Neurocrit Care DOI 10.1007/s12028-014-0073-x
ORIGINAL ARTICLE
Ethnic Disparities in End-of-Life Care After Subarachnoid Hemorrhage H. Alex Choi • Andres Fernandez • Sang-Beom Jeon • J. Michael Schmidt E. Sander Connolly • Stephan A. Mayer • Jan Claassen • Neeraj Badjatia • Kenneth M. Prager • Kiwon Lee
•
Ó Springer Science+Business Media New York 2014
Abstract Background It is common for patients who die from subarachnoid hemorrhage to have a focus on comfort measures at the end of life. The potential role of ethnicity in end-of-life decisions after brain injury has not been extensively studied. Methods Patients with subarachnoid hemorrhage were prospectively followed in an observational database. Demographic information including ethnicity was collected
H. A. Choi (&) S.-B. Jeon K. Lee Departments of Neurosurgery and Neurology, The University of Texas Health Science Center at Houston, 6431 Fannin St, MSB 7.154, Houston, TX 77030, USA e-mail: [email protected] A. Fernandez J. M. Schmidt J. Claassen Department of Neurology, Division of Neurocritical Care, Columbia University College of Physicians and Surgeons, New York, NY, USA S.-B. Jeon Department of Neurology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea
from medical records and self-reported by patients or their family. Significant in-hospital events including do-notresuscitate orders, comfort measures only orders (CMO; care withheld or withdrawn), and mortality were recorded prospectively. Results 1255 patients were included in our analysis: 650 (52 %) were White, 387 (31 %) Hispanic, and 218 (17 %) Black. Mortality was similar between the groups. CMO was more commonly observed in Whites (14 %) compared to either Blacks (10 %) or Hispanics (9 %) (p = 0.04). In a multivariate analysis controlling for age and Hunt-Hess grade, Hispanics were less likely to have CMO than Whites (OR, 0.6; 95 %CI, 0.4–0.9; p = 0.02). Of the 229 patients who died, 77 % of Whites had CMO compared to 54 % of Blacks and 49 % of Hispanics (p < 0.01). In a multivariate analysis, Blacks (OR, 0.3; 95 %CI, 0.2–0.7; p < 0.01) and Hispanics (OR, 0.3; 95 %CI, 0.2–0.6; p < 0.01) were less likely to die with CMO orders than Whites. Conclusion After subarachnoid hemorrhage, Blacks and Hispanics are less likely to die with CMO orders than Whites. Further research to confirm and investigate the causes of these ethnic differences should be performed.
E. S. Connolly Department of Neurosurgery, Columbia University College of Physicians and Surgeons, New York, NY, USA
Keywords Subarachnoid hemorrhage Ethnic disparities End-of-life care Hospice
S. A. Mayer Institute for Critical Care Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, USA
Introduction
N. Badjatia Department of Neurology, University of Maryland, Baltimore, MD, USA K. M. Prager Department of Internal Medicine, Columbia University College of Physicians and Surgeons, New York, NY, USA
Subarachnoid hemorrhage (SAH) is associated with a high risk of mortality. Advances in critical care management of patients with SAH have contributed to a decrease in overall mortality [1]. Critical care management has also brought to the forefront ethical issues regarding withdrawal or withholding of life-prolonging measures. As our treatment of
123
Neurocrit Care
severe neurologic disease has improved, we are sometimes able to save lives but unable to avoid an outcome of prolonged severe disability. Often life and death decisions are focused around quality of life and severe disability [2]. General population based and disease-specific studies in the US have consistently shown that patients of Black and Hispanic ethnicities, are less likely to choose hospice care and less likely to have Do Not Resuscitate (DNR) orders at the end of life. [3–6] The reasons for these differences are unclear but most likely involve a complex interaction of social, educational, religious, and cultural issues. Adding to this complexity is the often unclear functional prognosis in brain-injured patients and the inability of patients to express their own wishes after their injury. Few studies have addressed the ethnic disparities which influence decisions made by surrogates for patients at the end of life and there are no studies of this kind in the SAH population [5]. In this study, we examined potential ethnic disparities in end-of-life decisions made by surrogate decision makers after SAH. We hypothesized that Hispanic and Black patients died less frequently with a comfort care measures only orders (CMO) compared to non-Hispanic whites.
Methods We identified patients for this study from the SAH Outcomes Project, a single-center, prospective observational cohort. Included subjects were enrolled from August 1996 to January 2011. Study details have been previously described [7]. Inclusion criteria include age 18 or above, and diagnosis of SAH by computed tomography (CT) scan or cerebral spinal fluid. Demographic characteristics, hospital complications, and in-hospital mortality were collected prospectively. The study is approved by the Columbia University institutional review board. Race-ethnicity was determined by self-identification when possible or from family members or next-of-kin based on five categories: 1) White, non- Hispanic; 2) Hispanic; 3) Black; 4) Asian, or 5) Other/unknown. Only subjects who were White, Hispanic, or Black were included in this study, because there were significantly lower numbers of Asians and others/unknown. End-of-Life Decisions Decisions to discuss limitations of life-sustaining measures with family members were left up to the physician treating team. Three categories of limitation of treatment were collected. 1) DNR/DNI: no resuscitative measures performed in the event of cardiopulmonary arrest. 2) Withholding of life-sustaining measures: life-sustaining
123
interventions withheld with a focus on comfort measures e.g.no further increase in vasopressor use for hypotension; no intubation, or mechanical ventilation in case of respiratory failure. 3) Withdrawal of life-sustaining measures: life-sustaining measures withdrawn with a focus on comfort measures only. For the purpose of this study, given that there are only subtle differences between the withholding treatment category and the withdrawal of treatment category, they were combined into a CMO designation. The institutional withdrawal of life-sustaining measures policy from 1996 to 2010 stated that decisions for withdrawal of life-sustaining measures could be made by a surrogate when the patient lacked decisional capacity only if: 1) there was no chance for meaningful recovery; and 2) there was either a) written advance directive that the patient would not wish life support under the current circumstances, or b) a patient-designated health care agent who requested removal of life support based on the patient’s wishes, or c) a surrogate who presented clear and convincing evidence that the patient had verbally stated his/her wishes not to continue on life support under the current circumstances. The above items needed to be agreed upon by two separate attending physicians. In 2010, New York State passed the Family Health Care Decisions Act which allowed surrogates of incapacitated patients to make medical decisions on behalf of patients on the basis of their best interests even in the absence of a living will or without being a health care agent [8]. The law went into effect on June 1st, 2010 and the hospital changed its protocol accordingly. Mortality Cause of death was coded as cardiopulmonary death or brain death in accordance to hospital protocol and consistent with the American Academy of Neurology criteria for brain death [9, 10]. The institutional protocol at the time called for two brain death examinations at least six hours apart with an apnea test and confirmatory testing as needed. Statistical Analysis Demographic differences between ethnicities were analyzed using the ANOVA or Kruskal–Wallis test for continuous variables, and using the X2 test for categorical variables. Differences in-hospital complications and outcomes were analyzed using the X2 test. Univariate and multivariate analysis was used to examine associations between ethnicity and DNR status and CMO status. Factors to include in the multivariate model were chosen based on biological plausibly or known associations with mortality in SAH. In the subset of patients who died, univariate and multivariate analysis was performed to examine the
Neurocrit Care
Results
Table 1 Patient characteristics White Black Hispanic p Value (n = 650) (n = 218) (n = 387) Demographics Age, years
56 ± 15
52 ± 14
53 ± 15
0.01
Sex, female
436 (67)
157 (72)
243 (63)
0.07
1, 2
293 (46)
90 (42)
159 (42)
3 4
149 (23) 99 (15)
59 (27) 29 (13)
104 (27) 54 (14)
5
100 (16)
38 (18)
66 (17)
1
192 (31)
87 (41)
127 (33)
2
64 (10)
18 (8)
33 (9)
3
226 (36)
66 (31)
155 (41)
4
148 (24)
43 (20)
65 (17)
12 ± 7.9
13 ± 8.3
12 ± 8.2
Disease Severity Hunt-Hess grade
0.63
Modified Fisher Scale
APACHE II
0.03
0.2
Values are number (column %) or mean ± standard deviation, as appropriate Percentages exclude missing values from denominators APACHE II Acute Physiology and Chronic Health Evaluation II
association between ethnicity and CMO status. SPSS version 18 was used for statistical analyses. A two-tailed test of statistical significance was set at p < 0.05.
1255 patients were included in our analysis: 650 Whites, 387 Hispanics, and 218 Blacks. Age was the only baseline characteristic that was different between groups. A slight difference in age was observed between groups: Whites (56 ± 15 years), Blacks (52 ± 14 years), and Hispanics (53 ± 15 years) (p = 0.01). There was no difference in disease severity as measured by the APACHE II score or the Hunt-Hess grade. A difference in the modified Fisher scale (mFS) was seen, mainly driven by differences in mFS 3 and mFS 4 percentages across groups. [See Table 1] Table 2 shows the hospital complications and discharge disposition. No differences were seen in-hospital complications. Mortality was similar between the different ethnicities: 119 (18 %) Whites, 41 (19 %) Blacks, and 69 (18 %) Hispanics died while in-hospital. However, type of death was different: Whites (34 % of deaths) were significantly less likely to die from brain death compared to Blacks (60 %) and Hispanics (52 %). As far as aggressiveness of treatment at the end of life, DNR orders were more common in Whites (117, 18 %) compared to Blacks (27, 12 %) and Hispanics (49, 13 %) (p = 0.03). CMO was more commonly ordered for Whites (93, 14 %) than Blacks (22, 10 %) and Hispanics (36, 9 %) (p = 0.04). Total hospital length of stay was slightly higher in Blacks and Hispanics; however, ICU length of stay was not different. In a multivariate analysis controlling
Table 2 Hospital complications and outcomes
White (n = 650)
Black (n = 218)
Hispanic (n = 387)
p Value
Complications Mechanical ventilation
307 (47)
108 (50)
175 (45)
0.50
Sepsis
62 (10)
23 (11)
22 (9)
0.70
Pneumonia
148 (23)
43 (20)
73 (19)
0.30
Hydrocephalus, symptomatic
252 (39)
75 (34)
128 (33)
0.15
Delayed cerebral ischemia Symptomatic vasospasm
119 (18)
39 (18)
53 (14)
0.14
Vasospasm-related infarction
72 (12)
15 (7)
42 (11)
0.19
End-of-life care Do-not-resuscitate
117 (18)
27 (12)
49 (13)
0.03
Comfort measures only Hospital Length of Stay (median days, IQR)
93 (14) 12 (8,21)
22 (10) 14.5 (9,24)
36(9) 13 (9,24)
0.04 0.04
ICU Length of Stay (median days, IQR)
8 (5,13)
9 (6,13)
8 (5,12)
0.53
Disposition at discharge Values are number (column %) Percentages exclude missing values from denominators SNF skilled nursing facilities
Data in 229 dead patients
0.30
Home
313 (48)
102 (47)
Acute rehabilitation center
130 (20)
45 (21)
214 (55) 65 (17)
SNF or other hospital
85 (13)
30 (14)
39 (10)
Death
119 (18)
41 (19)
69 (18)
1.00
Brain Death
40 (34)
24 (60)
36 (52)
ORIGINAL ARTICLE
Ethnic Disparities in End-of-Life Care After Subarachnoid Hemorrhage H. Alex Choi • Andres Fernandez • Sang-Beom Jeon • J. Michael Schmidt E. Sander Connolly • Stephan A. Mayer • Jan Claassen • Neeraj Badjatia • Kenneth M. Prager • Kiwon Lee
•
Ó Springer Science+Business Media New York 2014
Abstract Background It is common for patients who die from subarachnoid hemorrhage to have a focus on comfort measures at the end of life. The potential role of ethnicity in end-of-life decisions after brain injury has not been extensively studied. Methods Patients with subarachnoid hemorrhage were prospectively followed in an observational database. Demographic information including ethnicity was collected
H. A. Choi (&) S.-B. Jeon K. Lee Departments of Neurosurgery and Neurology, The University of Texas Health Science Center at Houston, 6431 Fannin St, MSB 7.154, Houston, TX 77030, USA e-mail: [email protected] A. Fernandez J. M. Schmidt J. Claassen Department of Neurology, Division of Neurocritical Care, Columbia University College of Physicians and Surgeons, New York, NY, USA S.-B. Jeon Department of Neurology, Asan Medical Center, University of Ulsan College of Medicine, Seoul, Republic of Korea
from medical records and self-reported by patients or their family. Significant in-hospital events including do-notresuscitate orders, comfort measures only orders (CMO; care withheld or withdrawn), and mortality were recorded prospectively. Results 1255 patients were included in our analysis: 650 (52 %) were White, 387 (31 %) Hispanic, and 218 (17 %) Black. Mortality was similar between the groups. CMO was more commonly observed in Whites (14 %) compared to either Blacks (10 %) or Hispanics (9 %) (p = 0.04). In a multivariate analysis controlling for age and Hunt-Hess grade, Hispanics were less likely to have CMO than Whites (OR, 0.6; 95 %CI, 0.4–0.9; p = 0.02). Of the 229 patients who died, 77 % of Whites had CMO compared to 54 % of Blacks and 49 % of Hispanics (p < 0.01). In a multivariate analysis, Blacks (OR, 0.3; 95 %CI, 0.2–0.7; p < 0.01) and Hispanics (OR, 0.3; 95 %CI, 0.2–0.6; p < 0.01) were less likely to die with CMO orders than Whites. Conclusion After subarachnoid hemorrhage, Blacks and Hispanics are less likely to die with CMO orders than Whites. Further research to confirm and investigate the causes of these ethnic differences should be performed.
E. S. Connolly Department of Neurosurgery, Columbia University College of Physicians and Surgeons, New York, NY, USA
Keywords Subarachnoid hemorrhage Ethnic disparities End-of-life care Hospice
S. A. Mayer Institute for Critical Care Medicine, Icahn School of Medicine at Mount Sinai, New York, NY, USA
Introduction
N. Badjatia Department of Neurology, University of Maryland, Baltimore, MD, USA K. M. Prager Department of Internal Medicine, Columbia University College of Physicians and Surgeons, New York, NY, USA
Subarachnoid hemorrhage (SAH) is associated with a high risk of mortality. Advances in critical care management of patients with SAH have contributed to a decrease in overall mortality [1]. Critical care management has also brought to the forefront ethical issues regarding withdrawal or withholding of life-prolonging measures. As our treatment of
123
Neurocrit Care
severe neurologic disease has improved, we are sometimes able to save lives but unable to avoid an outcome of prolonged severe disability. Often life and death decisions are focused around quality of life and severe disability [2]. General population based and disease-specific studies in the US have consistently shown that patients of Black and Hispanic ethnicities, are less likely to choose hospice care and less likely to have Do Not Resuscitate (DNR) orders at the end of life. [3–6] The reasons for these differences are unclear but most likely involve a complex interaction of social, educational, religious, and cultural issues. Adding to this complexity is the often unclear functional prognosis in brain-injured patients and the inability of patients to express their own wishes after their injury. Few studies have addressed the ethnic disparities which influence decisions made by surrogates for patients at the end of life and there are no studies of this kind in the SAH population [5]. In this study, we examined potential ethnic disparities in end-of-life decisions made by surrogate decision makers after SAH. We hypothesized that Hispanic and Black patients died less frequently with a comfort care measures only orders (CMO) compared to non-Hispanic whites.
Methods We identified patients for this study from the SAH Outcomes Project, a single-center, prospective observational cohort. Included subjects were enrolled from August 1996 to January 2011. Study details have been previously described [7]. Inclusion criteria include age 18 or above, and diagnosis of SAH by computed tomography (CT) scan or cerebral spinal fluid. Demographic characteristics, hospital complications, and in-hospital mortality were collected prospectively. The study is approved by the Columbia University institutional review board. Race-ethnicity was determined by self-identification when possible or from family members or next-of-kin based on five categories: 1) White, non- Hispanic; 2) Hispanic; 3) Black; 4) Asian, or 5) Other/unknown. Only subjects who were White, Hispanic, or Black were included in this study, because there were significantly lower numbers of Asians and others/unknown. End-of-Life Decisions Decisions to discuss limitations of life-sustaining measures with family members were left up to the physician treating team. Three categories of limitation of treatment were collected. 1) DNR/DNI: no resuscitative measures performed in the event of cardiopulmonary arrest. 2) Withholding of life-sustaining measures: life-sustaining
123
interventions withheld with a focus on comfort measures e.g.no further increase in vasopressor use for hypotension; no intubation, or mechanical ventilation in case of respiratory failure. 3) Withdrawal of life-sustaining measures: life-sustaining measures withdrawn with a focus on comfort measures only. For the purpose of this study, given that there are only subtle differences between the withholding treatment category and the withdrawal of treatment category, they were combined into a CMO designation. The institutional withdrawal of life-sustaining measures policy from 1996 to 2010 stated that decisions for withdrawal of life-sustaining measures could be made by a surrogate when the patient lacked decisional capacity only if: 1) there was no chance for meaningful recovery; and 2) there was either a) written advance directive that the patient would not wish life support under the current circumstances, or b) a patient-designated health care agent who requested removal of life support based on the patient’s wishes, or c) a surrogate who presented clear and convincing evidence that the patient had verbally stated his/her wishes not to continue on life support under the current circumstances. The above items needed to be agreed upon by two separate attending physicians. In 2010, New York State passed the Family Health Care Decisions Act which allowed surrogates of incapacitated patients to make medical decisions on behalf of patients on the basis of their best interests even in the absence of a living will or without being a health care agent [8]. The law went into effect on June 1st, 2010 and the hospital changed its protocol accordingly. Mortality Cause of death was coded as cardiopulmonary death or brain death in accordance to hospital protocol and consistent with the American Academy of Neurology criteria for brain death [9, 10]. The institutional protocol at the time called for two brain death examinations at least six hours apart with an apnea test and confirmatory testing as needed. Statistical Analysis Demographic differences between ethnicities were analyzed using the ANOVA or Kruskal–Wallis test for continuous variables, and using the X2 test for categorical variables. Differences in-hospital complications and outcomes were analyzed using the X2 test. Univariate and multivariate analysis was used to examine associations between ethnicity and DNR status and CMO status. Factors to include in the multivariate model were chosen based on biological plausibly or known associations with mortality in SAH. In the subset of patients who died, univariate and multivariate analysis was performed to examine the
Neurocrit Care
Results
Table 1 Patient characteristics White Black Hispanic p Value (n = 650) (n = 218) (n = 387) Demographics Age, years
56 ± 15
52 ± 14
53 ± 15
0.01
Sex, female
436 (67)
157 (72)
243 (63)
0.07
1, 2
293 (46)
90 (42)
159 (42)
3 4
149 (23) 99 (15)
59 (27) 29 (13)
104 (27) 54 (14)
5
100 (16)
38 (18)
66 (17)
1
192 (31)
87 (41)
127 (33)
2
64 (10)
18 (8)
33 (9)
3
226 (36)
66 (31)
155 (41)
4
148 (24)
43 (20)
65 (17)
12 ± 7.9
13 ± 8.3
12 ± 8.2
Disease Severity Hunt-Hess grade
0.63
Modified Fisher Scale
APACHE II
0.03
0.2
Values are number (column %) or mean ± standard deviation, as appropriate Percentages exclude missing values from denominators APACHE II Acute Physiology and Chronic Health Evaluation II
association between ethnicity and CMO status. SPSS version 18 was used for statistical analyses. A two-tailed test of statistical significance was set at p < 0.05.
1255 patients were included in our analysis: 650 Whites, 387 Hispanics, and 218 Blacks. Age was the only baseline characteristic that was different between groups. A slight difference in age was observed between groups: Whites (56 ± 15 years), Blacks (52 ± 14 years), and Hispanics (53 ± 15 years) (p = 0.01). There was no difference in disease severity as measured by the APACHE II score or the Hunt-Hess grade. A difference in the modified Fisher scale (mFS) was seen, mainly driven by differences in mFS 3 and mFS 4 percentages across groups. [See Table 1] Table 2 shows the hospital complications and discharge disposition. No differences were seen in-hospital complications. Mortality was similar between the different ethnicities: 119 (18 %) Whites, 41 (19 %) Blacks, and 69 (18 %) Hispanics died while in-hospital. However, type of death was different: Whites (34 % of deaths) were significantly less likely to die from brain death compared to Blacks (60 %) and Hispanics (52 %). As far as aggressiveness of treatment at the end of life, DNR orders were more common in Whites (117, 18 %) compared to Blacks (27, 12 %) and Hispanics (49, 13 %) (p = 0.03). CMO was more commonly ordered for Whites (93, 14 %) than Blacks (22, 10 %) and Hispanics (36, 9 %) (p = 0.04). Total hospital length of stay was slightly higher in Blacks and Hispanics; however, ICU length of stay was not different. In a multivariate analysis controlling
Table 2 Hospital complications and outcomes
White (n = 650)
Black (n = 218)
Hispanic (n = 387)
p Value
Complications Mechanical ventilation
307 (47)
108 (50)
175 (45)
0.50
Sepsis
62 (10)
23 (11)
22 (9)
0.70
Pneumonia
148 (23)
43 (20)
73 (19)
0.30
Hydrocephalus, symptomatic
252 (39)
75 (34)
128 (33)
0.15
Delayed cerebral ischemia Symptomatic vasospasm
119 (18)
39 (18)
53 (14)
0.14
Vasospasm-related infarction
72 (12)
15 (7)
42 (11)
0.19
End-of-life care Do-not-resuscitate
117 (18)
27 (12)
49 (13)
0.03
Comfort measures only Hospital Length of Stay (median days, IQR)
93 (14) 12 (8,21)
22 (10) 14.5 (9,24)
36(9) 13 (9,24)
0.04 0.04
ICU Length of Stay (median days, IQR)
8 (5,13)
9 (6,13)
8 (5,12)
0.53
Disposition at discharge Values are number (column %) Percentages exclude missing values from denominators SNF skilled nursing facilities
Data in 229 dead patients
0.30
Home
313 (48)
102 (47)
Acute rehabilitation center
130 (20)
45 (21)
214 (55) 65 (17)
SNF or other hospital
85 (13)
30 (14)
39 (10)
Death
119 (18)
41 (19)
69 (18)
1.00
Brain Death
40 (34)
24 (60)
36 (52)
