Evaluation of automated specialty palliative care in the intensive care unit: A retrospective cohort study.

INTRODUCTION: Automated specialty palliative care consultation (SPC) has been proposed as an intervention to improve patient-centered care in the intensive care unit (ICU). Existing automated SPC trigger criteria are designed to identify patients at highest risk of in-hospital death. We sought to evaluate common mortality-based SPC triggers and determine whether these triggers reflect actual use of SPC consultation. We additionally aimed to characterize the population of patients who receive SPC without meeting mortality-based triggers.
METHODS: We conducted a retrospective cohort study of all adult ICU admissions from 2012-2017 at an academic medical center with five subspecialty ICUs to determine the sensitivity and specificity of the five most common SPC triggers for predicting receipt of SPC. Among ICU admissions receiving SPC, we assessed differences in patients who met any SPC trigger compared to those who met none.
RESULTS: Of 48,744 eligible admissions, 1,965 (4.03%) received SPC; 979 (49.82%) of consultations met at least 1 trigger. The sensitivity and specificity for any trigger predicting SPC was 49.82% and 79.61%, respectively. Patients who met no triggers but received SPC were younger (62.71 years vs 66.58 years, mean difference (MD) 3.87 years (95% confidence interval (CI) 2.44-5.30) p<0.001), had longer ICU length of stay (11.43 days vs 8.42 days, MD -3.01 days (95% CI -4.30 --1.72) p<0.001), and had a lower rate of in-hospital death (48.68% vs 58.12%, p<0.001).
CONCLUSION: Mortality-based triggers for specialty palliative care poorly reflect actual use of SPC in the ICU. Reliance on such triggers may unintentionally overlook an important population of patients with clinician-identified palliative care needs.

Introduction

The delivery of high-quality palliative care in the intensive care unit (ICU) is a cornerstone of critical care medicine and is widely recognized as necessary for the provision of patient- and family-centered ICU care [1-4]. The World Health Organization defines palliative care as an approach to healthcare that improves quality of life for patients with serious illness and their families by attending to physical, psychosocial, and spiritual needs [5]. The delivery of palliative care for patients with critical illness has been associated with increased rates of advanced directives, decreased ICU length of stay, increased use of hospice, and decreased use of non-beneficial life-sustaining therapies in observational studies [6-8]. Despite these benefits, the provision of palliative care in the ICU remains suboptimal [9].

Palliative care in the ICU includes both primary palliative care, which is provided by any clinician in the ICU, and specialty palliative care consultation (SPC), which entails consultation by palliative care specialists with advanced training. One-quarter of all hospital-based specialty palliative care consultations occur in the ICU [10]. Recent guidelines recommend an integrative palliative care approach wherein a patient receives palliative care concurrently with restorative/curative care from the time of ICU admission in an individualized manner [2]. However, the optimal role of SPC and the relationship between primary and specialty palliative care in the ICU remain poorly defined with few empiric data to guide clinicians’ decision making.

Recent efforts to improve palliative care for patients admitted to the ICU have focused on the use of screening criteria or “triggers” to prompt or automate SPC. An analysis of a national database estimated that 14–20% of ICU patients meet at least one of the commonly used, recommended triggers for SPC [11]. Because this need for palliative care cannot be met by the existing palliative care workforce in the United States [12], it is essential to define the optimal role of SPC and primary palliative care in the ICU. Existing triggers have been designed with the primary goal of identifying patients at highest risk of in-hospital death [6, 13–15]. However, it is unknown whether patients at the highest risk for imminent death are the most likely to derive benefit from SPC [16] and, alternatively, whether a mortality-based trigger strategy unintentionally overlooks an important population of patients who may benefit from SPC.

In this study, we sought to evaluate the most common and recommended mortality-based SPC triggers and determine whether these triggers reflect actual use of SPC in a large, urban, academic medical center with multiple subspecialty ICUs. In addition, we aimed to characterize the population of patients who receive SPC guided by clinician-identified need, without meeting mortality-based triggers.

Methods

Study design, setting, and participants

We conducted a retrospective cohort study of all adult (age > 18) ICU admissions from January 1st, 2012 to December 31st, 2017 at an 894-bed, urban academic medical center with 115 intensive care beds distributed over five subspecialty ICUs (Surgical Intensive Care Unit (SICU), Medical Intensive Care Unit (MICU), Cardiac Care Unit (CCU), Neurosciences-spine Intensive Care Unit (NSICU), and a Cardiothoracic/Transplant Intensive Care Unit (CTICU)). No hospital or unit protocols to standardize SPC consultation were used during the study period, apart from routine SPC consultations during evaluation for a cardiac Ventricular Assist Device (VAD). We excluded admissions with an ICU length of stay less than four hours and patients who received SPC consultation prior to the ICU admission during the study period. Deidentified patient-level data was obtained from the Northwestern Medicine Enterprise Data Warehouse, an integrated repository of all electronic health record data from the study site. The Northwestern University Institutional Review Board reviewed and approved the study, and a waiver of informed consent was obtained.

Variables

Admissions were defined as receiving SPC if an initial palliative care consultation report was documented in the electronic health record (EHR) during the ICU admission. Admissions were defined as having a SPC trigger if they met at least one of the five most recommended, published PC triggers: (1) ICU admission following a hospital stay greater than or equal to 10 days; (2) age greater than 80 with two or more life-threatening comorbidities (as defined by the Acute Physiology and Chronic Health Evaluation IV definitions of severe chronic organ insufficiency); (3) diagnosis of active stage IV malignancy; (4) status post cardiac arrest; or (5) diagnosis of intracerebral hemorrhage requiring mechanical ventilation [7, 11, 16].

The reasons for SPC were abstracted from the first palliative care consultation report in the EHR from the “Reason for Consult” or “Reason for Encounter” section of the note template. Reason for consult is selected by the note writer from a drop-down box within the note template; selection from this list is not mandatory for consult completion. If a note did not include information within this section, the entire content of the note was reviewed for additional free-text statements (“palliative care consulted for […]” or “asked to see patient to […]”) that explicitly stated the reasons for consultation. Two investigators independently reviewed and categorized all non-template, free-text consult reasons, and a third investigator arbitrated all discrepancies. If an SPC note included more than one reason for consultation, all reasons were abstracted and included in the analysis. If an SPC did not contain any stated reason for consultation, the encounter was designated as “no evident reason for consultation”, which was categorized as “other” during analyses. Reasons for consultation were grouped into nine categories based on the palliative care consult note template options: (1) goals of care, (2) symptom management, (3) end-of-life care management, (4) disposition planning, (5) medical decision making, (6) patient/family support, (7) advance care planning, (8) consideration of mechanical circulatory support, and (9) other, including no evident reason for consultation.

Statistical analysis

We assessed differences in demographics, clinical characteristics, and outcomes between the ICU admissions that received a new specialty palliative care consultation and those that did not, using t-tests and chi-square tests of independence. We repeated these analyses for ICU admissions that had received SPC, comparing those who had at least one SPC trigger with those that did not meet any trigger criteria. Of ICU admissions with a palliative care consultation, we conducted t-tests to evaluate for any differences in the reasons for SPC, again between those who had at least one SPC trigger and those without.

We evaluated the accuracy of using SPC trigger criteria to predict receipt of SPC when compared to common illness severity scores associated with in-hospital mortality. For all ICU admissions, a Receiver-Operating-Characteristic (ROC) curve analysis was conducted and area under the curve (AUC) was calculated to determine how accurately the Sepsis-related Organ Failure Assessment (SOFA) score [17-19] and Acute Physiology Score (APS) [20] at the time of ICU admission predicted new specialty palliative care consultation in the ICU. The calculated Youden index was used to determine cut-off values for the SOFA and APS scores that resulted in optimal sensitivity and specificity. Sensitivity and specificity for predicting actual receipt of SPC were calculated for: (1) one or more trigger criteria, (2) each individual trigger criterion, (3) admission APS score, and (4) admission SOFA score. Statistical significance was defined by a 2-tailed P-value of <0.05. SAS version 9.4 software (Cary, NC, USA) was used for all analyses [21].

Results

Of 48,744 eligible ICU admissions, 10,513 (21.57%) met one or more triggers for SPC, and 1,965 (4.03%) received specialty palliative care consultation (Fig 1). The trigger most frequently met was a diagnosis of active stage IV malignancy (38.88%), followed by ICU admission after a hospital stay greater than 10 days (27.36%), age greater than 80 years with two or more life-threatening comorbidities (24.26%), ICU admission after cardiac arrest (15.45%), and intracerebral hemorrhage requiring mechanical ventilation (5.10%). The rate of SPC by subspecialty ICU was: 1,018 (6.81%) of MICU admissions, 292 (4.77%) of CCU admissions, 319 (2.83%) of NSICU admissions, 205 (2.24%) of CTICU admissions, and 131 (1.81%) of SICU admissions (p <0.001). The table in S1 Table shows the demographics, clinical characteristics, and hospitalization outcomes of the entire cohort, comparing those who received SPC to those who did not.

Fig 1

The relationship between actual specialty palliative care delivery and recommended trigger criteria among all adults admitted to the intensive care unit.

Of 48,744 eligible intensive care unit admissions, 10,513 (21.57%) met one or more triggers for specialty palliative care consultations, and 1,965 (4.03%) received specialty palliative care consultation. Of all intensive care unit admissions, only 979 (2.01%) both received specialty palliative care and met one or more triggers for specialty palliative care.

Of the 1,965 ICU admissions that included specialty palliative care consultation, 979 (49.82%) met one or more SPC triggers (Table 1). Admissions that did not meet any SPC triggers but received SPC were younger than admissions that met triggers (62.71 years vs 66.58 years, mean difference (MD) 3.87 years (95% CI 2.44–5.30), p<0.0001). Illness severity by SOFA or APS score did not differ between groups. Admissions that did not meet any triggers had more days from ICU admission to specialty palliative care consultation (6.63 days vs 5.20 days, MD -1.43 days (95% CI -2.24 –-0.62), p<0.001) and longer ICU length of stay (11.43 days vs 8.42 days, MD -3.01 days (95% CI -4.30 –-1.72), p < 0.001) compared to those that met triggers. In the CCU and CTICU, there were more trigger-negative compared to trigger-positive admissions receiving SPC (18.97% vs 10.73%, p < 0.001, and 14.30% vs 6.45%, p < 0.001, respectively). Admissions that did not meet any triggers were less likely to die in the hospital than those meeting triggers (48.68% vs 58.12%, p < 0.001). Among survivors, hospital discharge destination differed for trigger-negative vs trigger-positive admissions, respectively: 48.42% vs 43.90% were discharged to home, 20.36% vs 30.00% to hospice, and 28.26% vs 23.90% to a post-acute care facility (p = 0.008).

Table 1

Demographics, clinical characteristics, and hospitalization outcomes of intensive care unit admissions who received specialty palliative care consultation in the intensive care unit.

	All admissions that received specialty palliative care	Admissions with one or more triggers	Admissions with no triggers	Mean difference or df, N	P-value
	N = 1965	N = 979	N = 986
Age in years, mean (95% CI)	64.64 (63.92–65.36)	66.58 (65.55–67.60)	62.71 (61.71–63.71)	3.87 (2.44–5.30)	<0.001a
Female, No./total No. (%)	918/1965 (46.72)	473/979 (48.31)	445/986 (45.13)	df = 1, N = 1965	0.157b
Race, No./total No. (%)				df = 4, N = 1965	0.167b
White or Caucasian	1031/1965 (52.47)	529/979 (54.03)	502/986 (50.91)		-
Black or African American	399/1965 (20.31)	186/979 (19.00)	213/986 (21.60)		-
Asian	67/1965 (3.41)	40/979 (4.09)	27/986 (2.74)		-
Other Racec	215/1965 (10.94)	107/979 (10.93)	108/986 (10.95)		-
Unknown Raced	253/1965 (12.88)	117/979 (11.95)	136/986 (13.79)
Hispanic or Latinx Ethnicity	145/1965 (7.38)	70/979 (7.15)	75/986 (7.61)	df = 2, N = 1965	0.656b
Illness severity, mean (95% CI)
APS Score	55.27 (54.07–56.46)	56.26 (54.54–57.99)	54.27 (52.62–55.92)	2.00 (-0.39–4.39)	0.101a
SOFA score	5.69 (5.51–5.86)	5.70 (5.45–5.94)	5.68 (5.43–5.93)	0.01 (-0.34–0.36)	0.939a
ICU type, No./total No. (%)				df = 4, N = 1965	<0.001b
MICU	1018/1965 (51.81)	558/979 (57.00)	460/986 (46.65)		-
NSICU	319/1965 (16.23)	185/979 (18.90)	134/986 (13.59)		-
CCU	205/1965 (10.43)	105/979 (10.73)	187/986 (18.97)		-
CTICU	131/1965 (6.67)	64/979 (6.54)	141/986 (14.30)		-
SICU	292/1965 (14.86)	67/979 (6.84)	64/986 (6.49)		-
ICU length of stay in days, mean (95% CI)	9.93 (9.29–10.57)	8.42 (7.67–9.16)	11.43 (10.38–12.47)	-3.01 (-4.30 –-1.72)	<0.001a
In-hospital death, No./total No. (%)	1049/1965 (53.38)	569/979 (58.12)	480/986 (48.68)	df = 1, N = 1965	<0.001b
Discharge disposition among survivorse				df = 3, N = 914	0.008b
Home	425/914 (46.40)	180/409 (43.90)	245/505 (48.42)		-
Hospicef	226/914 (24.67)	123/409 (30.00)	103/505 (20.36)		-
Post-acute Care Facilityg	212/914 (26.31)	98/409 (23.90)	143/505 (28.26)		-
Otherh	22/914 (2.40)	8/409 (1.95)	14/505 (2.77)		-

a T-test.

b χ2 test of independence.

c Other Race includes American Indian or Alaska Native, Hispanic or Latinx, Native Hawaiian or other Pacific Islander, or Other.

d Unknown Race includes Declined, Unable to answer, Unknown, or missing data.

e Statistics were calculated using available data.

f Hospice includes Expiration Hospice, Home with Hospice, or Inpatient Hospice.

g Post-acute Care Facility includes Acute Inpatient Rehabilitation, Long-term Acute Care hospital (LTAC), Nursing Home (Custodial), and Skilled nursing facility or subacute rehab.

h Other includes Acute Care Hospital, Against Medical Advice (AMA) or Elopement, Intermediate care facility including state, Other, or VA System Facility.

Abbreviations: APS = Acute Physiology Score, CCU = Cardiac Care Unit, CTICU = Cardiothoracic/Transplant Intensive Care Unit, ICU = Intensive Care Unit, MICU = Medical Intensive Care Unit, NSICU = Neurosciences-spine Intensive Care Unit, SICU = Surgical Intensive Care Unit, SOFA = Sequential Organ Failure Assessment, SPC = Specialty Palliative Care.

Among the 1,965 admissions that included SPC, the reasons for consultation differed by trigger status (Table 2). Goals of care, symptom management, and disposition planning were more frequent reasons for consultation in the trigger-positive admissions compared to trigger negative admissions (65.58% vs 57.51%, MD 0.08 (95% CI 0.04–0.12), p<0.001; 39.94% vs 32.76%, MD 0.07 (0.03–0.11), p<0.001; and 18.59% vs 14.60%, MD 0.04 (95% CI 0.01–0.07), p = 0.018, respectively). Automated consultation prior to VAD was more frequent in trigger-negative compared to trigger-positive admissions (9.03% vs 1.43%, MD -0.08 (95% CI -0.10 –-0.06), p < 0.001).

Table 2

Characteristics of specialty palliative care consultations in the intensive care unit.

	All admissions that received specialty palliative care	Admissions with one or more triggers	Admissions with no triggers	Mean difference	P-valuea
	N = 1965	N = 979	N = 986
Days from ICU admission to consult, mean (95% CI)	5.92 (5.51–6.32)	5.20 (4.70–5.70)	6.63 (5.99–7.26)	-1.43 (-2.24 –-0.62)	<0.001
Days from consult to hospital discharge, mean (95% CI)	4.01 (3.57–4.45)	3.22 (2.79–3.65)	4.80 (4.04–5.55)	-1.58 (-2.45 –-0.71)	<0.001
Reasons for consultation, No./Total No. (%)
Goals of care	1209/1965 (61.53)	642/979 (65.58)	567/986 (57.51)	0.08 (0.04–0.12)	<0.001
Symptom management	714/1965 (36.34)	391/979 (39.94)	323/986 (32.76)	0.07 (0.03–0.11)	<0.001
End-of-life management	413/1965 (21.02)	221/979 (22.57)	192/986 (19.47)	0.03 (-0.01–0.07)	0.092
Disposition planning	326/1965 (16.59)	182/979 (18.59)	144/986 (14.60)	0.04 (0.01–0.07)	0.018
Automated consult prior to VAD/cardiac mechanical support	103/1965 (5.24)	14/979 (1.43)	89/986 (9.03)	-0.08 (-0.10 –-0.06)	<0.001
Other	101/1965 (5.14)	43/979 (4.39)	58/986 (5.88)	-0.01 (-0.03–0.01)	0.135
Patient/family Support	46/1965 (2.34)	24/979 (2.45)	22/986 (2.23)	0.00 (-0.01–0.02)	0.747
Medical decision making	39/1965 (1.98)	17/979 (1.74)	22/986 (2.23)	-0.01 (-0.02–0.01)	0.432
Advanced care planning	38/1965 (1.93)	15/979 (1.53)	23/986 (2.33)	-0.00 (-0.02–0.00)	0.198

a T-test.

Abbreviations: VAD = Ventricular Assist Device.

The sensitivity and specificity for any trigger predicting actual specialty palliative care consultation was 49.82% and 79.61%, respectively (Table 3). The APS and SOFA scores at the time of ICU admission had sensitivities of 66.87% and 65.60% and specificities of 64.82% and 63.97%, respectively. The AUCs for any SPC trigger, APS score, and SOFA scores were 0.65, 0.72, and 0.69, respectively (S1 Fig demonstrates the ROC curves for APS score and SOFA score).

Table 3

Accuracy of mortality-based triggers and severity of illness scores for predicting specialty palliative care consultation in the intensive care unit.

	Sensitivity	Specificity	AUC
APS score at time of ICU admissiona	66.87%	64.82%	0.72
SOFA score at time of ICU admissiona	65.60%	63.97%	0.69
One or more triggers met	49.82%	79.61%	0.65
Individual triggers:
Metastatic cancer	25.39%	92.33%	0.59
Greater than 10 day hospital stay before ICU admission	13.54%	94.42%	0.54
Age greater than 80 plus two or more life threatening conditions	10.28%	94.98%	0.53
ICU admission following cardiac arrest	6.26%	96.79%	0.52
Intracerebral hemorrhage requiring mechanical ventilation	3.82%	99.01%	0.51

a The calculated Youden index was used to determine cut-off values for the SOFA and APS scores that resulted in optimal sensitivity and specificity. A cut-off of 39.07 was used for APS score and a cut-off of 4.00 was used for SOFA score.

Abbreviations: APS = Acute Physiology Score, AUC = Area Under the Curve, ICU = Intensive Care Unit, SOFA = Sequential Organ Failure Assessment.

Discussion

In this retrospective cohort study of almost 50,000 critically ill patients from a large academic medical center, we found that commonly recommended, mortality-based triggers for automated SPC are neither sensitive nor specific for predicting actual receipt of specialty palliative care consultation in the ICU. We also identified and characterized a population of younger patients with prolonged critical illness who are overlooked by mortality-based triggers but have a clinician-identified need for SPC. Because the ideal role of SPC in the ICU has not been established, definitive interpretation of the appropriateness of specialty palliative care utilization in this study cannot be determined. Nevertheless, our findings underscore the need to develop a more refined approach to optimize specialty palliative care delivery in the ICU and to overcome the common misperception that specialty palliative care is only relevant for patients facing imminent death.

The findings of this study build on and align with recent work that demonstrates the limitations of existing SPC triggers and seeks to define the optimal relationship between primary and specialty palliative care in the ICU. Hua and colleagues have focused on the need to develop novel triggers for SPC that are based on risk of longer-term outcomes instead of in-hospital mortality and have demonstrated that existing SPC triggers have low sensitivity and high specificity for predicting 6-month mortality [22]. In a large, multicenter survey of ICU clinicians, the vast majority of participants felt that that SPC was underutilized in the ICU and endorsed an automated-trigger based system to protocolize consultation [9]. While clinicians approved of the existing approach to SPC triggers that reflect a high risk of in-hospital mortality, they also endorsed the importance of non-mortality-based triggers such as unrealistic expectations and clinician-family conflict that have not been, to date, incorporated into SPC trigger criteria [9]. Clinician endorsement of the role of SPC in the ICU beyond the narrow focus of imminent death is consistent with our finding that over half of patients for whom clinicians request SPC consultation do not meet any of the existing mortality-based trigger criteria.

Patients in this study with a clinician-identified need for SPC but who did not meet any of the existing triggers differed in important ways from patients who met triggers. This population of patients are younger, have lengthy ICU admissions, are less likely to die, and more likely to receive SPC late into their ICU stay. The benefits of integrating palliative care early in the course of advanced cancer have been well-described [23, 24], and there is mounting evidence that the timing of in-hospital SPC is important. Early palliative care consultation in the ICU (defined as within 4 days of ICU admission) has been associated with reductions in hospital length of stay and direct costs [25]. A study of hospitalized oncologic patients found that late specialty palliative care consultations (defined as greater than one week into admission) were associated with a marked increase in health utilization outcomes [26]. Prior studies have also demonstrated a high burden of unmet palliative care needs for patients living in long-term care or nursing facilities [27-29], and in this study we found that more than a quarter of trigger-negative patients who receive SPC are discharged to post-acute care facilities. Taken together, our findings that a significant portion of trigger-negative patients receive SPC late in their course and are discharged to post-acute care facilities highlights an opportunity to develop novel triggers that proactively identify this group of patients with SPC needs. Our findings align with ongoing efforts to leverage the electronic health record to support the delivery of specialty palliative care by developing more nuanced predictive models and screening algorithms that move beyond in-patient mortality and combine patient characteristics, patient and family self-reported needs and symptoms, and risk for short- and long-term morbidity and mortality [30-34].

We found that in-hospital mortality-based triggers were particularly insensitive to the SPC needs among patients in the two cardiovascular-focused intensive care units included in this study. This finding may be explained, in part, because existing triggers were developed in medical, surgical, or neurological patient populations [16]. Nevertheless, prior literature demonstrates important unmet palliative care needs in this population; patients with advanced heart failure report similar or higher number of symptoms than patients with advanced cancer [35], are referred later to palliative care [36], and are more likely than those with cancer to be referred from a critical care unit [36]. Despite professional guideline recommendations for palliative care involvement for heart failure patients [37], there is no clear consensus regarding when and how to implement palliative care services. Slavin and Warraich suggest initiating specialty palliative care referral at critical moments in the trajectory of heart failure patients, such as at the time of hospitalization or evaluation for certain procedures [38]. Our work affirms the inadequacy of mortality-based triggers for identifying SPC needs among the population of patients with cardiovascular critical illness.

“Goals of care” was by far the most frequent stated reason for SPC, regardless of trigger status, which is consistent with previous work highlighting the need to establish realistic and patient-identified, overarching goals of medical treatments among hospitalized patients [39, 40]. However, the use and meaning of the specific phrase “goals of care” in clinical care is nonspecific [41], and is used in clinical documentation to indicate a variety of constructs unrelated to a patient’s individual values and preferences, including poor prognosis, to describe conflict with families, or to provide rationale for limitations on specific medical interventions [42]. The ambiguity and broad range of palliative care constructs associated with this phrase may partly explain the high frequency of “goals of care” as a reason for consultation and may also indicate that clinicians are challenged to clearly articulate their rationale for requesting the assistance of specialty palliative care consultants in the care of critically ill patients.

Our study has strengths and limitations. The study setting of a large, academic health center with five separate ICUs and a high volume of critically ill patients without standardized specialty palliative care consultation provides the opportunity to study and uncover clinician-derived practice patterns. However, hospital-level variability in specialty palliative care availability and practice patterns is well-described [43], thus our results from a single center may not be generalizable to other institutions who have differing palliative care resources, patient populations, or care delivery models. Future work could expand on our findings by exploring the use of SPC triggers across multiple medical centers. Because we were unable to measure whether patients’ palliative care needs were truly met during their ICU stays through either primary or specialty palliative care, we are unable to determine the appropriateness (including underutilization or overutilization) of SPC in this cohort or whether SPC was an effective mechanism to meet these needs. The documented reason for consultation in the EHR may not fully capture the intended purpose of the consultation and may not reflect all of the palliative care needs ultimately identified by the consultant. Future study using qualitative methods may provide a more rich and nuanced understanding of what motivates ICU clinicians to request SPC.

Conclusions

Mortality-based triggers designed to support automated specialty palliative care consultation in the ICU poorly reflect actual delivery of SPC guided by clinician-identified needs. In this study, patients who received SPC but did not meet any existing SPC triggers were younger, had longer ICU length of stay, were less likely to die in the hospital, and received SPC later in their course of illness. Reliance on automated, mortality-based triggers to improve the delivery of palliative care in the ICU may unintentionally overlook or delay SPC in an important population of patients with specialty palliative care needs but who are at lower risk for in-hospital mortality. To achieve optimal palliative care delivery in the ICU, system-level interventions such as automated triggers should focus on the broader role of SPC for patients with serious illness, beyond the narrow focus of imminent end of life.

Receiver Operating Characteristic curves for APS and SOFA severity scores for predicting specialty palliative care consultation among all ICU admissions.

Abbreviations: APS = Acute Physiology Score, SOFA = Sequential Organ Failure Assessment.

(TIF)

Click here for additional data file.

Demographics, clinical characteristics, and hospitalization outcomes of all intensive care unit admissions, stratified by receipt of specialty palliative care consultation.

a T-test. b χ2 test of independence. c Statistics were calculated using available data. d Other Race includes American Indian or Alaska Native, Hispanic or Latinx, Native Hawaiian or other Pacific Islander, or Other. e Unknown Race includes Declined, Unable to answer, Unknown, or missing data. Abbreviations: APS = Acute Physiology Score, CCU = Cardiac Care Unit, CTICU = Cardiothoracic/Transplant Intensive Care Unit, ICU = Intensive Care Unit, MICU = Medical Intensive Care Unit, NSICU = Neurosciences-spine Intensive Care Unit, SICU = Surgical Intensive Care Unit, SOFA = Sequential Organ Failure Assessment.

(DOCX)

Click here for additional data file.

17 May 2021

PONE-D-21-08904

Evaluation of Automated Specialty Palliative Care in the Intensive Care Unit: A Retrospective Cohort Study

PLOS ONE

Dear Dr. Secunda,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Jul 01 2021 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Tai-Heng Chen, M.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

3. We noted in your submission details that a portion of your manuscript may have been presented or published elsewhere.

[A portion of this work was published as an American Thoracic Society International Conference abstract. ]

Please clarify whether this conference proceeding was peer-reviewed and formally published. If this work was previously peer-reviewed and published, in the cover letter please provide the reason that this work does not constitute dual publication and should be included in the current manuscript.

4. Please include captions for your Supporting Information files at the end of your manuscript, and update any in-text citations to match accordingly. Please see our Supporting Information guidelines for more information: http://journals.plos.org/plosone/s/supporting-information.

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: very relevant data.Your study needs to include later other centers furthermore it is ,I think very interesting to study different categories of patients like patients COVID 19 infected as this subject is a big field of research and evaluation

Reviewer #2: The study is well conducted and the conclusions are appropriate. The authors show the need for new triggers for automated speciality palliative care in ICU. Further investigation could be useful in assessing the effectiveness of the implementation with these new triggers.

Reviewer #3: This manuscript by Secunda and coworkers reports the results of a retrospective cohort study in the intensive care unit (ICU) setting. The study aimed to examine whether the most commonly used Specialty Palliative Care (SPC) triggers accurately represent actual SPC consult in the ICU at an academic center. Using SPC triggers alone may miss subsets of patients who may otherwise benefit from SPC (e.g., younger patients). The authors state that the mortality-based SPC triggers used in this study are ones most commonly used and discussed in literature, which include: ICU admission following hospital stay >= 10 days, age >80 with 2 or more life-threatening comorbidities, stage 4 malignancy, post cardiac arrest, and intracerebral hemorrhage requiring mechanical ventilation.

The authors found that these conventional mortality-based criteria for SPC triggers have poor sensitivity and specificity for actual SPC consults. They did acknowledge that they cannot evaluate the appropriateness of the actual SPC utilization due to inability to measure whether the patient’s palliative care need was truly met.

The implications and future directions of this study were well stated in the discussion: to develop a more patient-centered predictive model to better and earlier identify patients who may benefit from SPC.

Minor comments:

Please follow the STROBE guideline for reporting more closely. Use the predefined headers.

The IRB approval statement typically appears in the first paragraph of the methods.

Line 101-102: “Multiple reasons for consultation could be used for single SPC encounter, and some notes did not document an explicit reason for consultation.” How were these data treated?

Table 1: Why is the denominator for race in all 3 columns not corresponding to the total number of subjects? For example, the first column under race shouldn’t the denominator be 1965 instead of 1896? What happened to the missing numbers? Could be due to “unknown” category? If so, would add an “unknown” category for race.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Pr Hanane EZZOUINE

Reviewer #2: No

Reviewer #3: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

7 Jul 2021

Reviewer comments

Reviewer's Responses to Questions

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: No

Author response: The raw data from this study cannot be shared publicly because they contain potentially identifying and sensitive patient information derived from electronic health records. Data requests by researchers who meet criteria for access to confidential data can be directed to the Northwestern University Institutional Review Board (contact via email at irb@northwestern.edu).

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

Reviewer #3: Yes

Reviewer Comments to the Author

Reviewer #1: very relevant data. Your study needs to include later other centers furthermore it is I think very interesting to study different categories of patients like patients COVID 19 infected as this subject is a big field of research and evaluation

Author response: Thank you for your comments. We agree that future research should examine the use of specialty palliative care triggers in different patient populations such as those with COVID-19. We also agree that future work should explore the use of specialty palliative care triggers across multiple medical centers.

Reviewer #2: The study is well-conducted and the conclusions are appropriate. The authors show the need for new triggers for automated specialty palliative care in ICU. Further investigation could be useful in assessing the effectiveness of the implementation with these new triggers.

Author response: Thank you for your comments. We agree that future work should involve development and implementation of more patient-centered triggers to better identify patients who may benefit from specialty palliative care.

Reviewer #3:

1. This manuscript by Secunda and coworkers reports the results of a retrospective cohort study in the intensive care unit (ICU) setting. The study aimed to examine whether the most commonly used Specialty Palliative Care (SPC) triggers accurately represent actual SPC consult in the ICU at an academic center. Using SPC triggers alone may miss subsets of patients who may otherwise benefit from SPC (e.g., younger patients). The authors state that the mortality-based SPC triggers used in this study are ones most commonly used and discussed in literature, which include: ICU admission following hospital stay >= 10 days, age >80 with 2 or more life-threatening comorbidities, stage 4 malignancy, post cardiac arrest, and intracerebral hemorrhage requiring mechanical ventilation.

The authors found that these conventional mortality-based criteria for SPC triggers have poor sensitivity and specificity for actual SPC consults. They did acknowledge that they cannot evaluate the appropriateness of the actual SPC utilization due to inability to measure whether the patient’s palliative care need was truly met.

The implications and future directions of this study were well stated in the discussion: to develop a more patient-centered predictive model to better and earlier identify patients who may benefit from SPC.

2. Minor comments

a.) Please follow the STROBE guideline for reporting more closely. Use the predefined headers.

Author response: We appreciate your comment and have updated the manuscript to follow the STROBE guidelines more closely.

b.) The IRB approval statement typically appears in the first paragraph of the methods.

Author response: We appreciate your comment and have moved the IRB approval statement to the first paragraph in the methods section in the revised manuscript.

c.) Line 101-102: “Multiple reasons for consultation could be used for single SPC encounter, and some notes did not document an explicit reason for consultation.” How were these data treated?

Author response: We appreciate the opportunity to further clarify how we categorized reasons for specialty palliative care consultation. We have updated the methods section in the revised manuscript with the following clarification statement:

If an SPC note included more than one reason for consultation, all reasons were abstracted and included in the analysis. If an SPC did not contain any stated reason for consultation, the encounter was designated as “no evident reason for consultation”, which was categorized as “other” during analyses.

d.) Table 1: Why is the denominator for race in all 3 columns not corresponding to the total number of subjects? For example, the first column under race shouldn’t the denominator be 1965 instead of 1896? What happened to the missing numbers? Could be due to “unknown” category? If so, would add an “unknown” category for race.

Author response: Thank you for pointing out this additional opportunity to clarify our results. In the dataset used in this study, “Unknown” is a selection made in the electronic medical record for the demographic variables race and ethnicity. “Missing” signifies that data is absent for these variables. We originally treated “unknown” differently than missing data which is why the original denominator was 1896 (which excluded the admissions missing race and/or ethnicity data). However, upon further consideration of these categories based on your comments, we agree that missing race and ethnicity data should be treated the same as “unknown” in a separate category. Thus, we have updated table 1 and the supplementary table to reflect the following new categories for race: (1) White or Caucasian, (2) Black or African American, (3) Asian, (4) Other Race (which includes American Indian or Alaska Native, Hispanic or Latinx Race, Native Hawaiian or other Pacific Islander, and Other), and (5) Unknown Race (which includes declined, unable to answer, unknown, or missing data). We have also updated the statistics to reflect this change which includes the unknown category (which is a combination of the available data stating that race is unknown and missing data). Please see the revised manuscript, Table 1, and Supplemental Table for specific tracked changes.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

28 Jul 2021

Evaluation of Automated Specialty Palliative Care in the Intensive Care Unit: A Retrospective Cohort Study

PONE-D-21-08904R1

Dear Dr. Secunda,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Tai-Heng Chen, M.D.

Academic Editor

PLOS ONE

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #3: All comments have been addressed

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #3: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #3: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #3: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #3: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #3: Thank you for the opportunity to re-review your work.

In this revision of the original submission, the authors appropriately addressed all comments and suggestions by the reviewers.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #3: No

2 Aug 2021

PONE-D-21-08904R1

Evaluation of Automated Specialty Palliative Care in the Intensive Care Unit: A Retrospective Cohort Study

Dear Dr. Secunda:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Tai-Heng Chen

Academic Editor

PLOS ONE