Article Tools
Supportive Care and Quality of Life
Article Tools
Outcomes After Hematopoietic Stem-Cell Transplantation for Hematologic Malignancies in Patients With or Without Advance Care Planning
Abstract
Engagement in advance care planning (ACP) is viewed as a way to prepare for possible death. In patients undergoing hematopoietic stem-cell transplantation (HSCT), an aggressive but possibly curative procedure for cancer, encouraging engagement in ACP is difficult. We conducted this analysis to determine if engagement in ACP among patients who undergo HSCT is associated with adverse outcomes.
Adult patients who were undergoing their first HSCT for hematologic malignancies between 2001 and 2003 were included. ACP was defined as having a living will, a power of attorney for health care, or life-support instructions. Outcomes assessed included the length of hospital stay, in-hospital mortality, and overall survival.
Of the 343 patients, 172 did not have ACP, whereas 171 did have ACP, and 127 of those were reviewable. Of those with reviewable ACP, 28 patients (22%) completed ACP before cancer diagnosis, 87 (68%) completed ACP after the cancer diagnosis but before HSCT, and 12 (10%) engaged in ACP after HSCT. Patients without ACP before HSCT had a significantly greater risk of death compared with patients with ACP (hazard ratio, 2.11; 95% CI, 1.34 to 3.33; P = .001) while adjusting for statistically significant factors.
Our study demonstrated that lack of engagement in ACP is associated with adverse outcomes after HSCT. Thus, the patients least likely to have planned for poor outcomes are the ones most likely to face them. Additional studies should evaluate the nature of this association and should seek modifiable explanatory factors that could be the target of interventions.
End-of-life care has become an important part of cancer therapy and affects a large number of Americans each year.1-3 End-of-life decisions, such as when to stop treatment directed toward the disease process and when to focus on palliative care, are common concerns of patients with cancer.4 In this setting, it is optimal that advance care directives be completed while the patient is competent, to ensure that care is consistent with the patient's wishes. The merits of having advance directives include autonomy in decision making, congruence between personal values and end-of-life actions, a decreased burden on family and health care providers as wishes are known, and a possible decrease in costs.5 However, recent estimates suggest that less than 10% of Americans have completed advance care directives.6 Even in the setting of incurable malignancies, in which death is anticipated, such discussions are relatively rare and occur in about one fourth of the patients.7 One of the possible reasons for this may be patient or family preferences. Another reason may be the unwillingness of the health care provider to broach this issue because of fear that a discussion of negative information could have an adverse effect on the patient.8-10
Hematopoietic stem-cell transplantation (HSCT) is a high-risk procedure that is performed with a curative intent in patients with certain hematologic malignancies, but it is one that is associated with significant short-term morbidity and mortality. Given the potential loss of decision-making capacity during HSCT, patients should be encouraged to engage in advance care planning (ACP). However, there are scarce data regarding the utilization of ACP in the context of HSCT.11 Little is known about the factors that affect engagement in ACP in the HSCT setting or about the most effective ways to raise these issues. Clinicians often have difficulty communicating with their patients about ACP, because it explicitly raises the possibility of death. They may fear that ACP discussions could heighten the patient's anxiety or, worse, could have an adverse effect on the outcome. In the setting of HSCT, discussions about ACP may be especially upsetting to patients, because they are about to undergo a procedure with very high risks of morbidity and mortality. Patients may fear that physicians bring up ACP because they expect the patient will not survive. This belief stems from findings among hemodialysis patients who viewed ACP as a way to prepare for death and dying and not just as a preparation for possibly being incapacitated in the future.12
Hence, we conducted this retrospective analysis to determine whether patients engage in ACP in the HSCT setting and to evaluate whether the presence of advance directives was associated with adverse clinical outcomes.
Data for the study were obtained using the Adult Oncology Stem Cell Transplant Database at the University of Nebraska Medical Center (UNMC; Omaha, NE). This database contains patient-, disease-, and treatment-related characteristics of all patients who have undergone HSCT since the initiation of the HSCT program at UNMC in the early 1980s. Systematic evaluation of outcomes after HSCT, including disease recurrence or progression and survival status, is updated annually on the patients' anniversary date by trained clinical research associates. Data are compliant with the standards set by the Center for International Blood and Marrow Transplant Registry (CIBMTR). Patients sign an informed consent that allows UNMC clinical personnel to contact them and that allows their physicians to record clinical events of interest.
Additionally, electronic medical records that contain the patients' psychosocial evaluations (performed on all patients undergoing HSCT at UNMC) by a social worker were also reviewed. These psychosocial evaluations used a semistructured questionnaire that inquired about the patient's social history, including place of residence; educational background; role in the family; completion or noncompletion of advance directives; levels of spirituality; health-related behaviors, such as the use of tobacco, alcohol or other illicit drugs; and emotional and financial needs during the transplantation process and follow-up care. This study was approved by the institutional review board at UNMC.
All patients who were at least 19 years of age (the age of majority in the state of Nebraska) at the time of HSCT and who underwent their first HSCT for hematologic malignancies between the years 2001 and 2003 were included in this analysis. Before the year 2001, psychosocial evaluations were not routinely done on all HSCT patients. Inclusion of HSCT patients until 2003 allowed us to have adequate follow-up on all patients. All patients were followed until December 2005.
For the purpose of this study, ACP was defined as having one of the following: a living will, a power of attorney for health care, or life support instructions. The presence of advance care directives was ascertained from a review of the patients' psychosocial evaluations by one of the authors (A.K.G.). When such evaluations indicated the presence of ACP, a copy of the plan was obtain from the medical records department. Review of available ACP documents was done by one of the authors (F.R.L.). Forty-four patients who had ACP but who had no reviewable ACP documents were excluded from the main analysis. Patients were classified into four groups according to the timing of ACP completion: no ACP, ACP before the diagnosis of malignancy, ACP after the diagnosis of malignancy but before HSCT, and ACP after HSCT. The patient's primary place of residence at the time of HSCT was classified as either rural or urban, based on the 2003 Urban Influence Codes.13 Table 1 lists the variables evaluated in the multivariate analyses and how the categories were stratified for each variable. Missing data were categorized as a separate stratum for each variable.
The main variable of interest was the presence or absence of ACP before HSCT. Other variables evaluated included age at transplantation, race, role in the family, level of education, residence, smoking history, alcohol history, level of spirituality, disease type, disease stage at transplantation, interval from diagnosis to transplantation, type of transplantation, and year of transplantation. The primary outcomes were 1-year and overall survivals; evaluation of events included death from any cause.
Univariate comparisons among groups were done using the Kruskal-Wallis test and the χ2 test for continuous and categoric data, respectively. Multivariate analyses to evaluate the risk of death at 1 year and the overall risk of death post-HSCT among the three groups (patients without ACP, patients with ACP before diagnosis of malignancy, and patients with ACP after the diagnosis of malignancy but before HSCT) were compared and adjusted for statistically significant patient- and disease-specific characteristics and for type of transplantation by using Cox proportional hazards regression analysis. Models for risk of death at 1-year survival were censored at 12 months. The main effect (ACP according to time) was forced in all the model building. Model selection was done in a forward stepwise fashion, and only covariates with P ≤ .05 were included in the final model. The three-way classification of the main effect was collapsed to compare groups without ACP versus those with ACP before HSCT after a test for equality (degrees of freedom test) showed no significant difference between the patients who completed ACP before and after their cancer diagnosis.14 First-order interactions between the main effect and the identified significant covariates were checked. No significant interactions were noted. Throughout model building, the assumption of proportionality was confirmed. Stepwise, multivariate logistic regression was also performed to determine the factors associated with the presence of ACP at the time of HSCT.
We performed several corollary multivariate analyses to test the robustness of our findings. In the analyses listed in Tables 2 and 3, the 12 patients who engaged in ACP after HSCT were excluded. Because the median time to obtain ACP in these 12 patients was 3 days post-HSCT, we felt it was too close to the actual HSCT for them to be classified truly as not having ACP. However, we also performed separate analyses, in which we included the 12 patients in the group without ACP. Additionally, the 44 patients with no reviewable ACP were added back to the group with ACP group in our sensitivity analyses.
Figure 1 shows the breakdown of the patients evaluated in our study. Three hundred eighty patients underwent HSCT between 2001 and 2003; psychosocial assessments were available for 343 patients. Of the 343 patients, 172 (50%) did not have ACP, and 171 patients (50%) had ACP. Copies of each patient's ACP were obtained from the filed medical records in 127 (74%) of the 171 patients with ACP. Although the characteristics of patients who had reviewable ACP were similar to those who were not reviewed, our study focused on the 299 patients (shaded boxes in Fig 1) without ACP (n = 172) and with ACP (n = 127). Of the 127 patients who had reviewable ACP, 28 (22%) had completed ACP before diagnosis of their malignancy (median, 44 months before cancer diagnosis; range, < 2 to 133 months); 87 (69%) had ACP after the diagnosis of malignancy but before HSCT (median, 2 months before HSCT; range, < 1 to 69 months); and 12 (9%) had ACP after HSCT (median, 3 days after HSCT; range, 0 to 40 months).
Table 1 lists the characteristics of the patients with or without ACP. Patients who had ACP completed at any time were more likely to be older, with a median age of 55 years (range, 24 to 78 years), compared with those without ACP (median age, 49 years; range, 19 to 72 years; P < .001). Patients who engaged in ACP were also more likely to be white. Sex, level of education, type of hematologic malignancy, role of the patient in the family, stage of disease at transplantation, type of transplantation, and time interval from diagnosis to transplantation did not differ according to the timing of ACP. Other factors, such as the area of primary residence (rural v urban), smoking or alcohol use, and use of spiritual faith to cope with illness, were also evaluated and did not differ among the patient groups (data not shown). By using multivariate analysis to determine factors associated with the odds of having ACP before HSCT, we found that patients of at least 40 years of age were more likely to have ACP at the time of transplantation compared with patients younger than 40 years (odds ratio, 2.43; 95% CI, 1.70 to 3.47; P < .001). No other factors listed in Table 1 were associated with the likelihood of having ACP at the time of transplantation.
Of the 127 patients with reviewable ACP documents, 82 (64%) had both a power of attorney and a living will; 21 (16%) had only a living will; and 24 (19%) had only a power of attorney. In a multivariate analysis, we failed to detect any association between the extent of ACP and overall survival (data not shown).
Tables 2 and 3 list the univariate and multivariate analyses, respectively, which were limited to the 172 patients without ACP and the 115 with ACP before transplantation. We failed to detect statistically significant differences in either the in-hospital mortality or the length of hospital stay between the two groups. Figure 2 shows the plot of survival probabilities according to presence or absence of ACP at the time of transplantation. Table 3 lists the multivariate analyses for the risk of death at 1 year and for overall survival. When adjusted for other statistically significant covariates (ie, age at transplantation, disease stage at transplantation, and type of transplantation), patients without ACP before HSCT were more likely to die within 1 year (hazard Rate [HR], 2.85; 95% CI, 1.48 to 5.48) compared with patients with ACP. Similarly, the overall risk of death for patients without ACP before HSCT was significantly higher compared with patients with ACP (HR, 2.11; 95% CI, 1.34 to 3.33; P = .001).
To test the robustness of our findings, we additionally performed the multivariate analysis that included the 44 patients who had ACP but did not have reviewable documents. In this analysis, patients who did not engage in ACP had a higher risk of death within 1 year (HR, 2.02; 95% CI, 1.20 to 3.39; P = .008) compared with patients with ACP. Similar results were obtained when multivariate analyses were done using overall risk of death as the outcome (data not shown). Reclassifying the 12 patients who obtained ACP after HSCT to the group without ACP also yielded similar findings.
Our study found that, in the setting of HSCT, only 50% of patients engage in ACP. This is within the approximate prevalence reported by Joffe et al11 in a population of HSCT patients (rates varied on components of ACP). We also found that older patients were more likely to engage in ACP. More important, our study observed that patients who engaged in ACP any time before HSCT had better 1-year and overall survivals compared with patients without ACP.
We believe that our study supports the contention that lack of engagement in ACP before HSCT is a marker for an adverse outcome. Indeed, patients who did not engage in ACP were the most likely to face a situation in which ACP might have helped. Despite the strong association between lack of ACP and increased mortality, we do not believe this relationship is causal. Clinicians and scientists in this field agree that outcomes after HSCT are largely dictated by patient characteristics (eg, age, other comorbidities) and the biology of the disease (ie, its response to various forms of cytotoxic or biologic therapy). Additionally, outcomes also are dependent largely on how complications are recognized, prevented, or treated. Our results are consistent with the hypothesis that multiple unmeasured characteristics are systematically different in patients who do or do not engage in ACP.
Several methodological limitations should be noted. The retrospective nature of our study and its dependence on information abstracted from medical records (psychological evaluations) may have introduced some degree of measurement error. For instance, some patients who had ACP after cancer was diagnosed may have supplanted an earlier ACP before developing cancer. Furthermore, because the study involves a single institution with an ethnically homogenous patient population that largely received autologous HSCT, the generalizability of our findings may be limited.
Our finding is important, because it supports the argument that discussing medical complications and thinking about end-of-life issues do not necessarily mean physicians and patients expect poor treatment results. The biggest barrier perceived by patients while discussing ACP is that they would rather concentrate on staying alive than on talking about death. Similarly the second most common barrier to a discussion of ACP perceived by physicians is the worry that discussing end-of-life care will take away the patient's hope.15 Previous studies that involved patients in the surgical intensive care unit have suggested that, for health care providers with a strong sense of accountability for patient outcomes and a sense of obligation to cure, it may be difficult to discuss such issues with patients.16,17 The findings of these studies raise the hypothesis that subjective expectations of adverse outcomes, either by physicians or by patients, paradoxically may decrease the use of ACP, perhaps out of a concern to avoid exacerbating psychologically difficult situations. In this context, our findings are critical, as they demonstrate that discussing ACP should not necessarily be interpreted by patients as indicative of adverse expectations.
We also found that patients who engaged in ACP were more likely to be older. This is in agreement with previous studies that found that the rates of completion of advance directives increased dramatically with age,11,18,19 although other studies did not find such an association.20-23 Cugliari et al20 found no effect of the level of education on completion of advance directives, whereas other studies found that individuals who had a college education had an increased rate of ACP completion.11,24,25 In contrast, Gordon and Shade23 found an inverse relation between the level of education and the completion of ACP. In a multivariate analysis in the present study, we failed to find an association between the level of education and engagement in ACP. Similar to other studies,20-24 we also did not find a sex effect in the completion of ACP. We did not find any effect of spiritual beliefs on engagement in ACP. These findings are in contrast to those of True et al,26 who found that patients with cancer who used spiritual coping to a greater extent were less likely to have a living will and were more likely to desire life-sustaining measures. The reasons for the discrepancy are unclear but may be reflective of the different sample sizes and patient populations of the two studies. When we evaluated the association of disease characteristics and the probability of engagement in ACP, we found no difference in the proportion of patients who had ACP based on the nature of their malignancy, the stage of the disease at transplantation, or the type of transplantation (autologous v allogeneic). In contrast, Joffe et al11 found that recipients of allogeneic HSCT were more likely to engage in ACP.
Our study demonstrated that, contrary to some patients' fears, engagement in ACP is not associated with adverse outcomes after HSCT. Instead, lack of engagement in ACP is associated with increased mortality, which suggests that the people who may benefit most from considering these issues are not doing so. Given their unexpected nature, our findings should be confirmed in other settings. We feel that more studies should be done to evaluate whether encouraging participation in ACP benefits or harms patients undergoing HSCT. Future studies also should evaluate how ACP contributes to or interacts with factors known to influence outcome. Similarly, specific contents of ACP should be examined, preferably with a prospective study design, to evaluate how they may impact clinical and psychosocial outcomes in the setting of HSCT.
The author(s) indicated no potential conflicts of interest.
Conception and design: Apar Kishor Ganti, Fausto R. Loberiza Jr
Provision of study materials or patients: Julie M. Vose, Marcel P. Devetten, R. Gregory Bociek, James O. Armitage, Philip J. Bierman, Lori J. Maness, Elizabeth C. Reed
Collection and assembly of data: Apar Kishor Ganti, Fausto R. Loberiza Jr
Data analysis and interpretation: Apar Kishor Ganti, Stephanie J. Lee, Julie M. Vose, Marcel P. Devetten, R. Gregory Bociek, James O. Armitage, Philip J. Bierman, Lori J. Maness, Elizabeth C. Reed, Fausto R. Loberiza Jr
Manuscript writing: Apar Kishor Ganti, Stephanie J. Lee, Julie M. Vose, Marcel P. Devetten, R. Gregory Bociek, James O. Armitage, Philip J. Bierman, Lori J. Maness, Elizabeth C. Reed, Fausto R. Loberiza Jr
Final approval of manuscript: Apar Kishor Ganti, Stephanie J. Lee, Julie M. Vose, Marcel P. Devetten, R. Gregory Bociek, James O. Armitage, Philip J. Bierman, Lori J. Maness, Elizabeth C. Reed, Fausto R. Loberiza Jr
Fig 1. Schema of patients selected and analyzed in the study. HSCT, hematopoietic stem-cell transplantation; ACP, advance care planning.
Fig 2. Probability of survival in patients undergoing hematopoetic stem-cell transplantation with or without advance care planning (ACP).
|
| Variable | Without ACP (n = 172) | With ACP (n = 127) | P | ||||
|---|---|---|---|---|---|---|---|
| No. of Patients | % | No. of Patients | % | ||||
| Age, years | < .001 | ||||||
| Median | 49 | 55 | |||||
| Range | 19-72 | 24-78 | |||||
| < 40 | 51 | 30 | 13 | 10 | < .001 | ||
| 40-55 | 76 | 44 | 52 | 41 | |||
| > 55 | 45 | 26 | 62 | 49 | |||
| Female sex | 75 | 44 | 55 | 43 | .96 | ||
| White | 161 | 94 | 125 | 98 | .04 | ||
| Level of education | .11 | ||||||
| Some high school/high school graduate | 62 | 36 | 32 | 25 | |||
| Some college/college graduate | 84 | 49 | 66 | 52 | |||
| Postgraduate | 15 | 9 | 20 | 16 | |||
| Unknown | 11 | 6 | 9 | 7 | |||
| Main wage earner in family | 110 | 64 | 81 | 64 | .98 | ||
| Disease type | .47 | ||||||
| Acute leukemia/MDS | 24 | 14 | 13 | 10 | |||
| Chronic leukemia | 9 | 5 | 6 | 5 | |||
| Lymphoma | 112 | 65 | 80 | 63 | |||
| Multiple myeloma | 27 | 16 | 28 | 22 | |||
| Disease stage at HSCT | .68 | ||||||
| Early | 49 | 28 | 32 | 25 | |||
| Intermediate | 33 | 19 | 22 | 17 | |||
| Advanced | 90 | 52 | 73 | 57 | |||
| Time from diagnosis to HSCT, months | .26 | ||||||
| Median | 12 | 13 | |||||
| Range | 1-174 | 2-231 | |||||
| ≤ 12 | 89 | 52 | 56 | 44 | .19 | ||
| > 12 | 83 | 48 | 71 | 56 | |||
| Type of HSCT | .91 | ||||||
| Autologous | 131 | 76 | 96 | 76 | |||
| Allogeneic | 41 | 24 | 31 | 24 | |||
| Referred for HSCT | 130 | 76 | 87 | 68 | .18 | ||
Abbreviations: HSCT, hematopoietic stem-cell transplantation; ACP, advance care planning; MDS, myelodysplastic syndrome.
|
| Variable | Univariate Analysis | P | ||
|---|---|---|---|---|
| Without ACP | With ACP | |||
| No. of patients | 172 | 115* | ||
| Probability of 1-year survival | 77 | 89 | .008† | |
| 95% CI | 70 to 82 | 82 to 93 | ||
| In-hospital deaths | .16 | |||
| No. | 4 | 3 | ||
| % | 2 | 2 | ||
| Length of hospital stay, days | .31 | |||
| Median | 13 | 13 | ||
| Range | 8-57 | 3-59 | ||
Abbreviations: HSCT, hematopoietic stem-cell transplantation; ACP, advance care planning.
*Twelve patients who had ACP after HSCT were removed from this analysis.
†Log-rank test.
|
| Variable | No. of Patients | Risk of Death | P | ||
|---|---|---|---|---|---|
| HR | 95% CI | ||||
| ACP status | |||||
| Completed prior to HSCT | 115 | 1.00 | .001 | ||
| Not completed | 172 | 2.11 | 1.34 to 3.33 | ||
| Age at transplantation, years | |||||
| < 40 | 61 | 1.00 | .002 | ||
| 40-55 | 123 | 0.66 | 0.34 to 0.96 | ||
| > 55 | 103 | 1.57 | 0.90 to 2.73 | ||
| Type of transplantation | |||||
| Autologous | 219 | 1.00 | |||
| Allogeneic | 68 | 3.22 | 2.09 to 4.94 | < .001 | |
| Disease stage at transplantation | |||||
| Early | 79 | 1.00 | < .001 | ||
| Intermediate | 54 | 2.25 | 1.11 to 4.56 | ||
| Advanced | 154 | 3.12 | 1.72 to 5.68 | ||
NOTE. A model that used a 1-year risk of death showed essentially the same findings and statistically significant covariates. A separate model that forced the level of education showed essentially the same findings. Additional analyses of models that categorized the 12 patients who obtained ACP after HSCT to the category without ACP completion showed essentially the same findings.
Abbreviations: HR, hazard rate; ACP, advance care planning; HSCT, hematopoietic stem-cell transplantation.
Authors' disclosures of potential conflicts of interest and author contributions are found at the end of this article.
| 1. | Hoyert DL, Kung HC, Smith BL: Deaths: Preliminary data for 2003. Natl Vital Stat Rep 53::1,2005-48, Google Scholar |
| 2. | Bruera E, Russell N, Sweeney C, et al: Place of death and its predictors for local patients registered at a comprehensive cancer center. J Clin Oncol 20::2127,2002-2133, Link, Google Scholar |
| 3. | Elsayem A, Swint K, Fisch MJ, et al: Palliative care inpatient service in a comprehensive cancer center: Clinical and financial outcomes. J Clin Oncol 22::2008,2004-2014, Crossref, Medline, Google Scholar |
| 4. | van der Heide A, Deliens L, Faisst K, et al: End-of-life decision-making in six European countries: Descriptive study. Lancet 362::345,2003-350, Crossref, Medline, Google Scholar |
| 5. | Salmond SW, David E: Attitudes toward advance directives and advance directive completion rates. Orthop Nurs 24::117,2005-127, Medline, Google Scholar |
| 6. | Johnston SC, Pfeifer MP, McNutt R: The discussion about advance directives: Patient and physician opinions regarding when and how it should be conducted. Arch Intern Med 155::1025,1995-1030, Crossref, Medline, Google Scholar |
| 7. | Butow PN, Kazemi JN, Beeney LJ, et al: When the diagnosis is cancer: Patient communication experiences and preferences. Cancer 77::2630,1996-2637, Crossref, Medline, Google Scholar |
| 8. | Tulsky JA, Fischer GS, Rose MR, et al: Opening the black box: How do physicians communicate about advance directives? Ann Intern Med 129::441,1998-449, Crossref, Medline, Google Scholar |
| 9. | Carrese JA, Rhodes LA: Western bioethics on the Navajo reservation. Benefit or harm? JAMA 274::826,1995-829, Crossref, Medline, Google Scholar |
| 10. | Schade SG, Muslin H: Do not resuscitate decisions: Discussions with patients. J Med Ethics 15::186,1989-190, Crossref, Medline, Google Scholar |
| 11. | Joffe S, Mello MM, Cook EF, et al: Advance care planning in patients undergoing hematopoietic cell transplantation. Biol Blood Marrow Transplant 13::65,2007-73, Crossref, Medline, Google Scholar |
| 12. | Singer PA, Martin DK, Lavery JV, et al: Reconceptualizing advance care planning from the patient's perspective. Arch Intern Med 158::879,1998-884, Crossref, Medline, Google Scholar |
| 13. | Economic Research Service: Measuring rurality: New definitions in 2003. http://www.ers.usda.gov/briefing/Rurality/NewDefinitions/ Google Scholar |
| 14. | Klein JK, Rizzo JD, Zhang MJ, et al: Statistical methods for the analysis and presentation of the results of bone marrow transplants: Part 2—Regression Modeling. Bone Marrow Transplant 28::1001,2001-1011, Crossref, Medline, Google Scholar |
| 15. | Knauft E, Nielsen EL, Engelberg RA, et al: Barriers and facilitators to end-of-life care communication for patients with COPD. Chest. 127::2188,2005-2196, Crossref, Medline, Google Scholar |
| 16. | Buchman TG, Cassell J, Ray SE, et al: Who should manage the dying patient? Rescue, shame, and the surgical ICU dilemma. J Am Coll Surg 194::665,2002-673, Crossref, Medline, Google Scholar |
| 17. | Cassell J, Buchman TG, Streat S, et al: Surgeons, intensivists, and the covenant of care: Administrative models and values affecting care at the end of life. Crit Care Med 31::1263,2003-1270, Crossref, Medline, Google Scholar |
| 18. | Braun KL, Onaka AT, Horiuchi BY: Advance directive completion rates and end-of-life preferences in Hawaii. J Am Geriatr Soc 49::1708,2001-1713, Crossref, Medline, Google Scholar |
| 19. | Eleazer GP, Hornung CA, Egbert CB, et al: The relationship between ethnicity and advance directives in a frail older population. J Am Geriatr Soc 44::938,1996-943, Crossref, Medline, Google Scholar |
| 20. | Cugliari AM, Miller T, Sobal J: Factors promoting completion of advance directives in the hospital. Arch Intern Med 155::1893,1995-1898, Crossref, Medline, Google Scholar |
| 21. | Despreaux MA, Coughlin DT, Drickamer MA: Discussing advance directives: A study of outpatients' attitudes. J Gen Int Med 11:114, 1996 (suppl) Google Scholar |
| 22. | Murphy ST, Palmer JM, Azen S, et al: Ethnicity and advance care directives. J Law Med Ethics 24::108,1996-117, Crossref, Medline, Google Scholar |
| 23. | Gordon NP, Shade SB: Advance directives are more likely among seniors asked about end-of-life care preferences. Arch Intern Med 159::701,1999-704, Crossref, Medline, Google Scholar |
| 24. | Hanson LC, Rodgman E: The use of living wills at the end of life: A national study. Arch Intern Med 156::1018,1996-1022, Crossref, Medline, Google Scholar |
| 25. | Mezey MD, Leitman R, Mitty EL, et al: Why hospital patients do and do not execute an advance directive. Nurs Outlook 48::165,2000-171, Crossref, Medline, Google Scholar |
| 26. | True G, Phipps EJ, Braitman LE, et al: Treatment preferences and advance care planning at end of life: The role of ethnicity and spiritual coping in cancer patients. Ann Behav Med 30::174,2005-179, Crossref, Medline, Google Scholar |
