Identifying Core Principles of Palliative Care Consultation in Surgical Patients and Potential Knowledge Gaps for Surgeons

Abstract

BACKGROUND:

Palliative medicine is an important component of care for patients with advanced cancer. Previous studies demonstrated that surgeons tend to underuse palliative care in comparison with medical services. Furthermore, little is known about the specific use of palliative care services among surgical oncology practices. Therefore, we designed and performed the present study to evaluate the use of palliative care in cancer patients undergoing major oncologic surgery.

STUDY DESIGN:

A single-institution retrospective review of consecutive palliative care consultations within a large National Cancer Institute-designated comprehensive cancer center in 2016–2017 was conducted.

RESULTS:

We analyzed 120 patients (60 surgical and 60 medical). Patient demographics in the two groups were similar. The surgical oncology patients were more likely to undergo consultation for advanced care planning (32% vs. 13%; p = 0.02). Medical oncology patients were more likely to undergo consultation for pain management (97% vs. 62%; p < 0.001). Symptom assessment scores for the medical patients more frequently demonstrated dyspnea and malignancy-related pain than in the surgical patients. Also, palliative care recommendations and interventions for the surgical patients more frequently included end of life discussions and transfer to the inpatient palliative care unit. For medical oncology patients, recommendations more often included changes in pain and bowel regimen medications. In addition, despite more frequent consults for advanced care planning in the surgical patients, code status was actually changed to “do-not-resuscitate” more frequently in the medical patient cohort.

CONCLUSIONS:

Surgical patients were less likely to undergo palliative care consultation for assistance with symptom management and more likely to undergo consultation for assistance with end-of-life discussions than were medical oncology patients. Advanced care planning and end of life discussions should be an area of focus in palliative care education for surgeons.

Précis:

The indication for Palliative Care consultation differs significantly between Medical Oncology and Surgical Oncology Patients, with Surgical Oncology patients undergoing consultation more often for end-of-life and advanced care planning discussions. These indications represent an area for focused palliative surgical education.

INTRODUCTION

Palliative care interventions for patients with advanced malignancies have had a number of benefits, including improved health-related quality of life (1), improved symptom control (2–4), and decreased cancer-related morbidity (5–7). Therefore, current American Society of Clinical Oncology guidelines recommend early integration of palliative care into the routine treatment of patients with advanced malignancies (8). However, the volume of consultations required to meet this recommendation is prohibitive based on the current number and availability of palliative care providers (9).

Given this critical need for palliative care services for cancer patients and a relative lack of available trained palliative care specialists, surgeons must be able to provide the critical components of palliative care as part of their core clinical practice (10). Training in palliative care is vital for surgical oncologists and an area of need for trainees as well as practitioners (11, 12). However, little research has focused specifically on the components of palliative care required for surgical oncology patients (13, 14). Identifying the palliative care needs of these patients has the potential to help direct trainee education, improve patient care, and meet current guidelines (15). Studies of palliative consultation in medical oncology are relatively well described in the literature, notably, that the complex medical, psychologic, and social problems in this population require multiple palliative interventions and that many important issues, such as do-not-resuscitate status, are frequently not addressed (16). However, surgical patients were rarely included in those studies. The literature contains some comparisons of medical and surgical patients undergoing palliative consultation in noncancer patients, but the important components of the consultations were not identified (17). Researchers showed that surgical patients in the U.S. Department of Veterans Affairs system receive palliative care services less frequently than medical patients do, although they did not investigate criteria for surgical patients needing palliative care (18). In summary, the literature has almost a complete lack of even descriptive studies of palliative care consultations in surgical patients.

In this context, we designed and performed the study described herein to better understand the key elements of palliative care consultations in patients undergoing surgery for solid organ malignancies. We sought to compare surgical and medical oncology patients undergoing inpatient palliative care consultation. We hypothesized that these two groups have important differences in baseline patient characteristics, indications for palliative care consultation, Edmonton Symptom Assessment Scale (ESAS) scores, functional status, and palliative care consultation recommendations.

METHODS

Data source and outcome measures

A retrospective review of patients undergoing inpatient palliative care consultation at The University of Texas MD Anderson Cancer Center, a large National Center Institute-designated comprehensive cancer center, in 2016–2017 was conducted. Patients with adequate documentation for assessment of our targeted variables were included in the final analysis. Only adult patients (age ≥18 years) with active malignancies were included in the study.

The patients’ clinicopathologic variables were extracted from their medical records. These variables included age, sex, functional performance status, location of consultation, length of hospital stay, and discharge disposition. Palliative care-specific variables were also collected, including indications for consultation (e.g., pain management, delirium management, code status changes) and recommendations for patient management.

Patient-reported symptoms were assessed using the ESAS, a well-validated symptom index used in palliative care (19). The ESAS is a routine component of initial palliative care consultations in our institution. It was included in this analysis to assess differences in symptom profiles between the surgical and medical oncology patients.

Statistical analysis

Descriptive statistics were used to calculate frequencies, percentages, and median ESAS scores. Continuous variables were analyzed using independent-sample t-tests. Categorical variables were analyzed using the Fisher exact test or a chi-square test as appropriate. Kaplan-Meier analyses were used to calculate overall survival from the day of discharge to most recent follow-up or day of death. Differences between the groups were assessed with the log rank test.

A power analysis was performed prior to data abstraction. Identification of about 100 patients in the medical and surgical services with adequate documentation of all variables of interest was anticipated. As a sample power calculation, we hypothetically compared 50 medical and 50 surgical patients with the aim of identifying a difference in the recommendation for a change in pain medication regimen. Given that this outcome is a dichotomous variable, chi-square testing was used for this variable. With 100 patients (divided evenly between the medical and surgical services), our study would have greater than 85% power to detect a medium effect size as defined by Cohen’s w of 0.3. Of note, this sample power calculation would be applicable to all of the dichotomous variables in our study.

All statistical analyses were conducted using STATA software (version 13.0; StataCorp). P values less than 0.05 were considered significant. All p values were two-sided. This study was approved by the MD Anderson Cancer Center Institutional Review Board.

RESULTS

We included a total of 120 patients (60 surgical oncology and 60 medical oncology) who had adequate documentation of the relative study variables in the final analysis. Patient characteristics are shown in Table 1. Patients in the two groups had similar mean ages (60.3 years and 56.3 years in the surgical and medical oncology groups, respectively; p = 0.11) and performance statuses (median Eastern Cooperative Oncology Group performance status score of 3 in both groups; p = 0.9). The two groups also had similar rates of metastatic or recurrent cancer (68% and 82% in the surgical oncology and medical oncology groups, respectively; p = 0.14). However, patients in the medical oncology group were more likely to be actively receiving systemic chemotherapy, defined by receipt of chemotherapy within the prior 6 weeks (52% vs. 22% of patients in the surgical oncology group; p < 0.01).

Table 1.

Clinicodemographic Variables of Patients with Cancer Undergoing Palliative Care Consultations Stratified by Medical and Surgical Oncology Status

	Number of patients (%)
Variable	Surgical oncology (n = 60)	Medical oncology (n = 60)	p
Mean age, years	60.3	56.3	0.11
Male sex	22 (37)	31 (52)	0.10
Median ECOG* performance status score	3	3	0.90
Metastatic or Recurrent Cancer	41 (68)	49 (82)	0.14
Systemic Chemotherapy within 6 weeks	13 (22)	31 (52)	<0.01
Reason for Admissionα
Planned Operation/Procedure	12 (20)	2 (3)	<0.01
GI**/GU***/biliary obstruction	30 (50)	24 (40)	0.36
Infection/Sepsis	11 (18)	16 (27)	0.38
Altered Mental Status	3 (5)	3 (5)	1.00
FTTϮ/dehydration/feeding issue	12 (20)	5 (8)	0.11
Shortness of Breath/Pulmonary	1 (2)	7 (12)	0.06
Pain	1 (2)	19 (32)	<0.01
GI** bleed	2 (3)	3 (5)	1.00
Ascites	0 (0)	8 (13)	<0.01
Other	3 (5)	1 (2)	0.62
Operation during Hospitalization
No operation	34 (57)	54 (90)	<0.01
Elective operation	12 (20)	1 (2)
Urgent/Emergent operation	14 (23)	5 (8)
Palliative (non-surgical) procedure	21 (35)	27 (45)	0.35
Indication for consultation
Pain	37 (62)	58 (97)	<0.01
Delirium	2 (3)	9 (15)	0.02
Transfer of care	8 (13)	2 (3)	0.05
Dyspnea	0	13 (22)	<0.01
Insomnia	2 (3)	9 (15)	0.03
Constipation	2 (3)	22 (37)	<0.01
Depression	11 (18)	9 (15)	0.62
Advanced care planning	19 (32)	8 (13)	0.02
Identified support
Spouse	38 (63)	35 (58)	0.65
Family	19 (32)	47 (78)	<0.01
Mean length of hospital stay, days	11.7	15.3	0.41
Patient discharge disposition
In-hospital death	4 (7)	10 (17)	0.09
Home	37 (62)	30 (50)	0.20
Home hospice	11 (18)	11 (18)	1.00
Inpatient hospice	4 (7)	7 (12)	0.34
Subacute and other hospital	4 (7)	2 (3)	0.40

^*ECOG, Eastern Cooperative Oncology Group.

^**GI, gastrointestinal

^***GU, genitourinary

^ϮFTT, failure to thrive

^αRates may not add up to 100%, as both primary and secondary admission diagnoses are included in this analysis.

The reasons for admission differed between the two groups. The rate of admission for a planned operation or procedure was higher in the surgical oncology group (20% vs 3%; p = <0.01), while admission rates for pain symptoms (2% vs 32%; p <0.01) and ascites (0% vs. 13%; p <0.01) were higher in the medical oncology group. Further, patients in the surgical oncology group were more likely to have undergone an operation during their hospitalization (43% vs. 10% of patients in the medical oncology group; p <0.01), but rates of palliative non-surgical procedures were similar between the two groups (35% of patients in the surgical oncology group vs. 45% in the medical oncology group; p = 0.35). Overall, there were a total of 75 palliative non-surgical procedures in 48 patients. Twenty-one surgical oncology patients (35%) underwent a total of 27 palliative procedures while 27 medical oncology patients (45%) underwent a total of 48 palliative procedures.

The mean length of hospital stay was 15.3 days in the medical oncology group and 11.7 days in the surgical oncology group (p = 0.41). The rate of in-hospital deaths was slightly higher in the medical oncology patients (17% vs. 7%); this difference was not statistically significant (p = 0.09). The two patient groups also had similar rates of discharge to inpatient hospice facilities or home hospice (30% in the medical oncology group and 25% in the surgical oncology group). At the time of data analysis, more patients in the medical oncology group than in the surgical oncology group had died (83% vs. 60%; p = 0.05). The median length of follow-up for the cohort was 50 days. The median (interquartile range) overall survival (OS) from day of discharge was 62 (IQR: 17–277) days for the entire cohort, 82 (IQR: 19–433) days for the surgical oncology group, and 50 (IQR: 14–139) days for the medical oncology group (p = 0.03).

The two groups of patients differed in terms of their indications for palliative care consultation. In the medical oncology patients, the most common indication was pain management (97%). In comparison, pain management was an indication for consultation in only 62% of the surgical oncology patients (p < 0.01). Also, palliative care consultation was more frequently initiated in medical than in surgical oncology patients for management of delirium (15% vs. 3%; p = 0.02), slow-transit constipation (37% vs. 3%; p < 0.01), and dyspnea (22% vs. 0%; p < 0.01). In the surgical oncology group, palliative care consultation was initiated for assistance with advanced care planning in 32% of the patients, whereas it was initiated for this reason in only 13% of the medical oncology patients (p = 0.02)

Patients in the two groups had relatively similar symptom profiles according to the ESAS (Table 2) except for dyspnea and pain, which were more common in the medical oncology patients. However, palliative care consultation recommendations and interventions differed between the groups (Table 3). Specifically, medical oncology patients more frequently had recommendations for changes in pain medications (98% vs. 75%; p < 0.01) and bowel regimen medications (77% vs. 53%; p < 0.01). Medical oncology patients were also more likely to have changes in code status to do-not-resuscitate following palliative care consultation (30% vs. 7%; p < 0.01). Surgical oncology patients more frequently had recommendations for supportive counseling, depression management, and advanced care planning (32% vs. 15%; p = 0.03).

Table 2.

ESAS^* Scores in Patients Undergoing Palliative Care Consultations Stratified by Medical and Surgical Oncology Status

	Median (25th, 75th quartile)
ESAS variable	Surgical oncology	Medical oncology	p
Fatigue	6 (2, 8)	7 (5, 8)	0.09
Anxiety	3 (0, 6)	3 (0, 6)	0.91
Wellbeing	4 (3, 6)	5 (3, 7)	0.07
Pain	5 (3, 8)	7 (5, 8)	0.04
Depression	2 (0, 5)	2 (0, 6)	0.38
Sleep	3 (0, 6)	4 (1, 7)	0.18
Appetite	5 (2, 8)	5 (3, 8)	0.92
Nausea	1 (0, 5)	2 (0, 5)	0.58
Drowsiness	2 (0, 5)	4 (0, 6)	0.16
Dyspnea	0 (0, 2)	1 (0, 5)	<0.01

^*ESAS, Edmonton Symptom Assessment Scale

Table 3.

Palliative Care Consultation Recommendations and Interventions for the Study Patients Stratified by Medical and Surgical Oncology Status

	Number of patients (%)
Recommendation/intervention	Surgical oncology (n = 60)	Medical oncology (n = 60)	p
Pain medication	45 (75)	59 (98)	<0.01
Bowel regimen medication	32 (53)	46 (77)	<0.01
Fatigue regimen	31 (52)	40 (67)	0.10
Antiemetics	32 (53)	33 (55)	0.86
Neuroleptics	14 (23)	18 (30)	0.41
Medication stoppage due to polypharmacy	6 (10)	3 (5)	0.30
Transfer to inpatient palliative care unit	9 (15)	1 (2)	<0.01
Advanced care planning/EOL* discussion	19 (32)	9 (15)	0.03
Code status changed to DNR**	4 (7)	18 (30)	<0.01

^*EOL, end-of-life;

^**DNR, do-not-resuscitate.

DISCUSSION

The importance of integrating palliative medicine into routine care for patients with advanced cancer has been well demonstrated in a number of studies as well as reinforced in clinical practice guidelines (8, 20). The surgical community has recognized the importance of palliative care as well, with consensus statements from the American College of Surgeons emphasizing the important role surgeons play in the provision of palliative care (21). Despite this, a number of studies have shown marked differences in the use of palliative care among surgeons, with surgical patients accounting for only a small minority of palliative care consultations (17, 18). In the present study, we sought to better understand the palliative care needs of surgical oncology patients to help address this disparity in utilization and identify areas where the surgical community can improve provision of palliative care.

In our analysis, we found that the common indications for palliative care consultation differed significantly between surgical and medical oncology patients. Patients on the surgical service more often had palliative care consultations for initiation of end-of-life or advanced care planning discussions, whereas medical oncology patients more commonly had consultations for assistance with management of symptoms, such as malignancy-related pain, slow-transit constipation, and dyspnea. Despite the lower likelihood of a surgical patient undergoing a palliative care consultation for symptom management, surgical patients reported symptom profiles similar to those of the medical oncology patients using the ESAS.

These findings have significant implications and are potential areas for future research and intervention. We found that in many surgical oncology patients, palliative care consultations were performed for end-of-life and goals-of-care discussions. One potential explanation for this is that surgeons may be uncomfortable initiating these discussions independently and seek the assistance of palliative care experts. Previous studies, including anthropologic work, identified a high level of personal responsibility for the life of the patient as a core part of the surgical ethos such that some surgeons view a patient’s death or the withdrawal of life-sustaining care as a personal failure (22–24). However, this difference in indication for consultation may also reflect inherent differences in the patient populations, as medical oncology patients were more likely to be admitted for symptoms/complications of their therapy than the surgical oncology patients. It is interesting to note that while surgical oncology patients more often underwent palliative care consultation for end-of-life issues and goals-of-care discussions, medical oncology patients were more likely to have a change in code status to “do-not-resuscitate.” These seemingly contradictory findings, while difficult to evaluate and understand in a retrospective study, are indicative of the advanced disease processes in both the surgical and medical oncology groups, and underscore the complexity of these issues in advanced oncology patients.

Surgical oncology patients were less likely to undergo palliative care consultations for symptom management, and palliative care providers uncovered symptom burdens that may have been unrecognized by the primary surgical team. This may reflect a relatively narrow understanding of the scope of palliative care services amongst surgical oncologists. A previous survey of trauma surgeons had similar findings, demonstrating that surgeons perceived assistance with end-of-life discussions as the primary indication for palliative care consultations (25).

Palliative care in surgical oncology is identified as an area of need for further training. In a recent national study, surgical oncology fellows identified palliative care as a “core responsibility” of their specialty but rated their own formal training in palliative care as poor in comparison with the other aspects of their training (12, 26). In that study, fellows identified a hesitancy to initiate conversations about death and dying, and in their most recent interactions with dying patients, more than half of the fellows avoided directly telling patients that they were dying. From a training program perspective, only a small number of surgical residency and fellowship programs offer formal training in core palliative care concepts (11). Addressing this gap in education at the trainee level may help guide surgeons in using palliative care services as well as increase their ability to provide core aspects of palliative medicine directly to their patients.

Our study had some important limitations. First, it was a retrospective study of practice within a single institution. Some of the variations in the results that we identified may reflect institutional practice patterns that are not applicable in other settings. Second, given the retrospective nature of the study, we are limited in our ability to abstract data from it, particularly in terms of understanding complex palliative care conversations that may not be adequately captured in medical records.

CONCLUSIONS

This study is the first to directly compare surgical and medical patients with advanced cancer and evaluate the use of palliative care services among them. We found that surgeons are more likely to engage palliative care services for assistance with end-of-life situations and planning and less likely to use them for management of symptoms, such as pain control. The reasons for these differences are complex and difficult to understand through retrospective review. This would be best addressed in future studies using mixed methods and qualitative studies to better evaluate attitudinal differences regarding the purpose of palliative care consultations. Our hope is that identifying these differences between surgical and medical oncology patients will help improve surgeons’ ability to deliver appropriate palliative care to their patients.

Abbreviations:

ESAS

Edmonton Symptom Assessment Scale

ECOG

Eastern Cooperative Oncology Group

EOL

end-of-life

DNR

do-not-resuscitate