Minority race and male sex as risk factors for non-beneficial gastrostomy tube placements after stroke
Minority race and male sex as risk factors for non-beneficial gastrostomy tube placements after stroke
Roland Faigle 1 2
Joseph A. Carrese 2
Lisa A. Cooper 0 2
Victor C. Urrutia 1 2
Rebecca F. Gottesman 1 2
0 Department of Medicine, Johns Hopkins University School of Medicine , Baltimore, MD , United States of America
1 Department of Neurology, Johns Hopkins University School of Medicine , Baltimore, MD , United States of America, 2 Department of Medicine, Johns Hopkins University School of Medicine and the Johns Hopkins Berman Institute of Bioethics, Johns Hopkins University , Baltimore, MD , United States of America
2 Editor: N.S. Duesbery , Spectrum Health , UNITED STATES
Data Availability Statement: All relevant data are
within the paper and its Supporting Information
Funding: Dr. Faigle is supported by an NINDS
Career Development Award (K23 NS101124). Dr.
Cooper is supported by a grant from the National
Heart, Lung, and Blood Institute (K24HL083113).
Dr Gottesman is supported by a grant from the
National Institute on Aging (K24AG052573). The
funders had no role in study design, data collection
Percutaneous endoscopic gastrostomy (PEG) tubes are widely used for enteral feeding
after stroke; however, PEG tubes placed in patients in whom death is imminent are
We sought to determine whether placement of non-beneficial PEG tubes differs by race and
Design and setting/participants
In this retrospective cohort study, inpatient admissions for stroke patients who underwent
palliative/withdrawal of care, were discharged to hospice, or died during the hospitalization,
were identified from the Nationwide Inpatient Sample between 2007 and 2011. Logistic
regression was used to evaluate the association between race and sex with PEG
Of 36,109 stroke admissions who underwent palliative/withdrawal of care, were discharge
to hospice, or experienced in-hospital death, a PEG was placed in 2,258 (6.3%). Among
PEG recipients 41.1% were of a race other than white, while only 22.0% of patients without
PEG were of a minority race (p<0.001). The proportion of men was higher among those with
compared to without a PEG tube (50.0% vs. 39.2%, p<0.001). Minority race was associated
with PEG placement compared to whites (OR 1.75, 95% CI 1.57±1.96), and men had 1.27
times higher odds of PEG compared to women (95% CI 1.16±1.40). Racial differences were
most pronounced among women: ethnic/racial minority women had over 2-fold higher odds
and analysis, decision to publish, or preparation of
Competing interests: Dr. Gottesman is an
Associate Editor for Neurology. The other authors
have declared that no competing interests exist.
of a PEG compared to their white counterparts (OR 2.09, 95% CI 1.81±2.41), while male
ethnic/racial minority patients had 1.44 increased odds of a PEG when compared to white
men (95% CI 1.24±1.67, p-value for interaction <0.001).
Minority race and male sex are risk factors for non-beneficial PEG tube placements after
Stroke is a leading causes of disability and mortality in the United States[
]. Dysphagia, or
difficulty swallowing, can be identified in up to 50% of acute stroke patients[
], and in some
patients persists beyond the acute care hospitalization[
]. While nasogastric (NG) tubes may
be used for enteral feeding for up to several weeks, percutaneous endoscopic gastrostomy
(PEG) tubes are commonly placed in stroke patients when dysphagia is expected to persist
beyond 4 to 6 weeks[6±9].
Procedures and interventions intended to benefit patients during their recovery process
after the acute care hospitalization, such as PEG tubes, are considered non-beneficial in
patients who die during their acute care hospitalization or in whom death is imminent[
Stroke patients are particularly susceptible to receive such non-beneficial interventions
because of the associated high rates of mortality, and variable use and timing of withdrawal of
life-sustaining treatments in this patient population[11±13]. Risk factors for withdrawal of
treatment and do not resuscitate (DNR) orders in stroke patients have previously been
11, 12, 14
]. African Americans are less likely to utilize withdrawal of treatment and DNR
orders compared to their white counterparts. Similarly, women are more likely to undergo
withdrawal of treatment compared to men[
11, 12, 14
The impact of withdrawal of treatment orders and in-hospital mortality on non-beneficial
procedures in stroke patients has not been studied systematically. Patients in whom death is
imminent may unnecessarily be subjected to procedures typically reserved for those who are
expected to survive beyond the acute care hospitalization. Such interventions, including PEG
tubes, may be appropriate only if the functional status at the time of the procedure is
sustainable and deemed acceptable by the patient and/or the patient's decision-maker(s). Since
clinical course and outcomes after stroke are frequently uncertain during the first few days after
hospital admission, placement of a PEG tube early on during the hospital course may be
premature in patients with uncertain prognosis[
]. Since NG tubes may serve as a viable
alternative to PEG tubes for up to several weeks, PEG tubes are typically placed during the later
course of the hospitalization, after patients and their decision-makers have had ample time to
deliberate prognosis and goals of care[
]. Because PEG placement constitutes a surgical
procedure associated with additional risks, costs, and no clear medical benefit over NG tubes during
the first few weeks, PEG tubes are discouraged unless it is anticipated that they will be utilized
beyond the acute hospitalization[
8, 10, 15
]. Therefore, PEG tubes placed in patients who
subsequently die while in the hospital, undergo palliative care or care withdrawal, and those who
are discharged to hospice are considered non-beneficial[
]. In the present study we aimed to
determine the role of race and sex as risk factors for non-beneficial PEG placements in stroke
patients, as defined by PEG placement in patients who underwent palliative/withdrawal of
care, were discharged to hospice, or failed to survive their hospitalization.
2 / 12
Data were obtained from the Nationwide Inpatient Sample (NIS), part of the Healthcare Cost
and Utilization Project (HCUP)[
]. The NIS is the largest all-payer inpatient database in the
US, representing a 20% stratified sample of all admissions to non-federal US hospitals. All
diagnoses and procedures are recorded using International Classification of Diseases version 9
Clinical Modification (ICD9-CM) codes. Detailed information on the design and contents of
the NIS is available at http://www.hcup-us.ahrq.gov. Because NIS data are publically available
and contain no personal identifying information, this study was exempt from institutional
review board approval.
We identified adult cases with primary diagnosis of non-traumatic ICH and ischemic stroke
by using ICD9-CM codes 431, 433.01, 433.11, 433.21, 433.31, 433.81, 433.91, 434.01, 434.11,
434.91, and 436 between 2007 and 2011[
]. Because the unit of observation in NIS is
discharge after hospitalization, cases transferred to another hospital were excluded in order to
prevent double counting of the same patient. This algorithm has been shown to identify acute
ischemic and hemorrhagic stroke with high sensitivity and specificity[
]. Cases with
missing information on race/ethnicity and sex, the primary exposures of interest, were excluded.
Definition of a non-beneficial PEG
We restricted the study population to cases that survived the first three days of their
hospitalization, but subsequently died while in the hospital, underwent palliative/withdrawal of care, or
were discharged to hospice; we considered placement of a PEG tube in this population as
nonbeneficial. Use of palliative/withdrawal of care was identified by ICD9-CM code V66.7. This
code identifies documented use of palliative care measures irrespective of the delivery mode
(i.e. via a palliative care consultation service or integrated into routine clinical practice by the
care team), and is coded when terms such as comfort care, end-of-life care, and hospice care
(or similar terms), are written in the patient record[
Since there is no previously uniformly accepted definition of a non-beneficial PEG, we
performed sensitivity analysis by analyzing PEG placement only among patients who died,
underwent withdrawal of care, or were discharged to hospice within 14 days of admission.
Primary exposures and outcome of interest
The primary exposures of interest were race/ethnicity and sex. The primary outcome of
interest was placement of a non-beneficial PEG as identified by ICD9-CM procedure code 43.11.
Comorbidity and severity adjustment
We calculated the Charlson comorbidity index, a weighted score of 17 different comorbidities
validated for outcome adjustment for analyses of administrative data sets using ICD9-CM
], for each patient. Case severity was determined using the all patient refined
diagnosis-related groups (APR-DRGs). The APR-DRG algorithm is a validated and reliable
indicator of mortality, and is commonly used as a severity indicator in studies relating to
3 / 12
Comparisons of sociodemographic, hospital-level, and clinical characteristics among patients
with and without PEG tube were made using Chi2 and Wilcoxon rank-sum tests for categorical
and continuous variables, respectively. Univariable logistic regression was performed to
determine the unadjusted association of PEG tube placement with race and sex, respectively.
Multivariable models were adjusted for age, hospital characteristics (teaching status, bed size,
location, region, and annual volume of stroke cases), discharge quarter, weekend admission,
modified Charlson Comorbidity Index, APR-DRG severity subclass, insurance status, median
household income per patient's ZIP code, hypertension, diabetes mellitus, dyslipidemia,
coronary artery disease, peripheral vascular disease, congestive heart failure, atrial fibrillation,
valvular disease, anemia, thrombocytopenia, alcohol abuse, drug abuse, chronic kidney disease,
and medical complications, such as pneumonia, urinary tract infection, sepsis, gastrointestinal
bleeding, deep vein thrombosis, and pulmonary embolism. For the primary analysis, we
excluded observations with missing information on the primary exposures of interest; since
the variable race had substantial missingness (15.1%), we performed sensitivity analysis
including imputed values for race via multiple imputation via chained equations (MICE) in order to
reduce potential bias as a result of missing data. We used a Generalized Estimation Equations
(GEE) approach to account for clustering of patients within hospitals. Statistical analysis was
performed using STATA version 13 (Stata Statistical Software: Release 13. College Station,
TX). A p-value of <0.05 was considered statistically significant. 95% confidence intervals are
reported. Statistical interaction between race and sex on PEG placement was explored.
Among the 36,109 cases that met all inclusion criteria (S1 Fig), 2,258 (6.3%) underwent PEG
placement, thus termed non-beneficial PEG placements. Patients who received a PEG tube
were more likely to be male (50.0% vs. 39.2%, p<0.001) than were patients who did not receive
a PEG. While 7.8% of men received a non-beneficial PEG, the percentage of women receiving
a non-beneficial PEG tube was 5.2% (p<0.001). While only 22.0% of patients without PEG
were of a minority race, 41.1% of all PEG recipients were of a race other than white (p<0.001).
Among black patients 12.5% received a PEG, compared to 4.8% receiving a PEG among white
patients (p<0.001). Among all ethnic/racial minorities, 11.1% received a PEG. Further baseline
characteristics are presented in Table 1.
Minority race and male sex are associated with increased risk of
In univariable analysis, the odds of PEG placement were significantly higher for patients of a
minority race compared to whites: OR 2.85, 95% CI 2.55±3.17 in blacks, OR 2.32, 95% CI
2.00±2.69 in Hispanics, and OR 1.66, 95% CI 1.30±2.12 in Asians/Pacific Islanders. These
results persisted after multivariable adjustment for age, comorbidities, medical complications
including dysphagia, and hospital characteristics (Table 2). Taken together, minority patients
had 1.75 times higher odds of PEG compared to whites in the adjusted model (95% CI 1.57±
1.96). Male sex was associated with PEG placement in univariable analysis (OR 1.55, 95% CI
1.43±1.69), and this association persisted in adjusted models (OR 1.27, 95% CI 1.16±1.40;
Table 2). To reduce the potential for bias from missing data, we performed sensitivity analysis
after multiple imputation of the missing race values. This sensitivity analysis yielded similar
4 / 12
PLOS ONE | https://doi.org/10.1371/journal.pone.0191293
No PEG (n = 33,851)
results as complete-case analysis with regard to effect size and statistical significance (S1
Racial differences in non-beneficial PEG tubes after stroke vary by sex
In the entire study population comprised of patients with in-hospital mortality, discharge to
hospice, or undergoing palliative/withdrawal of care, non-beneficial PEG tubes after any
stroke were most common among ethnic/racial minority men (11.9%), while only 3.8% of
white women received a PEG tube (Table 3). Similarly, the adjusted odds of PEG were highest
among ethnic/racial minority men (OR 2.10, 95% CI 1.78±2.48, compared to white women;
Table 3). Differences by race were most pronounced among women: ethnic/racial minority
Model adjusted for age, hospital teaching status, hospital bed size, hospital location, hospital region, and annual volume of stroke cases, discharge quarter, weekend
admission status, modified Charlson Comorbidity Index, APR-DRG severity subclass, insurance status, median household income per patient's ZIP code, hypertension,
diabetes mellitus, dyslipidemia, coronary artery disease, peripheral vascular disease, congestive heart failure, atrial fibrillation, valvular disease, anemia,
thrombocytopenia, alcohol abuse, drug abuse, chronic kidney disease, pneumonia, urinary tract infection, sepsis, gastrointestinal bleeding, deep vein thrombosis, and
6 / 12
Model adjusted for age, hospital teaching status, hospital bed size, hospital location, hospital region, and annual
volume of stroke cases, discharge quarter, weekend admission status, modified Charlson Comorbidity Index,
APR-DRG severity subclass, insurance status, median household income per patient's ZIP code, hypertension,
diabetes mellitus, dyslipidemia, coronary artery disease, peripheral vascular disease, congestive heart failure, atrial
fibrillation, valvular disease, anemia, thrombocytopenia, alcohol abuse, drug abuse, chronic kidney disease,
pneumonia, urinary tract infection, sepsis, gastrointestinal bleeding, deep vein thrombosis, and pulmonary
women had over 2-fold higher odds of PEG compared to their white counterparts (OR 2.08,
95% CI 1.81±2.41), while male ethnic/racial minority patients had 1.44-fold increased odds of
a PEG when compared to white men (95% CI 1.24±1.67, p-value for interaction <0.001; Fig
1). The odds of non-beneficial PEG tubes did not differ between ethnic/racial minority women
and men (OR 1.01, 95% CI 0.86±1.18 in ethnic/racial minority men compared to minority
Among patients with ischemic stroke, ethnic/racial minority men and women both had
similarly increased odds of non-beneficial PEG when compared to white women (OR 2.32,
95% CI 1.91±2.81, for ethnic/racial minority men vs. OR 2.17, 95% CI 1.85±2.54, for ethnic/
racial minority women; Table 3). Increased odds of PEG for ethnic/racial minorities vs. whites
were observed among both men and women after ischemic stroke, however, were more
pronounced among women (p-value for interaction 0.012; Fig 1). In patients with ICH, the odds
of PEG was significantly increased in ethnic/racial minority women (OR 1.91, 95% CI 1.39±
2.62), white men (OR 1.64, 95% CI 1.23±2.19), and ethnic/racial minority men (OR 1.71, 95%
CI 1.22±2.39), when compared to white women (Table 3). Racial differences for PEG
placement after ICH were seen among women, but not men (p-value for interaction 0.002; Fig 1).
Sensitivity analysis: Patients with death or discharge to hospice within 14
days of admission
There is no unified consensus as to what constitutes a non-beneficial PEG tube (i.e. a PEG
tube placed in a patient on hospital day 15 who weeks later succumbs to an unforeseen
7 / 12
Fig 1. Graphic representation of odds ratios of PEG among ethnic/racial minority women, white men, and ethnic/
racial minority men, compared to white women as reference. Data are presented for all stroke (top panel), and stratified
by stroke subtype (medium and lower panel), respectively. indicates a p-value <0.001 for comparison of PEG in ethnic/
racial minority vs. white for each sex. P-values for interaction following square brackets compare the odds of PEG in
ethnic/racial minorities vs. white between men and women. OR: odds ratio; CI: confidence interval; ICH: intracerebral
complication may not uniformly be regarded as ªnon-beneficialº). Therefore, we performed a
sensitivity analysis by investigating PEG placement only among stroke patients who were
discharged to hospice, died, or received end-of-life/palliative care within the first 14 days of their
hospitalization. Any PEG in this population would have been placed very close to the time of
death and/or relatively early in the hospital course.
Of all 2,258 PEG tubes in our original study population, 1,079 (47.8%) were placed in
patients who were discharged to hospice or died within 14 days of admission. The adjusted
odds of PEG in this patient population was higher in white men (OR 1.43, 95% CI 1.23±1.67),
ethnic/racial minority women (OR 2.00, 95% CI 1.64±2.45), and ethnic/racial minority men
(OR 2.17, 95% CI 1.74±2.72), when compared to white women (S2 Table).
In the present study we have identified minority race and male sex as risk factors for
non-beneficial PEG placement in stroke patients. Non-beneficial procedures or interventions in
patients who later transition to palliative care and treatment withdrawal and/or die during
their hospitalization may in part be due to continuation of aggressive therapy despite no
anticipated benefit (i.e. because of family wishes, or providers' reluctance to `give up'), or absence of
timely prediction of a poor outcome. Since timing and accuracy of outcome prediction should
not differ by race or sex, it is more likely that unnecessary continuation of aggressive therapy
among ethnic/racial minorities and men accounts for the increased rates of non-beneficial
8 / 12
PEG tubes placed in these populations. Therefore, the higher rates of PEG tubes placed in men
and ethnic/racial minority patients in our study suggests that ongoing aggressive care in some
patients may result in non-beneficial procedures.
Racial differences in utilization of interventions at the end of life have been described, i.e.
black patients are less likely to institute DNR orders or request withdrawal of active treatment
compared to their white counterparts[
]. Similarly, white patients more commonly have
]. Prolonged aggressive care despite a low likelihood of benefit may
result in higher rates of unnecessary procedures in ethnic/racial minorities. Among recipients
of non-beneficial PEG tubes, ethnic/racial minorities had longer length of stay and longer
time-to-PEG placement. This indicates that prolonged aggressive care may increase the risk of
non-beneficial procedures by providing additional time during which non-beneficial
treatment is administered.
The higher odds of non-beneficial PEG in men compared to women, as observed in our
study, may reflect that women overall receive less aggressive care after stroke compared to
their male counterparts[28±30]. In addition, our finding is in line with other studies suggesting
that women in general are less likely to receive life-sustaining interventions and are more likely
to have DNR orders[31±33]. We observed sex disparities only among white but not among
ethnic/racial minority patients, although both ethnic/racial minority men and women had
higher odds of PEG than white men or women. Although the reasons for this finding are not
entirely clear, it may suggest a hierarchical order of disparities, i.e. racial differences in PEG
placement after stroke may ªtrumpº differences by sex, and sex disparities may only be
relevant among whites.
We restricted our study population to patients in whom a PEG tube is considered
non-beneficial; 6.3% of those patients received a PEG, exceeding the rate of PEG tubes placed in our
total ischemic and hemorrhagic stroke population (5.1%). While some in-hospital deaths after
PEG may have been unexpected and therefore difficult to predict, timely identification of
patients in whom further procedures are unlikely to be beneficial may aid in reducing
nonbeneficial PEG placements. Although NG tubes are safe and sufficient to provide enteral
nutrition for up to about 2 weeks after stroke, more than half of the patients in the present study,
black and white alike, received their PEG within the first 10 days of hospitalization. This is
consistent with other recent reports[
]. This suggests that PEG tubes were placed
prematurely in a substantial number of patients, many of which likely had no meaningful chance of
recovery at the time of PEG placement as indicated by the median time from PEG to death/
hospice discharge of just 6 days. Formal and objective criteria for PEG placement after stroke
may help avoid PEG tubes in patients who are unlikely to benefit from it.
Our study has several limitations. The NIS does not contain clinical and physiological data
on stroke volume or location, level of consciousness, or associated laboratory parameters,
which may confound the described associations of race/sex with PEG placement. We
attempted to mitigate this shortcoming by adjusting all regression models for the Charlson
Comorbiditity Index, a validated measure of patient comorbidities in stroke[
and missing data may occur in large administrative datasets; however, it is unlikely that there
is differential miscoding by race or sex. To address missingness for race, we performed a
sensitivity analysis using multiple imputation for the race variable. Although ICD9-CM code V66.7
identifies palliative/withdrawal of care services use with high sensitivity and specificity,
there is a possibility of under-reporting or under-coding patients who actually received
palliative care or undergo care withdrawal. While we attempted a meaningful definition of a
ªnonbeneficial PEGº, our definition is subjective, yet not exhaustive; i.e. PEG tubes placed in
patients who regain swallow function and do not use their PEG after discharge may also be
considered non-beneficial, but were not examined in our study.
9 / 12
Despite these limitations, our data suggest that PEG tubes placed in stroke patients with
subsequent short-term mortality are relatively common, and that ethnic/racial minorities and
men are at particular risk. In the context of engaging all patients and their next of kin
decision-makers in conversations about goals of care and the appropriate use of available
interventions, we propose that health care professionals carefully examine their approaches to care of
ethnic/racial minorities and men when counseling families on neurological outcome and need
for PEG tubes after stroke. Standardized interdisciplinary approaches for evaluation and
timing of PEG placement in light of the underlying stroke prognosis may aid in minimizing
placement of non-beneficial PEG tubes.
S1 Fig. Flow diagram indicating selection of the study population. Not mutually exclusive.
S1 Table. Multivariable analysis for race and sex determinants of non-beneficial PEG
placement after stroke. Sensitivity analysis including imputed data for race; n = 42,235.
S2 Table. Sensitivity analysis: Odds of PEG among stroke patients who died or were
discharge to hospice within 14 days of admission (n = 32,560).
Data curation: Roland Faigle.
Formal analysis: Roland Faigle.
Methodology: Roland Faigle, Rebecca F. Gottesman.
Project administration: Roland Faigle.
Supervision: Rebecca F. Gottesman.
Writing ± original draft: Roland Faigle.
Conceptualization: Roland Faigle, Joseph A. Carrese, Lisa A. Cooper, Rebecca F. Gottesman.
Writing ± review & editing: Roland Faigle, Joseph A. Carrese, Lisa A. Cooper, Victor C.
Urrutia, Rebecca F. Gottesman.
10 / 12
11 / 12
1. Feigin VL , Lawes CM , Bennett DA , Barker-Collo SL , Parag V . Worldwide stroke incidence and early case fatality reported in 56 population-based studies: a systematic review . The LancetNeurology . 2009 ; 8 ( 4 ): 355 ± 69 . https://doi.org/10.1016/S1474- 4422 ( 09 ) 70025 - 0
2. Koton S , Schneider AL , Rosamond WD , Shahar E , Sang Y , Gottesman RF , et al. Stroke incidence and mortality trends in US communities, 1987 to 2011 . Jama. 2014 ; 312 ( 3 ): 259 ± 68 . https://doi.org/10.1001/ jama. 2014 .7692 PMID: 25027141
3. Mann G , Hankey GJ , Cameron D. Swallowing function after stroke: prognosis and prognostic factors at 6 months. Stroke; a journal of cerebral circulation . 1999 ; 30 ( 4 ): 744 ± 8 .
4. Martino R , Foley N , Bhogal S , Diamant N , Speechley M , Teasell R . Dysphagia after stroke: incidence, diagnosis, and pulmonary complications . Stroke; a journal of cerebral circulation . 2005 ; 36 ( 12 ): 2756 ± 63 . doi: 01.STR. 0000190056 .76543.eb [pii].
5. Smithard DG , O'Neill PA , England RE , Park CL , Wyatt R , Martin DF , et al. The natural history of dysphagia following a stroke . Dysphagia . 1997 ; 12 ( 4 ): 188 ± 93 . https://doi.org/10.1007/PL00009535 [pii]. PMID: 9294937
6. Corrigan ML , Escuro AA , Celestin J , Kirby DF . Nutrition in the stroke patient. Nutrition in clinical practice: official publication of the American Society for Parenteral and Enteral Nutrition . 2011 ; 26 ( 3 ): 242 ± 52 . https://doi.org/10.1177/0884533611405795 PMID: 21586409
7. James A , Kapur K , Hawthorne AB . Long-term outcome of percutaneous endoscopic gastrostomy feeding in patients with dysphagic stroke . Age and Ageing . 1998 ; 27 ( 6 ): 671 ± 6 . PMID: 10408659
8. Jauch EC , Saver JL , Adams HP Jr., Bruno A , Connors JJ , Demaerschalk BM , et al. Guidelines for the early management of patients with acute ischemic stroke: a guideline for healthcare professionals from the American Heart Association/American Stroke Association . Stroke; a journal of cerebral circulation . 2013 ; 44 ( 3 ): 870 ± 947 . https://doi.org/10.1161/STR.0b013e318284056a PMID: 23370205
9. Stroud M , Duncan H , Nightingale J , British Society of G. Guidelines for enteral feeding in adult hospital patients . Gut . 2003 ; 52 Suppl 7 :vii1± vii12 .
10. Johnston SD , Tham TC , Mason M. Death after PEG: results of the National Confidential Enquiry into Patient Outcome and Death . Gastrointestinal endoscopy. 2008 ; 68 ( 2 ): 223 ±7. https://doi.org/10.1016/j. gie. 2007 . 10 .019 PMID: 18329030
11. Diringer MN , Edwards DF , Aiyagari V , Hollingsworth H . Factors associated with withdrawal of mechanical ventilation in a neurology/neurosurgery intensive care unit . Critical Care Medicine . 2001 ; 29 ( 9 ): 1792 ± 7 . PMID: 11546988
12. Mayer SA , Kossoff SB . Withdrawal of life support in the neurological intensive care unit . Neurology . 1999 ; 52 ( 8 ): 1602 ± 9 . PMID: 10331685
13. Wijdicks EF , Rabinstein AA . Absolutely no hope? Some ambiguity of futility of care in devastating acute stroke . Critical Care Medicine . 2004 ; 32 ( 11 ): 2332 ± 42 . doi: 00003246- 200411000 -00024 [pii]. PMID: 15640651
14. Shepardson LB , Gordon HS , Ibrahim SA , Harper DL , Rosenthal GE . Racial variation in the use of donot-resuscitate orders . Journal of general internal medicine . 1999 ; 14 ( 1 ): 15 ± 20 . PMID: 9893086
15. Loser C , Aschl G , Hebuterne X , Mathus-Vliegen EM , Muscaritoli M , Niv Y , et al. ESPEN guidelines on artificial enteral nutritionÐpercutaneous endoscopic gastrostomy (PEG). Clinical nutrition (Edinburgh , Scotland). 2005 ; 24 ( 5 ): 848 ± 61 .
16. Steiner C , Elixhauser A , Schnaier J. The healthcare cost and utilization project: an overview . Effective clinical practice: ECP . 2002 ; 5 ( 3 ): 143 ± 51 . PMID: 12088294
17. Sacco RL , Kasner SE , Broderick JP , Caplan LR , Connors JJ , Culebras A , et al. An updated definition of stroke for the 21st century: a statement for healthcare professionals from the American Heart Association/American Stroke Association . Stroke; a journal of cerebral circulation . 2013 ; 44 ( 7 ): 2064 ±89. https:// doi.org/10.1161/STR.0b013e318296aeca PMID: 23652265
18. Kokotailo RA , Hill MD . Coding of stroke and stroke risk factors using international classification of diseases, revisions 9 and 10 . Stroke ; a journal of cerebral circulation . 2005 ; 36 ( 8 ): 1776 ± 81 . doi: 01. STR. 0000174293 .17959.a1 [pii].
19. Tirschwell DL , Longstreth WT Jr. Validating administrative data in stroke research . Stroke; a journal of cerebral circulation . 2002 ; 33 ( 10 ): 2465 ± 70 .
20. Brown F , Leon-Chisen N. ICD-9-CM Coding Handbook, With Answers, 2012 Revised Edition : AHA Press. 720 p.
21. Bar B , Hemphill JC 3rd. Charlson comorbidity index adjustment in intracerebral hemorrhage . Stroke; a journal of cerebral circulation . 2011 ; 42 ( 10 ): 2944 ±6. https://doi.org/10.1161/STROKEAHA.111.617639 [doi].
22. Deyo RA , Cherkin DC , Ciol MA . Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases . Journal of clinical epidemiology . 1992 ; 45 ( 6 ): 613 ±9. doi: 0895- 4356 ( 92 ) 90133 - 8 [pii]. PMID: 1607900
23. Edwards N , Honemann D , Burley D , Navarro M. Refinement of the Medicare diagnosis-related groups to incorporate a measure of severity . Health care financing review . 1994 ; 16 ( 2 ): 45 ± 64 . PMID: 10142374
24. Xian Y , Holloway RG , Pan W , Peterson ED. Challenges in assessing hospital-level stroke mortality as a quality measure: comparison of ischemic, intracerebral hemorrhage, and total stroke mortality rates . Stroke; a journal of cerebral circulation . 2012 ; 43 ( 6 ): 1687 ± 90 . https://doi.org/10.1161/STROKEAHA. 111.648600 PMID: 22535276
25. Mack JW , Paulk ME , Viswanath K , Prigerson HG . Racial disparities in the outcomes of communication on medical care received near death . Archives of Internal Medicine . 2010 ; 170 ( 17 ): 1533 ± 40 . https://doi. org/10.1001/archinternmed. 2010 .322 PMID: 20876403
26. Perkins HS , Geppert CM , Gonzales A , Cortez JD , Hazuda HP . Cross-cultural similarities and differences in attitudes about advance care planning . Journal of general internal medicine . 2002 ; 17 ( 1 ): 48 ± 57 . doi: 01032 [pii]. PMID: 11903775
27. Zaide GB , Pekmezaris R , Nouryan CN , Mir TP , Sison CP , Liberman T , et al. Ethnicity, race, and advance directives in an inpatient palliative care consultation service . Palliative & supportive care . 2013 ; 11 ( 1 ):5± 11 . https://doi.org/10.1017/S1478951512000417 PMID: 22874132
28. Kent JA , Patel V , Varela NA . Gender disparities in health care . The Mount Sinai journal of medicine , New York. 2012 ; 79 ( 5 ): 555 ±9. https://doi.org/10.1002/msj.21336 PMID: 22976361
29. Reeves M , Bhatt A , Jajou P , Brown M , Lisabeth L. Sex differences in the use of intravenous rt-PA thrombolysis treatment for acute ischemic stroke: a meta-analysis . Stroke; a journal of cerebral circulation . 2009 ; 40 ( 5 ): 1743 ±9. https://doi.org/10.1161/STROKEAHA.108.543181 PMID: 19228855
30. Towfighi A , Markovic D , Ovbiagele B . Sex differences in revascularization interventions after acute ischemic stroke . Journal of stroke and cerebrovascular diseases: the official journal of National Stroke Association . 2013 ; 22 ( 8 ):e347± 53 . https://doi.org/10.1016/j.jstrokecerebrovasdis. 2013 . 03 .018 PMID: 23660344
31. Bookwala J , Coppola KM , Fagerlin A , Ditto PH , Danks JH , Smucker WD . Gender differences in older adults' preferences for life-sustaining medical treatments and end-of-life values . Death studies . 2001 ; 25 ( 2 ): 127 ± 49 . https://doi.org/10.1080/07481180126202 PMID: 11708352
32. Frankl D , Oye RK , Bellamy PE . Attitudes of hospitalized patients toward life support: a survey of 200 medical inpatients . The American Journal of Medicine . 1989 ; 86 ( 6 ): 645 ± 8 . PMID: 11659207
33. Wenger NS , Pearson ML , Desmond KA , Harrison ER , Rubenstein LV , Rogers WH , et al. Epidemiology of do-not-resuscitate orders. Disparity by age, diagnosis, gender, race, and functional impairment . Archives of Internal Medicine . 1995 ; 155 ( 19 ): 2056 ± 62 . PMID: 7575064
34. George BP , Kelly AG , Albert GP , Hwang DY , Holloway RG . Timing of Percutaneous Endoscopic Gastrostomy for Acute Ischemic Stroke: An Observational Study From the US Nationwide Inpatient Sample . Stroke. 2017 ; 48 ( 2 ): 420 ±7. https://doi.org/10.1161/STROKEAHA.116.015119 PMID: 27965430
35. Wilmskoetter J , Simpson AN , Simpson KN , Bonilha HS . Practice Patterns of Percutaneous Endoscopic Gastrostomy Tube Placement in Acute Stroke: Are the Guidelines Achievable? Journal of stroke and cerebrovascular diseases: the official journal of National Stroke Association . 2016 ; 25 ( 11 ): 2694 ± 700 . doi: S1052- 3057 ( 16 ) 30202 - 6 [pii].
36. Goldstein LB , Samsa GP , Matchar DB , Horner RD . Charlson Index comorbidity adjustment for ischemic stroke outcome studies . Stroke; a journal of cerebral circulation . 2004 ; 35 ( 8 ): 1941 ± 5 . https://doi.org/10. 1161/01.STR. 0000135225 .80898.1c PMID: 15232123
37. Qureshi AI , Adil MM , Suri MF . Rate of utilization and determinants of withdrawal of care in acute ischemic stroke treated with thrombolytics in USA . Medical care. 2013 ; 51 ( 12 ): 1094 ± 100 . https://doi.org/10. 1097/MLR.0b013e3182a95db4 PMID: 24226307