care units in Korea: 6 month results of 2009 Korean Terminal Cancer Patient Information System. Korean J Hosp Palliat Care 2010; 13: National Hospice and Palliative Care Organization (NHPCO). NHPCO facts and figures: hospice care in America. NHPCO, Alexandria, VA, September 2010.