Patients with End-Stage Renal Disease and Acute Surgical Abdomen: Opportunities for Palliative Care

Abstract

Background:

End-stage renal disease (ESRD) is a life-limiting condition that is often complicated by acute abdominal emergency. Palliative care (PC) has been shown to improve the quality of life in patients with serious illness and yet is underutilized. We hypothesize that ESRD patients with abdominal emergency have high unmet PC needs.

Objective:

To characterize the outcomes of ESRD patients with acute surgical abdomen, define PC utilization patterns, and identify areas of unmet PC needs.

Design:

Retrospective study querying the National Inpatient Sample database (2009–2013).

Setting and Subjects:

Subjects were identified using ICD-9 codes for those aged ≥50 with preexisting diagnosis of ESRD with an acute abdominal emergency diagnosis of gastrointestinal perforation, obstruction, or ischemia.

Measurements:

Outcomes included PC rate, in-hospital mortality, discharge disposition, and intensity of care. Multivariable logistic regression analysis was used to identify predictors of PC.

Results:

A total of 9363 patients met the inclusion criteria; 24% underwent surgery, 16% died in hospital, and 43% were discharged to dependent living. Among in-hospital deaths, 23% received PC. Only 4% of survivors with dependent discharge received PC. Surgical mortality was 26%. PC was less utilized in surgical patients than nonsurgical patients. PC was associated with shorter hospital stay. Predictors of PC included increasing age, severity of underlying illness, white race, teaching hospitals, and the Western region.

Conclusions:

Patients with ESRD admitted for acute abdominal emergency have high risk for mortality and functional dependence. Despite this, few receive PC and have a high utilization of nonbeneficial life support at the end of life.

Background

More than 700,000 Americans suffer from end-stage renal disease (ESRD), a life-limiting condition that carries high cost of health care utilization and significant symptom burden and mortality.¹ Newly dialysis-dependent patients >75 years of age experience an annual mortality rate of 40%.^2–4 In addition, patients with ESRD are at increased risk of adverse surgical outcomes compared with their counterparts with normal renal function.^5,6 Despite these risks, 30% of Medicare beneficiaries with ESRD had inpatient surgical procedures during the last 90 days of life.¹

The ability to accurately prognosticate surgical outcomes, especially in emergent settings, is critical to ensuring that treatments are aligned with patient values. Palliative care (PC) centers around achieving this goal-concordant care, and it has been shown to result in better and earlier control of symptoms, shorter hospital stays, and fewer invasive interventions in patients with serious and life-limiting illnesses.^7,8 Research suggests that patients who clearly understand their prognosis are less likely to choose invasive treatments that prolong suffering and time away from home.⁹ PC can also assist with decision making in patients with acute abdominal emergency, regardless of whether surgical intervention is pursued.¹⁰ Although patients with ESRD have been identified as appropriate candidates for PC from the time of diagnosis or initiation of dialysis,^11,12 few are referred.

One of the barriers to PC in the emergency surgical setting is perceived uncertainty regarding prognosis. For patients with ESRD and acute abdominal surgical emergency, improving goal concordance and patient-centered care must, therefore, begin with a better understanding of expected outcomes to formulate treatment plans that best benefit the individual patient. However, surgical outcomes specifically related to acute abdominal emergencies in patients with ESRD remain poorly characterized. In addition, the extent of PC utilization in patients with ESRD and acute abdominal emergency is not yet defined. In this study, we sought to study the outcomes of ESRD patients with acute abdominal emergency and patterns of PC utilization.

Objective

The purpose of this study is (1) to characterize the outcomes of ESRD patients with acute surgical abdomen, (2) to define patterns of PC utilization, and (3) to identify specific areas of unmet PC needs in this population. We hypothesized that patients with ESRD who develop acute abdomen have poor outcomes yet rarely receive PC.

Design

This is a retrospective study that queried the National Inpatient Sample (NIS) from 2009 to 2013. NIS encapsulates nearly 20% of all hospitalizations in the United States, gathering information from 44 states and 4378 hospitals.^13,14

Setting and Study Population

Study subjects were identified using the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM). Patients aged 50 or older with preexisting ICD-9 diagnosis of ESRD who present emergently or urgently with an acute abdominal emergency diagnosis of gastrointestinal perforation, obstruction, or ischemia were included. Cases with acute renal failure or kidney transplant were excluded. Cases were excluded if they had incomplete or missing data on patient or hospital characteristics (Fig. 1). See Appendix Tables A1–A4 for pertinent ICD-9 codes.

FIG. 1.

Establishment of the cohort from the NIS based on ICD-9-CM codes. ESRD, end-stage renal disease; ICD-9-CM, International Classification of Diseases, 9th Revision, Clinical Modification; NIS, National Inpatient Sample.

Measurements

Outcomes

Outcome variables included in-hospital mortality, discharge disposition, PC utilization, intensity of care, hospital length of stay (LOS), and total charges. Discharge disposition was categorized as either independent or dependent living among those who survived the hospitalization. Discharge to independent living is defined as discharge home to self-care with no additional health care services, whereas dependent living refers to transfer to facility, transfer to short-term hospital, and discharge with home health care services. Utilization of PC was defined by the ICD-9 diagnosis code v66.7, which may be used for any patient determined to be the recipient of PC, comfort care, end-of-life care, or hospice care. Intensity of care was defined as following interventions: ventilator support, dialysis, tracheostomy, or percutaneous endoscopic gastrostomy (PEG) tube placement. See Appendix Tables A1–A4 for pertinent ICD-9 codes.

Covariates

Patient covariates included surgical diagnosis (gastrointestinal perforation, gastrointestinal obstruction, bowel ischemia, and mixed, which includes any combination of the three diagnoses), age, gender, race, payer status, and socioeconomic status (based on zip code income quartiles). Hospital characteristics included bed size (small, medium, and large), region (Northeast, Midwest, South, and West), and location/teaching status (rural, urban nonteaching, and urban teaching). We described the patients' comorbidity burden using a summary count of Elixhauser comorbidities (0, 1, 2, 3, or more comorbidities). We used the Elixhauser comorbidities available in the NIS Severity Files provided by Healthcare Cost and Utilization Project (HCUP). A weighted Elixhauser comorbidity score was generated using previously published weights as a covariate in the regression analyses. The Elixhauser comorbidity score was preferred over the routinely used Charlson Comorbidity index as it has been proven to be more discriminative for in-hospital deaths.¹⁵

Statistical analysis

We first described the patient-level and hospital-related characteristics of our cohort, using counts and proportions for categorical variables and means with standard deviations for continuous variables. Using cross tabulations, we then described primary outcomes by the type of diagnoses and treatment strategy (surgical vs. nonsurgical). Next, we performed subgroup analyses. Within each stratum of discharge disposition, we assessed the association between the treatment strategy and PC utilization using chi-square tests. We fit a multivariable logistic regression model using all study variables (full model) to identify independent predictors for PC. Adjusted odds ratios generated from the full multivariable logistic regression model were interpreted using a p-value of 0.05 as cutoff for assessing statistical significance. All statistical analyses were performed using SAS v. 9.4 (SAS Institute, NC). Given the complex multistage stratified design of the NIS, we used SAS's survey procedures to account to generate unbiased results using sampling weights provided by the HCUP.

Results

The NIS database from 2009 to 2013 linked 9363 hospitalizations to patients with ESRD who were admitted with our acute abdominal emergency diagnoses. Three patients were missing discharge disposition. Clinical and demographic characteristics are shown in Table 1. A majority of patients were <80 years old (82%), had Medicare (84%), and had four or more comorbidities (61%).

Table 1.

Patient Demographics and Hospital Characteristics

Variables	Total
Variables	n = 9363 (%)
Age, years
51–60	2061 (22)
61–70	2785 (30)
71–80	2845 (30)
81–89	1495 (16)
90+	177 (2)
Gender
Men	4440 (47)
Women	4923 (53)
Race
White	4400 (47)
Black	2673 (29)
Hispanic	1107 (12)
Asian or Pacific Islander	251 (3)
Other/unknown	932 (10)
Elixhauser comorbidity count
0	74 (1)
1	487 (5)
2	1237 (13)
3	1872 (20)
≥4	5693 (61)
Elixhauser comorbidity score^a, mean (SD)	6.5 (0.09)
Income by zip code
First quartile (lowest)	3076 (33)
Second quartile	2281 (24)
Third quartile	2071 (22)
Fourth quartile (highest)	1765 (19)
Payer status
Medicare	7902 (84)
Medicaid	466 (5)
Private	837 (9)
Other	158 (2)
Hospital bed size
Small	929 (10)
Medium	2189 (24)
Large	6183 (66)
Hospital teaching status/location
Rural	655 (6)
Urban nonteaching	3908 (42)
Urban teaching	4738 (51)
Hospital region
Northeast	1817 (20)
Midwest	2116 (23)
South	3600 (38)
West	1830 (19)

Missing data for payer status, median household income, hospital location/teaching status, hospital bed size, and discharge disposition were omitted; they account for <2% of the total for each category.

Elixhauser comorbidities excluded conditions associated with renal failure and only the Elixhauser comorbidity score was used for logistic regression.

To study the outcomes of patients with ESRD and acute abdomen, we characterized patient mortality and discharge disposition according to the treatment strategy (surgical vs. nonsurgical) and diagnosis (gastrointestinal obstruction, perforation, ischemia, or mixed pathology) (Table 2). Overall, 16% of patients died in hospital, and 43% were discharged to dependent living (26% were transferred to facilities, 14% were discharged with home health care services, and 3% were transferred to an acute hospital). Almost a quarter (24%) of patients underwent abdominal surgery. Mortality rate for the surgical patients was twice that of nonsurgical patients (26% vs. 13%; p < 0.0001), and fewer surgical patients were discharged to independent living (21% vs. 47%; p < 0.0001). Surgical patients had higher weighted Elixhauser comorbidity scores as compared with nonsurgical patients (median 7 vs. 5; p < 0.0001). Gastrointestinal obstruction (45%) was the most common surgical diagnosis among our patient cohort. Outcomes varied substantially based on surgical diagnosis, with bowel ischemia and mixed pathology diagnoses associated with increased mortality compared with perforation and obstruction (39% and 34% vs. 11% and 12%, respectively).

Table 2.

Discharge Disposition by Treatment Strategies and Surgical Diagnoses

n (%)	Discharge disposition ^a
	Died in hospital	Dependent living	Independent living
	n = 1482 (16%)	n = 4041 (43%)	n = 3837 (41%)
Treatment strategy
Surgical, n = 2205 (24%)	582 (26)	1171 (53)	452 (21)
Nonsurgical, n = 7158 (76%)	900 (13)	2873 (40)	3385 (47)
Diagnosis
Ischemia, n = 1028 (11%)	396 (39)	431 (41)	201 (20)
Perforation, n = 3663 (39%)	421 (11)	1532 (42)	1709 (47)
Obstruction, n = 4197 (45%)	506 (12)	1841 (44)	1848 (44)
Mixed pathology, n = 475 (5%)	159 (34)	237 (49)	79 (17)

Discharge to independent living refers to discharge home to self-care. Dependent living refers to transfer to facility, transfer to short-term hospital, discharge with home health care services, unknown discharge destination, or leaving against medical advice, as defined by National Inpatient Sample.

We examined associations between the treatment strategy and PC utilization stratified by discharge disposition (Table 3). Overall, only 5% of the cohort received PC. Among mortalities, 23% received PC, with a significantly more frequent PC in nonsurgical patients compared with surgical patients (24% vs. 20%; p = 0.029). Among patients who were discharged to dependent living, only 4% received PC. Nonsurgical patients utilized PC twice as frequently in this subcohort (4% vs. 2%; p = 0.003).

Table 3.

Palliative Care utilization and Treatment Strategies Stratified by Discharge Disposition

n (%)	Total	PC	No PC	p
n (%)	n = 9360	n = 503 (5%)	n = 8857 (95%)	p
Died in hospital	1482
Surgical	582 (39)	114 (20)	468 (80)	0.029
Nonsurgical	900 (61)	220 (24)	680 (76)	0.029
Discharge to dependent living	4041
Surgical	1171 (29)	28 (2)	1143 (98)	0.003
Nonsurgical	2870 (71)	124 (4)	2746 (96)	0.003
Discharge to independent living	3837
Surgical	452 (12)	1 (0.002)	451 (99)	0.76^a
Nonsurgical	3385 (88)	16 (0.005)	3369 (99)	0.76^a

Fisher's exact test used.

PC, palliative care.

The relationship of PC to intensity of care for the whole cohort was examined. (Table 4). Fifteen percent of the entire cohort received ventilatory support, and >80% underwent dialysis during the hospitalization. Only a small portion of the cohort underwent tracheostomy (1%) or PEG placement (1%). PC was associated with more ventilator utilization (35% vs. 14%), but less dialysis utilization than no PC (67% vs. 81%; p < 0.0001).

Table 4.

Intensity of Care and Palliative Care Utilization in the Entire Cohort

n (%)	Ventilator	Dialysis	Tracheostomy	PEG
n (%)	n = 1380 (15%)	n = 7541 (81%)	n = 128 (1%)	n = 95 (1%)
PC, n = 503 (5%)	178 (35)	339 (67)	10 (2)	7 (1)
No PC, n = 8860 (95%)	1202 (14)	7202 (81)	118 (1)	88 (1)
p	<0.001	<0.001	0.23	0.36

PEG, percutaneous endoscopic gastrostomy.

The relationship of PC to intensity of care and life support at the end of life was examined (Table 5). Among mortalities, patients who received PC had less ventilator utilization (45% vs. 53%; p = 0.02) and less surgery (34% vs. 40%; p = 0.03) than those who did not receive PC. However, PC had no association with utilization of dialysis or tracheostomy at the end of life. The mortalities who received PC had shorter hospital LOS (6 vs. 8 days; p = 0.057) and lower hospital charges than those without PC ($71,062.50 vs. $84,702.00; p = 0.006) (Table 6). Interestingly, PC was associated with decreased LOS and hospital charges among survivors who were discharged to dependent living. In addition, patients who received PC had a significantly shorter hospital stay and fewer charges, even though their Elixhauser comorbidity score was higher.

Table 5.

Intensity of Care and Palliative Care Utilization in Patients Who Died in Hospital (n = 1482)

n (%)	Ventilator	Dialysis	Tracheostomy	PEG
n (%)	n = 763 (51%)	n = 988 (67%)	n = 48 (3%)	n = 21 (1%)
PC, n = 334 (23%)	153 (45)	212 (63)	8 (2)	2 (1)
No PC, n = 1148 (77%)	610 (53)	776 (68)	40 (3)	19 (2)
p	0.0213	0.17	0.38	0.19

Table 6.

Palliative Care Utilization and Hospital Length of Stay and Charges Stratified by Discharge Disposition

	PC	No PC	p
Died in hospital
Median hospital LOS (days)	6	8	0.057
Median hospital charges ($)	71,062.50	84,702.00	0.006
Median Elixhauser comorbidity score	8	8	0.894
Discharge to dependent living
Median hospital LOS (days)	8	9	0.006
Median hospital charges ($)	59,997.00	70,867.00	0.037
Median Elixhauser comorbidity score	11	7	<0.0001
Discharge to independent living
Median hospital LOS (days)	7	4	0.033
Median hospital charges ($)	65,919.00	31,282.00	0.0051
Median Elixhauser comorbidity score	14	3	<0.0001

LOS, length of stay.

Lastly, we examined patterns of PC utilization with a multivariable logistic regression analysis (Table 7). The logistic regression analysis of patient characteristics revealed that patients who were older (age >70) and sicker (a higher weighted Elixhauser comorbidity score) were more likely to receive PC. Hispanics and blacks were less likely to receive PC compared with whites (adjusted odds ratio [aOR] 0.5; 95% confidence interval [CI] 0.4–0.8 and aOR 0.6; 95% CI 0.5–0.8, respectively). Patients with private or other payer status were more likely to receive PC than those on Medicare (aOR 1.6; 95% CI 1.1–2.2 and OR 2.2; 95% CI 1.1–4.5, respectively). Surgical diagnoses, such as bowel ischemia or mixed pathology, were predictors of PC, in comparison with obstruction (aOR 2.2; 95% CI 1.7–2.9 and aOR 2.1; 95% CI 1.4–2.9, respectively). The treatment strategy, whether surgical or nonsurgical, was not an independent predictor of PC utilization when adjusting for all other variables. Analyses of hospital characteristics revealed that patients treated in larger hospitals, hospitals in the Western region, or urban teaching hospitals were significantly more likely than their counterparts to utilize PC.

Table 7.

Predictors of Palliative Care Utilization Using Multivariate Logistic Regression

Variable	PC	No PC	OR (95% CI)	p
Age, n (%)^a
51–60	77 (4)	1984 (96)	1.0	Reference
61–70	120 (4)	2665 (96)	1.1 (0.8–1.5)	0.5
71–80	160 (6)	2685 (94)	1.6 (1.2–2.1)	0.004
81–89	123 (8)	1372 (92)	2.4 (1.7–3.2)	<0.0001
90+	23 (13)	154 (87)	4.2 (2.5–7.2)	<0.0001
Gender, n (%)
Men	230 (5)	4210 (95)	1.0	Reference
Women	273 (6)	4650 (94)	1.1 (0.9–1.4)	0.2
Race, n (%)^a
White	298 (7)	4102 (93)	1.0	Reference
Asian or Pacific Islander	17 (7)	234 (93)	0.7 (0.4–1.3)	0.3
Black	105 (4)	2568 (96)	0.6 (0.5–0.8)	0.0001
Hispanic	45 (4)	1062 (96)	0.5 (0.4–0.8)	0.002
Other	15 (6)	261 (94)	1.0 (0.5–1.9)	1.0
Unknown	23 (4)	633 (96)	0.5 (0.4–0.9)	0.007
Elixhauser comorbidity score, mean (SD)^a	10 (7)	6 (7)	1.1 (1.05–1.08)	<0.0001
Income by zip code, n (%)
First quartile	146 (5)	2930 (95)	1.0	Reference
Second quartile	108 (5)	2173 (95)	0.9 (0.7–1.1)	0.3
Third quartile	126 (6)	1945 (94)	1.1 (0.8–1.4)	0.6
Fourth quartile	116 (7)	1649 (93)	1.1 (0.8–1.4)	0.6
Payer status, n (%)^a
Medicare	416 (5)	7486 (95)	1.0	Reference
Medicaid	23 (5)	443 (95)	1.1 (0.7–1.8)	0.6
Private	49 (6)	788 (94)	1.6 (1.1–2.2)	0.01
Other	11 (9)	123 (91)	2.2 (1.1–4.5)	0.03
Hospital bed size, n (%)^a
Small	34 (4)	895 (96)	1.0	Reference
Medium	108 (5)	2081 (95)	1.3 (0.8–2.0)	0.3
Large	354 (6)	5829 (94)	1.7 (1.1–2.6)	0.009
Hospital teaching status/location, n (%)
Rural	25 (4)	630 (96)	1.0	Reference
Urban nonteaching	215 (5)	3693 (95)	1.4 (0.9–2.2)	0.1
Urban teaching	256 (5)	4482 (95)	1.6 (1.0–2.5)	0.04
Hospital region, n (%)^a
West	135 (7)	1695 (93)	1.0	Reference
Northeast	87 (5)	1730 (95)	0.6 (0.4–0.8)	0.002
Midwest	113 (5)	2003 (95)	0.7 (0.5–1.0)	0.02
South	168 (5)	3432 (95)	0.7 (0.5–1.0)	0.03
Surgical diagnosis, n (%)^a
Obstruction	206 (5)	3991 (95)	1.0	Reference
Ischemia	108 (10)	920 (90)	2.2 (1.7–2.9)	<0.0001
Perforation	140 (4)	3523 (96)	0.8 (0.7–1.0)	0.11
Mixed	49 (11)	426 (89)	2.1 (1.4–2.9)	<0.0001
Treatment strategy, n (%)
Nonsurgical	360 (5)	6798 (95)	1.0	Reference
Surgical	143 (6)	2062 (94)	1.1 (0.9–1.4)	0.5

Statistically significant variables (p < 0.05).

CI, confidence interval; OR, odds ratio.

Conclusions

One in six ESRD patients with acute abdominal emergency died in hospital, highlighting a vulnerable patient population with high PC needs. Moreover, for those requiring surgical intervention, 25% died, and only 20% were discharged to independent living. Despite these poor outcomes, again, few received PC. Mortality rates at this level would certainly be consistent with accepted indications for PC among hospitalized patients,¹⁶ suggesting that any patient with ESRD who presents for emergency abdominal surgery or acute abdomen should receive these services.

Predictors to receive PC were higher severity of underlying illness and older age, suggesting a general focus by physicians on traditional prognostic factors, or perceived futility and risk of death. The PC rate was significantly lower in patients who underwent surgery, which aligns with other studies that find that surgeons perceive a dichotomous relationship between surgery and PC¹⁷ and reflects the traditional model of physicians offering PC only after aggressive treatment has failed. Despite this, the NIS dataset only has data on presence or absence of PC, but no data on the timing of PC during the hospitalization, which would be a more relevant measure of practice patterns for this population. The variability across race, US regions, hospitals, and payer are consistent with other population studies on PC utilization¹⁸ and suggest that institutional and local practice culture has a significant impact on medical and surgical decisions for individual patients regarding PC. These all suggest potential target areas for efforts to improve access to and delivery of PC on the individual and systems levels.

Patients with ESRD, in general, face a high baseline mortality risk and significant health-related burdens. Despite these sequelae and most patients' desire to discuss their values, goals, and preferences with their treating physicians, these discussions occur rarely.⁴ Indeed, although decisions about the initiation of dialysis represent a pivotal point in the health care trajectory of ESRD patients, dialysis is often begun without appropriate advance care planning and consideration of alternative supportive care strategies.¹⁹ Barriers to such discussions include prognostic uncertainty, concerns about patients' or family's distress, and lack of clarity about the “right time” for initiating such a conversation.²⁰ In the context of an acute abdominal emergency, the challenges of acute surgical decision making are further compounded by the lack of advance directives and clear health care goals.¹² Our data provide another contribution to prognosis for ESRD patients when physicians are discussing goals of care. An acute surgical abdomen is one of the many emergencies that can arise in patients, others including cardiovascular, cerebrovascular, infectious, and traumatic cases. Although the outcomes evaluated in our study do not translate to other acute medical or surgical problems, the approach to patients with ESRD should be similar. The focus on prognosis, specifically with respect to outcomes that matter to patients, should be the goal and we should aim to engage PC as a conduit for allowing conversations to take place to ensure patients' goals remain the foundation of treatment plans. This study further delineates opportunities for advance care planning both before an emergency surgical intervention is contemplated and in the immediate perioperative period.

The National Consensus Project for Quality Palliative Care advocates for PC along the course of illness, concurrently with curative or life-prolonging therapy²¹ and the subspecialty of surgical palliative care also advocates for this approach. Our study suggests that patients with ESRD who present with an acute abdomen represent a clear opportunity for PC integration from the time of admission in the emergency department. Even in institutions in which PC team and consultation are not available, these data can be used by the surgeons to provide “primary PC” when communicating with patients and their families by underpinning the implications of the acute surgical problem in the context of the underlying ESRD to reach goal-concordant treatment plans.

As expected, the rate of PC was the highest among patients who died. Interestingly, among the deaths, PC was associated with less nonbeneficial ventilator support and lower LOS and cost. This is consistent with other literature that shows that PC is associated with less nonbeneficial life support in a variety of populations.^22,23 PC was not associated with less utilization of dialysis at the end of life in our study, and this is also consistent with other literature, which find that withdrawal of dialysis in ESRD patients involves complex decisions and is complicated by the Medicare benefits for dialysis.²⁴ The withdrawal from dialysis before death among patients with ESRD has increased for the past decade,^1,25 and a case-control study showed that patients with ESRD with advance care planning were more likely to withdraw from dialysis in accordance with their preferences at the end of life.²⁶ Only two-thirds of the in-hospital deaths underwent dialysis, less frequently than the entire cohort. This suggests that patients who were approaching end of life chose not to pursue dialysis during hospitalization, regardless of whether formal PC was involved or not. However, as NIS does not distinguish patients who were on dialysis before admission and then had it withdrawn from those who did not undergo dialysis during the hospitalization, it is hard to draw a meaningful conclusion.

Despite no reduction in utilization of dialysis, our study demonstrates that patients with PC had decreased hospital stay and lower charges compared with those without PC. This was true for both mortalities and survivors discharged to dependent living. Our findings corroborate studies in many populations that PC was associated with shorter LOS and reduced hospital charges,²⁷ especially in dying patients in the ICU.^8,22,28,29

There are limitations of this study due to the nature of the administrative dataset utilized. First, there is likely an overall underreporting of patients who received PC, of patients with an acute abdomen, and of patients who underwent surgery. Although the use of the v66.7 diagnosis code in large dataset analyses is limited by poor sensitivity (49% in a large single-center validation study³⁰), even doubling our detected rate to account for failure to capture would result in a rate of PC utilization that does not nearly approach the degree of demand. We were selective in the diagnosis codes used to define acute surgical abdomen, modeling our range after previously published studies,⁹ and excluding other surgical diagnoses such as gastrointestinal hemorrhage and trauma. Finally, NIS data include no information on patient-centered outcomes, patient preferences, advance directives, or posthospitalization outcomes, which are subject to significant effects from PC.

In summary, our study reveals important prognostic information about outcomes in emergency abdominal surgical conditions for patients with ESRD and defines unmet PC needs in this population. It also highlights opportunities for further studies to better understand the barriers and systems issues to integrating surgical PC into hospital care in the future.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

This study was funded by U.S. Department of Health and Human Services, Health Resources and Services Administration (D34HP26020) and Patient-Centered Outcomes Research Institute (PCORI-1502-27462).

Appendix

Appendix Table A4.

ICD-9-CM Procedure Codes to Represent Intensity of Care

Ventilator	96.7, 96.70, 96.71, 96.72
Dialysis	39.95, 54.98
Tracheostomy	31.1, 31.2, 31.21, 31.29
Percutaneous endoscopic gastrostomy	43.11

References

Saran

, Robinson

, Abbott

, et al.: US Renal Data System 2017 Annual Data Report: Epidemiology of kidney disease in the United States. Am J Kidney Dis, 2018; 71 (suppl 1):Svii, S1–S672.

Wong

, Kreuter

, O'Hare

: Treatment intensity at the end of life in older adults receiving long-term dialysis. Arch Intern Med, 2012; 172:661–663.

Wachterman

, Marcantonio

, Davis

, et al.: Relationship between the prognostic expectations of seriously ill patients undergoing hemodialysis and their nephrologists. JAMA Intern Med, 2013; 173:1206–1214.

Davison

: End-of-life care preferences and needs: Perceptions of patients with chronic kidney disease. Clin J Am Soc Nephrol, 2010; 5:195–204.

Schneider

, Cobb

, Patel

, et al.: Elective surgery in patients with end stage renal disease: What's the risk?. Am Surg, 2009; 75:790–793.

Gajdos

, Hawn

, Kile

, et al.: Risk of major nonemergent inpatient general surgical procedures in patients on long-term dialysis. JAMA Surg, 2013; 148:137–143.

Delgado-Guay

, Rodriguez-Nunez

, Shin

, et al.: Characteristics and outcomes of patients with advanced cancer evaluated by a palliative care team at an emergency center. A retrospective study. Support Care Cancer, 2016; 24:2287–2295.

Mosenthal

, Murphy

, Barker

, et al.: Changing the culture around end-of-life care in the trauma intensive care unit. J Trauma, 2008; 64:1587–1593.

Cauley

, Panizales

, Reznor

, et al.: Outcomes after emergency surgery in patients with advanced cancer: Opportunities to reduce complications and improve palliative care. J Trauma Care Surg, 2015; 79:399–406.

10.

Cooper

, Koritsanszky

, Cauley

, et al.: Recommendations for best communication practices to facilitate goal-concordant care for seriously ill older patients with emergency surgical conditions. Ann Surg, 2016; 263:1–6.

11.

Tamura

, Meier

: Five policies to promote palliative care for patients with ESRD. Clin J Am Soc Nephrol, 2013; 8:1783–1790.

12.

Grubbs

, Moss

, Cohen

, et al.: A palliative approach to dialysis care: A patient-centered transition to the end of life. Clin J Am Soc Nephrol, 2014; 9:2203–2209.

13.

HCUP National Inpatient Sample (NIS): Healthcare Cost and Utilization Project (HCUP). 2012. Rockville, MD: Agency for Healthcare Research and Quality. www.hcup-us-ahrq.gov/nisoverview.jsp (last accessed November 14, 2008).

14.

HCUP Nationwide Inpatient Sample (NIS): Healthcare Cost and Utilization Project (HCUP). 2011. Rockville, MD: Agency for Healthcare Research and Quality. www.hcup-us-ahrq.gov/nisoverview.jsp (last accessed November 14, 2008).

15.

van Walraven

, Austin

, Jennings

, et al.: A modification of the Elixhauser comorbidity measures into a point system for hospital death using administrative data. Med Care, 2009; 47:626–633.

16.

Morrison

, Meier

: Clinical practice. Palliative care. N Engl J Med, 2004; 350:2582–2590.

17.

Schwarze

, Bradley

, Brasel

: Surgical “buy-in”: The contractual relationship between surgeons and patients that influences decisions regarding life-supporting therapy. Crit Care Med, 2010; 38:843–848.

18.

Committee on Approaching Death: Addressing key end of life issues Institute of Medicine. Dying in America: Improving Quality and Honoring Individual Preferences near the End of Life. Washington, DC: National Academies Press, 2015.

19.

Greer

, Cooper

, Crews

, et al.: Quality of patient-physician discussions about CKD in primary care: A cross-sectional study. Am J Kidney Dis, 2011; 57:583–591.

20.

Javier

, Figueroa

, Siew

, et al.: Reliability and utility of the surprise question in CKD stages 4 to 5. Am J Kidney Dis, 2017; 70:93–101.

21.

American Academy of Hospice and Palliative Medicine; Center to Advance Palliative Care; Hospice and Palliative Nurses Association; Last Acts Partnership and the National Hospice and Palliative Care Organization: National Consensus Project for Quality Palliative Care: Clinical Practice Guidelines for quality palliative care, executive summary. J Palliat Med, 2004; 7:611–627.

22.

Keele

, Keenan

, Sheetz

, et al.: Differences in characteristics of dying children who receive and do not receive palliative care. Pediatrics, 2013; 132:72–78.

23.

O'Mahony

, Blank

, Zallman

, et al.: The benefits of a hospital-based inpatient palliative care consultation service: Preliminary outcome data. J Palliat Med, 2005; 8:1033–1039.

24.

Wachterman

, Hailpern

, Keating

, et al.: Association between hospice length of stay, health care utilization, and Medicare costs at the end of life among patients who received maintenance hemodialysis. JAMA Intern Med, 2018; 178:792–799.

25.

Cohen

, Germain

, Poppel

, et al.: Dialysis discontinuation and palliative care. Am J Kidney Dis, 2000; 36:140–144.

26.

Sellars

, Morton

, Clayton

, et al.: A case-control study of end-of-life treatment preferences and costs following advance care planning for adults with end stage kidney disease. Nephrology, 2018 [Epub ahead of print]; DOI: 10.1111/nep.13230.

27.

May

, Garrido

, Cassel

, et al.: Prospective cohort study of hospital palliative are teams for inpatients with advanced cancer: Earlier consultation is associated with large cost-saving effect. J Clin Oncol, 2015; 33:2745–2752.

28.

Campbell

, Guzman

: Impact of a proactive approach to improve end-of-life care in a medical ICU. Chest, 2003; 123:266–271.

29.

Norton

, Hogan

, Holloway

, et al.: Proactive palliative care in the medical intensive care unit: Effects on length of stay for selected high-risk patients. Crit Care Med, 2007; 35:1530–1535.

30.

Hua

, Li

, Clancy

, et al.: Validation of the v66.7 code for palliative are consultation in a single academic medical center. J Palliative Med, 2017; 20:372–377.